From 14c1c2b9351f16d43ba4e6b6c9062edad44a6bec Mon Sep 17 00:00:00 2001 From: Alexandre Simard Date: Wed, 19 Oct 2022 13:53:52 -0400 Subject: Show PB texts at same time and earlier For big tasks (1000+ steps), waiting 1 minute to see ETA is long and this changes it so the number of steps done plays a role in showing the text as well. --- modules/ui.py | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index a2dbd41e..0abd177a 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -261,14 +261,14 @@ def wrap_gradio_call(func, extra_outputs=None): return f -def calc_time_left(progress, threshold, label, force_display): +def calc_time_left(progress, threshold, label, force_display, showTime): if progress == 0: return "" else: time_since_start = time.time() - shared.state.time_start eta = (time_since_start/progress) eta_relative = eta-time_since_start - if (eta_relative > threshold and progress > 0.02) or force_display: + if (eta_relative > threshold and showTime) or force_display: if eta_relative > 3600: return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) elif eta_relative > 60: @@ -290,7 +290,10 @@ def check_progress_call(id_part): if shared.state.sampling_steps > 0: progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps - time_left = calc_time_left( progress, 1, " ETA: ", shared.state.time_left_force_display ) + # Show progress percentage and time left at the same moment, and base it also on steps done + showPBText = progress >= 0.01 or shared.state.sampling_step >= 10 + + time_left = calc_time_left( progress, 1, " ETA: ", shared.state.time_left_force_display, showPBText ) if time_left != "": shared.state.time_left_force_display = True @@ -298,7 +301,7 @@ def check_progress_call(id_part): progressbar = "" if opts.show_progressbar: - progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if progress > 0.01 else ""}
""" + progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if showPBText else ""}
""" image = gr_show(False) preview_visibility = gr_show(False) -- cgit v1.2.3 From 29e74d6e71826da9a3fe3c5790fed1329fc4d1e8 Mon Sep 17 00:00:00 2001 From: Melan Date: Thu, 20 Oct 2022 16:26:16 +0200 Subject: Add support for Tensorboard for training embeddings --- modules/shared.py | 4 ++++ modules/textual_inversion/textual_inversion.py | 31 +++++++++++++++++++++++++- 2 files changed, 34 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index faede821..2c6341f7 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -254,6 +254,10 @@ options_templates.update(options_section(('training', "Training"), { "dataset_filename_join_string": OptionInfo(" ", "Filename join string"), "training_image_repeats_per_epoch": OptionInfo(1, "Number of repeats for a single input image per epoch; used only for displaying epoch number", gr.Number, {"precision": 0}), "training_write_csv_every": OptionInfo(500, "Save an csv containing the loss to log directory every N steps, 0 to disable"), + "training_enable_tensorboard": OptionInfo(False, "Enable tensorboard logging."), + "training_tensorboard_save_images": OptionInfo(False, "Save generated images within tensorboard."), + "training_tensorboard_flush_every": OptionInfo(120, "How often, in seconds, to flush the pending tensorboard events and summaries to disk."), + })) options_templates.update(options_section(('sd', "Stable Diffusion"), { diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 3be69562..c57d3ace 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -7,9 +7,11 @@ import tqdm import html import datetime import csv +import numpy as np +import torchvision.transforms from PIL import Image, PngImagePlugin - +from torch.utils.tensorboard import SummaryWriter from modules import shared, devices, sd_hijack, processing, sd_models import modules.textual_inversion.dataset from modules.textual_inversion.learn_schedule import LearnRateScheduler @@ -199,6 +201,19 @@ def write_loss(log_directory, filename, step, epoch_len, values): **values, }) +def tensorboard_add_scaler(tensorboard_writer, tag, value, step): + if shared.opts.training_enable_tensorboard: + tensorboard_writer.add_scalar(tag=tag, + scalar_value=value, global_step=step) + +def tensorboard_add_image(tensorboard_writer, tag, pil_image, step): + if shared.opts.training_enable_tensorboard: + # Convert a pil image to a torch tensor + img_tensor = torch.as_tensor(np.array(pil_image, copy=True)) + img_tensor = img_tensor.view(pil_image.size[1], pil_image.size[0], len(pil_image.getbands())) + img_tensor = img_tensor.permute((2, 0, 1)) + + tensorboard_writer.add_image(tag, img_tensor, global_step=step) def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): assert embedding_name, 'embedding not selected' @@ -252,6 +267,12 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) optimizer = torch.optim.AdamW([embedding.vec], lr=scheduler.learn_rate) + if shared.opts.training_enable_tensorboard: + os.makedirs(os.path.join(log_directory, "tensorboard"), exist_ok=True) + tensorboard_writer = SummaryWriter( + log_dir=os.path.join(log_directory, "tensorboard"), + flush_secs=shared.opts.training_tensorboard_flush_every) + pbar = tqdm.tqdm(enumerate(ds), total=steps-ititial_step) for i, entries in pbar: embedding.step = i + ititial_step @@ -270,6 +291,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc del x losses[embedding.step % losses.shape[0]] = loss.item() + optimizer.zero_grad() loss.backward() @@ -285,6 +307,12 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc embedding.save(last_saved_file) embedding_yet_to_be_embedded = True + if shared.opts.training_enable_tensorboard: + tensorboard_add_scaler(tensorboard_writer, "Loss/train", losses.mean(), embedding.step) + tensorboard_add_scaler(tensorboard_writer, f"Loss/train/epoch-{epoch_num}", losses.mean(), epoch_step) + tensorboard_add_scaler(tensorboard_writer, "Learn rate/train", scheduler.learn_rate, embedding.step) + tensorboard_add_scaler(tensorboard_writer, f"Learn rate/train/epoch-{epoch_num}", scheduler.learn_rate, epoch_step) + write_loss(log_directory, "textual_inversion_loss.csv", embedding.step, len(ds), { "loss": f"{losses.mean():.7f}", "learn_rate": scheduler.learn_rate @@ -349,6 +377,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc embedding_yet_to_be_embedded = False image.save(last_saved_image) + tensorboard_add_image(tensorboard_writer, f"Validation at epoch {epoch_num}", image, embedding.step) last_saved_image += f", prompt: {preview_text}" -- cgit v1.2.3 From a6d593a6b51dc6a8443f2aa5c24caa391a04cd56 Mon Sep 17 00:00:00 2001 From: Melan Date: Thu, 20 Oct 2022 19:43:21 +0200 Subject: Fixed a typo in a variable --- modules/textual_inversion/textual_inversion.py | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index c57d3ace..ec8176bf 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -260,11 +260,11 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc last_saved_image = "" embedding_yet_to_be_embedded = False - ititial_step = embedding.step or 0 - if ititial_step > steps: + initial_step = embedding.step or 0 + if initial_step > steps: return embedding, filename - scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) + scheduler = LearnRateScheduler(learn_rate, steps, initial_step) optimizer = torch.optim.AdamW([embedding.vec], lr=scheduler.learn_rate) if shared.opts.training_enable_tensorboard: @@ -273,9 +273,9 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc log_dir=os.path.join(log_directory, "tensorboard"), flush_secs=shared.opts.training_tensorboard_flush_every) - pbar = tqdm.tqdm(enumerate(ds), total=steps-ititial_step) + pbar = tqdm.tqdm(enumerate(ds), total=steps-initial_step) for i, entries in pbar: - embedding.step = i + ititial_step + embedding.step = i + initial_step scheduler.apply(optimizer, embedding.step) if scheduler.finished: -- cgit v1.2.3 From 8f5912984794c4c69e429c4636e984854d911b6a Mon Sep 17 00:00:00 2001 From: Melan Date: Thu, 20 Oct 2022 22:37:16 +0200 Subject: Some changes to the tensorboard code and hypernetwork support --- modules/hypernetworks/hypernetwork.py | 18 ++++++++++- modules/textual_inversion/textual_inversion.py | 45 +++++++++++++++----------- 2 files changed, 44 insertions(+), 19 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 74300122..5e919775 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -4,6 +4,7 @@ import html import os import sys import traceback +import tensorboard import tqdm import csv @@ -18,7 +19,6 @@ import modules.textual_inversion.dataset from modules.textual_inversion import textual_inversion from modules.textual_inversion.learn_schedule import LearnRateScheduler - class HypernetworkModule(torch.nn.Module): multiplier = 1.0 @@ -291,6 +291,9 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) optimizer = torch.optim.AdamW(weights, lr=scheduler.learn_rate) + if shared.opts.training_enable_tensorboard: + tensorboard_writer = textual_inversion.tensorboard_setup(log_directory) + pbar = tqdm.tqdm(enumerate(ds), total=steps - ititial_step) for i, entries in pbar: hypernetwork.step = i + ititial_step @@ -315,6 +318,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log optimizer.zero_grad() loss.backward() optimizer.step() + mean_loss = losses.mean() if torch.isnan(mean_loss): raise RuntimeError("Loss diverged.") @@ -323,6 +327,14 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log if hypernetwork.step > 0 and hypernetwork_dir is not None and hypernetwork.step % save_hypernetwork_every == 0: last_saved_file = os.path.join(hypernetwork_dir, f'{hypernetwork_name}-{hypernetwork.step}.pt') hypernetwork.save(last_saved_file) + + if shared.opts.training_enable_tensorboard: + epoch_num = hypernetwork.step // len(ds) + epoch_step = hypernetwork.step - (epoch_num * len(ds)) + 1 + + textual_inversion.tensorboard_add(tensorboard_writer, loss=mean_loss, + global_step=hypernetwork.step, step=epoch_step, + learn_rate=scheduler.learn_rate, epoch_num=epoch_num) textual_inversion.write_loss(log_directory, "hypernetwork_loss.csv", hypernetwork.step, len(ds), { "loss": f"{mean_loss:.7f}", @@ -360,6 +372,10 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log processed = processing.process_images(p) image = processed.images[0] if len(processed.images)>0 else None + if shared.opts.training_enable_tensorboard and shared.opts.training_tensorboard_save_images: + textual_inversion.tensorboard_add_image(tensorboard_writer, f"Validation at epoch {epoch_num}", + image, hypernetwork.step) + if unload: shared.sd_model.cond_stage_model.to(devices.cpu) shared.sd_model.first_stage_model.to(devices.cpu) diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index ec8176bf..b1dc2596 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -201,19 +201,30 @@ def write_loss(log_directory, filename, step, epoch_len, values): **values, }) +def tensorboard_setup(log_directory): + os.makedirs(os.path.join(log_directory, "tensorboard"), exist_ok=True) + return SummaryWriter( + log_dir=os.path.join(log_directory, "tensorboard"), + flush_secs=shared.opts.training_tensorboard_flush_every) + +def tensorboard_add(tensorboard_writer, loss, global_step, step, learn_rate, epoch_num): + tensorboard_add_scaler(tensorboard_writer, "Loss/train", loss, global_step) + tensorboard_add_scaler(tensorboard_writer, f"Loss/train/epoch-{epoch_num}", loss, step) + tensorboard_add_scaler(tensorboard_writer, "Learn rate/train", learn_rate, global_step) + tensorboard_add_scaler(tensorboard_writer, f"Learn rate/train/epoch-{epoch_num}", learn_rate, step) + def tensorboard_add_scaler(tensorboard_writer, tag, value, step): - if shared.opts.training_enable_tensorboard: - tensorboard_writer.add_scalar(tag=tag, - scalar_value=value, global_step=step) + tensorboard_writer.add_scalar(tag=tag, + scalar_value=value, global_step=step) def tensorboard_add_image(tensorboard_writer, tag, pil_image, step): - if shared.opts.training_enable_tensorboard: - # Convert a pil image to a torch tensor - img_tensor = torch.as_tensor(np.array(pil_image, copy=True)) - img_tensor = img_tensor.view(pil_image.size[1], pil_image.size[0], len(pil_image.getbands())) - img_tensor = img_tensor.permute((2, 0, 1)) + # Convert a pil image to a torch tensor + img_tensor = torch.as_tensor(np.array(pil_image, copy=True)) + img_tensor = img_tensor.view(pil_image.size[1], pil_image.size[0], + len(pil_image.getbands())) + img_tensor = img_tensor.permute((2, 0, 1)) - tensorboard_writer.add_image(tag, img_tensor, global_step=step) + tensorboard_writer.add_image(tag, img_tensor, global_step=step) def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): assert embedding_name, 'embedding not selected' @@ -268,10 +279,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc optimizer = torch.optim.AdamW([embedding.vec], lr=scheduler.learn_rate) if shared.opts.training_enable_tensorboard: - os.makedirs(os.path.join(log_directory, "tensorboard"), exist_ok=True) - tensorboard_writer = SummaryWriter( - log_dir=os.path.join(log_directory, "tensorboard"), - flush_secs=shared.opts.training_tensorboard_flush_every) + tensorboard_writer = tensorboard_setup(log_directory) pbar = tqdm.tqdm(enumerate(ds), total=steps-initial_step) for i, entries in pbar: @@ -308,10 +316,8 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc embedding_yet_to_be_embedded = True if shared.opts.training_enable_tensorboard: - tensorboard_add_scaler(tensorboard_writer, "Loss/train", losses.mean(), embedding.step) - tensorboard_add_scaler(tensorboard_writer, f"Loss/train/epoch-{epoch_num}", losses.mean(), epoch_step) - tensorboard_add_scaler(tensorboard_writer, "Learn rate/train", scheduler.learn_rate, embedding.step) - tensorboard_add_scaler(tensorboard_writer, f"Learn rate/train/epoch-{epoch_num}", scheduler.learn_rate, epoch_step) + tensorboard_add(tensorboard_writer, loss=losses.mean(), global_step=embedding.step, + step=epoch_step, learn_rate=scheduler.learn_rate, epoch_num=epoch_num) write_loss(log_directory, "textual_inversion_loss.csv", embedding.step, len(ds), { "loss": f"{losses.mean():.7f}", @@ -377,7 +383,10 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc embedding_yet_to_be_embedded = False image.save(last_saved_image) - tensorboard_add_image(tensorboard_writer, f"Validation at epoch {epoch_num}", image, embedding.step) + + if shared.opts.training_enable_tensorboard and shared.opts.training_tensorboard_save_images: + tensorboard_add_image(tensorboard_writer, f"Validation at epoch {epoch_num}", + image, embedding.step) last_saved_image += f", prompt: {preview_text}" -- cgit v1.2.3 From 7543cf5e3b5eaced00582da257801227d1ff2a6e Mon Sep 17 00:00:00 2001 From: Melan Date: Thu, 20 Oct 2022 22:43:08 +0200 Subject: Fixed some typos in the code --- modules/hypernetworks/hypernetwork.py | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 5e919775..0cd94f49 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -284,19 +284,19 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log last_saved_file = "" last_saved_image = "" - ititial_step = hypernetwork.step or 0 - if ititial_step > steps: + initial_step = hypernetwork.step or 0 + if initial_step > steps: return hypernetwork, filename - scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) + scheduler = LearnRateScheduler(learn_rate, steps, initial_step) optimizer = torch.optim.AdamW(weights, lr=scheduler.learn_rate) if shared.opts.training_enable_tensorboard: tensorboard_writer = textual_inversion.tensorboard_setup(log_directory) - pbar = tqdm.tqdm(enumerate(ds), total=steps - ititial_step) + pbar = tqdm.tqdm(enumerate(ds), total=steps - initial_step) for i, entries in pbar: - hypernetwork.step = i + ititial_step + hypernetwork.step = i + initial_step scheduler.apply(optimizer, hypernetwork.step) if scheduler.finished: -- cgit v1.2.3 From 18f86e41f6f289042c075bff1498e620ab997b8c Mon Sep 17 00:00:00 2001 From: Melan Date: Mon, 24 Oct 2022 17:21:18 +0200 Subject: Removed two unused imports --- modules/hypernetworks/hypernetwork.py | 1 - modules/textual_inversion/textual_inversion.py | 1 - 2 files changed, 2 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 0cd94f49..2263e95e 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -4,7 +4,6 @@ import html import os import sys import traceback -import tensorboard import tqdm import csv diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index b1dc2596..589314fe 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -9,7 +9,6 @@ import datetime import csv import numpy as np -import torchvision.transforms from PIL import Image, PngImagePlugin from torch.utils.tensorboard import SummaryWriter from modules import shared, devices, sd_hijack, processing, sd_models -- cgit v1.2.3 From c4b5ca5778340b21288d84dfb8fe1d5773c886a8 Mon Sep 17 00:00:00 2001 From: Yuta Hayashibe Date: Thu, 27 Oct 2022 22:00:28 +0900 Subject: Truncate too long filename --- modules/images.py | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/images.py b/modules/images.py index 7870b5b7..42363ed3 100644 --- a/modules/images.py +++ b/modules/images.py @@ -416,6 +416,14 @@ def get_next_sequence_number(path, basename): return result + 1 +def truncate_fullpath(full_path, encoding='utf-8'): + dir_name, full_name = os.path.split(full_path) + file_name, file_ext = os.path.splitext(full_name) + max_length = os.statvfs(dir_name).f_namemax + file_name_truncated = file_name.encode(encoding)[:max_length - len(file_ext)].decode(encoding, 'ignore') + return os.path.join(dir_name , file_name_truncated + file_ext) + + def save_image(image, path, basename, seed=None, prompt=None, extension='png', info=None, short_filename=False, no_prompt=False, grid=False, pnginfo_section_name='parameters', p=None, existing_info=None, forced_filename=None, suffix="", save_to_dirs=None): """Save an image. @@ -456,7 +464,7 @@ def save_image(image, path, basename, seed=None, prompt=None, extension='png', i if save_to_dirs: dirname = namegen.apply(opts.directories_filename_pattern or "[prompt_words]").lstrip(' ').rstrip('\\ /') - path = os.path.join(path, dirname) + path = truncate_fullpath(os.path.join(path, dirname)) os.makedirs(path, exist_ok=True) @@ -480,13 +488,13 @@ def save_image(image, path, basename, seed=None, prompt=None, extension='png', i fullfn = None for i in range(500): fn = f"{basecount + i:05}" if basename == '' else f"{basename}-{basecount + i:04}" - fullfn = os.path.join(path, f"{fn}{file_decoration}.{extension}") + fullfn = truncate_fullpath(os.path.join(path, f"{fn}{file_decoration}.{extension}")) if not os.path.exists(fullfn): break else: - fullfn = os.path.join(path, f"{file_decoration}.{extension}") + fullfn = truncate_fullpath(os.path.join(path, f"{file_decoration}.{extension}")) else: - fullfn = os.path.join(path, f"{forced_filename}.{extension}") + fullfn = truncate_fullpath(os.path.join(path, f"{forced_filename}.{extension}")) pnginfo = existing_info or {} if info is not None: -- cgit v1.2.3 From 2a25729623717cc499e873752d9f4ebebd1e1078 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Fri, 28 Oct 2022 09:44:56 +0700 Subject: Gradient clipping in train tab --- modules/hypernetworks/hypernetwork.py | 10 +++++++++- modules/ui.py | 7 +++++++ 2 files changed, 16 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 8113b35b..c5d60654 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -327,7 +327,7 @@ def report_statistics(loss_info:dict): -def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): +def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, clip_grad_mode, clip_grad_value, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): # images allows training previews to have infotext. Importing it at the top causes a circular import problem. from modules import images @@ -384,6 +384,9 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log if ititial_step > steps: return hypernetwork, filename + clip_grad_mode_value = clip_grad_mode == "value" + clip_grad_mode_norm = clip_grad_mode == "norm" + scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) # if optimizer == "AdamW": or else Adam / AdamW / SGD, etc... optimizer = torch.optim.AdamW(weights, lr=scheduler.learn_rate) @@ -426,6 +429,11 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log steps_without_grad = 0 assert steps_without_grad < 10, 'no gradient found for the trained weight after backward() for 10 steps in a row; this is a bug; training cannot continue' + if clip_grad_mode_value: + torch.nn.utils.clip_grad_value_(weights, clip_value=clip_grad_value) + elif clip_grad_mode_norm: + torch.nn.utils.clip_grad_norm_(weights, max_norm=clip_grad_value) + optimizer.step() if torch.isnan(losses[hypernetwork.step % losses.shape[0]]): diff --git a/modules/ui.py b/modules/ui.py index 0a63e357..97de7da2 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1313,6 +1313,9 @@ def create_ui(wrap_gradio_gpu_call): training_width = gr.Slider(minimum=64, maximum=2048, step=64, label="Width", value=512) training_height = gr.Slider(minimum=64, maximum=2048, step=64, label="Height", value=512) steps = gr.Number(label='Max steps', value=100000, precision=0) + with gr.Row(): + clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) + clip_grad_value = gr.Number(value=1.0, show_label=False) create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0) save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0) save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True) @@ -1406,6 +1409,8 @@ def create_ui(wrap_gradio_gpu_call): training_width, training_height, steps, + clip_grad_mode, + clip_grad_value, create_image_every, save_embedding_every, template_file, @@ -1431,6 +1436,8 @@ def create_ui(wrap_gradio_gpu_call): training_width, training_height, steps, + clip_grad_mode, + clip_grad_value, create_image_every, save_embedding_every, template_file, -- cgit v1.2.3 From a133042c669f666763f5da0f4440abdc839db653 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Fri, 28 Oct 2022 10:01:46 +0700 Subject: Forgot to remove this from train_embedding --- modules/ui.py | 2 -- 1 file changed, 2 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 97de7da2..ba5e92a7 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1409,8 +1409,6 @@ def create_ui(wrap_gradio_gpu_call): training_width, training_height, steps, - clip_grad_mode, - clip_grad_value, create_image_every, save_embedding_every, template_file, -- cgit v1.2.3 From 1618df41bad092e068c61bf510b1e20856821ad5 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Fri, 28 Oct 2022 10:31:27 +0700 Subject: Gradient clipping for textual embedding --- modules/textual_inversion/textual_inversion.py | 11 ++++++++++- modules/ui.py | 2 ++ 2 files changed, 12 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index ff002d3e..7bad73a6 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -206,7 +206,7 @@ def write_loss(log_directory, filename, step, epoch_len, values): }) -def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): +def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, clip_grad_mode, clip_grad_value, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): assert embedding_name, 'embedding not selected' shared.state.textinfo = "Initializing textual inversion training..." @@ -256,6 +256,9 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc if ititial_step > steps: return embedding, filename + clip_grad_mode_value = clip_grad_mode == "value" + clip_grad_mode_norm = clip_grad_mode == "norm" + scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) optimizer = torch.optim.AdamW([embedding.vec], lr=scheduler.learn_rate) @@ -280,6 +283,12 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc optimizer.zero_grad() loss.backward() + + if clip_grad_mode_value: + torch.nn.utils.clip_grad_value_(embedding.vec, clip_value=clip_grad_value) + elif clip_grad_mode_norm: + torch.nn.utils.clip_grad_norm_(embedding.vec, max_norm=clip_grad_value) + optimizer.step() diff --git a/modules/ui.py b/modules/ui.py index ba5e92a7..97de7da2 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1409,6 +1409,8 @@ def create_ui(wrap_gradio_gpu_call): training_width, training_height, steps, + clip_grad_mode, + clip_grad_value, create_image_every, save_embedding_every, template_file, -- cgit v1.2.3 From 16451ca573220e49f2eaaab97580b6b91287c8c4 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Fri, 28 Oct 2022 17:16:23 +0700 Subject: Learning rate sched syntax support for grad clipping --- modules/hypernetworks/hypernetwork.py | 13 ++++++++++--- modules/textual_inversion/learn_schedule.py | 11 ++++++++--- modules/textual_inversion/textual_inversion.py | 12 +++++++++--- modules/ui.py | 7 +++---- 4 files changed, 30 insertions(+), 13 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index c5d60654..86532063 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -383,11 +383,15 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log ititial_step = hypernetwork.step or 0 if ititial_step > steps: return hypernetwork, filename - + clip_grad_mode_value = clip_grad_mode == "value" clip_grad_mode_norm = clip_grad_mode == "norm" + clip_grad_enabled = clip_grad_mode_value or clip_grad_mode_norm + if clip_grad_enabled: + clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, ititial_step, verbose=False) scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) + # if optimizer == "AdamW": or else Adam / AdamW / SGD, etc... optimizer = torch.optim.AdamW(weights, lr=scheduler.learn_rate) @@ -407,6 +411,9 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log if shared.state.interrupted: break + if clip_grad_enabled: + clip_grad_sched.step(hypernetwork.step) + with torch.autocast("cuda"): c = stack_conds([entry.cond for entry in entries]).to(devices.device) # c = torch.vstack([entry.cond for entry in entries]).to(devices.device) @@ -430,9 +437,9 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log assert steps_without_grad < 10, 'no gradient found for the trained weight after backward() for 10 steps in a row; this is a bug; training cannot continue' if clip_grad_mode_value: - torch.nn.utils.clip_grad_value_(weights, clip_value=clip_grad_value) + torch.nn.utils.clip_grad_value_(weights, clip_value=clip_grad_sched.learn_rate) elif clip_grad_mode_norm: - torch.nn.utils.clip_grad_norm_(weights, max_norm=clip_grad_value) + torch.nn.utils.clip_grad_norm_(weights, max_norm=clip_grad_sched.learn_rate) optimizer.step() diff --git a/modules/textual_inversion/learn_schedule.py b/modules/textual_inversion/learn_schedule.py index 2062726a..ffec3e1b 100644 --- a/modules/textual_inversion/learn_schedule.py +++ b/modules/textual_inversion/learn_schedule.py @@ -51,14 +51,19 @@ class LearnRateScheduler: self.finished = False - def apply(self, optimizer, step_number): + def step(self, step_number): if step_number <= self.end_step: - return + return False try: (self.learn_rate, self.end_step) = next(self.schedules) - except Exception: + except StopIteration: self.finished = True + return False + return True + + def apply(self, optimizer, step_number): + if not self.step(step_number): return if self.verbose: diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 7bad73a6..6b00c6a1 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -255,9 +255,12 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc ititial_step = embedding.step or 0 if ititial_step > steps: return embedding, filename - + clip_grad_mode_value = clip_grad_mode == "value" clip_grad_mode_norm = clip_grad_mode == "norm" + clip_grad_enabled = clip_grad_mode_value or clip_grad_mode_norm + if clip_grad_enabled: + clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, ititial_step, verbose=False) scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) optimizer = torch.optim.AdamW([embedding.vec], lr=scheduler.learn_rate) @@ -273,6 +276,9 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc if shared.state.interrupted: break + if clip_grad_enabled: + clip_grad_sched.step(embedding.step) + with torch.autocast("cuda"): c = cond_model([entry.cond_text for entry in entries]) x = torch.stack([entry.latent for entry in entries]).to(devices.device) @@ -285,9 +291,9 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc loss.backward() if clip_grad_mode_value: - torch.nn.utils.clip_grad_value_(embedding.vec, clip_value=clip_grad_value) + torch.nn.utils.clip_grad_value_(embedding.vec, clip_value=clip_grad_sched.learn_rate) elif clip_grad_mode_norm: - torch.nn.utils.clip_grad_norm_(embedding.vec, max_norm=clip_grad_value) + torch.nn.utils.clip_grad_norm_(embedding.vec, max_norm=clip_grad_sched.learn_rate) optimizer.step() diff --git a/modules/ui.py b/modules/ui.py index 97de7da2..47d16429 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1305,7 +1305,9 @@ def create_ui(wrap_gradio_gpu_call): with gr.Row(): embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005") hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001") - + with gr.Row(): + clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) + clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="1.0", show_label=False) batch_size = gr.Number(label='Batch size', value=1, precision=0) dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images") log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion") @@ -1313,9 +1315,6 @@ def create_ui(wrap_gradio_gpu_call): training_width = gr.Slider(minimum=64, maximum=2048, step=64, label="Width", value=512) training_height = gr.Slider(minimum=64, maximum=2048, step=64, label="Height", value=512) steps = gr.Number(label='Max steps', value=100000, precision=0) - with gr.Row(): - clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) - clip_grad_value = gr.Number(value=1.0, show_label=False) create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0) save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0) save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True) -- cgit v1.2.3 From 840307f23738c38f7ac3ad636e53ccec66e71f8b Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Mon, 31 Oct 2022 13:49:24 +0700 Subject: Change default clip grad value to 0.1 It still defaults to disabled. Ref for value: https://github.com/danielalcalde/stable-diffusion-webui/commit/732b15820a9bde9f47e075a6209c3d47d47acb08 --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 98f9565f..364953aa 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1256,7 +1256,7 @@ def create_ui(wrap_gradio_gpu_call): hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001") with gr.Row(): clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) - clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="1.0", show_label=False) + clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) batch_size = gr.Number(label='Batch size', value=1, precision=0) dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images") log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion") -- cgit v1.2.3 From 4123be632a98f70cda06e14c2f556f7ad38cd436 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Mon, 31 Oct 2022 13:53:22 +0700 Subject: Fix merge conflicts --- modules/hypernetworks/hypernetwork.py | 17 ++++++----------- 1 file changed, 6 insertions(+), 11 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 65a584bb..207808ee 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -373,6 +373,12 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) + clip_grad_mode_value = clip_grad_mode == "value" + clip_grad_mode_norm = clip_grad_mode == "norm" + clip_grad_enabled = clip_grad_mode_value or clip_grad_mode_norm + if clip_grad_enabled: + clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, ititial_step, verbose=False) + # dataset loading may take a while, so input validations and early returns should be done before this shared.state.textinfo = f"Preparing dataset from {html.escape(data_root)}..." with torch.autocast("cuda"): @@ -389,21 +395,10 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log previous_mean_loss = 0 print("Mean loss of {} elements".format(size)) - last_saved_file = "" - last_saved_image = "" - forced_filename = "" - ititial_step = hypernetwork.step or 0 if ititial_step > steps: return hypernetwork, filename - clip_grad_mode_value = clip_grad_mode == "value" - clip_grad_mode_norm = clip_grad_mode == "norm" - clip_grad_enabled = clip_grad_mode_value or clip_grad_mode_norm - if clip_grad_enabled: - clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, ititial_step, verbose=False) - - scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) weights = hypernetwork.weights() for weight in weights: -- cgit v1.2.3 From d5ea878b2aa117588d85287cbd8983aa52177df5 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Mon, 31 Oct 2022 13:54:40 +0700 Subject: Fix merge conflicts --- modules/hypernetworks/hypernetwork.py | 5 ----- 1 file changed, 5 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 207808ee..2df38c70 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -395,11 +395,6 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log previous_mean_loss = 0 print("Mean loss of {} elements".format(size)) - ititial_step = hypernetwork.step or 0 - if ititial_step > steps: - return hypernetwork, filename - - weights = hypernetwork.weights() for weight in weights: weight.requires_grad = True -- cgit v1.2.3 From cffc240a7327ae60671ff533469fc4ed4bf605de Mon Sep 17 00:00:00 2001 From: Nerogar Date: Sun, 23 Oct 2022 14:05:25 +0200 Subject: fixed textual inversion training with inpainting models --- modules/textual_inversion/textual_inversion.py | 27 +++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 0aeb0459..2630c7c9 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -224,6 +224,26 @@ def validate_train_inputs(model_name, learn_rate, batch_size, data_root, templat if save_model_every or create_image_every: assert log_directory, "Log directory is empty" +def create_dummy_mask(x, width=None, height=None): + if shared.sd_model.model.conditioning_key in {'hybrid', 'concat'}: + + # The "masked-image" in this case will just be all zeros since the entire image is masked. + image_conditioning = torch.zeros(x.shape[0], 3, height, width, device=x.device) + image_conditioning = shared.sd_model.get_first_stage_encoding(shared.sd_model.encode_first_stage(image_conditioning)) + + # Add the fake full 1s mask to the first dimension. + image_conditioning = torch.nn.functional.pad(image_conditioning, (0, 0, 0, 0, 1, 0), value=1.0) + image_conditioning = image_conditioning.to(x.dtype) + + else: + # Dummy zero conditioning if we're not using inpainting model. + # Still takes up a bit of memory, but no encoder call. + # Pretty sure we can just make this a 1x1 image since its not going to be used besides its batch size. + image_conditioning = torch.zeros(x.shape[0], 5, 1, 1, dtype=x.dtype, device=x.device) + + return image_conditioning + + def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): save_embedding_every = save_embedding_every or 0 create_image_every = create_image_every or 0 @@ -286,6 +306,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc forced_filename = "" embedding_yet_to_be_embedded = False + img_c = None pbar = tqdm.tqdm(enumerate(ds), total=steps-ititial_step) for i, entries in pbar: embedding.step = i + ititial_step @@ -299,8 +320,12 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc with torch.autocast("cuda"): c = cond_model([entry.cond_text for entry in entries]) + if img_c is None: + img_c = create_dummy_mask(c, training_width, training_height) + x = torch.stack([entry.latent for entry in entries]).to(devices.device) - loss = shared.sd_model(x, c)[0] + cond = {"c_concat": [img_c], "c_crossattn": [c]} + loss = shared.sd_model(x, cond)[0] del x losses[embedding.step % losses.shape[0]] = loss.item() -- cgit v1.2.3 From bb832d7725187f8a8ab44faa6ee1b38cb5f600aa Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Sat, 5 Nov 2022 11:48:38 +0700 Subject: Simplify grad clip --- modules/hypernetworks/hypernetwork.py | 16 +++++++--------- modules/textual_inversion/textual_inversion.py | 16 +++++++--------- 2 files changed, 14 insertions(+), 18 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index f4c2668f..02b624e1 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -385,10 +385,10 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) - clip_grad_mode_value = clip_grad_mode == "value" - clip_grad_mode_norm = clip_grad_mode == "norm" - clip_grad_enabled = clip_grad_mode_value or clip_grad_mode_norm - if clip_grad_enabled: + clip_grad = torch.nn.utils.clip_grad_value_ if clip_grad_mode == "value" else \ + torch.nn.utils.clip_grad_norm_ if clip_grad_mode == "norm" else \ + None + if clip_grad: clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, ititial_step, verbose=False) # dataset loading may take a while, so input validations and early returns should be done before this @@ -433,7 +433,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log if shared.state.interrupted: break - if clip_grad_enabled: + if clip_grad: clip_grad_sched.step(hypernetwork.step) with torch.autocast("cuda"): @@ -458,10 +458,8 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, data_root, log steps_without_grad = 0 assert steps_without_grad < 10, 'no gradient found for the trained weight after backward() for 10 steps in a row; this is a bug; training cannot continue' - if clip_grad_mode_value: - torch.nn.utils.clip_grad_value_(weights, clip_value=clip_grad_sched.learn_rate) - elif clip_grad_mode_norm: - torch.nn.utils.clip_grad_norm_(weights, max_norm=clip_grad_sched.learn_rate) + if clip_grad: + clip_grad(weights, clip_grad_sched.learn_rate) optimizer.step() diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index c567ec3f..687d97bb 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -269,10 +269,10 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc scheduler = LearnRateScheduler(learn_rate, steps, ititial_step) - clip_grad_mode_value = clip_grad_mode == "value" - clip_grad_mode_norm = clip_grad_mode == "norm" - clip_grad_enabled = clip_grad_mode_value or clip_grad_mode_norm - if clip_grad_enabled: + clip_grad = torch.nn.utils.clip_grad_value_ if clip_grad_mode == "value" else \ + torch.nn.utils.clip_grad_norm_ if clip_grad_mode == "norm" else \ + None + if clip_grad: clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, ititial_step, verbose=False) # dataset loading may take a while, so input validations and early returns should be done before this shared.state.textinfo = f"Preparing dataset from {html.escape(data_root)}..." @@ -302,7 +302,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc if shared.state.interrupted: break - if clip_grad_enabled: + if clip_grad: clip_grad_sched.step(embedding.step) with torch.autocast("cuda"): @@ -316,10 +316,8 @@ def train_embedding(embedding_name, learn_rate, batch_size, data_root, log_direc optimizer.zero_grad() loss.backward() - if clip_grad_mode_value: - torch.nn.utils.clip_grad_value_(embedding.vec, clip_value=clip_grad_sched.learn_rate) - elif clip_grad_mode_norm: - torch.nn.utils.clip_grad_norm_(embedding.vec, max_norm=clip_grad_sched.learn_rate) + if clip_grad: + clip_grad(embedding.vec, clip_grad_sched.learn_rate) optimizer.step() -- cgit v1.2.3 From 75c4511e6b81ae8fb0dbd932043e8eb35cd09f72 Mon Sep 17 00:00:00 2001 From: zhaohu xing <920232796@qq.com> Date: Tue, 29 Nov 2022 10:28:41 +0800 Subject: add AltDiffusion to webui Signed-off-by: zhaohu xing <920232796@qq.com> --- modules/devices.py | 4 ++-- modules/sd_hijack.py | 23 +++++++++++++++++------ modules/shared.py | 6 +++++- 3 files changed, 24 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/devices.py b/modules/devices.py index 67165bf6..f30b6ebc 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -36,8 +36,8 @@ def get_optimal_device(): else: return torch.device("cuda") - if has_mps(): - return torch.device("mps") + # if has_mps(): + # return torch.device("mps") return cpu diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index eaedac13..26280fe4 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -70,14 +70,19 @@ class StableDiffusionModelHijack: embedding_db = modules.textual_inversion.textual_inversion.EmbeddingDatabase(cmd_opts.embeddings_dir) def hijack(self, m): - model_embeddings = m.cond_stage_model.transformer.text_model.embeddings + + if shared.text_model_name == "XLMR-Large": + model_embeddings = m.cond_stage_model.roberta.embeddings + model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.word_embeddings, self) + else : + model_embeddings = m.cond_stage_model.transformer.text_model.embeddings + model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.token_embeddings, self) - model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.token_embedding, self) m.cond_stage_model = FrozenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) self.clip = m.cond_stage_model - apply_optimizations() + # apply_optimizations() def flatten(el): flattened = [flatten(children) for children in el.children()] @@ -125,8 +130,11 @@ class FrozenCLIPEmbedderWithCustomWords(torch.nn.Module): self.tokenizer = wrapped.tokenizer self.token_mults = {} - self.comma_token = [v for k, v in self.tokenizer.get_vocab().items() if k == ','][0] - + try: + self.comma_token = [v for k, v in self.tokenizer.get_vocab().items() if k == ','][0] + except: + self.comma_token = None + tokens_with_parens = [(k, v) for k, v in self.tokenizer.get_vocab().items() if '(' in k or ')' in k or '[' in k or ']' in k] for text, ident in tokens_with_parens: mult = 1.0 @@ -298,6 +306,9 @@ class FrozenCLIPEmbedderWithCustomWords(torch.nn.Module): return batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count def forward(self, text): + if shared.text_model_name == "XLMR-Large": + return self.wrapped.encode(text) + use_old = opts.use_old_emphasis_implementation if use_old: batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count = self.process_text_old(text) @@ -359,7 +370,7 @@ class FrozenCLIPEmbedderWithCustomWords(torch.nn.Module): z = self.wrapped.transformer.text_model.final_layer_norm(z) else: z = outputs.last_hidden_state - + # restoring original mean is likely not correct, but it seems to work well to prevent artifacts that happen otherwise batch_multipliers_of_same_length = [x + [1.0] * (75 - len(x)) for x in batch_multipliers] batch_multipliers = torch.asarray(batch_multipliers_of_same_length).to(device) diff --git a/modules/shared.py b/modules/shared.py index c93ae2a3..9941d2f4 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -21,7 +21,7 @@ from modules.paths import models_path, script_path, sd_path sd_model_file = os.path.join(script_path, 'model.ckpt') default_sd_model_file = sd_model_file parser = argparse.ArgumentParser() -parser.add_argument("--config", type=str, default=os.path.join(sd_path, "configs/stable-diffusion/v1-inference.yaml"), help="path to config which constructs model",) +parser.add_argument("--config", type=str, default="configs/altdiffusion/ad-inference.yaml", help="path to config which constructs model",) parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; if specified, this checkpoint will be added to the list of checkpoints and loaded",) parser.add_argument("--ckpt-dir", type=str, default=None, help="Path to directory with stable diffusion checkpoints") parser.add_argument("--gfpgan-dir", type=str, help="GFPGAN directory", default=('./src/gfpgan' if os.path.exists('./src/gfpgan') else './GFPGAN')) @@ -106,6 +106,10 @@ restricted_opts = { "outdir_txt2img_grids", "outdir_save", } +from omegaconf import OmegaConf +config = OmegaConf.load(f"{cmd_opts.config}") +# XLMR-Large +text_model_name = config.model.params.cond_stage_config.params.name cmd_opts.disable_extension_access = (cmd_opts.share or cmd_opts.listen or cmd_opts.server_name) and not cmd_opts.enable_insecure_extension_access -- cgit v1.2.3 From 52cc83d36b7663a77b79fd2258d2ca871af73e55 Mon Sep 17 00:00:00 2001 From: zhaohu xing <920232796@qq.com> Date: Wed, 30 Nov 2022 14:56:12 +0800 Subject: fix bugs Signed-off-by: zhaohu xing <920232796@qq.com> --- modules/sd_hijack.py | 15 ++--- modules/sd_hijack_clip.py | 10 +++- modules/xlmr.py | 137 ++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 153 insertions(+), 9 deletions(-) create mode 100644 modules/xlmr.py (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index 3ec3f98a..edb8b420 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -28,7 +28,7 @@ diffusionmodules_model_AttnBlock_forward = ldm.modules.diffusionmodules.model.At # new memory efficient cross attention blocks do not support hypernets and we already # have memory efficient cross attention anyway, so this disables SD2.0's memory efficient cross attention ldm.modules.attention.MemoryEfficientCrossAttention = ldm.modules.attention.CrossAttention -ldm.modules.attention.BasicTransformerBlock.ATTENTION_MODES["softmax-xformers"] = ldm.modules.attention.CrossAttention +# ldm.modules.attention.BasicTransformerBlock.ATTENTION_MODES["softmax-xformers"] = ldm.modules.attention.CrossAttention # silence new console spam from SD2 ldm.modules.attention.print = lambda *args: None @@ -82,7 +82,12 @@ class StableDiffusionModelHijack: def hijack(self, m): - if type(m.cond_stage_model) == ldm.modules.encoders.modules.FrozenCLIPEmbedder: + if shared.text_model_name == "XLMR-Large": + model_embeddings = m.cond_stage_model.roberta.embeddings + model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.word_embeddings, self) + m.cond_stage_model = sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) + + elif type(m.cond_stage_model) == ldm.modules.encoders.modules.FrozenCLIPEmbedder: model_embeddings = m.cond_stage_model.transformer.text_model.embeddings model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.token_embedding, self) m.cond_stage_model = sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) @@ -91,11 +96,7 @@ class StableDiffusionModelHijack: m.cond_stage_model.model.token_embedding = EmbeddingsWithFixes(m.cond_stage_model.model.token_embedding, self) m.cond_stage_model = sd_hijack_open_clip.FrozenOpenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) apply_optimizations() - elif shared.text_model_name == "XLMR-Large": - model_embeddings = m.cond_stage_model.roberta.embeddings - model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.word_embeddings, self) - m.cond_stage_model = sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) - + self.clip = m.cond_stage_model fix_checkpoint() diff --git a/modules/sd_hijack_clip.py b/modules/sd_hijack_clip.py index b451d1cf..9ea6e1ce 100644 --- a/modules/sd_hijack_clip.py +++ b/modules/sd_hijack_clip.py @@ -4,7 +4,7 @@ import torch from modules import prompt_parser, devices from modules.shared import opts - +import modules.shared as shared def get_target_prompt_token_count(token_count): return math.ceil(max(token_count, 1) / 75) * 75 @@ -177,6 +177,9 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): return batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count def forward(self, text): + if shared.text_model_name == "XLMR-Large": + return self.wrapped.encode(text) + use_old = opts.use_old_emphasis_implementation if use_old: batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count = self.process_text_old(text) @@ -254,7 +257,10 @@ class FrozenCLIPEmbedderWithCustomWords(FrozenCLIPEmbedderWithCustomWordsBase): def __init__(self, wrapped, hijack): super().__init__(wrapped, hijack) self.tokenizer = wrapped.tokenizer - self.comma_token = [v for k, v in self.tokenizer.get_vocab().items() if k == ','][0] + if shared.text_model_name == "XLMR-Large": + self.comma_token = None + else : + self.comma_token = [v for k, v in self.tokenizer.get_vocab().items() if k == ','][0] self.token_mults = {} tokens_with_parens = [(k, v) for k, v in self.tokenizer.get_vocab().items() if '(' in k or ')' in k or '[' in k or ']' in k] diff --git a/modules/xlmr.py b/modules/xlmr.py new file mode 100644 index 00000000..beab3fdf --- /dev/null +++ b/modules/xlmr.py @@ -0,0 +1,137 @@ +from transformers import BertPreTrainedModel,BertModel,BertConfig +import torch.nn as nn +import torch +from transformers.models.xlm_roberta.configuration_xlm_roberta import XLMRobertaConfig +from transformers import XLMRobertaModel,XLMRobertaTokenizer +from typing import Optional + +class BertSeriesConfig(BertConfig): + def __init__(self, vocab_size=30522, hidden_size=768, num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072, hidden_act="gelu", hidden_dropout_prob=0.1, attention_probs_dropout_prob=0.1, max_position_embeddings=512, type_vocab_size=2, initializer_range=0.02, layer_norm_eps=1e-12, pad_token_id=0, position_embedding_type="absolute", use_cache=True, classifier_dropout=None,project_dim=512, pooler_fn="average",learn_encoder=False,model_type='bert',**kwargs): + + super().__init__(vocab_size, hidden_size, num_hidden_layers, num_attention_heads, intermediate_size, hidden_act, hidden_dropout_prob, attention_probs_dropout_prob, max_position_embeddings, type_vocab_size, initializer_range, layer_norm_eps, pad_token_id, position_embedding_type, use_cache, classifier_dropout, **kwargs) + self.project_dim = project_dim + self.pooler_fn = pooler_fn + self.learn_encoder = learn_encoder + +class RobertaSeriesConfig(XLMRobertaConfig): + def __init__(self, pad_token_id=1, bos_token_id=0, eos_token_id=2,project_dim=512,pooler_fn='cls',learn_encoder=False, **kwargs): + super().__init__(pad_token_id=pad_token_id, bos_token_id=bos_token_id, eos_token_id=eos_token_id, **kwargs) + self.project_dim = project_dim + self.pooler_fn = pooler_fn + self.learn_encoder = learn_encoder + + +class BertSeriesModelWithTransformation(BertPreTrainedModel): + + _keys_to_ignore_on_load_unexpected = [r"pooler"] + _keys_to_ignore_on_load_missing = [r"position_ids", r"predictions.decoder.bias"] + config_class = BertSeriesConfig + + def __init__(self, config=None, **kargs): + # modify initialization for autoloading + if config is None: + config = XLMRobertaConfig() + config.attention_probs_dropout_prob= 0.1 + config.bos_token_id=0 + config.eos_token_id=2 + config.hidden_act='gelu' + config.hidden_dropout_prob=0.1 + config.hidden_size=1024 + config.initializer_range=0.02 + config.intermediate_size=4096 + config.layer_norm_eps=1e-05 + config.max_position_embeddings=514 + + config.num_attention_heads=16 + config.num_hidden_layers=24 + config.output_past=True + config.pad_token_id=1 + config.position_embedding_type= "absolute" + + config.type_vocab_size= 1 + config.use_cache=True + config.vocab_size= 250002 + config.project_dim = 768 + config.learn_encoder = False + super().__init__(config) + self.roberta = XLMRobertaModel(config) + self.transformation = nn.Linear(config.hidden_size,config.project_dim) + self.pre_LN=nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps) + self.tokenizer = XLMRobertaTokenizer.from_pretrained('xlm-roberta-large') + self.pooler = lambda x: x[:,0] + self.post_init() + + def encode(self,c): + device = next(self.parameters()).device + text = self.tokenizer(c, + truncation=True, + max_length=77, + return_length=False, + return_overflowing_tokens=False, + padding="max_length", + return_tensors="pt") + text["input_ids"] = torch.tensor(text["input_ids"]).to(device) + text["attention_mask"] = torch.tensor( + text['attention_mask']).to(device) + features = self(**text) + return features['projection_state'] + + def forward( + self, + input_ids: Optional[torch.Tensor] = None, + attention_mask: Optional[torch.Tensor] = None, + token_type_ids: Optional[torch.Tensor] = None, + position_ids: Optional[torch.Tensor] = None, + head_mask: Optional[torch.Tensor] = None, + inputs_embeds: Optional[torch.Tensor] = None, + encoder_hidden_states: Optional[torch.Tensor] = None, + encoder_attention_mask: Optional[torch.Tensor] = None, + output_attentions: Optional[bool] = None, + return_dict: Optional[bool] = None, + output_hidden_states: Optional[bool] = None, + ) : + r""" + """ + + return_dict = return_dict if return_dict is not None else self.config.use_return_dict + + + outputs = self.roberta( + input_ids=input_ids, + attention_mask=attention_mask, + token_type_ids=token_type_ids, + position_ids=position_ids, + head_mask=head_mask, + inputs_embeds=inputs_embeds, + encoder_hidden_states=encoder_hidden_states, + encoder_attention_mask=encoder_attention_mask, + output_attentions=output_attentions, + output_hidden_states=True, + return_dict=return_dict, + ) + + # last module outputs + sequence_output = outputs[0] + + + # project every module + sequence_output_ln = self.pre_LN(sequence_output) + + # pooler + pooler_output = self.pooler(sequence_output_ln) + pooler_output = self.transformation(pooler_output) + projection_state = self.transformation(outputs.last_hidden_state) + + return { + 'pooler_output':pooler_output, + 'last_hidden_state':outputs.last_hidden_state, + 'hidden_states':outputs.hidden_states, + 'attentions':outputs.attentions, + 'projection_state':projection_state, + 'sequence_out': sequence_output + } + + +class RobertaSeriesModelWithTransformation(BertSeriesModelWithTransformation): + base_model_prefix = 'roberta' + config_class= RobertaSeriesConfig \ No newline at end of file -- cgit v1.2.3 From 9c86fb8cace6d8ac0843e0ddad0ba5ae7f3148c9 Mon Sep 17 00:00:00 2001 From: zhaohu xing <920232796@qq.com> Date: Fri, 2 Dec 2022 16:08:46 +0800 Subject: fix bug Signed-off-by: zhaohu xing <920232796@qq.com> --- modules/shared.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 1408dee3..ac7678c3 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -111,7 +111,11 @@ restricted_opts = { from omegaconf import OmegaConf config = OmegaConf.load(f"{cmd_opts.config}") # XLMR-Large -text_model_name = config.model.params.cond_stage_config.params.name +try: + text_model_name = config.model.params.cond_stage_config.params.name + +except : + text_model_name = "stable_diffusion" cmd_opts.disable_extension_access = (cmd_opts.share or cmd_opts.listen or cmd_opts.server_name) and not cmd_opts.enable_insecure_extension_access -- cgit v1.2.3 From 4929503258d80abbc4b5f40da034298fe3803906 Mon Sep 17 00:00:00 2001 From: zhaohu xing <920232796@qq.com> Date: Tue, 6 Dec 2022 09:03:55 +0800 Subject: fix bugs Signed-off-by: zhaohu xing <920232796@qq.com> --- modules/devices.py | 4 ++-- modules/sd_hijack.py | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/devices.py b/modules/devices.py index e69c1fe3..f00079c6 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -38,8 +38,8 @@ def get_optimal_device(): if torch.cuda.is_available(): return torch.device(get_cuda_device_string()) - # if has_mps(): - # return torch.device("mps") + if has_mps(): + return torch.device("mps") return cpu diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index edb8b420..cd65d356 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -28,7 +28,7 @@ diffusionmodules_model_AttnBlock_forward = ldm.modules.diffusionmodules.model.At # new memory efficient cross attention blocks do not support hypernets and we already # have memory efficient cross attention anyway, so this disables SD2.0's memory efficient cross attention ldm.modules.attention.MemoryEfficientCrossAttention = ldm.modules.attention.CrossAttention -# ldm.modules.attention.BasicTransformerBlock.ATTENTION_MODES["softmax-xformers"] = ldm.modules.attention.CrossAttention +ldm.modules.attention.BasicTransformerBlock.ATTENTION_MODES["softmax-xformers"] = ldm.modules.attention.CrossAttention # silence new console spam from SD2 ldm.modules.attention.print = lambda *args: None -- cgit v1.2.3 From 5dcc22606d05ebe5ae89c990bd83a3eb068fcb78 Mon Sep 17 00:00:00 2001 From: zhaohu xing <920232796@qq.com> Date: Tue, 6 Dec 2022 16:04:50 +0800 Subject: add hash and fix undo hijack bug Signed-off-by: zhaohu xing <920232796@qq.com> --- modules/sd_hijack.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index 9b5890e7..9fed1b6f 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -112,7 +112,11 @@ class StableDiffusionModelHijack: self.layers = flatten(m) def undo_hijack(self, m): - if type(m.cond_stage_model) == sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords: + + if shared.text_model_name == "XLMR-Large": + m.cond_stage_model = m.cond_stage_model.wrapped + + elif type(m.cond_stage_model) == sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords: m.cond_stage_model = m.cond_stage_model.wrapped model_embeddings = m.cond_stage_model.transformer.text_model.embeddings -- cgit v1.2.3 From 965fc5ac5a6ccdf38342e21c97183011a04e799e Mon Sep 17 00:00:00 2001 From: zhaohu xing <920232796@qq.com> Date: Tue, 6 Dec 2022 16:15:15 +0800 Subject: delete a file Signed-off-by: zhaohu xing <920232796@qq.com> --- modules/shared.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 522c56c1..8419b531 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -22,7 +22,7 @@ demo = None sd_model_file = os.path.join(script_path, 'model.ckpt') default_sd_model_file = sd_model_file parser = argparse.ArgumentParser() -parser.add_argument("--config", type=str, default="configs/altdiffusion/ad-inference.yaml", help="path to config which constructs model",) +parser.add_argument("--config", type=str, default=os.path.join(script_path, "v1-inference.yaml"), help="path to config which constructs model",) parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; if specified, this checkpoint will be added to the list of checkpoints and loaded",) parser.add_argument("--ckpt-dir", type=str, default=None, help="Path to directory with stable diffusion checkpoints") parser.add_argument("--gfpgan-dir", type=str, help="GFPGAN directory", default=('./src/gfpgan' if os.path.exists('./src/gfpgan') else './GFPGAN')) -- cgit v1.2.3 From f23a822f1c9cb3bd2e8772c75af429e06515eaef Mon Sep 17 00:00:00 2001 From: Philpax Date: Sat, 24 Dec 2022 20:45:16 +1100 Subject: feat(api): include job_timestamp in progress --- modules/shared.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 8ea3b441..f356dbf7 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -171,6 +171,7 @@ class State: "interrupted": self.skipped, "job": self.job, "job_count": self.job_count, + "job_timestamp": self.job_timestamp, "job_no": self.job_no, "sampling_step": self.sampling_step, "sampling_steps": self.sampling_steps, -- cgit v1.2.3 From fa931733f6acc94e058a1d3d4655846e33ae34be Mon Sep 17 00:00:00 2001 From: Philpax Date: Sun, 25 Dec 2022 20:17:49 +1100 Subject: fix(api): assign sd_model after settings change --- modules/api/api.py | 2 -- modules/processing.py | 6 ++++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 1ceba75d..0a1a1905 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -121,7 +121,6 @@ class Api: def text2imgapi(self, txt2imgreq: StableDiffusionTxt2ImgProcessingAPI): populate = txt2imgreq.copy(update={ # Override __init__ params - "sd_model": shared.sd_model, "sampler_name": validate_sampler_name(txt2imgreq.sampler_name or txt2imgreq.sampler_index), "do_not_save_samples": True, "do_not_save_grid": True @@ -153,7 +152,6 @@ class Api: mask = decode_base64_to_image(mask) populate = img2imgreq.copy(update={ # Override __init__ params - "sd_model": shared.sd_model, "sampler_name": validate_sampler_name(img2imgreq.sampler_name or img2imgreq.sampler_index), "do_not_save_samples": True, "do_not_save_grid": True, diff --git a/modules/processing.py b/modules/processing.py index 4a406084..0b270278 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -50,9 +50,9 @@ def apply_color_correction(correction, original_image): correction, channel_axis=2 ), cv2.COLOR_LAB2RGB).astype("uint8")) - + image = blendLayers(image, original_image, BlendType.LUMINOSITY) - + return image @@ -466,6 +466,8 @@ def process_images(p: StableDiffusionProcessing) -> Processed: if k == 'sd_model_checkpoint': sd_models.reload_model_weights() # make onchange call for changing SD model if k == 'sd_vae': sd_vae.reload_vae_weights() # make onchange call for changing VAE + # Assign sd_model here to ensure that it reflects the model after any changes + p.sd_model = shared.sd_model res = process_images_inner(p) finally: -- cgit v1.2.3 From 5be9387b230794a8c771120577cb213490c940c0 Mon Sep 17 00:00:00 2001 From: Philpax Date: Sun, 25 Dec 2022 21:45:44 +1100 Subject: fix(api): only begin/end state in lock --- modules/api/api.py | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 1ceba75d..59b81c93 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -130,14 +130,12 @@ class Api: if populate.sampler_name: populate.sampler_index = None # prevent a warning later on p = StableDiffusionProcessingTxt2Img(**vars(populate)) - # Override object param - - shared.state.begin() with self.queue_lock: + shared.state.begin() processed = process_images(p) + shared.state.end() - shared.state.end() b64images = list(map(encode_pil_to_base64, processed.images)) @@ -169,12 +167,10 @@ class Api: p.init_images = [decode_base64_to_image(x) for x in init_images] - shared.state.begin() - with self.queue_lock: + shared.state.begin() processed = process_images(p) - - shared.state.end() + shared.state.end() b64images = list(map(encode_pil_to_base64, processed.images)) -- cgit v1.2.3 From 893933e05ad267778111b4fad6d1ecb80937afdf Mon Sep 17 00:00:00 2001 From: hitomi Date: Sun, 25 Dec 2022 20:49:25 +0800 Subject: Add memory cache for VAE weights --- modules/sd_vae.py | 31 +++++++++++++++++++++++++------ modules/shared.py | 1 + 2 files changed, 26 insertions(+), 6 deletions(-) (limited to 'modules') diff --git a/modules/sd_vae.py b/modules/sd_vae.py index 3856418e..ac71d62d 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -1,5 +1,6 @@ import torch import os +import collections from collections import namedtuple from modules import shared, devices, script_callbacks from modules.paths import models_path @@ -30,6 +31,7 @@ base_vae = None loaded_vae_file = None checkpoint_info = None +checkpoints_loaded = collections.OrderedDict() def get_base_vae(model): if base_vae is not None and checkpoint_info == model.sd_checkpoint_info and model: @@ -149,13 +151,30 @@ def load_vae(model, vae_file=None): global first_load, vae_dict, vae_list, loaded_vae_file # save_settings = False + cache_enabled = shared.opts.sd_vae_checkpoint_cache > 0 + if vae_file: - assert os.path.isfile(vae_file), f"VAE file doesn't exist: {vae_file}" - print(f"Loading VAE weights from: {vae_file}") - store_base_vae(model) - vae_ckpt = torch.load(vae_file, map_location=shared.weight_load_location) - vae_dict_1 = {k: v for k, v in vae_ckpt["state_dict"].items() if k[0:4] != "loss" and k not in vae_ignore_keys} - _load_vae_dict(model, vae_dict_1) + if cache_enabled and vae_file in checkpoints_loaded: + # use vae checkpoint cache + print(f"Loading VAE weights [{get_filename(vae_file)}] from cache") + store_base_vae(model) + _load_vae_dict(model, checkpoints_loaded[vae_file]) + else: + assert os.path.isfile(vae_file), f"VAE file doesn't exist: {vae_file}" + print(f"Loading VAE weights from: {vae_file}") + store_base_vae(model) + vae_ckpt = torch.load(vae_file, map_location=shared.weight_load_location) + vae_dict_1 = {k: v for k, v in vae_ckpt["state_dict"].items() if k[0:4] != "loss" and k not in vae_ignore_keys} + _load_vae_dict(model, vae_dict_1) + + if cache_enabled: + # cache newly loaded vae + checkpoints_loaded[vae_file] = vae_dict_1.copy() + + # clean up cache if limit is reached + if cache_enabled: + while len(checkpoints_loaded) > shared.opts.sd_vae_checkpoint_cache + 1: # we need to count the current model + checkpoints_loaded.popitem(last=False) # LRU # If vae used is not in dict, update it # It will be removed on refresh though diff --git a/modules/shared.py b/modules/shared.py index d4ddeea0..671d30e1 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -356,6 +356,7 @@ options_templates.update(options_section(('training', "Training"), { options_templates.update(options_section(('sd', "Stable Diffusion"), { "sd_model_checkpoint": OptionInfo(None, "Stable Diffusion checkpoint", gr.Dropdown, lambda: {"choices": list_checkpoint_tiles()}, refresh=refresh_checkpoints), "sd_checkpoint_cache": OptionInfo(0, "Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), + "sd_vae_checkpoint_cache": OptionInfo(0, "VAE Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), "sd_vae": OptionInfo("auto", "SD VAE", gr.Dropdown, lambda: {"choices": sd_vae.vae_list}, refresh=sd_vae.refresh_vae_list), "sd_vae_as_default": OptionInfo(False, "Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them"), "sd_hypernetwork": OptionInfo("None", "Hypernetwork", gr.Dropdown, lambda: {"choices": ["None"] + [x for x in hypernetworks.keys()]}, refresh=reload_hypernetworks), -- cgit v1.2.3 From 4af3ca5393151d61363c30eef4965e694eeac15e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 26 Dec 2022 10:11:28 +0300 Subject: make it so that blank ENSD does not break image generation --- modules/processing.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 4a406084..0a9a8f95 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -338,13 +338,14 @@ def slerp(val, low, high): def create_random_tensors(shape, seeds, subseeds=None, subseed_strength=0.0, seed_resize_from_h=0, seed_resize_from_w=0, p=None): + eta_noise_seed_delta = opts.eta_noise_seed_delta or 0 xs = [] # if we have multiple seeds, this means we are working with batch size>1; this then # enables the generation of additional tensors with noise that the sampler will use during its processing. # Using those pre-generated tensors instead of simple torch.randn allows a batch with seeds [100, 101] to # produce the same images as with two batches [100], [101]. - if p is not None and p.sampler is not None and (len(seeds) > 1 and opts.enable_batch_seeds or opts.eta_noise_seed_delta > 0): + if p is not None and p.sampler is not None and (len(seeds) > 1 and opts.enable_batch_seeds or eta_noise_seed_delta > 0): sampler_noises = [[] for _ in range(p.sampler.number_of_needed_noises(p))] else: sampler_noises = None @@ -384,8 +385,8 @@ def create_random_tensors(shape, seeds, subseeds=None, subseed_strength=0.0, see if sampler_noises is not None: cnt = p.sampler.number_of_needed_noises(p) - if opts.eta_noise_seed_delta > 0: - torch.manual_seed(seed + opts.eta_noise_seed_delta) + if eta_noise_seed_delta > 0: + torch.manual_seed(seed + eta_noise_seed_delta) for j in range(cnt): sampler_noises[j].append(devices.randn_without_seed(tuple(noise_shape))) -- cgit v1.2.3 From ae955b0146a52ea2474c79655ede0d361829ef63 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Mon, 26 Dec 2022 09:53:26 -0500 Subject: fix rgba to rgb when using jpeg output --- modules/images.py | 3 +++ 1 file changed, 3 insertions(+) (limited to 'modules') diff --git a/modules/images.py b/modules/images.py index 31d4528d..962a955d 100644 --- a/modules/images.py +++ b/modules/images.py @@ -525,6 +525,9 @@ def save_image(image, path, basename, seed=None, prompt=None, extension='png', i image_to_save.save(temp_file_path, format=image_format, quality=opts.jpeg_quality, pnginfo=pnginfo_data) elif extension.lower() in (".jpg", ".jpeg", ".webp"): + if image_to_save.mode == 'RGBA': + image_to_save = image_to_save.convert("RGB") + image_to_save.save(temp_file_path, format=image_format, quality=opts.jpeg_quality) if opts.enable_pnginfo and info is not None: -- cgit v1.2.3 From 4df5009acb6832daef1ff5949404b5aadc8f8fa4 Mon Sep 17 00:00:00 2001 From: hentailord85ez <112723046+hentailord85ez@users.noreply.github.com> Date: Mon, 26 Dec 2022 20:49:13 +0000 Subject: Update sd_samplers.py --- modules/sd_samplers.py | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py index 177b5338..f4473832 100644 --- a/modules/sd_samplers.py +++ b/modules/sd_samplers.py @@ -462,6 +462,9 @@ class KDiffusionSampler: return extra_params_kwargs def get_sigmas(self, p, steps): + disc = opts.always_discard_next_to_last_sigma or (self.config is not None and self.config.options.get('discard_next_to_last_sigma', False)) + steps += 1 if disc else 0 + if p.sampler_noise_scheduler_override: sigmas = p.sampler_noise_scheduler_override(steps) elif self.config is not None and self.config.options.get('scheduler', None) == 'karras': @@ -469,7 +472,7 @@ class KDiffusionSampler: else: sigmas = self.model_wrap.get_sigmas(steps) - if self.config is not None and self.config.options.get('discard_next_to_last_sigma', False): + if disc: sigmas = torch.cat([sigmas[:-2], sigmas[-1:]]) return sigmas -- cgit v1.2.3 From 03f486a2399df0a2b24c7aeea72e64f106a87297 Mon Sep 17 00:00:00 2001 From: hentailord85ez <112723046+hentailord85ez@users.noreply.github.com> Date: Mon, 26 Dec 2022 20:49:33 +0000 Subject: Update shared.py --- modules/shared.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index d4ddeea0..5edb316c 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -418,6 +418,7 @@ options_templates.update(options_section(('sampler-params', "Sampler parameters" 's_tmin': OptionInfo(0.0, "sigma tmin", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), 's_noise': OptionInfo(1.0, "sigma noise", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), 'eta_noise_seed_delta': OptionInfo(0, "Eta noise seed delta", gr.Number, {"precision": 0}), + 'always_discard_next_to_last_sigma': OptionInfo(False, "Always discard next-to-last sigma"), })) options_templates.update(options_section((None, "Hidden options"), { -- cgit v1.2.3 From 5ba04f9ec050a66e918571f07e8863f157f05b44 Mon Sep 17 00:00:00 2001 From: Nicolas Patry Date: Wed, 21 Dec 2022 13:45:58 +0100 Subject: Attempting to solve slow loads for `safetensors`. Fixes #5893 --- modules/sd_models.py | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index ecdd91c5..cd938656 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -168,7 +168,10 @@ def get_state_dict_from_checkpoint(pl_sd): def read_state_dict(checkpoint_file, print_global_state=False, map_location=None): _, extension = os.path.splitext(checkpoint_file) if extension.lower() == ".safetensors": - pl_sd = safetensors.torch.load_file(checkpoint_file, device=map_location or shared.weight_load_location) + device = map_location or shared.weight_load_location + if device is None: + device = "cuda:0" if torch.cuda.is_available() else "cpu" + pl_sd = safetensors.torch.load_file(checkpoint_file, device=device) else: pl_sd = torch.load(checkpoint_file, map_location=map_location or shared.weight_load_location) -- cgit v1.2.3 From 5958bbd244703f7c248a91e86dea5d52acc85505 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Fri, 30 Dec 2022 19:36:36 -0500 Subject: add additional memory states --- modules/memmon.py | 3 +++ 1 file changed, 3 insertions(+) (limited to 'modules') diff --git a/modules/memmon.py b/modules/memmon.py index 9fb9b687..a7060f58 100644 --- a/modules/memmon.py +++ b/modules/memmon.py @@ -71,10 +71,13 @@ class MemUsageMonitor(threading.Thread): def read(self): if not self.disabled: free, total = torch.cuda.mem_get_info() + self.data["free"] = free self.data["total"] = total torch_stats = torch.cuda.memory_stats(self.device) + self.data["active"] = torch_stats["active.all.current"] self.data["active_peak"] = torch_stats["active_bytes.all.peak"] + self.data["reserved"] = torch_stats["reserved_bytes.all.current"] self.data["reserved_peak"] = torch_stats["reserved_bytes.all.peak"] self.data["system_peak"] = total - self.data["min_free"] -- cgit v1.2.3 From d3aa2a48e1e896b6ffafda5367200a4bbd46b0d7 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Fri, 30 Dec 2022 19:38:53 -0500 Subject: remove unnecessary console message --- modules/sd_hijack_inpainting.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_hijack_inpainting.py b/modules/sd_hijack_inpainting.py index bb5499b3..06b75772 100644 --- a/modules/sd_hijack_inpainting.py +++ b/modules/sd_hijack_inpainting.py @@ -178,7 +178,7 @@ def sample_plms(self, # sampling C, H, W = shape size = (batch_size, C, H, W) - print(f'Data shape for PLMS sampling is {size}') + # print(f'Data shape for PLMS sampling is {size}') # remove unnecessary message samples, intermediates = self.plms_sampling(conditioning, size, callback=callback, -- cgit v1.2.3 From 463048344fc036b262aa132584b65ee6e9fec6cf Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Fri, 30 Dec 2022 19:41:47 -0500 Subject: fix shared state dictionary --- modules/shared.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index d4ddeea0..9a13fb60 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -168,7 +168,7 @@ class State: def dict(self): obj = { "skipped": self.skipped, - "interrupted": self.skipped, + "interrupted": self.interrupted, "job": self.job, "job_count": self.job_count, "job_no": self.job_no, -- cgit v1.2.3 From fef98723b2b1c7a9893ead41bbefcb36192babd6 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 31 Dec 2022 12:44:26 +0300 Subject: set sd_model for API later, inside the lock, to prevent multiple requests with different models ending up with incorrect results #5877 #6012 --- modules/api/api.py | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 59b81c93..11daff0d 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -121,7 +121,6 @@ class Api: def text2imgapi(self, txt2imgreq: StableDiffusionTxt2ImgProcessingAPI): populate = txt2imgreq.copy(update={ # Override __init__ params - "sd_model": shared.sd_model, "sampler_name": validate_sampler_name(txt2imgreq.sampler_name or txt2imgreq.sampler_index), "do_not_save_samples": True, "do_not_save_grid": True @@ -129,9 +128,10 @@ class Api: ) if populate.sampler_name: populate.sampler_index = None # prevent a warning later on - p = StableDiffusionProcessingTxt2Img(**vars(populate)) with self.queue_lock: + p = StableDiffusionProcessingTxt2Img(sd_model=shared.sd_model, **vars(populate)) + shared.state.begin() processed = process_images(p) shared.state.end() @@ -151,7 +151,6 @@ class Api: mask = decode_base64_to_image(mask) populate = img2imgreq.copy(update={ # Override __init__ params - "sd_model": shared.sd_model, "sampler_name": validate_sampler_name(img2imgreq.sampler_name or img2imgreq.sampler_index), "do_not_save_samples": True, "do_not_save_grid": True, @@ -163,11 +162,11 @@ class Api: args = vars(populate) args.pop('include_init_images', None) # this is meant to be done by "exclude": True in model, but it's for a reason that I cannot determine. - p = StableDiffusionProcessingImg2Img(**args) - - p.init_images = [decode_base64_to_image(x) for x in init_images] with self.queue_lock: + p = StableDiffusionProcessingImg2Img(sd_model=shared.sd_model, **args) + p.init_images = [decode_base64_to_image(x) for x in init_images] + shared.state.begin() processed = process_images(p) shared.state.end() -- cgit v1.2.3 From 65be1df7bb55b21a3d76630a397c820218cbd12a Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Sat, 31 Dec 2022 07:46:04 -0500 Subject: initialize result so not to cause exception on empty results --- modules/interrogate.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/interrogate.py b/modules/interrogate.py index 46935210..6f761c5a 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -135,7 +135,7 @@ class InterrogateModels: return caption[0] def interrogate(self, pil_image): - res = None + res = "" try: -- cgit v1.2.3 From f34c7341720fb2059992926c9f9ae6ff25f7385b Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 31 Dec 2022 18:06:35 +0300 Subject: alt-diffusion integration --- modules/sd_hijack.py | 18 ++++++++++-------- modules/sd_hijack_clip.py | 14 +++++--------- modules/sd_hijack_xlmr.py | 34 ++++++++++++++++++++++++++++++++++ modules/shared.py | 10 +--------- 4 files changed, 50 insertions(+), 26 deletions(-) create mode 100644 modules/sd_hijack_xlmr.py (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index bce23b03..edcbaf52 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -5,7 +5,7 @@ import modules.textual_inversion.textual_inversion from modules import devices, sd_hijack_optimizations, shared, sd_hijack_checkpoint from modules.hypernetworks import hypernetwork from modules.shared import cmd_opts -from modules import sd_hijack_clip, sd_hijack_open_clip, sd_hijack_unet +from modules import sd_hijack_clip, sd_hijack_open_clip, sd_hijack_unet, sd_hijack_xlmr, xlmr from modules.sd_hijack_optimizations import invokeAI_mps_available @@ -68,6 +68,7 @@ def fix_checkpoint(): ldm.modules.diffusionmodules.openaimodel.ResBlock.forward = sd_hijack_checkpoint.ResBlock_forward ldm.modules.diffusionmodules.openaimodel.AttentionBlock.forward = sd_hijack_checkpoint.AttentionBlock_forward + class StableDiffusionModelHijack: fixes = None comments = [] @@ -79,21 +80,22 @@ class StableDiffusionModelHijack: def hijack(self, m): - if shared.text_model_name == "XLMR-Large": + if type(m.cond_stage_model) == xlmr.BertSeriesModelWithTransformation: model_embeddings = m.cond_stage_model.roberta.embeddings model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.word_embeddings, self) - m.cond_stage_model = sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) - + m.cond_stage_model = sd_hijack_xlmr.FrozenXLMREmbedderWithCustomWords(m.cond_stage_model, self) + elif type(m.cond_stage_model) == ldm.modules.encoders.modules.FrozenCLIPEmbedder: model_embeddings = m.cond_stage_model.transformer.text_model.embeddings model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.token_embedding, self) m.cond_stage_model = sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) - apply_optimizations() + elif type(m.cond_stage_model) == ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder: m.cond_stage_model.model.token_embedding = EmbeddingsWithFixes(m.cond_stage_model.model.token_embedding, self) m.cond_stage_model = sd_hijack_open_clip.FrozenOpenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) - apply_optimizations() - + + apply_optimizations() + self.clip = m.cond_stage_model fix_checkpoint() @@ -109,7 +111,7 @@ class StableDiffusionModelHijack: def undo_hijack(self, m): - if shared.text_model_name == "XLMR-Large": + if type(m.cond_stage_model) == xlmr.BertSeriesModelWithTransformation: m.cond_stage_model = m.cond_stage_model.wrapped elif type(m.cond_stage_model) == sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords: diff --git a/modules/sd_hijack_clip.py b/modules/sd_hijack_clip.py index 9ea6e1ce..6ec50cca 100644 --- a/modules/sd_hijack_clip.py +++ b/modules/sd_hijack_clip.py @@ -4,7 +4,6 @@ import torch from modules import prompt_parser, devices from modules.shared import opts -import modules.shared as shared def get_target_prompt_token_count(token_count): return math.ceil(max(token_count, 1) / 75) * 75 @@ -177,9 +176,6 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): return batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count def forward(self, text): - if shared.text_model_name == "XLMR-Large": - return self.wrapped.encode(text) - use_old = opts.use_old_emphasis_implementation if use_old: batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count = self.process_text_old(text) @@ -257,13 +253,13 @@ class FrozenCLIPEmbedderWithCustomWords(FrozenCLIPEmbedderWithCustomWordsBase): def __init__(self, wrapped, hijack): super().__init__(wrapped, hijack) self.tokenizer = wrapped.tokenizer - if shared.text_model_name == "XLMR-Large": - self.comma_token = None - else : - self.comma_token = [v for k, v in self.tokenizer.get_vocab().items() if k == ','][0] + + vocab = self.tokenizer.get_vocab() + + self.comma_token = vocab.get(',', None) self.token_mults = {} - tokens_with_parens = [(k, v) for k, v in self.tokenizer.get_vocab().items() if '(' in k or ')' in k or '[' in k or ']' in k] + tokens_with_parens = [(k, v) for k, v in vocab.items() if '(' in k or ')' in k or '[' in k or ']' in k] for text, ident in tokens_with_parens: mult = 1.0 for c in text: diff --git a/modules/sd_hijack_xlmr.py b/modules/sd_hijack_xlmr.py new file mode 100644 index 00000000..4ac51c38 --- /dev/null +++ b/modules/sd_hijack_xlmr.py @@ -0,0 +1,34 @@ +import open_clip.tokenizer +import torch + +from modules import sd_hijack_clip, devices +from modules.shared import opts + + +class FrozenXLMREmbedderWithCustomWords(sd_hijack_clip.FrozenCLIPEmbedderWithCustomWords): + def __init__(self, wrapped, hijack): + super().__init__(wrapped, hijack) + + self.id_start = wrapped.config.bos_token_id + self.id_end = wrapped.config.eos_token_id + self.id_pad = wrapped.config.pad_token_id + + self.comma_token = self.tokenizer.get_vocab().get(',', None) # alt diffusion doesn't have bits for comma + + def encode_with_transformers(self, tokens): + # there's no CLIP Skip here because all hidden layers have size of 1024 and the last one uses a + # trained layer to transform those 1024 into 768 for unet; so you can't choose which transformer + # layer to work with - you have to use the last + + attention_mask = (tokens != self.id_pad).to(device=tokens.device, dtype=torch.int64) + features = self.wrapped(input_ids=tokens, attention_mask=attention_mask) + z = features['projection_state'] + + return z + + def encode_embedding_init_text(self, init_text, nvpt): + embedding_layer = self.wrapped.roberta.embeddings + ids = self.wrapped.tokenizer(init_text, max_length=nvpt, return_tensors="pt", add_special_tokens=False)["input_ids"] + embedded = embedding_layer.token_embedding.wrapped(ids.to(devices.device)).squeeze(0) + + return embedded diff --git a/modules/shared.py b/modules/shared.py index 2b31e717..715b9169 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -23,7 +23,7 @@ demo = None sd_model_file = os.path.join(script_path, 'model.ckpt') default_sd_model_file = sd_model_file parser = argparse.ArgumentParser() -parser.add_argument("--config", type=str, default=os.path.join(script_path, "v1-inference.yaml"), help="path to config which constructs model",) +parser.add_argument("--config", type=str, default=os.path.join(script_path, "configs/v1-inference.yaml"), help="path to config which constructs model",) parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; if specified, this checkpoint will be added to the list of checkpoints and loaded",) parser.add_argument("--ckpt-dir", type=str, default=None, help="Path to directory with stable diffusion checkpoints") parser.add_argument("--gfpgan-dir", type=str, help="GFPGAN directory", default=('./src/gfpgan' if os.path.exists('./src/gfpgan') else './GFPGAN')) @@ -108,14 +108,6 @@ restricted_opts = { "outdir_txt2img_grids", "outdir_save", } -from omegaconf import OmegaConf -config = OmegaConf.load(f"{cmd_opts.config}") -# XLMR-Large -try: - text_model_name = config.model.params.cond_stage_config.params.name - -except : - text_model_name = "stable_diffusion" cmd_opts.disable_extension_access = (cmd_opts.share or cmd_opts.listen or cmd_opts.server_name) and not cmd_opts.enable_insecure_extension_access -- cgit v1.2.3 From f55ac33d446185680604e872ceda2ae858821d5c Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Sat, 31 Dec 2022 11:27:02 -0500 Subject: validate textual inversion embeddings --- modules/sd_models.py | 3 ++ modules/textual_inversion/textual_inversion.py | 43 +++++++++++++++++++++++--- modules/ui.py | 2 -- 3 files changed, 41 insertions(+), 7 deletions(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index ecdd91c5..ebd4dff7 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -325,6 +325,9 @@ def load_model(checkpoint_info=None): script_callbacks.model_loaded_callback(sd_model) print("Model loaded.") + + sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings(force_reload = True) # Reload embeddings after model load as they may or may not fit the model + return sd_model diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index f6112578..103ace60 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -23,6 +23,8 @@ class Embedding: self.vec = vec self.name = name self.step = step + self.shape = None + self.vectors = 0 self.cached_checksum = None self.sd_checkpoint = None self.sd_checkpoint_name = None @@ -57,8 +59,10 @@ class EmbeddingDatabase: def __init__(self, embeddings_dir): self.ids_lookup = {} self.word_embeddings = {} + self.skipped_embeddings = [] self.dir_mtime = None self.embeddings_dir = embeddings_dir + self.expected_shape = -1 def register_embedding(self, embedding, model): @@ -75,14 +79,35 @@ class EmbeddingDatabase: return embedding - def load_textual_inversion_embeddings(self): + def get_expected_shape(self): + expected_shape = -1 # initialize with unknown + idx = torch.tensor(0).to(shared.device) + if expected_shape == -1: + try: # matches sd15 signature + first_embedding = shared.sd_model.cond_stage_model.wrapped.transformer.text_model.embeddings.token_embedding.wrapped(idx) + expected_shape = first_embedding.shape[0] + except: + pass + if expected_shape == -1: + try: # matches sd20 signature + first_embedding = shared.sd_model.cond_stage_model.wrapped.model.token_embedding.wrapped(idx) + expected_shape = first_embedding.shape[0] + except: + pass + if expected_shape == -1: + print('Could not determine expected embeddings shape from model') + return expected_shape + + def load_textual_inversion_embeddings(self, force_reload = False): mt = os.path.getmtime(self.embeddings_dir) - if self.dir_mtime is not None and mt <= self.dir_mtime: + if not force_reload and self.dir_mtime is not None and mt <= self.dir_mtime: return self.dir_mtime = mt self.ids_lookup.clear() self.word_embeddings.clear() + self.skipped_embeddings = [] + self.expected_shape = self.get_expected_shape() def process_file(path, filename): name = os.path.splitext(filename)[0] @@ -122,7 +147,14 @@ class EmbeddingDatabase: embedding.step = data.get('step', None) embedding.sd_checkpoint = data.get('sd_checkpoint', None) embedding.sd_checkpoint_name = data.get('sd_checkpoint_name', None) - self.register_embedding(embedding, shared.sd_model) + embedding.vectors = vec.shape[0] + embedding.shape = vec.shape[-1] + + if (self.expected_shape == -1) or (self.expected_shape == embedding.shape): + self.register_embedding(embedding, shared.sd_model) + else: + self.skipped_embeddings.append(name) + # print('Skipping embedding {name}: shape was {shape} expected {expected}'.format(name = name, shape = embedding.shape, expected = self.expected_shape)) for fn in os.listdir(self.embeddings_dir): try: @@ -137,8 +169,9 @@ class EmbeddingDatabase: print(traceback.format_exc(), file=sys.stderr) continue - print(f"Loaded a total of {len(self.word_embeddings)} textual inversion embeddings.") - print("Embeddings:", ', '.join(self.word_embeddings.keys())) + print("Textual inversion embeddings {num} loaded: {val}".format(num = len(self.word_embeddings), val = ', '.join(self.word_embeddings.keys()))) + if (len(self.skipped_embeddings) > 0): + print("Textual inversion embeddings {num} skipped: {val}".format(num = len(self.skipped_embeddings), val = ', '.join(self.skipped_embeddings))) def find_embedding_at_position(self, tokens, offset): token = tokens[offset] diff --git a/modules/ui.py b/modules/ui.py index 57ee0465..397dd804 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1157,8 +1157,6 @@ def create_ui(): with gr.Column(variant='panel'): submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) - sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings() - with gr.Blocks(analytics_enabled=False) as train_interface: with gr.Row().style(equal_height=False): gr.HTML(value="

See wiki for detailed explanation.

") -- cgit v1.2.3 From bdbe09827b39be63c9c0b3636132ca58da38ebf6 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 31 Dec 2022 22:49:09 +0300 Subject: changed embedding accepted shape detection to use existing code and support the new alt-diffusion model, and reformatted messages a bit #6149 --- modules/textual_inversion/textual_inversion.py | 30 ++++++-------------------- 1 file changed, 6 insertions(+), 24 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 103ace60..66f40367 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -80,23 +80,8 @@ class EmbeddingDatabase: return embedding def get_expected_shape(self): - expected_shape = -1 # initialize with unknown - idx = torch.tensor(0).to(shared.device) - if expected_shape == -1: - try: # matches sd15 signature - first_embedding = shared.sd_model.cond_stage_model.wrapped.transformer.text_model.embeddings.token_embedding.wrapped(idx) - expected_shape = first_embedding.shape[0] - except: - pass - if expected_shape == -1: - try: # matches sd20 signature - first_embedding = shared.sd_model.cond_stage_model.wrapped.model.token_embedding.wrapped(idx) - expected_shape = first_embedding.shape[0] - except: - pass - if expected_shape == -1: - print('Could not determine expected embeddings shape from model') - return expected_shape + vec = shared.sd_model.cond_stage_model.encode_embedding_init_text(",", 1) + return vec.shape[1] def load_textual_inversion_embeddings(self, force_reload = False): mt = os.path.getmtime(self.embeddings_dir) @@ -112,8 +97,6 @@ class EmbeddingDatabase: def process_file(path, filename): name = os.path.splitext(filename)[0] - data = [] - if os.path.splitext(filename.upper())[-1] in ['.PNG', '.WEBP', '.JXL', '.AVIF']: embed_image = Image.open(path) if hasattr(embed_image, 'text') and 'sd-ti-embedding' in embed_image.text: @@ -150,11 +133,10 @@ class EmbeddingDatabase: embedding.vectors = vec.shape[0] embedding.shape = vec.shape[-1] - if (self.expected_shape == -1) or (self.expected_shape == embedding.shape): + if self.expected_shape == -1 or self.expected_shape == embedding.shape: self.register_embedding(embedding, shared.sd_model) else: self.skipped_embeddings.append(name) - # print('Skipping embedding {name}: shape was {shape} expected {expected}'.format(name = name, shape = embedding.shape, expected = self.expected_shape)) for fn in os.listdir(self.embeddings_dir): try: @@ -169,9 +151,9 @@ class EmbeddingDatabase: print(traceback.format_exc(), file=sys.stderr) continue - print("Textual inversion embeddings {num} loaded: {val}".format(num = len(self.word_embeddings), val = ', '.join(self.word_embeddings.keys()))) - if (len(self.skipped_embeddings) > 0): - print("Textual inversion embeddings {num} skipped: {val}".format(num = len(self.skipped_embeddings), val = ', '.join(self.skipped_embeddings))) + print(f"Textual inversion embeddings loaded({len(self.word_embeddings)}): {', '.join(self.word_embeddings.keys())}") + if len(self.skipped_embeddings) > 0: + print(f"Textual inversion embeddings skipped({len(self.skipped_embeddings)}): {', '.join(self.skipped_embeddings)}") def find_embedding_at_position(self, tokens, offset): token = tokens[offset] -- cgit v1.2.3 From f4535f6e4f001314bd155bc6e1b6908e02792b9a Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 31 Dec 2022 23:40:55 +0300 Subject: make it so that memory/embeddings info is displayed in a separate UI element from generation parameters, and is preserved when you change the displayed infotext by clicking on gallery images --- modules/img2img.py | 2 +- modules/processing.py | 5 +++-- modules/txt2img.py | 2 +- modules/ui.py | 31 +++++++++++++++++-------------- 4 files changed, 22 insertions(+), 18 deletions(-) (limited to 'modules') diff --git a/modules/img2img.py b/modules/img2img.py index 81da4b13..ca58b5d8 100644 --- a/modules/img2img.py +++ b/modules/img2img.py @@ -162,4 +162,4 @@ def img2img(mode: int, prompt: str, negative_prompt: str, prompt_style: str, pro if opts.do_not_show_images: processed.images = [] - return processed.images, generation_info_js, plaintext_to_html(processed.info) + return processed.images, generation_info_js, plaintext_to_html(processed.info), plaintext_to_html(processed.comments) diff --git a/modules/processing.py b/modules/processing.py index 0a9a8f95..42dc19ea 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -239,7 +239,7 @@ class StableDiffusionProcessing(): class Processed: - def __init__(self, p: StableDiffusionProcessing, images_list, seed=-1, info="", subseed=None, all_prompts=None, all_negative_prompts=None, all_seeds=None, all_subseeds=None, index_of_first_image=0, infotexts=None): + def __init__(self, p: StableDiffusionProcessing, images_list, seed=-1, info="", subseed=None, all_prompts=None, all_negative_prompts=None, all_seeds=None, all_subseeds=None, index_of_first_image=0, infotexts=None, comments=""): self.images = images_list self.prompt = p.prompt self.negative_prompt = p.negative_prompt @@ -247,6 +247,7 @@ class Processed: self.subseed = subseed self.subseed_strength = p.subseed_strength self.info = info + self.comments = comments self.width = p.width self.height = p.height self.sampler_name = p.sampler_name @@ -646,7 +647,7 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: devices.torch_gc() - res = Processed(p, output_images, p.all_seeds[0], infotext() + "".join(["\n\n" + x for x in comments]), subseed=p.all_subseeds[0], index_of_first_image=index_of_first_image, infotexts=infotexts) + res = Processed(p, output_images, p.all_seeds[0], infotext(), comments="".join(["\n\n" + x for x in comments]), subseed=p.all_subseeds[0], index_of_first_image=index_of_first_image, infotexts=infotexts) if p.scripts is not None: p.scripts.postprocess(p, res) diff --git a/modules/txt2img.py b/modules/txt2img.py index c8f81176..7f61e19a 100644 --- a/modules/txt2img.py +++ b/modules/txt2img.py @@ -59,4 +59,4 @@ def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: if opts.do_not_show_images: processed.images = [] - return processed.images, generation_info_js, plaintext_to_html(processed.info) + return processed.images, generation_info_js, plaintext_to_html(processed.info), plaintext_to_html(processed.comments) diff --git a/modules/ui.py b/modules/ui.py index 397dd804..f550ad00 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -159,7 +159,7 @@ def save_files(js_data, images, do_make_zip, index): zip_file.writestr(filenames[i], f.read()) fullfns.insert(0, zip_filepath) - return gr.File.update(value=fullfns, visible=True), '', '', plaintext_to_html(f"Saved: {filenames[0]}") + return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") @@ -593,6 +593,8 @@ Requested path was: {f} with gr.Group(): html_info = gr.HTML() + html_log = gr.HTML() + generation_info = gr.Textbox(visible=False) if tabname == 'txt2img' or tabname == 'img2img': generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") @@ -615,16 +617,16 @@ Requested path was: {f} ], outputs=[ download_files, - html_info, - html_info, - html_info, + html_log, ] ) else: html_info_x = gr.HTML() html_info = gr.HTML() + html_log = gr.HTML() + parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) - return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info + return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log def create_ui(): @@ -686,14 +688,14 @@ def create_ui(): with gr.Group(): custom_inputs = modules.scripts.scripts_txt2img.setup_ui() - txt2img_gallery, generation_info, html_info = create_output_panel("txt2img", opts.outdir_txt2img_samples) + txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) txt2img_args = dict( - fn=wrap_gradio_gpu_call(modules.txt2img.txt2img), + fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), _js="submit", inputs=[ txt2img_prompt, @@ -720,7 +722,8 @@ def create_ui(): outputs=[ txt2img_gallery, generation_info, - html_info + html_info, + html_log, ], show_progress=False, ) @@ -799,7 +802,6 @@ def create_ui(): with gr.Blocks(analytics_enabled=False) as img2img_interface: img2img_prompt, roll, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) - with gr.Row(elem_id='img2img_progress_row'): img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) @@ -883,7 +885,7 @@ def create_ui(): with gr.Group(): custom_inputs = modules.scripts.scripts_img2img.setup_ui() - img2img_gallery, generation_info, html_info = create_output_panel("img2img", opts.outdir_img2img_samples) + img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) @@ -915,7 +917,7 @@ def create_ui(): ) img2img_args = dict( - fn=wrap_gradio_gpu_call(modules.img2img.img2img), + fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), _js="submit_img2img", inputs=[ dummy_component, @@ -954,7 +956,8 @@ def create_ui(): outputs=[ img2img_gallery, generation_info, - html_info + html_info, + html_log, ], show_progress=False, ) @@ -1078,10 +1081,10 @@ def create_ui(): with gr.Group(): upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False) - result_images, html_info_x, html_info = create_output_panel("extras", opts.outdir_extras_samples) + result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) submit.click( - fn=wrap_gradio_gpu_call(modules.extras.run_extras), + fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), _js="get_extras_tab_index", inputs=[ dummy_component, -- cgit v1.2.3 From 360feed9b55fb03060c236773867b08b4265645d Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 00:38:58 +0300 Subject: HAPPY NEW YEAR make save to zip into its own button instead of a checkbox --- modules/ui.py | 30 ++++++++++++++++++++++-------- 1 file changed, 22 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index f550ad00..279b5110 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -570,13 +570,14 @@ Requested path was: {f} generation_info = None with gr.Column(): - with gr.Row(): + with gr.Row(elem_id=f"image_buttons_{tabname}"): + open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else 'open_folder') + if tabname != "extras": save = gr.Button('Save', elem_id=f'save_{tabname}') + save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) - button_id = "hidden_element" if shared.cmd_opts.hide_ui_dir_config else 'open_folder' - open_folder_button = gr.Button(folder_symbol, elem_id=button_id) open_folder_button.click( fn=lambda: open_folder(opts.outdir_samples or outdir), @@ -585,9 +586,6 @@ Requested path was: {f} ) if tabname != "extras": - with gr.Row(): - do_make_zip = gr.Checkbox(label="Make Zip when Save?", value=False) - with gr.Row(): download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False) @@ -608,11 +606,11 @@ Requested path was: {f} save.click( fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, z, selected_gallery_index()]", + _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", inputs=[ generation_info, result_gallery, - do_make_zip, + html_info, html_info, ], outputs=[ @@ -620,6 +618,22 @@ Requested path was: {f} html_log, ] ) + + save_zip.click( + fn=wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ] + ) + else: html_info_x = gr.HTML() html_info = gr.HTML() -- cgit v1.2.3 From 29a3a7eb13478297bc7093971b48827ab8246f45 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 01:19:10 +0300 Subject: show sampler selection in dropdown, add option selection to revert to old radio group --- modules/shared.py | 1 + modules/ui.py | 22 +++++++++++++++------- 2 files changed, 16 insertions(+), 7 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 715b9169..948b9542 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -406,6 +406,7 @@ options_templates.update(options_section(('ui', "User interface"), { "js_modal_lightbox": OptionInfo(True, "Enable full page image viewer"), "js_modal_lightbox_initially_zoomed": OptionInfo(True, "Show images zoomed in by default in full page image viewer"), "show_progress_in_title": OptionInfo(True, "Show generation progress in window title."), + "samplers_in_dropdown": OptionInfo(True, "Use dropdown for sampler selection instead of radio group"), 'quicksettings': OptionInfo("sd_model_checkpoint", "Quicksettings list"), 'localization': OptionInfo("None", "Localization (requires restart)", gr.Dropdown, lambda: {"choices": ["None"] + list(localization.localizations.keys())}, refresh=lambda: localization.list_localizations(cmd_opts.localizations_dir)), })) diff --git a/modules/ui.py b/modules/ui.py index 279b5110..c7b8ea5d 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -643,6 +643,19 @@ Requested path was: {f} return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log +def create_sampler_and_steps_selection(choices, tabname): + if opts.samplers_in_dropdown: + with gr.Row(elem_id=f"sampler_selection_{tabname}"): + sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling Steps", value=20) + else: + with gr.Group(elem_id=f"sampler_selection_{tabname}"): + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling Steps", value=20) + sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + + return steps, sampler_index + + def create_ui(): import modules.img2img import modules.txt2img @@ -660,9 +673,6 @@ def create_ui(): dummy_component = gr.Label(visible=False) txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) - - - with gr.Row(elem_id='txt2img_progress_row'): with gr.Column(scale=1): pass @@ -674,8 +684,7 @@ def create_ui(): with gr.Row().style(equal_height=False): with gr.Column(variant='panel', elem_id="txt2img_settings"): - steps = gr.Slider(minimum=1, maximum=150, step=1, label="Sampling Steps", value=20) - sampler_index = gr.Radio(label='Sampling method', elem_id="txt2img_sampling", choices=[x.name for x in samplers], value=samplers[0].name, type="index") + steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") with gr.Group(): width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512) @@ -875,8 +884,7 @@ def create_ui(): with gr.Row(): resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", show_label=False, choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") - steps = gr.Slider(minimum=1, maximum=150, step=1, label="Sampling Steps", value=20) - sampler_index = gr.Radio(label='Sampling method', choices=[x.name for x in samplers_for_img2img], value=samplers_for_img2img[0].name, type="index") + steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") with gr.Group(): width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") -- cgit v1.2.3 From 210449b374d522c94a67fe54289a9eb515933a9f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 02:41:15 +0300 Subject: fix 'RuntimeError: Expected all tensors to be on the same device' error preventing models from loading on lowvram/medvram. --- modules/sd_hijack_clip.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_hijack_clip.py b/modules/sd_hijack_clip.py index 6ec50cca..ca92b142 100644 --- a/modules/sd_hijack_clip.py +++ b/modules/sd_hijack_clip.py @@ -298,6 +298,6 @@ class FrozenCLIPEmbedderWithCustomWords(FrozenCLIPEmbedderWithCustomWordsBase): def encode_embedding_init_text(self, init_text, nvpt): embedding_layer = self.wrapped.transformer.text_model.embeddings ids = self.wrapped.tokenizer(init_text, max_length=nvpt, return_tensors="pt", add_special_tokens=False)["input_ids"] - embedded = embedding_layer.token_embedding.wrapped(ids.to(devices.device)).squeeze(0) + embedded = embedding_layer.token_embedding.wrapped(ids.to(embedding_layer.token_embedding.wrapped.weight.device)).squeeze(0) return embedded -- cgit v1.2.3 From 16b9661d2741b241c3964fcbd56559c078b84822 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 09:51:37 +0300 Subject: change karras scheduler sigmas to values recommended by SD from old 0.1 to 10 with an option to revert to old --- modules/sd_samplers.py | 4 +++- modules/shared.py | 6 +++++- 2 files changed, 8 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py index 177b5338..e904d860 100644 --- a/modules/sd_samplers.py +++ b/modules/sd_samplers.py @@ -465,7 +465,9 @@ class KDiffusionSampler: if p.sampler_noise_scheduler_override: sigmas = p.sampler_noise_scheduler_override(steps) elif self.config is not None and self.config.options.get('scheduler', None) == 'karras': - sigmas = k_diffusion.sampling.get_sigmas_karras(n=steps, sigma_min=0.1, sigma_max=10, device=shared.device) + sigma_min, sigma_max = (0.1, 10) if opts.use_old_karras_scheduler_sigmas else (self.model_wrap.sigmas[0].item(), self.model_wrap.sigmas[-1].item()) + + sigmas = k_diffusion.sampling.get_sigmas_karras(n=steps, sigma_min=sigma_min, sigma_max=sigma_max, device=shared.device) else: sigmas = self.model_wrap.get_sigmas(steps) diff --git a/modules/shared.py b/modules/shared.py index 948b9542..7f430b93 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -368,13 +368,17 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "img2img_background_color": OptionInfo("#ffffff", "With img2img, fill image's transparent parts with this color.", gr.ColorPicker, {}), "enable_quantization": OptionInfo(False, "Enable quantization in K samplers for sharper and cleaner results. This may change existing seeds. Requires restart to apply."), "enable_emphasis": OptionInfo(True, "Emphasis: use (text) to make model pay more attention to text and [text] to make it pay less attention"), - "use_old_emphasis_implementation": OptionInfo(False, "Use old emphasis implementation. Can be useful to reproduce old seeds."), "enable_batch_seeds": OptionInfo(True, "Make K-diffusion samplers produce same images in a batch as when making a single image"), "comma_padding_backtrack": OptionInfo(20, "Increase coherency by padding from the last comma within n tokens when using more than 75 tokens", gr.Slider, {"minimum": 0, "maximum": 74, "step": 1 }), 'CLIP_stop_at_last_layers': OptionInfo(1, "Clip skip", gr.Slider, {"minimum": 1, "maximum": 12, "step": 1}), "random_artist_categories": OptionInfo([], "Allowed categories for random artists selection when using the Roll button", gr.CheckboxGroup, {"choices": artist_db.categories()}), })) +options_templates.update(options_section(('compatibility', "Compatibility"), { + "use_old_emphasis_implementation": OptionInfo(False, "Use old emphasis implementation. Can be useful to reproduce old seeds."), + "use_old_karras_scheduler_sigmas": OptionInfo(False, "Use old karras scheduler sigmas (0.1 to 10)."), +})) + options_templates.update(options_section(('interrogate', "Interrogate Options"), { "interrogate_keep_models_in_memory": OptionInfo(False, "Interrogate: keep models in VRAM"), "interrogate_use_builtin_artists": OptionInfo(True, "Interrogate: use artists from artists.csv"), -- cgit v1.2.3 From 11d432d92d63660c516540dcb48faac87669b4f0 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 10:35:38 +0300 Subject: add refresh buttons to checkpoint merger --- modules/ui.py | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index c7b8ea5d..4cc2ce4f 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1167,8 +1167,14 @@ def create_ui(): with gr.Row(): primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") + create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") + secondary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_secondary_model_name", label="Secondary model (B)") + create_refresh_button(secondary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_B") + tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") + create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") + custom_name = gr.Textbox(label="Custom Name (Optional)") interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3) interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method") -- cgit v1.2.3 From 76f256fe8f844641f4e9b41f35c7dd2cba5090d6 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 11:08:39 +0300 Subject: Bump gradio version #YOLO --- modules/ui_tempdir.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui_tempdir.py b/modules/ui_tempdir.py index 07210d14..8d519310 100644 --- a/modules/ui_tempdir.py +++ b/modules/ui_tempdir.py @@ -15,7 +15,8 @@ Savedfile = namedtuple("Savedfile", ["name"]) def save_pil_to_file(pil_image, dir=None): already_saved_as = getattr(pil_image, 'already_saved_as', None) if already_saved_as and os.path.isfile(already_saved_as): - shared.demo.temp_dirs = shared.demo.temp_dirs | {os.path.abspath(os.path.dirname(already_saved_as))} + shared.demo.temp_file_sets[0] = shared.demo.temp_file_sets[0] | {os.path.abspath(already_saved_as)} + file_obj = Savedfile(already_saved_as) return file_obj -- cgit v1.2.3 From b46b97fa297b3a4a654da77cf98a775a2bcab4c7 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 11:38:17 +0300 Subject: more fixes for gradio update --- modules/generation_parameters_copypaste.py | 2 +- modules/ui_tempdir.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index fbd91300..54b3372d 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -38,7 +38,7 @@ def quote(text): def image_from_url_text(filedata): if type(filedata) == dict and filedata["is_file"]: filename = filedata["name"] - is_in_right_dir = any(Path(temp_dir).resolve() in Path(filename).resolve().parents for temp_dir in shared.demo.temp_dirs) + is_in_right_dir = any([filename in fileset for fileset in shared.demo.temp_file_sets]) assert is_in_right_dir, 'trying to open image file outside of allowed directories' return Image.open(filename) diff --git a/modules/ui_tempdir.py b/modules/ui_tempdir.py index 8d519310..363d449d 100644 --- a/modules/ui_tempdir.py +++ b/modules/ui_tempdir.py @@ -45,7 +45,7 @@ def on_tmpdir_changed(): os.makedirs(shared.opts.temp_dir, exist_ok=True) - shared.demo.temp_dirs = shared.demo.temp_dirs | {os.path.abspath(shared.opts.temp_dir)} + shared.demo.temp_file_sets[0] = shared.demo.temp_file_sets[0] | {os.path.abspath(shared.opts.temp_dir)} def cleanup_tmpdr(): -- cgit v1.2.3 From e5f1a37cb9b537d95b2df47c96b4a4f7242fd294 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 13:08:40 +0300 Subject: make refresh buttons look more nice --- modules/ui.py | 6 +++--- modules/ui_components.py | 18 ++++++++++++++++++ 2 files changed, 21 insertions(+), 3 deletions(-) create mode 100644 modules/ui_components.py (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 4cc2ce4f..32fa80d1 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -19,7 +19,7 @@ import numpy as np from PIL import Image, PngImagePlugin from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, ui_components from modules.paths import script_path from modules.shared import opts, cmd_opts, restricted_opts @@ -532,7 +532,7 @@ def create_refresh_button(refresh_component, refresh_method, refreshed_args, ele return gr.update(**(args or {})) - refresh_button = gr.Button(value=refresh_symbol, elem_id=elem_id) + refresh_button = ui_components.ToolButton(value=refresh_symbol, elem_id=elem_id) refresh_button.click( fn=refresh, inputs=[], @@ -1476,7 +1476,7 @@ def create_ui(): res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) else: - with gr.Row(variant="compact"): + with ui_components.FormRow(): res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) else: diff --git a/modules/ui_components.py b/modules/ui_components.py new file mode 100644 index 00000000..d0519d2d --- /dev/null +++ b/modules/ui_components.py @@ -0,0 +1,18 @@ +import gradio as gr + + +class ToolButton(gr.Button, gr.components.FormComponent): + """Small button with single emoji as text, fits inside gradio forms""" + + def __init__(self, **kwargs): + super().__init__(variant="tool", **kwargs) + + def get_block_name(self): + return "button" + + +class FormRow(gr.Row, gr.components.FormComponent): + """Same as gr.Row but fits inside gradio forms""" + + def get_block_name(self): + return "row" -- cgit v1.2.3 From 5f12b23b8bb7fca585a3a1e844881d06f171364e Mon Sep 17 00:00:00 2001 From: AlUlkesh <99896447+AlUlkesh@users.noreply.github.com> Date: Wed, 28 Dec 2022 22:18:19 +0100 Subject: Adding image numbers on grids New grid option in settings enables adding of image numbers on grids. This makes identifying the images, especially in larger batches, much easier. Revert "Adding image numbers on grids" This reverts commit 3530c283b4b1d3a3cab40efbffe4cf2697938b6f. Implements Callback for image grid loop Necessary to make "Add image's number to its picture in the grid" extension possible. --- modules/images.py | 1 + modules/script_callbacks.py | 20 ++++++++++++++++++++ 2 files changed, 21 insertions(+) (limited to 'modules') diff --git a/modules/images.py b/modules/images.py index 31d4528d..5afd3891 100644 --- a/modules/images.py +++ b/modules/images.py @@ -43,6 +43,7 @@ def image_grid(imgs, batch_size=1, rows=None): grid = Image.new('RGB', size=(cols * w, rows * h), color='black') for i, img in enumerate(imgs): + script_callbacks.image_grid_loop_callback(img) grid.paste(img, box=(i % cols * w, i // cols * h)) return grid diff --git a/modules/script_callbacks.py b/modules/script_callbacks.py index 8e22f875..0c854407 100644 --- a/modules/script_callbacks.py +++ b/modules/script_callbacks.py @@ -51,6 +51,11 @@ class UiTrainTabParams: self.txt2img_preview_params = txt2img_preview_params +class ImageGridLoopParams: + def __init__(self, img): + self.img = img + + ScriptCallback = namedtuple("ScriptCallback", ["script", "callback"]) callback_map = dict( callbacks_app_started=[], @@ -63,6 +68,7 @@ callback_map = dict( callbacks_cfg_denoiser=[], callbacks_before_component=[], callbacks_after_component=[], + callbacks_image_grid_loop=[], ) @@ -154,6 +160,12 @@ def after_component_callback(component, **kwargs): except Exception: report_exception(c, 'after_component_callback') +def image_grid_loop_callback(component, **kwargs): + for c in callback_map['callbacks_image_grid_loop']: + try: + c.callback(component, **kwargs) + except Exception: + report_exception(c, 'image_grid_loop') def add_callback(callbacks, fun): stack = [x for x in inspect.stack() if x.filename != __file__] @@ -255,3 +267,11 @@ def on_before_component(callback): def on_after_component(callback): """register a function to be called after a component is created. See on_before_component for more.""" add_callback(callback_map['callbacks_after_component'], callback) + + +def on_image_grid_loop(callback): + """register a function to be called inside the image grid loop. + The callback is called with one argument: + - params: ImageGridLoopParams - parameters to be used inside the image grid loop. + """ + add_callback(callback_map['callbacks_image_grid_loop'], callback) -- cgit v1.2.3 From 524d532b387732d4d32f237e792c7f201a934400 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 14:07:40 +0300 Subject: moved roll artist to built-in extensions --- modules/ui.py | 37 +++---------------------------------- 1 file changed, 3 insertions(+), 34 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 32fa80d1..27da2c2c 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -80,7 +80,6 @@ css_hide_progressbar = """ # Important that they exactly match script.js for tooltip to work. random_symbol = '\U0001f3b2\ufe0f' # 🎲️ reuse_symbol = '\u267b\ufe0f' # ♻️ -art_symbol = '\U0001f3a8' # 🎨 paste_symbol = '\u2199\ufe0f' # ↙ folder_symbol = '\U0001f4c2' # 📂 refresh_symbol = '\U0001f504' # 🔄 @@ -234,13 +233,6 @@ def check_progress_call_initial(id_part): return check_progress_call(id_part) -def roll_artist(prompt): - allowed_cats = set([x for x in shared.artist_db.categories() if len(opts.random_artist_categories)==0 or x in opts.random_artist_categories]) - artist = random.choice([x for x in shared.artist_db.artists if x.category in allowed_cats]) - - return prompt + ", " + artist.name if prompt != '' else artist.name - - def visit(x, func, path=""): if hasattr(x, 'children'): for c in x.children: @@ -403,7 +395,6 @@ def create_toprow(is_img2img): ) with gr.Column(scale=1, elem_id="roll_col"): - roll = gr.Button(value=art_symbol, elem_id="roll", visible=len(shared.artist_db.artists) > 0) paste = gr.Button(value=paste_symbol, elem_id="paste") save_style = gr.Button(value=save_style_symbol, elem_id="style_create") prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") @@ -452,7 +443,7 @@ def create_toprow(is_img2img): prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) prompt_style2.save_to_config = True - return prompt, roll, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button + return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button def setup_progressbar(progressbar, preview, id_part, textinfo=None): @@ -668,7 +659,7 @@ def create_ui(): modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) with gr.Blocks(analytics_enabled=False) as txt2img_interface: - txt2img_prompt, roll, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) + txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) dummy_component = gr.Label(visible=False) txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) @@ -771,16 +762,6 @@ def create_ui(): outputs=[hr_options], ) - roll.click( - fn=roll_artist, - _js="update_txt2img_tokens", - inputs=[ - txt2img_prompt, - ], - outputs=[ - txt2img_prompt, - ] - ) txt2img_paste_fields = [ (txt2img_prompt, "Prompt"), @@ -823,7 +804,7 @@ def create_ui(): modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) with gr.Blocks(analytics_enabled=False) as img2img_interface: - img2img_prompt, roll, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) + img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) with gr.Row(elem_id='img2img_progress_row'): img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) @@ -999,18 +980,6 @@ def create_ui(): outputs=[img2img_prompt], ) - - roll.click( - fn=roll_artist, - _js="update_img2img_tokens", - inputs=[ - img2img_prompt, - ], - outputs=[ - img2img_prompt, - ] - ) - prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] -- cgit v1.2.3 From e672cfb07418a1a3130d3bf21c14a0d3819f81fb Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 1 Jan 2023 18:37:37 +0300 Subject: rework of callback for #6094 --- modules/images.py | 10 ++++++---- modules/script_callbacks.py | 26 +++++++++++++++----------- 2 files changed, 21 insertions(+), 15 deletions(-) (limited to 'modules') diff --git a/modules/images.py b/modules/images.py index 719aaf3b..f84fd485 100644 --- a/modules/images.py +++ b/modules/images.py @@ -39,12 +39,14 @@ def image_grid(imgs, batch_size=1, rows=None): cols = math.ceil(len(imgs) / rows) + params = script_callbacks.ImageGridLoopParams(imgs, cols, rows) + script_callbacks.image_grid_callback(params) + w, h = imgs[0].size - grid = Image.new('RGB', size=(cols * w, rows * h), color='black') + grid = Image.new('RGB', size=(params.cols * w, params.rows * h), color='black') - for i, img in enumerate(imgs): - script_callbacks.image_grid_loop_callback(img) - grid.paste(img, box=(i % cols * w, i // cols * h)) + for i, img in enumerate(params.imgs): + grid.paste(img, box=(i % params.cols * w, i // params.cols * h)) return grid diff --git a/modules/script_callbacks.py b/modules/script_callbacks.py index 0c854407..de69fd9f 100644 --- a/modules/script_callbacks.py +++ b/modules/script_callbacks.py @@ -52,8 +52,10 @@ class UiTrainTabParams: class ImageGridLoopParams: - def __init__(self, img): - self.img = img + def __init__(self, imgs, cols, rows): + self.imgs = imgs + self.cols = cols + self.rows = rows ScriptCallback = namedtuple("ScriptCallback", ["script", "callback"]) @@ -68,7 +70,7 @@ callback_map = dict( callbacks_cfg_denoiser=[], callbacks_before_component=[], callbacks_after_component=[], - callbacks_image_grid_loop=[], + callbacks_image_grid=[], ) @@ -160,12 +162,14 @@ def after_component_callback(component, **kwargs): except Exception: report_exception(c, 'after_component_callback') -def image_grid_loop_callback(component, **kwargs): - for c in callback_map['callbacks_image_grid_loop']: + +def image_grid_callback(params: ImageGridLoopParams): + for c in callback_map['callbacks_image_grid']: try: - c.callback(component, **kwargs) + c.callback(params) except Exception: - report_exception(c, 'image_grid_loop') + report_exception(c, 'image_grid') + def add_callback(callbacks, fun): stack = [x for x in inspect.stack() if x.filename != __file__] @@ -269,9 +273,9 @@ def on_after_component(callback): add_callback(callback_map['callbacks_after_component'], callback) -def on_image_grid_loop(callback): - """register a function to be called inside the image grid loop. +def on_image_grid(callback): + """register a function to be called before making an image grid. The callback is called with one argument: - - params: ImageGridLoopParams - parameters to be used inside the image grid loop. + - params: ImageGridLoopParams - parameters to be used for grid creation. Can be modified. """ - add_callback(callback_map['callbacks_image_grid_loop'], callback) + add_callback(callback_map['callbacks_image_grid'], callback) -- cgit v1.2.3 From a005fccddd5a37c57f1afe5234660b59b9a41508 Mon Sep 17 00:00:00 2001 From: me <25877290+Kryptortio@users.noreply.github.com> Date: Sun, 1 Jan 2023 14:51:12 +0100 Subject: Add a lot more elem_id/HTML id, modified some that were duplicates for seed section --- modules/generation_parameters_copypaste.py | 2 +- modules/ui.py | 254 ++++++++++++++--------------- 2 files changed, 128 insertions(+), 128 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 54b3372d..8e7f0df0 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -93,7 +93,7 @@ def integrate_settings_paste_fields(component_dict): def create_buttons(tabs_list): buttons = {} for tab in tabs_list: - buttons[tab] = gr.Button(f"Send to {tab}") + buttons[tab] = gr.Button(f"Send to {tab}", elem_id=f"{tab}_tab") return buttons diff --git a/modules/ui.py b/modules/ui.py index 27da2c2c..7070ea15 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -272,17 +272,17 @@ def interrogate_deepbooru(image): return gr_show(True) if prompt is None else prompt -def create_seed_inputs(): +def create_seed_inputs(target_interface): with gr.Row(): with gr.Box(): - with gr.Row(elem_id='seed_row'): - seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1) + with gr.Row(elem_id=target_interface + '_seed_row'): + seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') seed.style(container=False) - random_seed = gr.Button(random_symbol, elem_id='random_seed') - reuse_seed = gr.Button(reuse_symbol, elem_id='reuse_seed') + random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') + reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') - with gr.Box(elem_id='subseed_show_box'): - seed_checkbox = gr.Checkbox(label='Extra', elem_id='subseed_show', value=False) + with gr.Box(elem_id=target_interface + '_subseed_show_box'): + seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) # Components to show/hide based on the 'Extra' checkbox seed_extras = [] @@ -290,17 +290,17 @@ def create_seed_inputs(): with gr.Row(visible=False) as seed_extra_row_1: seed_extras.append(seed_extra_row_1) with gr.Box(): - with gr.Row(elem_id='subseed_row'): - subseed = gr.Number(label='Variation seed', value=-1) + with gr.Row(elem_id=target_interface + '_subseed_row'): + subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') subseed.style(container=False) - random_subseed = gr.Button(random_symbol, elem_id='random_subseed') - reuse_subseed = gr.Button(reuse_symbol, elem_id='reuse_subseed') - subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01) + random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') + reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') + subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') with gr.Row(visible=False) as seed_extra_row_2: seed_extras.append(seed_extra_row_2) - seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0) - seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0) + seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') + seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') random_seed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[seed]) random_subseed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[subseed]) @@ -678,28 +678,28 @@ def create_ui(): steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") with gr.Group(): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512) - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512) + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") with gr.Row(): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1) - tiling = gr.Checkbox(label='Tiling', value=False) - enable_hr = gr.Checkbox(label='Highres. fix', value=False) + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") + enable_hr = gr.Checkbox(label='Highres. fix', value=False, elem_id="txt2img_enable_hr") with gr.Row(visible=False) as hr_options: - firstphase_width = gr.Slider(minimum=0, maximum=1024, step=8, label="Firstpass width", value=0) - firstphase_height = gr.Slider(minimum=0, maximum=1024, step=8, label="Firstpass height", value=0) - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7) + firstphase_width = gr.Slider(minimum=0, maximum=1024, step=8, label="Firstpass width", value=0, elem_id="txt2img_firstphase_width") + firstphase_height = gr.Slider(minimum=0, maximum=1024, step=8, label="Firstpass height", value=0, elem_id="txt2img_firstphase_height") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") with gr.Row(equal_height=True): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1) - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1) + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0) + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs() + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') - with gr.Group(): + with gr.Group(elem_id="txt2img_script_container"): custom_inputs = modules.scripts.scripts_txt2img.setup_ui() txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) @@ -821,10 +821,10 @@ def create_ui(): with gr.Column(variant='panel', elem_id="img2img_settings"): with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: - with gr.TabItem('img2img', id='img2img'): + with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) - with gr.TabItem('Inpaint', id='inpaint'): + with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) init_img_with_mask_orig = gr.State(None) @@ -843,24 +843,24 @@ def create_ui(): init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") with gr.Row(): - mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4) - mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch) + mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") + mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") with gr.Row(): mask_mode = gr.Radio(label="Mask mode", show_label=False, choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") - inpainting_mask_invert = gr.Radio(label='Masking mode', show_label=False, choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index") + inpainting_mask_invert = gr.Radio(label='Masking mode', show_label=False, choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") - inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index") + inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") with gr.Row(): - inpaint_full_res = gr.Checkbox(label='Inpaint at full resolution', value=False) - inpaint_full_res_padding = gr.Slider(label='Inpaint at full resolution padding, pixels', minimum=0, maximum=256, step=4, value=32) + inpaint_full_res = gr.Checkbox(label='Inpaint at full resolution', value=False, elem_id="img2img_inpaint_full_res") + inpaint_full_res_padding = gr.Slider(label='Inpaint at full resolution padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") - with gr.TabItem('Batch img2img', id='batch'): + with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") - img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs) - img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs) + img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") + img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") with gr.Row(): resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", show_label=False, choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") @@ -872,20 +872,20 @@ def create_ui(): height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") with gr.Row(): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1) - tiling = gr.Checkbox(label='Tiling', value=False) + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") with gr.Row(): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1) - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1) + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") with gr.Group(): - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0) - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75) + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs() + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') - with gr.Group(): + with gr.Group(elem_id="img2img_script_container"): custom_inputs = modules.scripts.scripts_img2img.setup_ui() img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) @@ -1032,45 +1032,45 @@ def create_ui(): with gr.Row().style(equal_height=False): with gr.Column(variant='panel'): with gr.Tabs(elem_id="mode_extras"): - with gr.TabItem('Single Image'): - extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil") + with gr.TabItem('Single Image', elem_id="extras_single_tab"): + extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") - with gr.TabItem('Batch Process'): - image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file") + with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): + image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") - with gr.TabItem('Batch from Directory'): - extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.") - extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.") - show_extras_results = gr.Checkbox(label='Show result images', value=True) + with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): + extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") + extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") + show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') with gr.Tabs(elem_id="extras_resize_mode"): - with gr.TabItem('Scale by'): - upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4) - with gr.TabItem('Scale to'): + with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): + upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") + with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): with gr.Group(): with gr.Row(): - upscaling_resize_w = gr.Number(label="Width", value=512, precision=0) - upscaling_resize_h = gr.Number(label="Height", value=512, precision=0) - upscaling_crop = gr.Checkbox(label='Crop to fit', value=True) + upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") + upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") + upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") with gr.Group(): extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") with gr.Group(): extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1) + extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") with gr.Group(): - gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan) + gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") with gr.Group(): - codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer) - codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer) + codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") + codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") with gr.Group(): - upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False) + upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) @@ -1117,7 +1117,7 @@ def create_ui(): with gr.Column(variant='panel'): html = gr.HTML() - generation_info = gr.Textbox(visible=False) + generation_info = gr.Textbox(visible=False, elem_id="pnginfo_generation_info") html2 = gr.HTML() with gr.Row(): buttons = parameters_copypaste.create_buttons(["txt2img", "img2img", "inpaint", "extras"]) @@ -1144,13 +1144,13 @@ def create_ui(): tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") - custom_name = gr.Textbox(label="Custom Name (Optional)") - interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3) - interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method") + custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") + interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") + interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") with gr.Row(): - checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format") - save_as_half = gr.Checkbox(value=False, label="Save as float16") + checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") + save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') @@ -1165,58 +1165,58 @@ def create_ui(): with gr.Tabs(elem_id="train_tabs"): with gr.Tab(label="Create embedding"): - new_embedding_name = gr.Textbox(label="Name") - initialization_text = gr.Textbox(label="Initialization text", value="*") - nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1) - overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding") + new_embedding_name = gr.Textbox(label="Name", elem_id="train_new_embedding_name") + initialization_text = gr.Textbox(label="Initialization text", value="*", elem_id="train_initialization_text") + nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1, elem_id="train_nvpt") + overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding", elem_id="train_overwrite_old_embedding") with gr.Row(): with gr.Column(scale=3): gr.HTML(value="") with gr.Column(): - create_embedding = gr.Button(value="Create embedding", variant='primary') + create_embedding = gr.Button(value="Create embedding", variant='primary', elem_id="train_create_embedding") with gr.Tab(label="Create hypernetwork"): - new_hypernetwork_name = gr.Textbox(label="Name") - new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"]) - new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'") - new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys) - new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"]) - new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization") - new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout") - overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork") + new_hypernetwork_name = gr.Textbox(label="Name", elem_id="train_new_hypernetwork_name") + new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"], elem_id="train_new_hypernetwork_sizes") + new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'", elem_id="train_new_hypernetwork_layer_structure") + new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys, elem_id="train_new_hypernetwork_activation_func") + new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") + new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") + new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") + overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") with gr.Row(): with gr.Column(scale=3): gr.HTML(value="") with gr.Column(): - create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary') + create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary', elem_id="train_create_hypernetwork") with gr.Tab(label="Preprocess images"): - process_src = gr.Textbox(label='Source directory') - process_dst = gr.Textbox(label='Destination directory') - process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512) - process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512) - preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"]) + process_src = gr.Textbox(label='Source directory', elem_id="train_process_src") + process_dst = gr.Textbox(label='Destination directory', elem_id="train_process_dst") + process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_process_width") + process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_process_height") + preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"], elem_id="train_preprocess_txt_action") with gr.Row(): - process_flip = gr.Checkbox(label='Create flipped copies') - process_split = gr.Checkbox(label='Split oversized images') - process_focal_crop = gr.Checkbox(label='Auto focal point crop') - process_caption = gr.Checkbox(label='Use BLIP for caption') - process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True) + process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") + process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") + process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") + process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") + process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") with gr.Row(visible=False) as process_split_extra_row: - process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05) - process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05) + process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_split_threshold") + process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05, elem_id="train_process_overlap_ratio") with gr.Row(visible=False) as process_focal_crop_row: - process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05) - process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05) - process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05) - process_focal_crop_debug = gr.Checkbox(label='Create debug image') + process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_face_weight") + process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") + process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") + process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") with gr.Row(): with gr.Column(scale=3): @@ -1224,8 +1224,8 @@ def create_ui(): with gr.Column(): with gr.Row(): - interrupt_preprocessing = gr.Button("Interrupt") - run_preprocess = gr.Button(value="Preprocess", variant='primary') + interrupt_preprocessing = gr.Button("Interrupt", elem_id="train_interrupt_preprocessing") + run_preprocess = gr.Button(value="Preprocess", variant='primary', elem_id="train_run_preprocess") process_split.change( fn=lambda show: gr_show(show), @@ -1248,31 +1248,31 @@ def create_ui(): train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") with gr.Row(): - embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005") - hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001") - - batch_size = gr.Number(label='Batch size', value=1, precision=0) - gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0) - dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images") - log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion") - template_file = gr.Textbox(label='Prompt template file', value=os.path.join(script_path, "textual_inversion_templates", "style_filewords.txt")) - training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512) - training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512) - steps = gr.Number(label='Max steps', value=100000, precision=0) - create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0) - save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0) - save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True) - preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False) + embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") + hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") + + batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") + gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") + dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") + log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") + template_file = gr.Textbox(label='Prompt template file', value=os.path.join(script_path, "textual_inversion_templates", "style_filewords.txt"), elem_id="train_template_file") + training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") + training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") + steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") + create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") + save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") + save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") + preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") with gr.Row(): - shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False) - tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0) + shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") + tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") with gr.Row(): - latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random']) + latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") with gr.Row(): - interrupt_training = gr.Button(value="Interrupt") - train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary') - train_embedding = gr.Button(value="Train Embedding", variant='primary') + interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") + train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") + train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") params = script_callbacks.UiTrainTabParams(txt2img_preview_params) @@ -1490,7 +1490,7 @@ def create_ui(): return gr.update(value=value), opts.dumpjson() with gr.Blocks(analytics_enabled=False) as settings_interface: - settings_submit = gr.Button(value="Apply settings", variant='primary') + settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") result = gr.HTML() settings_cols = 3 @@ -1541,8 +1541,8 @@ def create_ui(): download_localization = gr.Button(value='Download localization template', elem_id="download_localization") with gr.Row(): - reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary') - restart_gradio = gr.Button(value='Restart Gradio and Refresh components (Custom Scripts, ui.py, js and css only)', variant='primary') + reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") + restart_gradio = gr.Button(value='Restart Gradio and Refresh components (Custom Scripts, ui.py, js and css only)', variant='primary', elem_id="settings_restart_gradio") request_notifications.click( fn=lambda: None, -- cgit v1.2.3 From 311354c0bb8930ea939d6aa6b3edd50c69301320 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 2 Jan 2023 00:38:09 +0300 Subject: fix the issue with training on SD2.0 --- modules/sd_models.py | 2 ++ modules/textual_inversion/textual_inversion.py | 3 +-- 2 files changed, 3 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index ebd4dff7..bff8d6c9 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -228,6 +228,8 @@ def load_model_weights(model, checkpoint_info, vae_file="auto"): model.sd_model_checkpoint = checkpoint_file model.sd_checkpoint_info = checkpoint_info + model.logvar = model.logvar.to(devices.device) # fix for training + sd_vae.delete_base_vae() sd_vae.clear_loaded_vae() vae_file = sd_vae.resolve_vae(checkpoint_file, vae_file=vae_file) diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 66f40367..1e5722e7 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -282,7 +282,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ return embedding, filename scheduler = LearnRateScheduler(learn_rate, steps, initial_step) - # dataset loading may take a while, so input validations and early returns should be done before this + # dataset loading may take a while, so input validations and early returns should be done before this shared.state.textinfo = f"Preparing dataset from {html.escape(data_root)}..." old_parallel_processing_allowed = shared.parallel_processing_allowed @@ -310,7 +310,6 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ loss_step = 0 _loss_step = 0 #internal - last_saved_file = "" last_saved_image = "" forced_filename = "" -- cgit v1.2.3 From b5819d9bf1794071139c640b5f1e72c84a0e051a Mon Sep 17 00:00:00 2001 From: Philpax Date: Mon, 2 Jan 2023 10:17:33 +1100 Subject: feat(api): add /sdapi/v1/embeddings --- modules/api/api.py | 8 ++++++++ modules/api/models.py | 3 +++ 2 files changed, 11 insertions(+) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 11daff0d..30bf3dac 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -100,6 +100,7 @@ class Api: self.add_api_route("/sdapi/v1/prompt-styles", self.get_prompt_styles, methods=["GET"], response_model=List[PromptStyleItem]) self.add_api_route("/sdapi/v1/artist-categories", self.get_artists_categories, methods=["GET"], response_model=List[str]) self.add_api_route("/sdapi/v1/artists", self.get_artists, methods=["GET"], response_model=List[ArtistItem]) + self.add_api_route("/sdapi/v1/embeddings", self.get_embeddings, methods=["GET"], response_model=EmbeddingsResponse) self.add_api_route("/sdapi/v1/refresh-checkpoints", self.refresh_checkpoints, methods=["POST"]) self.add_api_route("/sdapi/v1/create/embedding", self.create_embedding, methods=["POST"], response_model=CreateResponse) self.add_api_route("/sdapi/v1/create/hypernetwork", self.create_hypernetwork, methods=["POST"], response_model=CreateResponse) @@ -327,6 +328,13 @@ class Api: def get_artists(self): return [{"name":x[0], "score":x[1], "category":x[2]} for x in shared.artist_db.artists] + def get_embeddings(self): + db = sd_hijack.model_hijack.embedding_db + return { + "loaded": sorted(db.word_embeddings.keys()), + "skipped": sorted(db.skipped_embeddings), + } + def refresh_checkpoints(self): shared.refresh_checkpoints() diff --git a/modules/api/models.py b/modules/api/models.py index c446ce7a..a8472dc9 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -249,3 +249,6 @@ class ArtistItem(BaseModel): score: float = Field(title="Score") category: str = Field(title="Category") +class EmbeddingsResponse(BaseModel): + loaded: List[str] = Field(title="Loaded", description="Embeddings loaded for the current model") + skipped: List[str] = Field(title="Skipped", description="Embeddings skipped for the current model (likely due to architecture incompatibility)") \ No newline at end of file -- cgit v1.2.3 From c65909ad16a1962129114c6251de092f49479b06 Mon Sep 17 00:00:00 2001 From: Philpax Date: Mon, 2 Jan 2023 12:21:22 +1100 Subject: feat(api): return more data for embeddings --- modules/api/api.py | 17 +++++++++++++++-- modules/api/models.py | 11 +++++++++-- modules/textual_inversion/textual_inversion.py | 8 ++++---- 3 files changed, 28 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 30bf3dac..9c670f00 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -330,9 +330,22 @@ class Api: def get_embeddings(self): db = sd_hijack.model_hijack.embedding_db + + def convert_embedding(embedding): + return { + "step": embedding.step, + "sd_checkpoint": embedding.sd_checkpoint, + "sd_checkpoint_name": embedding.sd_checkpoint_name, + "shape": embedding.shape, + "vectors": embedding.vectors, + } + + def convert_embeddings(embeddings): + return {embedding.name: convert_embedding(embedding) for embedding in embeddings.values()} + return { - "loaded": sorted(db.word_embeddings.keys()), - "skipped": sorted(db.skipped_embeddings), + "loaded": convert_embeddings(db.word_embeddings), + "skipped": convert_embeddings(db.skipped_embeddings), } def refresh_checkpoints(self): diff --git a/modules/api/models.py b/modules/api/models.py index a8472dc9..4a632c68 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -249,6 +249,13 @@ class ArtistItem(BaseModel): score: float = Field(title="Score") category: str = Field(title="Category") +class EmbeddingItem(BaseModel): + step: Optional[int] = Field(title="Step", description="The number of steps that were used to train this embedding, if available") + sd_checkpoint: Optional[str] = Field(title="SD Checkpoint", description="The hash of the checkpoint this embedding was trained on, if available") + sd_checkpoint_name: Optional[str] = Field(title="SD Checkpoint Name", description="The name of the checkpoint this embedding was trained on, if available. Note that this is the name that was used by the trainer; for a stable identifier, use `sd_checkpoint` instead") + shape: int = Field(title="Shape", description="The length of each individual vector in the embedding") + vectors: int = Field(title="Vectors", description="The number of vectors in the embedding") + class EmbeddingsResponse(BaseModel): - loaded: List[str] = Field(title="Loaded", description="Embeddings loaded for the current model") - skipped: List[str] = Field(title="Skipped", description="Embeddings skipped for the current model (likely due to architecture incompatibility)") \ No newline at end of file + loaded: Dict[str, EmbeddingItem] = Field(title="Loaded", description="Embeddings loaded for the current model") + skipped: Dict[str, EmbeddingItem] = Field(title="Skipped", description="Embeddings skipped for the current model (likely due to architecture incompatibility)") \ No newline at end of file diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 1e5722e7..fd253477 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -59,7 +59,7 @@ class EmbeddingDatabase: def __init__(self, embeddings_dir): self.ids_lookup = {} self.word_embeddings = {} - self.skipped_embeddings = [] + self.skipped_embeddings = {} self.dir_mtime = None self.embeddings_dir = embeddings_dir self.expected_shape = -1 @@ -91,7 +91,7 @@ class EmbeddingDatabase: self.dir_mtime = mt self.ids_lookup.clear() self.word_embeddings.clear() - self.skipped_embeddings = [] + self.skipped_embeddings.clear() self.expected_shape = self.get_expected_shape() def process_file(path, filename): @@ -136,7 +136,7 @@ class EmbeddingDatabase: if self.expected_shape == -1 or self.expected_shape == embedding.shape: self.register_embedding(embedding, shared.sd_model) else: - self.skipped_embeddings.append(name) + self.skipped_embeddings[name] = embedding for fn in os.listdir(self.embeddings_dir): try: @@ -153,7 +153,7 @@ class EmbeddingDatabase: print(f"Textual inversion embeddings loaded({len(self.word_embeddings)}): {', '.join(self.word_embeddings.keys())}") if len(self.skipped_embeddings) > 0: - print(f"Textual inversion embeddings skipped({len(self.skipped_embeddings)}): {', '.join(self.skipped_embeddings)}") + print(f"Textual inversion embeddings skipped({len(self.skipped_embeddings)}): {', '.join(self.skipped_embeddings.keys())}") def find_embedding_at_position(self, tokens, offset): token = tokens[offset] -- cgit v1.2.3 From ef27a18b6b7cb1a8eebdc9b2e88d25baf2c2414d Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 2 Jan 2023 19:42:10 +0300 Subject: Hires fix rework --- modules/generation_parameters_copypaste.py | 32 ++++++++++++++ modules/images.py | 24 +++++++++-- modules/processing.py | 68 ++++++++++++------------------ modules/shared.py | 7 ++- modules/txt2img.py | 6 +-- modules/ui.py | 15 +++---- 6 files changed, 95 insertions(+), 57 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 8e7f0df0..d6fa822b 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -1,5 +1,6 @@ import base64 import io +import math import os import re from pathlib import Path @@ -164,6 +165,35 @@ def find_hypernetwork_key(hypernet_name, hypernet_hash=None): return None +def restore_old_hires_fix_params(res): + """for infotexts that specify old First pass size parameter, convert it into + width, height, and hr scale""" + + firstpass_width = res.get('First pass size-1', None) + firstpass_height = res.get('First pass size-2', None) + + if firstpass_width is None or firstpass_height is None: + return + + firstpass_width, firstpass_height = int(firstpass_width), int(firstpass_height) + width = int(res.get("Size-1", 512)) + height = int(res.get("Size-2", 512)) + + if firstpass_width == 0 or firstpass_height == 0: + # old algorithm for auto-calculating first pass size + desired_pixel_count = 512 * 512 + actual_pixel_count = width * height + scale = math.sqrt(desired_pixel_count / actual_pixel_count) + firstpass_width = math.ceil(scale * width / 64) * 64 + firstpass_height = math.ceil(scale * height / 64) * 64 + + hr_scale = width / firstpass_width if firstpass_width > 0 else height / firstpass_height + + res['Size-1'] = firstpass_width + res['Size-2'] = firstpass_height + res['Hires upscale'] = hr_scale + + def parse_generation_parameters(x: str): """parses generation parameters string, the one you see in text field under the picture in UI: ``` @@ -221,6 +251,8 @@ Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 965400086, Size: 512x512, Model hypernet_hash = res.get("Hypernet hash", None) res["Hypernet"] = find_hypernetwork_key(hypernet_name, hypernet_hash) + restore_old_hires_fix_params(res) + return res diff --git a/modules/images.py b/modules/images.py index f84fd485..c3a5fc8b 100644 --- a/modules/images.py +++ b/modules/images.py @@ -230,16 +230,32 @@ def draw_prompt_matrix(im, width, height, all_prompts): return draw_grid_annotations(im, width, height, hor_texts, ver_texts) -def resize_image(resize_mode, im, width, height): +def resize_image(resize_mode, im, width, height, upscaler_name=None): + """ + Resizes an image with the specified resize_mode, width, and height. + + Args: + resize_mode: The mode to use when resizing the image. + 0: Resize the image to the specified width and height. + 1: Resize the image to fill the specified width and height, maintaining the aspect ratio, and then center the image within the dimensions, cropping the excess. + 2: Resize the image to fit within the specified width and height, maintaining the aspect ratio, and then center the image within the dimensions, filling empty with data from image. + im: The image to resize. + width: The width to resize the image to. + height: The height to resize the image to. + upscaler_name: The name of the upscaler to use. If not provided, defaults to opts.upscaler_for_img2img. + """ + + upscaler_name = upscaler_name or opts.upscaler_for_img2img + def resize(im, w, h): - if opts.upscaler_for_img2img is None or opts.upscaler_for_img2img == "None" or im.mode == 'L': + if upscaler_name is None or upscaler_name == "None" or im.mode == 'L': return im.resize((w, h), resample=LANCZOS) scale = max(w / im.width, h / im.height) if scale > 1.0: - upscalers = [x for x in shared.sd_upscalers if x.name == opts.upscaler_for_img2img] - assert len(upscalers) > 0, f"could not find upscaler named {opts.upscaler_for_img2img}" + upscalers = [x for x in shared.sd_upscalers if x.name == upscaler_name] + assert len(upscalers) > 0, f"could not find upscaler named {upscaler_name}" upscaler = upscalers[0] im = upscaler.scaler.upscale(im, scale, upscaler.data_path) diff --git a/modules/processing.py b/modules/processing.py index 42dc19ea..4654570c 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -658,14 +658,18 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): sampler = None - def __init__(self, enable_hr: bool=False, denoising_strength: float=0.75, firstphase_width: int=0, firstphase_height: int=0, **kwargs): + def __init__(self, enable_hr: bool = False, denoising_strength: float = 0.75, firstphase_width: int = 0, firstphase_height: int = 0, hr_scale: float = 2.0, hr_upscaler: str = None, **kwargs): super().__init__(**kwargs) self.enable_hr = enable_hr self.denoising_strength = denoising_strength - self.firstphase_width = firstphase_width - self.firstphase_height = firstphase_height - self.truncate_x = 0 - self.truncate_y = 0 + self.hr_scale = hr_scale + self.hr_upscaler = hr_upscaler + + if firstphase_width != 0 or firstphase_height != 0: + print("firstphase_width/firstphase_height no longer supported; use hr_scale", file=sys.stderr) + self.hr_scale = self.width / firstphase_width + self.width = firstphase_width + self.height = firstphase_height def init(self, all_prompts, all_seeds, all_subseeds): if self.enable_hr: @@ -674,47 +678,29 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): else: state.job_count = state.job_count * 2 - self.extra_generation_params["First pass size"] = f"{self.firstphase_width}x{self.firstphase_height}" - - if self.firstphase_width == 0 or self.firstphase_height == 0: - desired_pixel_count = 512 * 512 - actual_pixel_count = self.width * self.height - scale = math.sqrt(desired_pixel_count / actual_pixel_count) - self.firstphase_width = math.ceil(scale * self.width / 64) * 64 - self.firstphase_height = math.ceil(scale * self.height / 64) * 64 - firstphase_width_truncated = int(scale * self.width) - firstphase_height_truncated = int(scale * self.height) - - else: - - width_ratio = self.width / self.firstphase_width - height_ratio = self.height / self.firstphase_height - - if width_ratio > height_ratio: - firstphase_width_truncated = self.firstphase_width - firstphase_height_truncated = self.firstphase_width * self.height / self.width - else: - firstphase_width_truncated = self.firstphase_height * self.width / self.height - firstphase_height_truncated = self.firstphase_height - - self.truncate_x = int(self.firstphase_width - firstphase_width_truncated) // opt_f - self.truncate_y = int(self.firstphase_height - firstphase_height_truncated) // opt_f + self.extra_generation_params["Hires upscale"] = self.hr_scale + if self.hr_upscaler is not None: + self.extra_generation_params["Hires upscaler"] = self.hr_upscaler def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength, prompts): self.sampler = sd_samplers.create_sampler(self.sampler_name, self.sd_model) + latent_scale_mode = shared.latent_upscale_modes.get(self.hr_upscaler, None) if self.hr_upscaler is not None else shared.latent_upscale_default_mode + if self.enable_hr and latent_scale_mode is None: + assert len([x for x in shared.sd_upscalers if x.name == self.hr_upscaler]) > 0, f"could not find upscaler named {self.hr_upscaler}" + + x = create_random_tensors([opt_C, self.height // opt_f, self.width // opt_f], seeds=seeds, subseeds=subseeds, subseed_strength=self.subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self) + samples = self.sampler.sample(self, x, conditioning, unconditional_conditioning, image_conditioning=self.txt2img_image_conditioning(x)) + if not self.enable_hr: - x = create_random_tensors([opt_C, self.height // opt_f, self.width // opt_f], seeds=seeds, subseeds=subseeds, subseed_strength=self.subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self) - samples = self.sampler.sample(self, x, conditioning, unconditional_conditioning, image_conditioning=self.txt2img_image_conditioning(x)) return samples - x = create_random_tensors([opt_C, self.firstphase_height // opt_f, self.firstphase_width // opt_f], seeds=seeds, subseeds=subseeds, subseed_strength=self.subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self) - samples = self.sampler.sample(self, x, conditioning, unconditional_conditioning, image_conditioning=self.txt2img_image_conditioning(x, self.firstphase_width, self.firstphase_height)) - - samples = samples[:, :, self.truncate_y//2:samples.shape[2]-self.truncate_y//2, self.truncate_x//2:samples.shape[3]-self.truncate_x//2] + target_width = int(self.width * self.hr_scale) + target_height = int(self.height * self.hr_scale) - """saves image before applying hires fix, if enabled in options; takes as an argument either an image or batch with latent space images""" def save_intermediate(image, index): + """saves image before applying hires fix, if enabled in options; takes as an argument either an image or batch with latent space images""" + if not opts.save or self.do_not_save_samples or not opts.save_images_before_highres_fix: return @@ -723,11 +709,11 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): images.save_image(image, self.outpath_samples, "", seeds[index], prompts[index], opts.samples_format, suffix="-before-highres-fix") - if opts.use_scale_latent_for_hires_fix: + if latent_scale_mode is not None: for i in range(samples.shape[0]): save_intermediate(samples, i) - samples = torch.nn.functional.interpolate(samples, size=(self.height // opt_f, self.width // opt_f), mode="bilinear") + samples = torch.nn.functional.interpolate(samples, size=(target_height // opt_f, target_width // opt_f), mode=latent_scale_mode) # Avoid making the inpainting conditioning unless necessary as # this does need some extra compute to decode / encode the image again. @@ -747,7 +733,7 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): save_intermediate(image, i) - image = images.resize_image(0, image, self.width, self.height) + image = images.resize_image(0, image, target_width, target_height, upscaler_name=self.hr_upscaler) image = np.array(image).astype(np.float32) / 255.0 image = np.moveaxis(image, 2, 0) batch_images.append(image) @@ -764,7 +750,7 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): self.sampler = sd_samplers.create_sampler(self.sampler_name, self.sd_model) - noise = create_random_tensors(samples.shape[1:], seeds=seeds, subseeds=subseeds, subseed_strength=subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self) + noise = create_random_tensors(samples.shape[1:], seeds=seeds, subseeds=subseeds, subseed_strength=subseed_strength, p=self) # GC now before running the next img2img to prevent running out of memory x = None diff --git a/modules/shared.py b/modules/shared.py index 7f430b93..b65559ee 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -327,7 +327,6 @@ options_templates.update(options_section(('upscaling', "Upscaling"), { "ESRGAN_tile_overlap": OptionInfo(8, "Tile overlap, in pixels for ESRGAN upscalers. Low values = visible seam.", gr.Slider, {"minimum": 0, "maximum": 48, "step": 1}), "realesrgan_enabled_models": OptionInfo(["R-ESRGAN 4x+", "R-ESRGAN 4x+ Anime6B"], "Select which Real-ESRGAN models to show in the web UI. (Requires restart)", gr.CheckboxGroup, lambda: {"choices": realesrgan_models_names()}), "upscaler_for_img2img": OptionInfo(None, "Upscaler for img2img", gr.Dropdown, lambda: {"choices": [x.name for x in sd_upscalers]}), - "use_scale_latent_for_hires_fix": OptionInfo(False, "Upscale latent space image when doing hires. fix"), })) options_templates.update(options_section(('face-restoration', "Face restoration"), { @@ -545,6 +544,12 @@ opts = Options() if os.path.exists(config_filename): opts.load(config_filename) +latent_upscale_default_mode = "Latent" +latent_upscale_modes = { + "Latent": "bilinear", + "Latent (nearest)": "nearest", +} + sd_upscalers = [] sd_model = None diff --git a/modules/txt2img.py b/modules/txt2img.py index 7f61e19a..e189a899 100644 --- a/modules/txt2img.py +++ b/modules/txt2img.py @@ -8,7 +8,7 @@ import modules.processing as processing from modules.ui import plaintext_to_html -def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, firstphase_width: int, firstphase_height: int, *args): +def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, *args): p = StableDiffusionProcessingTxt2Img( sd_model=shared.sd_model, outpath_samples=opts.outdir_samples or opts.outdir_txt2img_samples, @@ -33,8 +33,8 @@ def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: tiling=tiling, enable_hr=enable_hr, denoising_strength=denoising_strength if enable_hr else None, - firstphase_width=firstphase_width if enable_hr else None, - firstphase_height=firstphase_height if enable_hr else None, + hr_scale=hr_scale, + hr_upscaler=hr_upscaler, ) p.scripts = modules.scripts.scripts_txt2img diff --git a/modules/ui.py b/modules/ui.py index 7070ea15..27cd9ddd 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -684,11 +684,11 @@ def create_ui(): with gr.Row(): restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") - enable_hr = gr.Checkbox(label='Highres. fix', value=False, elem_id="txt2img_enable_hr") + enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") with gr.Row(visible=False) as hr_options: - firstphase_width = gr.Slider(minimum=0, maximum=1024, step=8, label="Firstpass width", value=0, elem_id="txt2img_firstphase_width") - firstphase_height = gr.Slider(minimum=0, maximum=1024, step=8, label="Firstpass height", value=0, elem_id="txt2img_firstphase_height") + hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) + hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") with gr.Row(equal_height=True): @@ -729,8 +729,8 @@ def create_ui(): width, enable_hr, denoising_strength, - firstphase_width, - firstphase_height, + hr_scale, + hr_upscaler, ] + custom_inputs, outputs=[ @@ -762,7 +762,6 @@ def create_ui(): outputs=[hr_options], ) - txt2img_paste_fields = [ (txt2img_prompt, "Prompt"), (txt2img_negative_prompt, "Negative prompt"), @@ -781,8 +780,8 @@ def create_ui(): (denoising_strength, "Denoising strength"), (enable_hr, lambda d: "Denoising strength" in d), (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), - (firstphase_width, "First pass size-1"), - (firstphase_height, "First pass size-2"), + (hr_scale, "Hires upscale"), + (hr_upscaler, "Hires upscaler"), *modules.scripts.scripts_txt2img.infotext_fields ] parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) -- cgit v1.2.3 From 4dbde228ff48dbb105241b1ed25c21ce3f87d182 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 2 Jan 2023 20:01:16 +0300 Subject: make it possible to use fractional values for SD upscale. --- modules/upscaler.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/upscaler.py b/modules/upscaler.py index c4e6e6bd..231680cb 100644 --- a/modules/upscaler.py +++ b/modules/upscaler.py @@ -53,10 +53,10 @@ class Upscaler: def do_upscale(self, img: PIL.Image, selected_model: str): return img - def upscale(self, img: PIL.Image, scale: int, selected_model: str = None): + def upscale(self, img: PIL.Image, scale, selected_model: str = None): self.scale = scale - dest_w = img.width * scale - dest_h = img.height * scale + dest_w = int(img.width * scale) + dest_h = int(img.height * scale) for i in range(3): shape = (img.width, img.height) -- cgit v1.2.3 From 84dd7e8e2495c4fc2997e97f8267aa831eb90d11 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 2 Jan 2023 20:30:02 +0300 Subject: error out with a readable message in chwewckpoint merger for incompatible tensor shapes (ie when trying to merge SD1.5 with SD2.0) --- modules/extras.py | 2 ++ modules/ui.py | 2 +- 2 files changed, 3 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 68939dea..5e270250 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -303,6 +303,8 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) result_is_inpainting_model = True else: + assert a.shape == b.shape, f'Incompatible shapes for layer {key}: A is {a.shape}, and B is {b.shape}' + theta_0[key] = theta_func2(a, b, multiplier) if save_as_half: diff --git a/modules/ui.py b/modules/ui.py index 27cd9ddd..67a51888 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1663,7 +1663,7 @@ def create_ui(): print("Error loading/saving model file:", file=sys.stderr) print(traceback.format_exc(), file=sys.stderr) modules.sd_models.list_models() # to remove the potentially missing models from the list - return ["Error loading/saving model file. It doesn't exist or the name contains illegal characters"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(3)] + return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] return results modelmerger_merge.click( -- cgit v1.2.3 From 8d12a729b8b036cb765cf2d87576d5ae256135c8 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 2 Jan 2023 20:46:51 +0300 Subject: fix possible error with accessing nonexistent setting --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 67a51888..9350a80f 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -491,7 +491,7 @@ def apply_setting(key, value): return valtype = type(opts.data_labels[key].default) - oldval = opts.data[key] + oldval = opts.data.get(key, None) opts.data[key] = valtype(value) if valtype != type(None) else value if oldval != value and opts.data_labels[key].onchange is not None: opts.data_labels[key].onchange() -- cgit v1.2.3 From 251ecee6949c36e9df1d99a950b3e1af2b5fa2b6 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 2 Jan 2023 22:44:46 +0300 Subject: make "send to" buttons send actual dimension of the sent image rather than fields --- modules/generation_parameters_copypaste.py | 58 ++++++++++++++++++++---------- 1 file changed, 40 insertions(+), 18 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index d6fa822b..ec60319a 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -103,35 +103,57 @@ def bind_buttons(buttons, send_image, send_generate_info): bind_list.append([buttons, send_image, send_generate_info]) +def send_image_and_dimensions(x): + if isinstance(x, Image.Image): + img = x + else: + img = image_from_url_text(x) + + if shared.opts.send_size and isinstance(img, Image.Image): + w = img.width + h = img.height + else: + w = gr.update() + h = gr.update() + + return img, w, h + + def run_bind(): - for buttons, send_image, send_generate_info in bind_list: + for buttons, source_image_component, send_generate_info in bind_list: for tab in buttons: button = buttons[tab] - if send_image and paste_fields[tab]["init_img"]: - if type(send_image) == gr.Gallery: - button.click( - fn=lambda x: image_from_url_text(x), - _js="extract_image_from_gallery", - inputs=[send_image], - outputs=[paste_fields[tab]["init_img"]], - ) + destination_image_component = paste_fields[tab]["init_img"] + fields = paste_fields[tab]["fields"] + + destination_width_component = next(iter([field for field, name in fields if name == "Size-1"] if fields else []), None) + destination_height_component = next(iter([field for field, name in fields if name == "Size-2"] if fields else []), None) + + if source_image_component and destination_image_component: + if isinstance(source_image_component, gr.Gallery): + func = send_image_and_dimensions if destination_width_component else image_from_url_text + jsfunc = "extract_image_from_gallery" else: - button.click( - fn=lambda x: x, - inputs=[send_image], - outputs=[paste_fields[tab]["init_img"]], - ) + func = send_image_and_dimensions if destination_width_component else lambda x: x + jsfunc = None + + button.click( + fn=func, + _js=jsfunc, + inputs=[source_image_component], + outputs=[destination_image_component, destination_width_component, destination_height_component] if destination_width_component else [destination_image_component], + ) - if send_generate_info and paste_fields[tab]["fields"] is not None: + if send_generate_info and fields is not None: if send_generate_info in paste_fields: - paste_field_names = ['Prompt', 'Negative prompt', 'Steps', 'Face restoration'] + (['Size-1', 'Size-2'] if shared.opts.send_size else []) + (["Seed"] if shared.opts.send_seed else []) + paste_field_names = ['Prompt', 'Negative prompt', 'Steps', 'Face restoration'] + (["Seed"] if shared.opts.send_seed else []) button.click( fn=lambda *x: x, inputs=[field for field, name in paste_fields[send_generate_info]["fields"] if name in paste_field_names], - outputs=[field for field, name in paste_fields[tab]["fields"] if name in paste_field_names], + outputs=[field for field, name in fields if name in paste_field_names], ) else: - connect_paste(button, paste_fields[tab]["fields"], send_generate_info) + connect_paste(button, fields, send_generate_info) button.click( fn=None, -- cgit v1.2.3 From 269f6e867651cadef40d2c939a79d13291280bcd Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 07:20:20 +0300 Subject: change settings UI to use vertical tabs --- modules/ui.py | 45 +++++++++++++++++---------------------------- 1 file changed, 17 insertions(+), 28 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 9350a80f..f8c973ba 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1489,41 +1489,34 @@ def create_ui(): return gr.update(value=value), opts.dumpjson() with gr.Blocks(analytics_enabled=False) as settings_interface: - settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") - result = gr.HTML() + with gr.Row(): + settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") + restart_gradio = gr.Button(value='Restart UI', variant='primary', elem_id="settings_restart_gradio") - settings_cols = 3 - items_per_col = int(len(opts.data_labels) * 0.9 / settings_cols) + result = gr.HTML(elem_id="settings_result") quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] quicksettings_names = set(x for x in quicksettings_names if x != 'quicksettings') quicksettings_list = [] - cols_displayed = 0 - items_displayed = 0 previous_section = None - column = None - with gr.Row(elem_id="settings").style(equal_height=False): + current_tab = None + with gr.Tabs(elem_id="settings"): for i, (k, item) in enumerate(opts.data_labels.items()): section_must_be_skipped = item.section[0] is None if previous_section != item.section and not section_must_be_skipped: - if cols_displayed < settings_cols and (items_displayed >= items_per_col or previous_section is None): - if column is not None: - column.__exit__() + elem_id, text = item.section - column = gr.Column(variant='panel') - column.__enter__() + if current_tab is not None: + current_tab.__exit__() - items_displayed = 0 - cols_displayed += 1 + current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) + current_tab.__enter__() previous_section = item.section - elem_id, text = item.section - gr.HTML(elem_id="settings_header_text_{}".format(elem_id), value='

{}

'.format(text)) - if k in quicksettings_names and not shared.cmd_opts.freeze_settings: quicksettings_list.append((i, k, item)) components.append(dummy_component) @@ -1533,15 +1526,14 @@ def create_ui(): component = create_setting_component(k) component_dict[k] = component components.append(component) - items_displayed += 1 - with gr.Row(): - request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") - download_localization = gr.Button(value='Download localization template', elem_id="download_localization") + if current_tab is not None: + current_tab.__exit__() - with gr.Row(): - reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") - restart_gradio = gr.Button(value='Restart Gradio and Refresh components (Custom Scripts, ui.py, js and css only)', variant='primary', elem_id="settings_restart_gradio") + with gr.TabItem("Actions"): + request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") + download_localization = gr.Button(value='Download localization template', elem_id="download_localization") + reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") request_notifications.click( fn=lambda: None, @@ -1578,9 +1570,6 @@ def create_ui(): outputs=[], ) - if column is not None: - column.__exit__() - interfaces = [ (txt2img_interface, "txt2img", "txt2img"), (img2img_interface, "img2img", "img2img"), -- cgit v1.2.3 From 18c03cdeac6272734b0c09afd3fbe47d1372dd07 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 09:04:29 +0300 Subject: styling rework to make things more compact --- modules/ui.py | 121 ++++++++++++++++++++++++----------------------- modules/ui_components.py | 7 +++ 2 files changed, 68 insertions(+), 60 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index f8c973ba..f787b518 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -19,7 +19,8 @@ import numpy as np from PIL import Image, PngImagePlugin from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, ui_components +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru +from modules.ui_components import FormRow, FormGroup, ToolButton from modules.paths import script_path from modules.shared import opts, cmd_opts, restricted_opts @@ -273,31 +274,27 @@ def interrogate_deepbooru(image): def create_seed_inputs(target_interface): - with gr.Row(): - with gr.Box(): - with gr.Row(elem_id=target_interface + '_seed_row'): - seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') - seed.style(container=False) - random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') - reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') - - with gr.Box(elem_id=target_interface + '_subseed_show_box'): + with FormRow(elem_id=target_interface + '_seed_row'): + seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') + seed.style(container=False) + random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') + reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') + + with gr.Group(elem_id=target_interface + '_subseed_show_box'): seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) # Components to show/hide based on the 'Extra' checkbox seed_extras = [] - with gr.Row(visible=False) as seed_extra_row_1: + with FormRow(visible=False, elem_id=target_interface + '_subseed_row') as seed_extra_row_1: seed_extras.append(seed_extra_row_1) - with gr.Box(): - with gr.Row(elem_id=target_interface + '_subseed_row'): - subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') - subseed.style(container=False) - random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') - reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') + subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') + subseed.style(container=False) + random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') + reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') - with gr.Row(visible=False) as seed_extra_row_2: + with FormRow(visible=False) as seed_extra_row_2: seed_extras.append(seed_extra_row_2) seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') @@ -523,7 +520,7 @@ def create_refresh_button(refresh_component, refresh_method, refreshed_args, ele return gr.update(**(args or {})) - refresh_button = ui_components.ToolButton(value=refresh_symbol, elem_id=elem_id) + refresh_button = ToolButton(value=refresh_symbol, elem_id=elem_id) refresh_button.click( fn=refresh, inputs=[], @@ -636,11 +633,11 @@ Requested path was: {f} def create_sampler_and_steps_selection(choices, tabname): if opts.samplers_in_dropdown: - with gr.Row(elem_id=f"sampler_selection_{tabname}"): + with FormRow(elem_id=f"sampler_selection_{tabname}"): sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling Steps", value=20) else: - with gr.Group(elem_id=f"sampler_selection_{tabname}"): + with FormGroup(elem_id=f"sampler_selection_{tabname}"): steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling Steps", value=20) sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") @@ -677,29 +674,29 @@ def create_ui(): with gr.Column(variant='panel', elem_id="txt2img_settings"): steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") - with gr.Group(): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") + with FormRow(): + with gr.Column(elem_id="txt2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") + with gr.Column(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - with gr.Row(): + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") + + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') + + with FormRow(elem_id="txt2img_checkboxes"): restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") - with gr.Row(visible=False) as hr_options: + with FormRow(visible=False) as hr_options: hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") - with gr.Row(equal_height=True): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") - - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') - - with gr.Group(elem_id="txt2img_script_container"): + with FormGroup(elem_id="txt2img_script_container"): custom_inputs = modules.scripts.scripts_txt2img.setup_ui() txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) @@ -816,7 +813,7 @@ def create_ui(): img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) setup_progressbar(progressbar, img2img_preview, 'img2img') - with gr.Row().style(equal_height=False): + with FormRow().style(equal_height=False): with gr.Column(variant='panel', elem_id="img2img_settings"): with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: @@ -841,19 +838,23 @@ def create_ui(): init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") - with gr.Row(): + with FormRow(): mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") - with gr.Row(): - mask_mode = gr.Radio(label="Mask mode", show_label=False, choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") - inpainting_mask_invert = gr.Radio(label='Masking mode', show_label=False, choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") + with FormRow(): + mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") + inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") - inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") + with FormRow(): + inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") - with gr.Row(): - inpaint_full_res = gr.Checkbox(label='Inpaint at full resolution', value=False, elem_id="img2img_inpaint_full_res") - inpaint_full_res_padding = gr.Slider(label='Inpaint at full resolution padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") + with FormRow(): + with gr.Column(): + inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") + + with gr.Column(scale=4): + inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' @@ -861,30 +862,30 @@ def create_ui(): img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") - with gr.Row(): - resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", show_label=False, choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") + with FormRow(): + resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") - with gr.Group(): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") - - with gr.Row(): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") + with FormRow(): + with gr.Column(elem_id="img2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") + with gr.Column(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - with gr.Row(): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - with gr.Group(): + with FormGroup(): cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') - with gr.Group(elem_id="img2img_script_container"): + with FormRow(elem_id="img2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") + + with FormGroup(elem_id="img2img_script_container"): custom_inputs = modules.scripts.scripts_img2img.setup_ui() img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) @@ -1444,7 +1445,7 @@ def create_ui(): res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) else: - with ui_components.FormRow(): + with FormRow(): res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) else: diff --git a/modules/ui_components.py b/modules/ui_components.py index d0519d2d..91eb0e3d 100644 --- a/modules/ui_components.py +++ b/modules/ui_components.py @@ -16,3 +16,10 @@ class FormRow(gr.Row, gr.components.FormComponent): def get_block_name(self): return "row" + + +class FormGroup(gr.Group, gr.components.FormComponent): + """Same as gr.Row but fits inside gradio forms""" + + def get_block_name(self): + return "group" -- cgit v1.2.3 From 2bc86712ec16cada01a2353f1d978c1aabc84dbb Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 09:13:35 +0300 Subject: make quicksettings UI elements appear in same order as they are listed in the setting --- modules/ui.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index f787b518..d7b911da 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1497,7 +1497,7 @@ def create_ui(): result = gr.HTML(elem_id="settings_result") quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] - quicksettings_names = set(x for x in quicksettings_names if x != 'quicksettings') + quicksettings_names = {x: i for i, x in enumerate(quicksettings_names) if x != 'quicksettings'} quicksettings_list = [] @@ -1604,7 +1604,7 @@ def create_ui(): with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: with gr.Row(elem_id="quicksettings"): - for i, k, item in quicksettings_list: + for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): component = create_setting_component(k, is_quicksettings=True) component_dict[k] = component -- cgit v1.2.3 From 9d4eff097deff6153c4023f158bd9fbd4f3e88b3 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 10:01:06 +0300 Subject: add a button to show all setting pages --- modules/ui.py | 2 ++ 1 file changed, 2 insertions(+) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index d7b911da..2c92c422 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1536,6 +1536,8 @@ def create_ui(): download_localization = gr.Button(value='Download localization template', elem_id="download_localization") reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") + gr.Button(value="Show all pages", elem_id="settings_show_all_pages") + request_notifications.click( fn=lambda: None, inputs=[], -- cgit v1.2.3 From a1cf55a9d1c82f8e56c00d549bca5c8fa069f412 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 10:39:21 +0300 Subject: add option to reorder items in main UI --- modules/shared.py | 13 ++++++ modules/ui.py | 130 +++++++++++++++++++++++++++++++++++------------------- 2 files changed, 97 insertions(+), 46 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index b65559ee..23657a93 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -109,6 +109,17 @@ restricted_opts = { "outdir_save", } +ui_reorder_categories = [ + "sampler", + "dimensions", + "cfg", + "seed", + "checkboxes", + "hires_fix", + "batch", + "scripts", +] + cmd_opts.disable_extension_access = (cmd_opts.share or cmd_opts.listen or cmd_opts.server_name) and not cmd_opts.enable_insecure_extension_access devices.device, devices.device_interrogate, devices.device_gfpgan, devices.device_esrgan, devices.device_codeformer = \ @@ -410,7 +421,9 @@ options_templates.update(options_section(('ui', "User interface"), { "js_modal_lightbox_initially_zoomed": OptionInfo(True, "Show images zoomed in by default in full page image viewer"), "show_progress_in_title": OptionInfo(True, "Show generation progress in window title."), "samplers_in_dropdown": OptionInfo(True, "Use dropdown for sampler selection instead of radio group"), + "dimensions_and_batch_together": OptionInfo(True, "Show Witdth/Height and Batch sliders in same row"), 'quicksettings': OptionInfo("sd_model_checkpoint", "Quicksettings list"), + 'ui_reorder': OptionInfo(", ".join(ui_reorder_categories), "txt2img/ing2img UI item order"), 'localization': OptionInfo("None", "Localization (requires restart)", gr.Dropdown, lambda: {"choices": ["None"] + list(localization.localizations.keys())}, refresh=lambda: localization.list_localizations(cmd_opts.localizations_dir)), })) diff --git a/modules/ui.py b/modules/ui.py index 2c92c422..f2e7c0d6 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -644,6 +644,13 @@ def create_sampler_and_steps_selection(choices, tabname): return steps, sampler_index +def ordered_ui_categories(): + user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} + + for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): + yield category + + def create_ui(): import modules.img2img import modules.txt2img @@ -672,32 +679,48 @@ def create_ui(): with gr.Row().style(equal_height=False): with gr.Column(variant='panel', elem_id="txt2img_settings"): - steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") - - with FormRow(): - with gr.Column(elem_id="txt2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") - with gr.Column(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") - - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') - - with FormRow(elem_id="txt2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") - enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") + for category in ordered_ui_categories(): + if category == "sampler": + steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") - with FormRow(visible=False) as hr_options: - hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) - hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") - - with FormGroup(elem_id="txt2img_script_container"): - custom_inputs = modules.scripts.scripts_txt2img.setup_ui() + elif category == "dimensions": + with FormRow(): + with gr.Column(elem_id="txt2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") + + if opts.dimensions_and_batch_together: + with gr.Column(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") + + elif category == "cfg": + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") + + elif category == "seed": + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') + + elif category == "checkboxes": + with FormRow(elem_id="txt2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") + enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") + + elif category == "hires_fix": + with FormRow(visible=False, elem_id="txt2img_hires_fix") as hr_options: + hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) + hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") + + elif category == "batch": + if not opts.dimensions_and_batch_together: + with FormRow(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") + + elif category == "scripts": + with FormGroup(elem_id="txt2img_script_container"): + custom_inputs = modules.scripts.scripts_txt2img.setup_ui() txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) @@ -865,28 +888,43 @@ def create_ui(): with FormRow(): resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") - steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") - - with FormRow(): - with gr.Column(elem_id="img2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") - with gr.Column(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - with FormGroup(): - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") - - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') + for category in ordered_ui_categories(): + if category == "sampler": + steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") - with FormRow(elem_id="img2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") - - with FormGroup(elem_id="img2img_script_container"): - custom_inputs = modules.scripts.scripts_img2img.setup_ui() + elif category == "dimensions": + with FormRow(): + with gr.Column(elem_id="img2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") + + if opts.dimensions_and_batch_together: + with gr.Column(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") + + elif category == "cfg": + with FormGroup(): + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") + + elif category == "seed": + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') + + elif category == "checkboxes": + with FormRow(elem_id="img2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") + + elif category == "batch": + if not opts.dimensions_and_batch_together: + with FormRow(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") + + elif category == "scripts": + with FormGroup(elem_id="img2img_script_container"): + custom_inputs = modules.scripts.scripts_img2img.setup_ui() img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) -- cgit v1.2.3 From c0ee1488702d5a6ae35fbf7e0422f9f685394920 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 14:18:48 +0300 Subject: add support for running with gradio 3.9 installed --- modules/generation_parameters_copypaste.py | 4 ++-- modules/ui_tempdir.py | 23 +++++++++++++++++++++-- 2 files changed, 23 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index ec60319a..d94f11a3 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -7,7 +7,7 @@ from pathlib import Path import gradio as gr from modules.shared import script_path -from modules import shared +from modules import shared, ui_tempdir import tempfile from PIL import Image @@ -39,7 +39,7 @@ def quote(text): def image_from_url_text(filedata): if type(filedata) == dict and filedata["is_file"]: filename = filedata["name"] - is_in_right_dir = any([filename in fileset for fileset in shared.demo.temp_file_sets]) + is_in_right_dir = ui_tempdir.check_tmp_file(shared.demo, filename) assert is_in_right_dir, 'trying to open image file outside of allowed directories' return Image.open(filename) diff --git a/modules/ui_tempdir.py b/modules/ui_tempdir.py index 363d449d..21945235 100644 --- a/modules/ui_tempdir.py +++ b/modules/ui_tempdir.py @@ -1,6 +1,7 @@ import os import tempfile from collections import namedtuple +from pathlib import Path import gradio as gr @@ -12,10 +13,28 @@ from modules import shared Savedfile = namedtuple("Savedfile", ["name"]) +def register_tmp_file(gradio, filename): + if hasattr(gradio, 'temp_file_sets'): # gradio 3.15 + gradio.temp_file_sets[0] = gradio.temp_file_sets[0] | {os.path.abspath(filename)} + + if hasattr(gradio, 'temp_dirs'): # gradio 3.9 + gradio.temp_dirs = gradio.temp_dirs | {os.path.abspath(os.path.dirname(filename))} + + +def check_tmp_file(gradio, filename): + if hasattr(gradio, 'temp_file_sets'): + return any([filename in fileset for fileset in gradio.temp_file_sets]) + + if hasattr(gradio, 'temp_dirs'): + return any(Path(temp_dir).resolve() in Path(filename).resolve().parents for temp_dir in gradio.temp_dirs) + + return False + + def save_pil_to_file(pil_image, dir=None): already_saved_as = getattr(pil_image, 'already_saved_as', None) if already_saved_as and os.path.isfile(already_saved_as): - shared.demo.temp_file_sets[0] = shared.demo.temp_file_sets[0] | {os.path.abspath(already_saved_as)} + register_tmp_file(shared.demo, already_saved_as) file_obj = Savedfile(already_saved_as) return file_obj @@ -45,7 +64,7 @@ def on_tmpdir_changed(): os.makedirs(shared.opts.temp_dir, exist_ok=True) - shared.demo.temp_file_sets[0] = shared.demo.temp_file_sets[0] | {os.path.abspath(shared.opts.temp_dir)} + register_tmp_file(shared.demo, os.path.join(shared.opts.temp_dir, "x")) def cleanup_tmpdr(): -- cgit v1.2.3 From bddebe09edeb6a18f2c06986d5658a7be3a563ea Mon Sep 17 00:00:00 2001 From: Shondoit Date: Tue, 3 Jan 2023 10:26:37 +0100 Subject: Save Optimizer next to TI embedding Also add check to load only .PT and .BIN files as embeddings. (since we add .optim files in the same directory) --- modules/shared.py | 2 +- modules/textual_inversion/textual_inversion.py | 40 ++++++++++++++++++++------ 2 files changed, 33 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 23657a93..c541d18c 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -355,7 +355,7 @@ options_templates.update(options_section(('system', "System"), { options_templates.update(options_section(('training', "Training"), { "unload_models_when_training": OptionInfo(False, "Move VAE and CLIP to RAM when training if possible. Saves VRAM."), "pin_memory": OptionInfo(False, "Turn on pin_memory for DataLoader. Makes training slightly faster but can increase memory usage."), - "save_optimizer_state": OptionInfo(False, "Saves Optimizer state as separate *.optim file. Training can be resumed with HN itself and matching optim file."), + "save_optimizer_state": OptionInfo(False, "Saves Optimizer state as separate *.optim file. Training of embedding or HN can be resumed with the matching optim file."), "dataset_filename_word_regex": OptionInfo("", "Filename word regex"), "dataset_filename_join_string": OptionInfo(" ", "Filename join string"), "training_image_repeats_per_epoch": OptionInfo(1, "Number of repeats for a single input image per epoch; used only for displaying epoch number", gr.Number, {"precision": 0}), diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index fd253477..16176e90 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -28,6 +28,7 @@ class Embedding: self.cached_checksum = None self.sd_checkpoint = None self.sd_checkpoint_name = None + self.optimizer_state_dict = None def save(self, filename): embedding_data = { @@ -41,6 +42,13 @@ class Embedding: torch.save(embedding_data, filename) + if shared.opts.save_optimizer_state and self.optimizer_state_dict is not None: + optimizer_saved_dict = { + 'hash': self.checksum(), + 'optimizer_state_dict': self.optimizer_state_dict, + } + torch.save(optimizer_saved_dict, filename + '.optim') + def checksum(self): if self.cached_checksum is not None: return self.cached_checksum @@ -95,9 +103,10 @@ class EmbeddingDatabase: self.expected_shape = self.get_expected_shape() def process_file(path, filename): - name = os.path.splitext(filename)[0] + name, ext = os.path.splitext(filename) + ext = ext.upper() - if os.path.splitext(filename.upper())[-1] in ['.PNG', '.WEBP', '.JXL', '.AVIF']: + if ext in ['.PNG', '.WEBP', '.JXL', '.AVIF']: embed_image = Image.open(path) if hasattr(embed_image, 'text') and 'sd-ti-embedding' in embed_image.text: data = embedding_from_b64(embed_image.text['sd-ti-embedding']) @@ -105,8 +114,10 @@ class EmbeddingDatabase: else: data = extract_image_data_embed(embed_image) name = data.get('name', name) - else: + elif ext in ['.BIN', '.PT']: data = torch.load(path, map_location="cpu") + else: + return # textual inversion embeddings if 'string_to_param' in data: @@ -300,6 +311,20 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ embedding.vec.requires_grad = True optimizer = torch.optim.AdamW([embedding.vec], lr=scheduler.learn_rate, weight_decay=0.0) + if shared.opts.save_optimizer_state: + optimizer_state_dict = None + if os.path.exists(filename + '.optim'): + optimizer_saved_dict = torch.load(filename + '.optim', map_location='cpu') + if embedding.checksum() == optimizer_saved_dict.get('hash', None): + optimizer_state_dict = optimizer_saved_dict.get('optimizer_state_dict', None) + + if optimizer_state_dict is not None: + optimizer.load_state_dict(optimizer_state_dict) + print("Loaded existing optimizer from checkpoint") + else: + print("No saved optimizer exists in checkpoint") + + scaler = torch.cuda.amp.GradScaler() batch_size = ds.batch_size @@ -366,9 +391,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ # Before saving, change name to match current checkpoint. embedding_name_every = f'{embedding_name}-{steps_done}' last_saved_file = os.path.join(embedding_dir, f'{embedding_name_every}.pt') - #if shared.opts.save_optimizer_state: - #embedding.optimizer_state_dict = optimizer.state_dict() - save_embedding(embedding, checkpoint, embedding_name_every, last_saved_file, remove_cached_checksum=True) + save_embedding(embedding, optimizer, checkpoint, embedding_name_every, last_saved_file, remove_cached_checksum=True) embedding_yet_to_be_embedded = True write_loss(log_directory, "textual_inversion_loss.csv", embedding.step, steps_per_epoch, { @@ -458,7 +481,7 @@ Last saved image: {html.escape(last_saved_image)}

""" filename = os.path.join(shared.cmd_opts.embeddings_dir, f'{embedding_name}.pt') - save_embedding(embedding, checkpoint, embedding_name, filename, remove_cached_checksum=True) + save_embedding(embedding, optimizer, checkpoint, embedding_name, filename, remove_cached_checksum=True) except Exception: print(traceback.format_exc(), file=sys.stderr) pass @@ -470,7 +493,7 @@ Last saved image: {html.escape(last_saved_image)}
return embedding, filename -def save_embedding(embedding, checkpoint, embedding_name, filename, remove_cached_checksum=True): +def save_embedding(embedding, optimizer, checkpoint, embedding_name, filename, remove_cached_checksum=True): old_embedding_name = embedding.name old_sd_checkpoint = embedding.sd_checkpoint if hasattr(embedding, "sd_checkpoint") else None old_sd_checkpoint_name = embedding.sd_checkpoint_name if hasattr(embedding, "sd_checkpoint_name") else None @@ -481,6 +504,7 @@ def save_embedding(embedding, checkpoint, embedding_name, filename, remove_cache if remove_cached_checksum: embedding.cached_checksum = None embedding.name = embedding_name + embedding.optimizer_state_dict = optimizer.state_dict() embedding.save(filename) except: embedding.sd_checkpoint = old_sd_checkpoint -- cgit v1.2.3 From e9fb9bb0c25f59109a816fc53c385bed58965c24 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 17:40:20 +0300 Subject: fix hires fix not working in API when user does not specify upscaler --- modules/processing.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 4654570c..a172af0b 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -685,7 +685,7 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength, prompts): self.sampler = sd_samplers.create_sampler(self.sampler_name, self.sd_model) - latent_scale_mode = shared.latent_upscale_modes.get(self.hr_upscaler, None) if self.hr_upscaler is not None else shared.latent_upscale_default_mode + latent_scale_mode = shared.latent_upscale_modes.get(self.hr_upscaler, None) if self.hr_upscaler is not None else shared.latent_upscale_modes.get(shared.latent_upscale_default_mode, "nearest") if self.enable_hr and latent_scale_mode is None: assert len([x for x in shared.sd_upscalers if x.name == self.hr_upscaler]) > 0, f"could not find upscaler named {self.hr_upscaler}" -- cgit v1.2.3 From aaa4c2aacbb6523077334093c81bd475d757f7a1 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 09:45:16 -0500 Subject: add api logging --- modules/api/api.py | 24 +++++++++++++++++++++++- modules/shared.py | 1 + 2 files changed, 24 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 9c670f00..53135470 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -1,11 +1,12 @@ import base64 import io import time +import datetime import uvicorn from threading import Lock from io import BytesIO from gradio.processing_utils import decode_base64_to_file -from fastapi import APIRouter, Depends, FastAPI, HTTPException +from fastapi import APIRouter, Depends, FastAPI, HTTPException, Request, Response from fastapi.security import HTTPBasic, HTTPBasicCredentials from secrets import compare_digest @@ -67,6 +68,26 @@ def encode_pil_to_base64(image): bytes_data = output_bytes.getvalue() return base64.b64encode(bytes_data) +def init_api_middleware(app: FastAPI): + @app.middleware("http") + async def log_and_time(req: Request, call_next): + ts = time.time() + res: Response = await call_next(req) + duration = str(round(time.time() - ts, 4)) + res.headers["X-Process-Time"] = duration + if shared.cmd_opts.api_log: + print('API {t} {code} {prot}/{ver} {method} {p} {cli} {duration}'.format( + t = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"), + code = res.status_code, + ver = req.scope.get('http_version', '0.0'), + cli = req.scope.get('client', ('0:0.0.0', 0))[0], + prot = req.scope.get('scheme', 'err'), + method = req.scope.get('method', 'err'), + p = req.scope.get('path', 'err'), + duration = duration, + )) + return res + class Api: def __init__(self, app: FastAPI, queue_lock: Lock): @@ -78,6 +99,7 @@ class Api: self.router = APIRouter() self.app = app + init_api_middleware(self.app) self.queue_lock = queue_lock self.add_api_route("/sdapi/v1/txt2img", self.text2imgapi, methods=["POST"], response_model=TextToImageResponse) self.add_api_route("/sdapi/v1/img2img", self.img2imgapi, methods=["POST"], response_model=ImageToImageResponse) diff --git a/modules/shared.py b/modules/shared.py index 23657a93..2a03d716 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -82,6 +82,7 @@ parser.add_argument('--vae-path', type=str, help='Path to Variational Autoencode parser.add_argument("--disable-safe-unpickle", action='store_true', help="disable checking pytorch models for malicious code", default=False) parser.add_argument("--api", action='store_true', help="use api=True to launch the API together with the webui (use --nowebui instead for only the API)") parser.add_argument("--api-auth", type=str, help='Set authentication for API like "username:password"; or comma-delimit multiple like "u1:p1,u2:p2,u3:p3"', default=None) +parser.add_argument("--api-log", action='store_true', help="use api-log=True to enable logging of all API requests") parser.add_argument("--nowebui", action='store_true', help="use api=True to launch the API instead of the webui") parser.add_argument("--ui-debug-mode", action='store_true', help="Don't load model to quickly launch UI") parser.add_argument("--device-id", type=str, help="Select the default CUDA device to use (export CUDA_VISIBLE_DEVICES=0,1,etc might be needed before)", default=None) -- cgit v1.2.3 From 1d9dc48efda2e8da6d13fc62e65500198a9b041c Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 10:21:51 -0500 Subject: init job and add info to model merge --- modules/extras.py | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 5e270250..7e222313 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -242,6 +242,9 @@ def run_pnginfo(image): def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format): + shared.state.begin() + shared.state.job = 'model-merge' + def weighted_sum(theta0, theta1, alpha): return ((1 - alpha) * theta0) + (alpha * theta1) @@ -263,8 +266,11 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam theta_func1, theta_func2 = theta_funcs[interp_method] if theta_func1 and not tertiary_model_info: + shared.state.textinfo = "Failed: Interpolation method requires a tertiary model." + shared.state.end() return ["Failed: Interpolation method requires a tertiary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + shared.state.textinfo = f"Loading {secondary_model_info.filename}..." print(f"Loading {secondary_model_info.filename}...") theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') @@ -281,6 +287,7 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam theta_1[key] = torch.zeros_like(theta_1[key]) del theta_2 + shared.state.textinfo = f"Loading {primary_model_info.filename}..." print(f"Loading {primary_model_info.filename}...") theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') @@ -291,6 +298,7 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam a = theta_0[key] b = theta_1[key] + shared.state.textinfo = f'Merging layer {key}' # this enables merging an inpainting model (A) with another one (B); # where normal model would have 4 channels, for latenst space, inpainting model would # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 @@ -303,8 +311,6 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) result_is_inpainting_model = True else: - assert a.shape == b.shape, f'Incompatible shapes for layer {key}: A is {a.shape}, and B is {b.shape}' - theta_0[key] = theta_func2(a, b, multiplier) if save_as_half: @@ -332,6 +338,7 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam output_modelname = os.path.join(ckpt_dir, filename) + shared.state.textinfo = f"Saving to {output_modelname}..." print(f"Saving to {output_modelname}...") _, extension = os.path.splitext(output_modelname) @@ -343,4 +350,7 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam sd_models.list_models() print("Checkpoint saved.") + shared.state.textinfo = "Checkpoint saved to " + output_modelname + shared.state.end() + return ["Checkpoint saved to " + output_modelname] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] -- cgit v1.2.3 From 192ddc04d6de0d780f73aa5fbaa8c66cd4642e1c Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 10:34:51 -0500 Subject: add job info to modules --- modules/extras.py | 17 +++++++++++++---- modules/hypernetworks/hypernetwork.py | 1 + modules/textual_inversion/preprocess.py | 1 + modules/textual_inversion/textual_inversion.py | 1 + 4 files changed, 16 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 7e222313..d665440a 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -58,6 +58,9 @@ cached_images: LruCache = LruCache(max_size=5) def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): devices.torch_gc() + shared.state.begin() + shared.state.job = 'extras' + imageArr = [] # Also keep track of original file names imageNameArr = [] @@ -94,6 +97,7 @@ def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_ # Extra operation definitions def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + shared.state.job = 'extras-gfpgan' restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) res = Image.fromarray(restored_img) @@ -104,6 +108,7 @@ def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_ return (res, info) def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + shared.state.job = 'extras-codeformer' restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) res = Image.fromarray(restored_img) @@ -114,6 +119,7 @@ def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_ return (res, info) def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): + shared.state.job = 'extras-upscale' upscaler = shared.sd_upscalers[scaler_index] res = upscaler.scaler.upscale(image, resize, upscaler.data_path) if mode == 1 and crop: @@ -180,6 +186,9 @@ def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_ for image, image_name in zip(imageArr, imageNameArr): if image is None: return outputs, "Please select an input image.", '' + + shared.state.textinfo = f'Processing image {image_name}' + existing_pnginfo = image.info or {} image = image.convert("RGB") @@ -193,6 +202,10 @@ def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_ else: basename = '' + if opts.enable_pnginfo: # append info before save + image.info = existing_pnginfo + image.info["extras"] = info + if save_output: # Add upscaler name as a suffix. suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" @@ -203,10 +216,6 @@ def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_ images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) - if opts.enable_pnginfo: - image.info = existing_pnginfo - image.info["extras"] = info - if extras_mode != 2 or show_extras_results : outputs.append(image) diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 109e8078..450fecac 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -417,6 +417,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, shared.loaded_hypernetwork = Hypernetwork() shared.loaded_hypernetwork.load(path) + shared.state.job = "train-hypernetwork" shared.state.textinfo = "Initializing hypernetwork training..." shared.state.job_count = steps diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py index 56b9b2eb..feb876c6 100644 --- a/modules/textual_inversion/preprocess.py +++ b/modules/textual_inversion/preprocess.py @@ -124,6 +124,7 @@ def preprocess_work(process_src, process_dst, process_width, process_height, pre files = listfiles(src) + shared.state.job = "preprocess" shared.state.textinfo = "Preprocessing..." shared.state.job_count = len(files) diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index fd253477..2c1251d6 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -245,6 +245,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ create_image_every = create_image_every or 0 validate_train_inputs(embedding_name, learn_rate, batch_size, gradient_step, data_root, template_file, steps, save_embedding_every, create_image_every, log_directory, name="embedding") + shared.state.job = "train-embedding" shared.state.textinfo = "Initializing textual inversion training..." shared.state.job_count = steps -- cgit v1.2.3 From 2d5a5076bb2a0c05cc27d75a1bcadab7f32a46d0 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 18:38:21 +0300 Subject: Make it so that upscalers are not repeated when restarting UI. --- modules/modelloader.py | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) (limited to 'modules') diff --git a/modules/modelloader.py b/modules/modelloader.py index e647f6fa..6a1a7ac8 100644 --- a/modules/modelloader.py +++ b/modules/modelloader.py @@ -123,6 +123,23 @@ def move_files(src_path: str, dest_path: str, ext_filter: str = None): pass +builtin_upscaler_classes = [] +forbidden_upscaler_classes = set() + + +def list_builtin_upscalers(): + load_upscalers() + + builtin_upscaler_classes.clear() + builtin_upscaler_classes.extend(Upscaler.__subclasses__()) + + +def forbid_loaded_nonbuiltin_upscalers(): + for cls in Upscaler.__subclasses__(): + if cls not in builtin_upscaler_classes: + forbidden_upscaler_classes.add(cls) + + def load_upscalers(): # We can only do this 'magic' method to dynamically load upscalers if they are referenced, # so we'll try to import any _model.py files before looking in __subclasses__ @@ -139,6 +156,9 @@ def load_upscalers(): datas = [] commandline_options = vars(shared.cmd_opts) for cls in Upscaler.__subclasses__(): + if cls in forbidden_upscaler_classes: + continue + name = cls.__name__ cmd_name = f"{name.lower().replace('upscaler', '')}_models_path" scaler = cls(commandline_options.get(cmd_name, None)) -- cgit v1.2.3 From 8f96f9289981a66741ba770d14f3d27ce335a0fb Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 18:39:14 +0300 Subject: call script callbacks for reloaded model after loading embeddings --- modules/sd_models.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index bff8d6c9..b98b05fc 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -324,12 +324,12 @@ def load_model(checkpoint_info=None): sd_model.eval() shared.sd_model = sd_model + sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings(force_reload=True) # Reload embeddings after model load as they may or may not fit the model + script_callbacks.model_loaded_callback(sd_model) print("Model loaded.") - sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings(force_reload = True) # Reload embeddings after model load as they may or may not fit the model - return sd_model -- cgit v1.2.3 From cec209981ee988536c2521297baf9bc1b256005f Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 10:58:52 -0500 Subject: log only sdapi --- modules/api/api.py | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 53135470..78751c57 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -68,22 +68,23 @@ def encode_pil_to_base64(image): bytes_data = output_bytes.getvalue() return base64.b64encode(bytes_data) -def init_api_middleware(app: FastAPI): +def api_middleware(app: FastAPI): @app.middleware("http") async def log_and_time(req: Request, call_next): ts = time.time() res: Response = await call_next(req) duration = str(round(time.time() - ts, 4)) res.headers["X-Process-Time"] = duration - if shared.cmd_opts.api_log: - print('API {t} {code} {prot}/{ver} {method} {p} {cli} {duration}'.format( + endpoint = req.scope.get('path', 'err') + if shared.cmd_opts.api_log and endpoint.startswith('/sdapi'): + print('API {t} {code} {prot}/{ver} {method} {endpoint} {cli} {duration}'.format( t = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"), code = res.status_code, ver = req.scope.get('http_version', '0.0'), cli = req.scope.get('client', ('0:0.0.0', 0))[0], prot = req.scope.get('scheme', 'err'), method = req.scope.get('method', 'err'), - p = req.scope.get('path', 'err'), + endpoint = endpoint, duration = duration, )) return res -- cgit v1.2.3 From d8d206c1685d1e7027d4af82ed18d106f41d1cc4 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 11:01:04 -0500 Subject: add state to interrogate --- modules/interrogate.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/interrogate.py b/modules/interrogate.py index 6f761c5a..738d8ff7 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -136,7 +136,8 @@ class InterrogateModels: def interrogate(self, pil_image): res = "" - + shared.state.begin() + shared.state.job = 'interrogate' try: if shared.cmd_opts.lowvram or shared.cmd_opts.medvram: @@ -177,5 +178,6 @@ class InterrogateModels: res += "" self.unload() + shared.state.end() return res -- cgit v1.2.3 From 82cfc227d735c140447d5b8dca29a71ee9bde127 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 20:23:17 +0300 Subject: added licenses screen to settings added footer removed unused inpainting code --- modules/sd_hijack_inpainting.py | 232 ---------------------------------------- modules/ui.py | 15 ++- 2 files changed, 13 insertions(+), 234 deletions(-) (limited to 'modules') diff --git a/modules/sd_hijack_inpainting.py b/modules/sd_hijack_inpainting.py index 06b75772..3c214a35 100644 --- a/modules/sd_hijack_inpainting.py +++ b/modules/sd_hijack_inpainting.py @@ -12,191 +12,6 @@ from ldm.models.diffusion.ddpm import LatentDiffusion from ldm.models.diffusion.plms import PLMSSampler from ldm.models.diffusion.ddim import DDIMSampler, noise_like -# ================================================================================================= -# Monkey patch DDIMSampler methods from RunwayML repo directly. -# Adapted from: -# https://github.com/runwayml/stable-diffusion/blob/main/ldm/models/diffusion/ddim.py -# ================================================================================================= -@torch.no_grad() -def sample_ddim(self, - S, - batch_size, - shape, - conditioning=None, - callback=None, - normals_sequence=None, - img_callback=None, - quantize_x0=False, - eta=0., - mask=None, - x0=None, - temperature=1., - noise_dropout=0., - score_corrector=None, - corrector_kwargs=None, - verbose=True, - x_T=None, - log_every_t=100, - unconditional_guidance_scale=1., - unconditional_conditioning=None, - # this has to come in the same format as the conditioning, # e.g. as encoded tokens, ... - **kwargs - ): - if conditioning is not None: - if isinstance(conditioning, dict): - ctmp = conditioning[list(conditioning.keys())[0]] - while isinstance(ctmp, list): - ctmp = ctmp[0] - cbs = ctmp.shape[0] - if cbs != batch_size: - print(f"Warning: Got {cbs} conditionings but batch-size is {batch_size}") - else: - if conditioning.shape[0] != batch_size: - print(f"Warning: Got {conditioning.shape[0]} conditionings but batch-size is {batch_size}") - - self.make_schedule(ddim_num_steps=S, ddim_eta=eta, verbose=verbose) - # sampling - C, H, W = shape - size = (batch_size, C, H, W) - print(f'Data shape for DDIM sampling is {size}, eta {eta}') - - samples, intermediates = self.ddim_sampling(conditioning, size, - callback=callback, - img_callback=img_callback, - quantize_denoised=quantize_x0, - mask=mask, x0=x0, - ddim_use_original_steps=False, - noise_dropout=noise_dropout, - temperature=temperature, - score_corrector=score_corrector, - corrector_kwargs=corrector_kwargs, - x_T=x_T, - log_every_t=log_every_t, - unconditional_guidance_scale=unconditional_guidance_scale, - unconditional_conditioning=unconditional_conditioning, - ) - return samples, intermediates - -@torch.no_grad() -def p_sample_ddim(self, x, c, t, index, repeat_noise=False, use_original_steps=False, quantize_denoised=False, - temperature=1., noise_dropout=0., score_corrector=None, corrector_kwargs=None, - unconditional_guidance_scale=1., unconditional_conditioning=None): - b, *_, device = *x.shape, x.device - - if unconditional_conditioning is None or unconditional_guidance_scale == 1.: - e_t = self.model.apply_model(x, t, c) - else: - x_in = torch.cat([x] * 2) - t_in = torch.cat([t] * 2) - if isinstance(c, dict): - assert isinstance(unconditional_conditioning, dict) - c_in = dict() - for k in c: - if isinstance(c[k], list): - c_in[k] = [ - torch.cat([unconditional_conditioning[k][i], c[k][i]]) - for i in range(len(c[k])) - ] - else: - c_in[k] = torch.cat([unconditional_conditioning[k], c[k]]) - else: - c_in = torch.cat([unconditional_conditioning, c]) - e_t_uncond, e_t = self.model.apply_model(x_in, t_in, c_in).chunk(2) - e_t = e_t_uncond + unconditional_guidance_scale * (e_t - e_t_uncond) - - if score_corrector is not None: - assert self.model.parameterization == "eps" - e_t = score_corrector.modify_score(self.model, e_t, x, t, c, **corrector_kwargs) - - alphas = self.model.alphas_cumprod if use_original_steps else self.ddim_alphas - alphas_prev = self.model.alphas_cumprod_prev if use_original_steps else self.ddim_alphas_prev - sqrt_one_minus_alphas = self.model.sqrt_one_minus_alphas_cumprod if use_original_steps else self.ddim_sqrt_one_minus_alphas - sigmas = self.model.ddim_sigmas_for_original_num_steps if use_original_steps else self.ddim_sigmas - # select parameters corresponding to the currently considered timestep - a_t = torch.full((b, 1, 1, 1), alphas[index], device=device) - a_prev = torch.full((b, 1, 1, 1), alphas_prev[index], device=device) - sigma_t = torch.full((b, 1, 1, 1), sigmas[index], device=device) - sqrt_one_minus_at = torch.full((b, 1, 1, 1), sqrt_one_minus_alphas[index],device=device) - - # current prediction for x_0 - pred_x0 = (x - sqrt_one_minus_at * e_t) / a_t.sqrt() - if quantize_denoised: - pred_x0, _, *_ = self.model.first_stage_model.quantize(pred_x0) - # direction pointing to x_t - dir_xt = (1. - a_prev - sigma_t**2).sqrt() * e_t - noise = sigma_t * noise_like(x.shape, device, repeat_noise) * temperature - if noise_dropout > 0.: - noise = torch.nn.functional.dropout(noise, p=noise_dropout) - x_prev = a_prev.sqrt() * pred_x0 + dir_xt + noise - return x_prev, pred_x0 - - -# ================================================================================================= -# Monkey patch PLMSSampler methods. -# This one was not actually patched correctly in the RunwayML repo, but we can replicate the changes. -# Adapted from: -# https://github.com/CompVis/stable-diffusion/blob/main/ldm/models/diffusion/plms.py -# ================================================================================================= -@torch.no_grad() -def sample_plms(self, - S, - batch_size, - shape, - conditioning=None, - callback=None, - normals_sequence=None, - img_callback=None, - quantize_x0=False, - eta=0., - mask=None, - x0=None, - temperature=1., - noise_dropout=0., - score_corrector=None, - corrector_kwargs=None, - verbose=True, - x_T=None, - log_every_t=100, - unconditional_guidance_scale=1., - unconditional_conditioning=None, - # this has to come in the same format as the conditioning, # e.g. as encoded tokens, ... - **kwargs - ): - if conditioning is not None: - if isinstance(conditioning, dict): - ctmp = conditioning[list(conditioning.keys())[0]] - while isinstance(ctmp, list): - ctmp = ctmp[0] - cbs = ctmp.shape[0] - if cbs != batch_size: - print(f"Warning: Got {cbs} conditionings but batch-size is {batch_size}") - else: - if conditioning.shape[0] != batch_size: - print(f"Warning: Got {conditioning.shape[0]} conditionings but batch-size is {batch_size}") - - self.make_schedule(ddim_num_steps=S, ddim_eta=eta, verbose=verbose) - # sampling - C, H, W = shape - size = (batch_size, C, H, W) - # print(f'Data shape for PLMS sampling is {size}') # remove unnecessary message - - samples, intermediates = self.plms_sampling(conditioning, size, - callback=callback, - img_callback=img_callback, - quantize_denoised=quantize_x0, - mask=mask, x0=x0, - ddim_use_original_steps=False, - noise_dropout=noise_dropout, - temperature=temperature, - score_corrector=score_corrector, - corrector_kwargs=corrector_kwargs, - x_T=x_T, - log_every_t=log_every_t, - unconditional_guidance_scale=unconditional_guidance_scale, - unconditional_conditioning=unconditional_conditioning, - ) - return samples, intermediates - @torch.no_grad() def p_sample_plms(self, x, c, t, index, repeat_noise=False, use_original_steps=False, quantize_denoised=False, @@ -280,44 +95,6 @@ def p_sample_plms(self, x, c, t, index, repeat_noise=False, use_original_steps=F return x_prev, pred_x0, e_t -# ================================================================================================= -# Monkey patch LatentInpaintDiffusion to load the checkpoint with a proper config. -# Adapted from: -# https://github.com/runwayml/stable-diffusion/blob/main/ldm/models/diffusion/ddpm.py -# ================================================================================================= - -@torch.no_grad() -def get_unconditional_conditioning(self, batch_size, null_label=None): - if null_label is not None: - xc = null_label - if isinstance(xc, ListConfig): - xc = list(xc) - if isinstance(xc, dict) or isinstance(xc, list): - c = self.get_learned_conditioning(xc) - else: - if hasattr(xc, "to"): - xc = xc.to(self.device) - c = self.get_learned_conditioning(xc) - else: - # todo: get null label from cond_stage_model - raise NotImplementedError() - c = repeat(c, "1 ... -> b ...", b=batch_size).to(self.device) - return c - - -class LatentInpaintDiffusion(LatentDiffusion): - def __init__( - self, - concat_keys=("mask", "masked_image"), - masked_image_key="masked_image", - *args, - **kwargs, - ): - super().__init__(*args, **kwargs) - self.masked_image_key = masked_image_key - assert self.masked_image_key in concat_keys - self.concat_keys = concat_keys - def should_hijack_inpainting(checkpoint_info): ckpt_basename = os.path.basename(checkpoint_info.filename).lower() @@ -326,15 +103,6 @@ def should_hijack_inpainting(checkpoint_info): def do_inpainting_hijack(): - # most of this stuff seems to no longer be needed because it is already included into SD2.0 # p_sample_plms is needed because PLMS can't work with dicts as conditionings - # this file should be cleaned up later if everything turns out to work fine - - # ldm.models.diffusion.ddpm.get_unconditional_conditioning = get_unconditional_conditioning - # ldm.models.diffusion.ddpm.LatentInpaintDiffusion = LatentInpaintDiffusion - - # ldm.models.diffusion.ddim.DDIMSampler.p_sample_ddim = p_sample_ddim - # ldm.models.diffusion.ddim.DDIMSampler.sample = sample_ddim ldm.models.diffusion.plms.PLMSSampler.p_sample_plms = p_sample_plms - # ldm.models.diffusion.plms.PLMSSampler.sample = sample_plms diff --git a/modules/ui.py b/modules/ui.py index f2e7c0d6..d941cb5f 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1529,8 +1529,10 @@ def create_ui(): with gr.Blocks(analytics_enabled=False) as settings_interface: with gr.Row(): - settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") - restart_gradio = gr.Button(value='Restart UI', variant='primary', elem_id="settings_restart_gradio") + with gr.Column(scale=6): + settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") + with gr.Column(): + restart_gradio = gr.Button(value='Reload UI', variant='primary', elem_id="settings_restart_gradio") result = gr.HTML(elem_id="settings_result") @@ -1574,6 +1576,11 @@ def create_ui(): download_localization = gr.Button(value='Download localization template', elem_id="download_localization") reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") + if os.path.exists("html/licenses.html"): + with open("html/licenses.html", encoding="utf8") as file: + with gr.TabItem("Licenses"): + gr.HTML(file.read(), elem_id="licenses") + gr.Button(value="Show all pages", elem_id="settings_show_all_pages") request_notifications.click( @@ -1659,6 +1666,10 @@ def create_ui(): if os.path.exists(os.path.join(script_path, "notification.mp3")): audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) + if os.path.exists("html/footer.html"): + with open("html/footer.html", encoding="utf8") as file: + gr.HTML(file.read(), elem_id="footer") + text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) settings_submit.click( fn=wrap_gradio_call(run_settings, extra_outputs=[gr.update()]), -- cgit v1.2.3 From 3e22e294135ed0327ce9d9738655ff03c53df3c0 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 3 Jan 2023 21:49:24 +0300 Subject: fix broken send to extras button --- modules/generation_parameters_copypaste.py | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index d94f11a3..4baf4d9a 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -37,7 +37,10 @@ def quote(text): def image_from_url_text(filedata): - if type(filedata) == dict and filedata["is_file"]: + if type(filedata) == list and len(filedata) > 0 and type(filedata[0]) == dict and filedata[0].get("is_file", False): + filedata = filedata[0] + + if type(filedata) == dict and filedata.get("is_file", False): filename = filedata["name"] is_in_right_dir = ui_tempdir.check_tmp_file(shared.demo, filename) assert is_in_right_dir, 'trying to open image file outside of allowed directories' -- cgit v1.2.3 From 917b5bd8d0cd47c9dc241c1852ccd440a8c61668 Mon Sep 17 00:00:00 2001 From: Max Weber Date: Tue, 3 Jan 2023 18:19:56 -0700 Subject: ui: save dropdown sampling method to the ui-config --- modules/ui.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index d941cb5f..bfc93634 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -635,6 +635,7 @@ def create_sampler_and_steps_selection(choices, tabname): if opts.samplers_in_dropdown: with FormRow(elem_id=f"sampler_selection_{tabname}"): sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + sampler_index.save_to_config = True steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling Steps", value=20) else: with FormGroup(elem_id=f"sampler_selection_{tabname}"): -- cgit v1.2.3 From e5b7ee910e7bb88f08e8876b5732cb034c6fe529 Mon Sep 17 00:00:00 2001 From: MMaker Date: Wed, 4 Jan 2023 04:22:01 -0500 Subject: fix: Save full res of intermediate step --- modules/processing.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index a172af0b..93e75ba6 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -705,7 +705,7 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): return if not isinstance(image, Image.Image): - image = sd_samplers.sample_to_image(image, index) + image = sd_samplers.sample_to_image(image, index, approximation=0) images.save_image(image, self.outpath_samples, "", seeds[index], prompts[index], opts.samples_format, suffix="-before-highres-fix") -- cgit v1.2.3 From 02d7abf5141431b9a3a8a189bb3136c71abd5e79 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 12:35:07 +0300 Subject: helpful error message when trying to load 2.0 without config failing to load model weights from settings won't break generation for currently loaded model anymore --- modules/errors.py | 25 +++++++++++++++++++++++-- modules/sd_models.py | 26 ++++++++++++++++++-------- modules/shared.py | 9 +++++++-- 3 files changed, 48 insertions(+), 12 deletions(-) (limited to 'modules') diff --git a/modules/errors.py b/modules/errors.py index 372dc51a..a668c014 100644 --- a/modules/errors.py +++ b/modules/errors.py @@ -2,9 +2,30 @@ import sys import traceback +def print_error_explanation(message): + lines = message.strip().split("\n") + max_len = max([len(x) for x in lines]) + + print('=' * max_len, file=sys.stderr) + for line in lines: + print(line, file=sys.stderr) + print('=' * max_len, file=sys.stderr) + + +def display(e: Exception, task): + print(f"{task or 'error'}: {type(e).__name__}", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + + message = str(e) + if "copying a param with shape torch.Size([640, 1024]) from checkpoint, the shape in current model is torch.Size([640, 768])" in message: + print_error_explanation(""" +The most likely cause of this is you are trying to load Stable Diffusion 2.0 model without specifying its connfig file. +See https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Features#stable-diffusion-20 for how to solve this. + """) + + def run(code, task): try: code() except Exception as e: - print(f"{task}: {type(e).__name__}", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) + display(task, e) diff --git a/modules/sd_models.py b/modules/sd_models.py index b98b05fc..6846b74a 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -278,6 +278,7 @@ def enable_midas_autodownload(): midas.api.load_model = load_model_wrapper + def load_model(checkpoint_info=None): from modules import lowvram, sd_hijack checkpoint_info = checkpoint_info or select_checkpoint() @@ -312,6 +313,7 @@ def load_model(checkpoint_info=None): sd_config.model.params.unet_config.params.use_fp16 = False sd_model = instantiate_from_config(sd_config.model) + load_model_weights(sd_model, checkpoint_info) if shared.cmd_opts.lowvram or shared.cmd_opts.medvram: @@ -336,10 +338,12 @@ def load_model(checkpoint_info=None): def reload_model_weights(sd_model=None, info=None): from modules import lowvram, devices, sd_hijack checkpoint_info = info or select_checkpoint() - + if not sd_model: sd_model = shared.sd_model + current_checkpoint_info = sd_model.sd_checkpoint_info + if sd_model.sd_model_checkpoint == checkpoint_info.filename: return @@ -356,13 +360,19 @@ def reload_model_weights(sd_model=None, info=None): sd_hijack.model_hijack.undo_hijack(sd_model) - load_model_weights(sd_model, checkpoint_info) - - sd_hijack.model_hijack.hijack(sd_model) - script_callbacks.model_loaded_callback(sd_model) - - if not shared.cmd_opts.lowvram and not shared.cmd_opts.medvram: - sd_model.to(devices.device) + try: + load_model_weights(sd_model, checkpoint_info) + except Exception as e: + print("Failed to load checkpoint, restoring previous") + load_model_weights(sd_model, current_checkpoint_info) + raise + finally: + sd_hijack.model_hijack.hijack(sd_model) + script_callbacks.model_loaded_callback(sd_model) + + if not shared.cmd_opts.lowvram and not shared.cmd_opts.medvram: + sd_model.to(devices.device) print("Weights loaded.") + return sd_model diff --git a/modules/shared.py b/modules/shared.py index 23657a93..7588c47b 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -14,7 +14,7 @@ import modules.interrogate import modules.memmon import modules.styles import modules.devices as devices -from modules import localization, sd_vae, extensions, script_loading +from modules import localization, sd_vae, extensions, script_loading, errors from modules.paths import models_path, script_path, sd_path @@ -494,7 +494,12 @@ class Options: return False if self.data_labels[key].onchange is not None: - self.data_labels[key].onchange() + try: + self.data_labels[key].onchange() + except Exception as e: + errors.display(e, f"changing setting {key} to {value}") + setattr(self, key, oldval) + return False return True -- cgit v1.2.3 From 8d8a05a3bbb50fdfeab51679a919d2487bd97976 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 12:47:42 +0300 Subject: find configs for models at runtime rather than when starting --- modules/sd_hijack_inpainting.py | 5 ++++- modules/sd_models.py | 31 ++++++++++++++++++------------- 2 files changed, 22 insertions(+), 14 deletions(-) (limited to 'modules') diff --git a/modules/sd_hijack_inpainting.py b/modules/sd_hijack_inpainting.py index 3c214a35..31d2c898 100644 --- a/modules/sd_hijack_inpainting.py +++ b/modules/sd_hijack_inpainting.py @@ -97,8 +97,11 @@ def p_sample_plms(self, x, c, t, index, repeat_noise=False, use_original_steps=F def should_hijack_inpainting(checkpoint_info): + from modules import sd_models + ckpt_basename = os.path.basename(checkpoint_info.filename).lower() - cfg_basename = os.path.basename(checkpoint_info.config).lower() + cfg_basename = os.path.basename(sd_models.find_checkpoint_config(checkpoint_info)).lower() + return "inpainting" in ckpt_basename and not "inpainting" in cfg_basename diff --git a/modules/sd_models.py b/modules/sd_models.py index 6846b74a..6dca4ddf 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -20,7 +20,7 @@ from modules.sd_hijack_inpainting import do_inpainting_hijack, should_hijack_inp model_dir = "Stable-diffusion" model_path = os.path.abspath(os.path.join(models_path, model_dir)) -CheckpointInfo = namedtuple("CheckpointInfo", ['filename', 'title', 'hash', 'model_name', 'config']) +CheckpointInfo = namedtuple("CheckpointInfo", ['filename', 'title', 'hash', 'model_name']) checkpoints_list = {} checkpoints_loaded = collections.OrderedDict() @@ -48,6 +48,14 @@ def checkpoint_tiles(): return sorted([x.title for x in checkpoints_list.values()], key = alphanumeric_key) +def find_checkpoint_config(info): + config = os.path.splitext(info.filename)[0] + ".yaml" + if os.path.exists(config): + return config + + return shared.cmd_opts.config + + def list_models(): checkpoints_list.clear() model_list = modelloader.load_models(model_path=model_path, command_path=shared.cmd_opts.ckpt_dir, ext_filter=[".ckpt", ".safetensors"]) @@ -73,7 +81,7 @@ def list_models(): if os.path.exists(cmd_ckpt): h = model_hash(cmd_ckpt) title, short_model_name = modeltitle(cmd_ckpt, h) - checkpoints_list[title] = CheckpointInfo(cmd_ckpt, title, h, short_model_name, shared.cmd_opts.config) + checkpoints_list[title] = CheckpointInfo(cmd_ckpt, title, h, short_model_name) shared.opts.data['sd_model_checkpoint'] = title elif cmd_ckpt is not None and cmd_ckpt != shared.default_sd_model_file: print(f"Checkpoint in --ckpt argument not found (Possible it was moved to {model_path}: {cmd_ckpt}", file=sys.stderr) @@ -81,12 +89,7 @@ def list_models(): h = model_hash(filename) title, short_model_name = modeltitle(filename, h) - basename, _ = os.path.splitext(filename) - config = basename + ".yaml" - if not os.path.exists(config): - config = shared.cmd_opts.config - - checkpoints_list[title] = CheckpointInfo(filename, title, h, short_model_name, config) + checkpoints_list[title] = CheckpointInfo(filename, title, h, short_model_name) def get_closet_checkpoint_match(searchString): @@ -282,9 +285,10 @@ def enable_midas_autodownload(): def load_model(checkpoint_info=None): from modules import lowvram, sd_hijack checkpoint_info = checkpoint_info or select_checkpoint() + checkpoint_config = find_checkpoint_config(checkpoint_info) - if checkpoint_info.config != shared.cmd_opts.config: - print(f"Loading config from: {checkpoint_info.config}") + if checkpoint_config != shared.cmd_opts.config: + print(f"Loading config from: {checkpoint_config}") if shared.sd_model: sd_hijack.model_hijack.undo_hijack(shared.sd_model) @@ -292,7 +296,7 @@ def load_model(checkpoint_info=None): gc.collect() devices.torch_gc() - sd_config = OmegaConf.load(checkpoint_info.config) + sd_config = OmegaConf.load(checkpoint_config) if should_hijack_inpainting(checkpoint_info): # Hardcoded config for now... @@ -302,7 +306,7 @@ def load_model(checkpoint_info=None): sd_config.model.params.finetune_keys = None # Create a "fake" config with a different name so that we know to unload it when switching models. - checkpoint_info = checkpoint_info._replace(config=checkpoint_info.config.replace(".yaml", "-inpainting.yaml")) + checkpoint_info = checkpoint_info._replace(config=checkpoint_config.replace(".yaml", "-inpainting.yaml")) if not hasattr(sd_config.model.params, "use_ema"): sd_config.model.params.use_ema = False @@ -343,11 +347,12 @@ def reload_model_weights(sd_model=None, info=None): sd_model = shared.sd_model current_checkpoint_info = sd_model.sd_checkpoint_info + checkpoint_config = find_checkpoint_config(current_checkpoint_info) if sd_model.sd_model_checkpoint == checkpoint_info.filename: return - if sd_model.sd_checkpoint_info.config != checkpoint_info.config or should_hijack_inpainting(checkpoint_info) != should_hijack_inpainting(sd_model.sd_checkpoint_info): + if checkpoint_config != find_checkpoint_config(checkpoint_info) or should_hijack_inpainting(checkpoint_info) != should_hijack_inpainting(sd_model.sd_checkpoint_info): del sd_model checkpoints_loaded.clear() load_model(checkpoint_info) -- cgit v1.2.3 From 96cf15bedecbed97ef9b70b8413d543a9aee5adf Mon Sep 17 00:00:00 2001 From: MMaker Date: Wed, 4 Jan 2023 05:12:06 -0500 Subject: Add new latent upscale modes --- modules/shared.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 7588c47b..a10f69a9 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -564,8 +564,11 @@ if os.path.exists(config_filename): latent_upscale_default_mode = "Latent" latent_upscale_modes = { - "Latent": "bilinear", - "Latent (nearest)": "nearest", + "Latent": {"mode": "bilinear", "antialias": False}, + "Latent (antialiased)": {"mode": "bilinear", "antialias": True}, + "Latent (bicubic)": {"mode": "bicubic", "antialias": False}, + "Latent (bicubic, antialiased)": {"mode": "bicubic", "antialias": True}, + "Latent (nearest)": {"mode": "nearest", "antialias": False}, } sd_upscalers = [] -- cgit v1.2.3 From 15fd0b8bc4734ea85bca1acfb12b51465ab9817d Mon Sep 17 00:00:00 2001 From: MMaker Date: Wed, 4 Jan 2023 05:12:54 -0500 Subject: Update processing.py --- modules/processing.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index a172af0b..7c72b56a 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -713,7 +713,7 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): for i in range(samples.shape[0]): save_intermediate(samples, i) - samples = torch.nn.functional.interpolate(samples, size=(target_height // opt_f, target_width // opt_f), mode=latent_scale_mode) + samples = torch.nn.functional.interpolate(samples, size=(target_height // opt_f, target_width // opt_f), mode=latent_scale_mode["mode"], antialias=latent_scale_mode["antialias"]) # Avoid making the inpainting conditioning unless necessary as # this does need some extra compute to decode / encode the image again. -- cgit v1.2.3 From 4ec6470a1a2d9430b91266426f995e48f59564e1 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 13:26:23 +0300 Subject: fix checkpoint list API --- modules/api/api.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 9c670f00..2b1f180c 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -18,7 +18,7 @@ from modules.textual_inversion.textual_inversion import create_embedding, train_ from modules.textual_inversion.preprocess import preprocess from modules.hypernetworks.hypernetwork import create_hypernetwork, train_hypernetwork from PIL import PngImagePlugin,Image -from modules.sd_models import checkpoints_list +from modules.sd_models import checkpoints_list, find_checkpoint_config from modules.realesrgan_model import get_realesrgan_models from modules import devices from typing import List @@ -303,7 +303,7 @@ class Api: return upscalers def get_sd_models(self): - return [{"title":x.title, "model_name":x.model_name, "hash":x.hash, "filename": x.filename, "config": x.config} for x in checkpoints_list.values()] + return [{"title":x.title, "model_name":x.model_name, "hash":x.hash, "filename": x.filename, "config": find_checkpoint_config(x)} for x in checkpoints_list.values()] def get_hypernetworks(self): return [{"name": name, "path": shared.hypernetworks[name]} for name in shared.hypernetworks] -- cgit v1.2.3 From b2151b934fe0a3613570c6abd7615d3788fd1c8f Mon Sep 17 00:00:00 2001 From: MMaker Date: Wed, 4 Jan 2023 05:36:18 -0500 Subject: Rename bicubic antialiased option Comma was causing the the value in PNG info to be quoted, which causes the upscaler dropdown option to be blank when sending to UI --- modules/shared.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index a10f69a9..c1b20081 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -567,7 +567,7 @@ latent_upscale_modes = { "Latent": {"mode": "bilinear", "antialias": False}, "Latent (antialiased)": {"mode": "bilinear", "antialias": True}, "Latent (bicubic)": {"mode": "bicubic", "antialias": False}, - "Latent (bicubic, antialiased)": {"mode": "bicubic", "antialias": True}, + "Latent (bicubic antialiased)": {"mode": "bicubic", "antialias": True}, "Latent (nearest)": {"mode": "nearest", "antialias": False}, } -- cgit v1.2.3 From 3bd737767b071878ea980e94b8705f603bcf545e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 14:20:32 +0300 Subject: disable broken API logging --- modules/api/api.py | 1 - 1 file changed, 1 deletion(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index a6c1d6ed..6267afdc 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -100,7 +100,6 @@ class Api: self.router = APIRouter() self.app = app - init_api_middleware(self.app) self.queue_lock = queue_lock self.add_api_route("/sdapi/v1/txt2img", self.text2imgapi, methods=["POST"], response_model=TextToImageResponse) self.add_api_route("/sdapi/v1/img2img", self.img2imgapi, methods=["POST"], response_model=ImageToImageResponse) -- cgit v1.2.3 From 0cd6399b8b1699b8b7acad6f0ad2988111fe618e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 14:29:13 +0300 Subject: fix broken inpainting model --- modules/sd_models.py | 3 --- 1 file changed, 3 deletions(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index 6dca4ddf..a568823d 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -305,9 +305,6 @@ def load_model(checkpoint_info=None): sd_config.model.params.unet_config.params.in_channels = 9 sd_config.model.params.finetune_keys = None - # Create a "fake" config with a different name so that we know to unload it when switching models. - checkpoint_info = checkpoint_info._replace(config=checkpoint_config.replace(".yaml", "-inpainting.yaml")) - if not hasattr(sd_config.model.params, "use_ema"): sd_config.model.params.use_ema = False -- cgit v1.2.3 From 11b8160a086c434d5baf4971edda46e6d2126800 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 4 Jan 2023 06:36:57 -0500 Subject: fix typo --- modules/api/api.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 6267afdc..48a70a44 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -101,6 +101,7 @@ class Api: self.router = APIRouter() self.app = app self.queue_lock = queue_lock + api_middleware(self.app) self.add_api_route("/sdapi/v1/txt2img", self.text2imgapi, methods=["POST"], response_model=TextToImageResponse) self.add_api_route("/sdapi/v1/img2img", self.img2imgapi, methods=["POST"], response_model=ImageToImageResponse) self.add_api_route("/sdapi/v1/extra-single-image", self.extras_single_image_api, methods=["POST"], response_model=ExtrasSingleImageResponse) -- cgit v1.2.3 From 642142556d8ecdea9beb86d7618b628b1803ab98 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 15:09:53 +0300 Subject: use commandline-supplied cuda device name instead of cuda:0 for safetensors PR that doesn't fix anything --- modules/sd_models.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index ee918f24..76a89e88 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -173,7 +173,7 @@ def read_state_dict(checkpoint_file, print_global_state=False, map_location=None if extension.lower() == ".safetensors": device = map_location or shared.weight_load_location if device is None: - device = "cuda:0" if torch.cuda.is_available() else "cpu" + device = devices.get_cuda_device_string() if torch.cuda.is_available() else "cpu" pl_sd = safetensors.torch.load_file(checkpoint_file, device=device) else: pl_sd = torch.load(checkpoint_file, map_location=map_location or shared.weight_load_location) -- cgit v1.2.3 From 21ee77db314ede7ccbb18787962347c09a4df0c7 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 4 Jan 2023 08:04:38 -0500 Subject: add cross-attention info --- modules/sd_hijack.py | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index edcbaf52..fa2cd4bb 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -35,26 +35,35 @@ def apply_optimizations(): ldm.modules.diffusionmodules.model.nonlinearity = silu ldm.modules.diffusionmodules.openaimodel.th = sd_hijack_unet.th + + optimization_method = None if cmd_opts.force_enable_xformers or (cmd_opts.xformers and shared.xformers_available and torch.version.cuda and (6, 0) <= torch.cuda.get_device_capability(shared.device) <= (9, 0)): print("Applying xformers cross attention optimization.") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.xformers_attention_forward ldm.modules.diffusionmodules.model.AttnBlock.forward = sd_hijack_optimizations.xformers_attnblock_forward + optimization_method = 'xformers' elif cmd_opts.opt_split_attention_v1: print("Applying v1 cross attention optimization.") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_v1 + optimization_method = 'V1' elif not cmd_opts.disable_opt_split_attention and (cmd_opts.opt_split_attention_invokeai or not torch.cuda.is_available()): if not invokeAI_mps_available and shared.device.type == 'mps': print("The InvokeAI cross attention optimization for MPS requires the psutil package which is not installed.") print("Applying v1 cross attention optimization.") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_v1 + optimization_method = 'V1' else: print("Applying cross attention optimization (InvokeAI).") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_invokeAI + optimization_method = 'InvokeAI' elif not cmd_opts.disable_opt_split_attention and (cmd_opts.opt_split_attention or torch.cuda.is_available()): print("Applying cross attention optimization (Doggettx).") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward ldm.modules.diffusionmodules.model.AttnBlock.forward = sd_hijack_optimizations.cross_attention_attnblock_forward + optimization_method = 'Doggettx' + + return optimization_method def undo_optimizations(): @@ -75,6 +84,7 @@ class StableDiffusionModelHijack: layers = None circular_enabled = False clip = None + optimization_method = None embedding_db = modules.textual_inversion.textual_inversion.EmbeddingDatabase(cmd_opts.embeddings_dir) @@ -94,7 +104,7 @@ class StableDiffusionModelHijack: m.cond_stage_model.model.token_embedding = EmbeddingsWithFixes(m.cond_stage_model.model.token_embedding, self) m.cond_stage_model = sd_hijack_open_clip.FrozenOpenCLIPEmbedderWithCustomWords(m.cond_stage_model, self) - apply_optimizations() + self.optimization_method = apply_optimizations() self.clip = m.cond_stage_model -- cgit v1.2.3 From 1cfd8aec4ae5a6ca1afd67b44cb4ef6dd14d8c34 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 16:05:42 +0300 Subject: make it possible to work with opts.show_progress_every_n_steps = -1 with medvram --- modules/shared.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 4fcc6edd..54a6ba23 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -214,12 +214,13 @@ class State: """sets self.current_image from self.current_latent if enough sampling steps have been made after the last call to this""" def set_current_image(self): + if not parallel_processing_allowed: + return + if self.sampling_step - self.current_image_sampling_step >= opts.show_progress_every_n_steps and opts.show_progress_every_n_steps > 0: self.do_set_current_image() def do_set_current_image(self): - if not parallel_processing_allowed: - return if self.current_latent is None: return @@ -231,6 +232,7 @@ class State: self.current_image_sampling_step = self.sampling_step + state = State() artist_db = modules.artists.ArtistsDatabase(os.path.join(script_path, 'artists.csv')) -- cgit v1.2.3 From 79c682ad4f2d982b26fa1a15044582d1005134f9 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 4 Jan 2023 08:20:42 -0500 Subject: fix jpeg --- modules/extras.py | 2 -- modules/images.py | 2 ++ 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index d665440a..7407bfe3 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -19,8 +19,6 @@ from modules.shared import opts import modules.gfpgan_model from modules.ui import plaintext_to_html import modules.codeformer_model -import piexif -import piexif.helper import gradio as gr import safetensors.torch diff --git a/modules/images.py b/modules/images.py index c3a5fc8b..a73be3fa 100644 --- a/modules/images.py +++ b/modules/images.py @@ -22,6 +22,8 @@ from modules.shared import opts, cmd_opts LANCZOS = (Image.Resampling.LANCZOS if hasattr(Image, 'Resampling') else Image.LANCZOS) +Image.init() # initialize once all known file format handlers + def image_grid(imgs, batch_size=1, rows=None): if rows is None: -- cgit v1.2.3 From 4d66bf2c0d27702cc83b9cc57ebb1f359d18d938 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 17:24:46 +0300 Subject: add infotext to "-before-highres-fix" images --- modules/processing.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index fd7c7015..c03e77e7 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -136,6 +136,7 @@ class StableDiffusionProcessing(): self.all_negative_prompts = None self.all_seeds = None self.all_subseeds = None + self.iteration = 0 def txt2img_image_conditioning(self, x, width=None, height=None): if self.sampler.conditioning_key not in {'hybrid', 'concat'}: @@ -544,6 +545,8 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: state.job_count = p.n_iter for n in range(p.n_iter): + p.iteration = n + if state.skipped: state.skipped = False @@ -707,7 +710,8 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): if not isinstance(image, Image.Image): image = sd_samplers.sample_to_image(image, index, approximation=0) - images.save_image(image, self.outpath_samples, "", seeds[index], prompts[index], opts.samples_format, suffix="-before-highres-fix") + info = create_infotext(self, self.all_prompts, self.all_seeds, self.all_subseeds, [], iteration=self.iteration, position_in_batch=index) + images.save_image(image, self.outpath_samples, "", seeds[index], prompts[index], opts.samples_format, info=info, suffix="-before-highres-fix") if latent_scale_mode is not None: for i in range(samples.shape[0]): -- cgit v1.2.3 From 184e670126f5fc50ba56fa0fedcf0cf60e45ed7e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 17:45:01 +0300 Subject: fix the merge --- modules/textual_inversion/textual_inversion.py | 14 +++++--------- 1 file changed, 5 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 5421a758..8731ea5d 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -251,6 +251,7 @@ def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, dat if save_model_every or create_image_every: assert log_directory, "Log directory is empty" + def create_dummy_mask(x, width=None, height=None): if shared.sd_model.model.conditioning_key in {'hybrid', 'concat'}: @@ -380,17 +381,12 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ break with devices.autocast(): - # c = stack_conds(batch.cond).to(devices.device) - # mask = torch.tensor(batch.emb_index).to(devices.device, non_blocking=pin_memory) - # print(mask) - # c[:, 1:1+embedding.vec.shape[0]] = embedding.vec.to(devices.device, non_blocking=pin_memory) - - - if img_c is None: - img_c = create_dummy_mask(c, training_width, training_height) - x = batch.latent_sample.to(devices.device, non_blocking=pin_memory) c = shared.sd_model.cond_stage_model(batch.cond_text) + + if img_c is None: + img_c = create_dummy_mask(c, training_width, training_height) + cond = {"c_concat": [img_c], "c_crossattn": [c]} loss = shared.sd_model(x, cond)[0] / gradient_step del x -- cgit v1.2.3 From 590c5ae016ae494f4873ca20079b30684ea3060c Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 4 Jan 2023 09:48:54 -0500 Subject: update pillow --- modules/images.py | 2 -- 1 file changed, 2 deletions(-) (limited to 'modules') diff --git a/modules/images.py b/modules/images.py index a73be3fa..c3a5fc8b 100644 --- a/modules/images.py +++ b/modules/images.py @@ -22,8 +22,6 @@ from modules.shared import opts, cmd_opts LANCZOS = (Image.Resampling.LANCZOS if hasattr(Image, 'Resampling') else Image.LANCZOS) -Image.init() # initialize once all known file format handlers - def image_grid(imgs, batch_size=1, rows=None): if rows is None: -- cgit v1.2.3 From 525cea924562afd676f55470095268a0f6fca59e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 17:58:07 +0300 Subject: use shared function from processing for creating dummy mask when training inpainting model --- modules/processing.py | 39 +++++++++++++------------- modules/textual_inversion/textual_inversion.py | 33 ++++++---------------- 2 files changed, 29 insertions(+), 43 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index c03e77e7..c7264aff 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -76,6 +76,24 @@ def apply_overlay(image, paste_loc, index, overlays): return image +def txt2img_image_conditioning(sd_model, x, width, height): + if sd_model.model.conditioning_key not in {'hybrid', 'concat'}: + # Dummy zero conditioning if we're not using inpainting model. + # Still takes up a bit of memory, but no encoder call. + # Pretty sure we can just make this a 1x1 image since its not going to be used besides its batch size. + return x.new_zeros(x.shape[0], 5, 1, 1, dtype=x.dtype, device=x.device) + + # The "masked-image" in this case will just be all zeros since the entire image is masked. + image_conditioning = torch.zeros(x.shape[0], 3, height, width, device=x.device) + image_conditioning = sd_model.get_first_stage_encoding(sd_model.encode_first_stage(image_conditioning)) + + # Add the fake full 1s mask to the first dimension. + image_conditioning = torch.nn.functional.pad(image_conditioning, (0, 0, 0, 0, 1, 0), value=1.0) + image_conditioning = image_conditioning.to(x.dtype) + + return image_conditioning + + class StableDiffusionProcessing(): """ The first set of paramaters: sd_models -> do_not_reload_embeddings represent the minimum required to create a StableDiffusionProcessing @@ -139,26 +157,9 @@ class StableDiffusionProcessing(): self.iteration = 0 def txt2img_image_conditioning(self, x, width=None, height=None): - if self.sampler.conditioning_key not in {'hybrid', 'concat'}: - # Dummy zero conditioning if we're not using inpainting model. - # Still takes up a bit of memory, but no encoder call. - # Pretty sure we can just make this a 1x1 image since its not going to be used besides its batch size. - return x.new_zeros(x.shape[0], 5, 1, 1) + self.is_using_inpainting_conditioning = self.sd_model.model.conditioning_key in {'hybrid', 'concat'} - self.is_using_inpainting_conditioning = True - - height = height or self.height - width = width or self.width - - # The "masked-image" in this case will just be all zeros since the entire image is masked. - image_conditioning = torch.zeros(x.shape[0], 3, height, width, device=x.device) - image_conditioning = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(image_conditioning)) - - # Add the fake full 1s mask to the first dimension. - image_conditioning = torch.nn.functional.pad(image_conditioning, (0, 0, 0, 0, 1, 0), value=1.0) - image_conditioning = image_conditioning.to(x.dtype) - - return image_conditioning + return txt2img_image_conditioning(self.sd_model, x, width or self.width, height or self.height) def depth2img_image_conditioning(self, source_image): # Use the AddMiDaS helper to Format our source image to suit the MiDaS model diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 8731ea5d..2250e41b 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -252,26 +252,6 @@ def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, dat assert log_directory, "Log directory is empty" -def create_dummy_mask(x, width=None, height=None): - if shared.sd_model.model.conditioning_key in {'hybrid', 'concat'}: - - # The "masked-image" in this case will just be all zeros since the entire image is masked. - image_conditioning = torch.zeros(x.shape[0], 3, height, width, device=x.device) - image_conditioning = shared.sd_model.get_first_stage_encoding(shared.sd_model.encode_first_stage(image_conditioning)) - - # Add the fake full 1s mask to the first dimension. - image_conditioning = torch.nn.functional.pad(image_conditioning, (0, 0, 0, 0, 1, 0), value=1.0) - image_conditioning = image_conditioning.to(x.dtype) - - else: - # Dummy zero conditioning if we're not using inpainting model. - # Still takes up a bit of memory, but no encoder call. - # Pretty sure we can just make this a 1x1 image since its not going to be used besides its batch size. - image_conditioning = torch.zeros(x.shape[0], 5, 1, 1, dtype=x.dtype, device=x.device) - - return image_conditioning - - def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, steps, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): save_embedding_every = save_embedding_every or 0 create_image_every = create_image_every or 0 @@ -346,7 +326,6 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ else: print("No saved optimizer exists in checkpoint") - scaler = torch.cuda.amp.GradScaler() batch_size = ds.batch_size @@ -362,7 +341,9 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ forced_filename = "" embedding_yet_to_be_embedded = False + is_training_inpainting_model = shared.sd_model.model.conditioning_key in {'hybrid', 'concat'} img_c = None + pbar = tqdm.tqdm(total=steps - initial_step) try: for i in range((steps-initial_step) * gradient_step): @@ -384,10 +365,14 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ x = batch.latent_sample.to(devices.device, non_blocking=pin_memory) c = shared.sd_model.cond_stage_model(batch.cond_text) - if img_c is None: - img_c = create_dummy_mask(c, training_width, training_height) + if is_training_inpainting_model: + if img_c is None: + img_c = processing.txt2img_image_conditioning(shared.sd_model, c, training_width, training_height) + + cond = {"c_concat": [img_c], "c_crossattn": [c]} + else: + cond = c - cond = {"c_concat": [img_c], "c_crossattn": [c]} loss = shared.sd_model(x, cond)[0] / gradient_step del x -- cgit v1.2.3 From a8eb9e3bf814f72293e474c11e9ff0098859a942 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 18:20:38 +0300 Subject: Revert "Merge pull request #3791 from shirayu/fix/filename" This reverts commit eed58279e7cb0e873ebd88a29609f9bab0f1f3af, reversing changes made to 4ae960b01c6711c66985479f14809dc7fa549fc2. --- modules/images.py | 16 ++++------------ 1 file changed, 4 insertions(+), 12 deletions(-) (limited to 'modules') diff --git a/modules/images.py b/modules/images.py index 2967fa9a..c3a5fc8b 100644 --- a/modules/images.py +++ b/modules/images.py @@ -447,14 +447,6 @@ def get_next_sequence_number(path, basename): return result + 1 -def truncate_fullpath(full_path, encoding='utf-8'): - dir_name, full_name = os.path.split(full_path) - file_name, file_ext = os.path.splitext(full_name) - max_length = os.statvfs(dir_name).f_namemax - file_name_truncated = file_name.encode(encoding)[:max_length - len(file_ext)].decode(encoding, 'ignore') - return os.path.join(dir_name , file_name_truncated + file_ext) - - def save_image(image, path, basename, seed=None, prompt=None, extension='png', info=None, short_filename=False, no_prompt=False, grid=False, pnginfo_section_name='parameters', p=None, existing_info=None, forced_filename=None, suffix="", save_to_dirs=None): """Save an image. @@ -495,7 +487,7 @@ def save_image(image, path, basename, seed=None, prompt=None, extension='png', i if save_to_dirs: dirname = namegen.apply(opts.directories_filename_pattern or "[prompt_words]").lstrip(' ').rstrip('\\ /') - path = truncate_fullpath(os.path.join(path, dirname)) + path = os.path.join(path, dirname) os.makedirs(path, exist_ok=True) @@ -519,13 +511,13 @@ def save_image(image, path, basename, seed=None, prompt=None, extension='png', i fullfn = None for i in range(500): fn = f"{basecount + i:05}" if basename == '' else f"{basename}-{basecount + i:04}" - fullfn = truncate_fullpath(os.path.join(path, f"{fn}{file_decoration}.{extension}")) + fullfn = os.path.join(path, f"{fn}{file_decoration}.{extension}") if not os.path.exists(fullfn): break else: - fullfn = truncate_fullpath(os.path.join(path, f"{file_decoration}.{extension}")) + fullfn = os.path.join(path, f"{file_decoration}.{extension}") else: - fullfn = truncate_fullpath(os.path.join(path, f"{forced_filename}.{extension}")) + fullfn = os.path.join(path, f"{forced_filename}.{extension}") pnginfo = existing_info or {} if info is not None: -- cgit v1.2.3 From 3dae545a03f5102ba5d9c3f27bb6241824c5a916 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 18:42:51 +0300 Subject: rename weirdly named variables from #3176 --- modules/ui.py | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index e4859020..184af7ad 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -162,16 +162,14 @@ def save_files(js_data, images, do_make_zip, index): return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") - - -def calc_time_left(progress, threshold, label, force_display, showTime): +def calc_time_left(progress, threshold, label, force_display, show_eta): if progress == 0: return "" else: time_since_start = time.time() - shared.state.time_start eta = (time_since_start/progress) eta_relative = eta-time_since_start - if (eta_relative > threshold and showTime) or force_display: + if (eta_relative > threshold and show_eta) or force_display: if eta_relative > 3600: return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) elif eta_relative > 60: @@ -194,9 +192,9 @@ def check_progress_call(id_part): progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps # Show progress percentage and time left at the same moment, and base it also on steps done - showPBText = progress >= 0.01 or shared.state.sampling_step >= 10 + show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 - time_left = calc_time_left( progress, 1, " ETA: ", shared.state.time_left_force_display, showPBText ) + time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) if time_left != "": shared.state.time_left_force_display = True @@ -204,7 +202,7 @@ def check_progress_call(id_part): progressbar = "" if opts.show_progressbar: - progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if showPBText else ""}
""" + progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" image = gr_show(False) preview_visibility = gr_show(False) -- cgit v1.2.3 From 097a90b88bb92878cf435c513b4757b5b82ae299 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 19:19:11 +0300 Subject: add XY plot parameters to grid image and do not add them to individual images --- modules/processing.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index c7264aff..47712159 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -422,7 +422,7 @@ def fix_seed(p): p.subseed = get_fixed_seed(p.subseed) -def create_infotext(p, all_prompts, all_seeds, all_subseeds, comments, iteration=0, position_in_batch=0): +def create_infotext(p, all_prompts, all_seeds, all_subseeds, comments=None, iteration=0, position_in_batch=0): index = position_in_batch + iteration * p.batch_size clip_skip = getattr(p, 'clip_skip', opts.CLIP_stop_at_last_layers) -- cgit v1.2.3 From 24d4a0841d3cc0e5908b098f65a9caa3fa889af8 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 20:10:40 +0300 Subject: train tab visual updates allow setting train tab values from ui-config.json --- modules/ui.py | 35 +++++++++++++++++++++-------------- 1 file changed, 21 insertions(+), 14 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 72e7b7d2..44f4f3a4 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1281,42 +1281,48 @@ def create_ui(): with gr.Tab(label="Train"): gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") - with gr.Row(): + with FormRow(): train_embedding_name = gr.Dropdown(label='Embedding', elem_id="train_embedding", choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())) create_refresh_button(train_embedding_name, sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings, lambda: {"choices": sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())}, "refresh_train_embedding_name") - with gr.Row(): + train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") - with gr.Row(): + + with FormRow(): embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") - with gr.Row(): + with FormRow(): clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) - batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") - gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") + with FormRow(): + batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") + gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") + dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") template_file = gr.Textbox(label='Prompt template file', value=os.path.join(script_path, "textual_inversion_templates", "style_filewords.txt"), elem_id="train_template_file") training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") - create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") - save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") + + with FormRow(): + create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") + save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") + save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") - with gr.Row(): - shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") - tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") - with gr.Row(): - latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") + + shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") + tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") + + latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") with gr.Row(): + train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") - train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") params = script_callbacks.UiTrainTabParams(txt2img_preview_params) @@ -1803,6 +1809,7 @@ def create_ui(): visit(img2img_interface, loadsave, "img2img") visit(extras_interface, loadsave, "extras") visit(modelmerger_interface, loadsave, "modelmerger") + visit(train_interface, loadsave, "train") if not error_loading and (not os.path.exists(ui_config_file) or settings_count != len(ui_settings)): with open(ui_config_file, "w", encoding="utf8") as file: -- cgit v1.2.3 From 81490780949fffed77493b4bd741e96ec737fe27 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 22:04:40 +0300 Subject: added the option to specify target resolution with possibility of truncating for hires fix; also sampling steps --- modules/generation_parameters_copypaste.py | 9 ++++-- modules/processing.py | 51 +++++++++++++++++++++++++++--- modules/txt2img.py | 5 ++- modules/ui.py | 24 ++++++++++---- 4 files changed, 74 insertions(+), 15 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 4baf4d9a..12a9de3d 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -212,11 +212,10 @@ def restore_old_hires_fix_params(res): firstpass_width = math.ceil(scale * width / 64) * 64 firstpass_height = math.ceil(scale * height / 64) * 64 - hr_scale = width / firstpass_width if firstpass_width > 0 else height / firstpass_height - res['Size-1'] = firstpass_width res['Size-2'] = firstpass_height - res['Hires upscale'] = hr_scale + res['Hires resize-1'] = width + res['Hires resize-2'] = height def parse_generation_parameters(x: str): @@ -276,6 +275,10 @@ Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 965400086, Size: 512x512, Model hypernet_hash = res.get("Hypernet hash", None) res["Hypernet"] = find_hypernetwork_key(hypernet_name, hypernet_hash) + if "Hires resize-1" not in res: + res["Hires resize-1"] = 0 + res["Hires resize-2"] = 0 + restore_old_hires_fix_params(res) return res diff --git a/modules/processing.py b/modules/processing.py index 47712159..9cad05f2 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -662,12 +662,17 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): sampler = None - def __init__(self, enable_hr: bool = False, denoising_strength: float = 0.75, firstphase_width: int = 0, firstphase_height: int = 0, hr_scale: float = 2.0, hr_upscaler: str = None, **kwargs): + def __init__(self, enable_hr: bool = False, denoising_strength: float = 0.75, firstphase_width: int = 0, firstphase_height: int = 0, hr_scale: float = 2.0, hr_upscaler: str = None, hr_second_pass_steps: int = 0, hr_resize_x: int = 0, hr_resize_y: int = 0, **kwargs): super().__init__(**kwargs) self.enable_hr = enable_hr self.denoising_strength = denoising_strength self.hr_scale = hr_scale self.hr_upscaler = hr_upscaler + self.hr_second_pass_steps = hr_second_pass_steps + self.hr_resize_x = hr_resize_x + self.hr_resize_y = hr_resize_y + self.hr_upscale_to_x = hr_resize_x + self.hr_upscale_to_y = hr_resize_y if firstphase_width != 0 or firstphase_height != 0: print("firstphase_width/firstphase_height no longer supported; use hr_scale", file=sys.stderr) @@ -675,6 +680,9 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): self.width = firstphase_width self.height = firstphase_height + self.truncate_x = 0 + self.truncate_y = 0 + def init(self, all_prompts, all_seeds, all_subseeds): if self.enable_hr: if state.job_count == -1: @@ -682,7 +690,38 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): else: state.job_count = state.job_count * 2 - self.extra_generation_params["Hires upscale"] = self.hr_scale + if self.hr_resize_x == 0 and self.hr_resize_y == 0: + self.extra_generation_params["Hires upscale"] = self.hr_scale + self.hr_upscale_to_x = int(self.width * self.hr_scale) + self.hr_upscale_to_y = int(self.height * self.hr_scale) + else: + self.extra_generation_params["Hires resize"] = f"{self.hr_resize_x}x{self.hr_resize_y}" + + if self.hr_resize_y == 0: + self.hr_upscale_to_x = self.hr_resize_x + self.hr_upscale_to_y = self.hr_resize_x * self.height // self.width + elif self.hr_resize_x == 0: + self.hr_upscale_to_x = self.hr_resize_y * self.width // self.height + self.hr_upscale_to_y = self.hr_resize_y + else: + target_w = self.hr_resize_x + target_h = self.hr_resize_y + src_ratio = self.width / self.height + dst_ratio = self.hr_resize_x / self.hr_resize_y + + if src_ratio < dst_ratio: + self.hr_upscale_to_x = self.hr_resize_x + self.hr_upscale_to_y = self.hr_resize_x * self.height // self.width + else: + self.hr_upscale_to_x = self.hr_resize_y * self.width // self.height + self.hr_upscale_to_y = self.hr_resize_y + + self.truncate_x = (self.hr_upscale_to_x - target_w) // opt_f + self.truncate_y = (self.hr_upscale_to_y - target_h) // opt_f + + if self.hr_second_pass_steps: + self.extra_generation_params["Hires steps"] = self.hr_second_pass_steps + if self.hr_upscaler is not None: self.extra_generation_params["Hires upscaler"] = self.hr_upscaler @@ -699,8 +738,8 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): if not self.enable_hr: return samples - target_width = int(self.width * self.hr_scale) - target_height = int(self.height * self.hr_scale) + target_width = self.hr_upscale_to_x + target_height = self.hr_upscale_to_y def save_intermediate(image, index): """saves image before applying hires fix, if enabled in options; takes as an argument either an image or batch with latent space images""" @@ -755,13 +794,15 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): self.sampler = sd_samplers.create_sampler(self.sampler_name, self.sd_model) + samples = samples[:, :, self.truncate_y//2:samples.shape[2]-(self.truncate_y+1)//2, self.truncate_x//2:samples.shape[3]-(self.truncate_x+1)//2] + noise = create_random_tensors(samples.shape[1:], seeds=seeds, subseeds=subseeds, subseed_strength=subseed_strength, p=self) # GC now before running the next img2img to prevent running out of memory x = None devices.torch_gc() - samples = self.sampler.sample_img2img(self, samples, noise, conditioning, unconditional_conditioning, steps=self.steps, image_conditioning=image_conditioning) + samples = self.sampler.sample_img2img(self, samples, noise, conditioning, unconditional_conditioning, steps=self.hr_second_pass_steps or self.steps, image_conditioning=image_conditioning) return samples diff --git a/modules/txt2img.py b/modules/txt2img.py index e189a899..38b5f591 100644 --- a/modules/txt2img.py +++ b/modules/txt2img.py @@ -8,7 +8,7 @@ import modules.processing as processing from modules.ui import plaintext_to_html -def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, *args): +def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, hr_second_pass_steps: int, hr_resize_x: int, hr_resize_y: int, *args): p = StableDiffusionProcessingTxt2Img( sd_model=shared.sd_model, outpath_samples=opts.outdir_samples or opts.outdir_txt2img_samples, @@ -35,6 +35,9 @@ def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: denoising_strength=denoising_strength if enable_hr else None, hr_scale=hr_scale, hr_upscaler=hr_upscaler, + hr_second_pass_steps=hr_second_pass_steps, + hr_resize_x=hr_resize_x, + hr_resize_y=hr_resize_y, ) p.scripts = modules.scripts.scripts_txt2img diff --git a/modules/ui.py b/modules/ui.py index 44f4f3a4..04091e67 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -637,10 +637,10 @@ def create_sampler_and_steps_selection(choices, tabname): with FormRow(elem_id=f"sampler_selection_{tabname}"): sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") sampler_index.save_to_config = True - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling Steps", value=20) + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) else: with FormGroup(elem_id=f"sampler_selection_{tabname}"): - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling Steps", value=20) + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") return steps, sampler_index @@ -709,10 +709,16 @@ def create_ui(): enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") elif category == "hires_fix": - with FormRow(visible=False, elem_id="txt2img_hires_fix") as hr_options: - hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) - hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") + with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: + with FormRow(elem_id="txt2img_hires_fix_row1"): + hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) + hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") + + with FormRow(elem_id="txt2img_hires_fix_row2"): + hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") + hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") + hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") elif category == "batch": if not opts.dimensions_and_batch_together: @@ -753,6 +759,9 @@ def create_ui(): denoising_strength, hr_scale, hr_upscaler, + hr_second_pass_steps, + hr_resize_x, + hr_resize_y, ] + custom_inputs, outputs=[ @@ -804,6 +813,9 @@ def create_ui(): (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), (hr_scale, "Hires upscale"), (hr_upscaler, "Hires upscaler"), + (hr_second_pass_steps, "Hires steps"), + (hr_resize_x, "Hires resize-1"), + (hr_resize_y, "Hires resize-2"), *modules.scripts.scripts_txt2img.infotext_fields ] parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) -- cgit v1.2.3 From 1288a3bb7d21064e5bd0af7158a3840886027c51 Mon Sep 17 00:00:00 2001 From: Suffocate <70031311+lolsuffocate@users.noreply.github.com> Date: Wed, 4 Jan 2023 20:36:30 +0000 Subject: Use the read_info_from_image function directly --- modules/api/api.py | 16 ++++++++++++---- modules/api/models.py | 5 +++-- 2 files changed, 15 insertions(+), 6 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 48a70a44..2103709b 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -11,10 +11,10 @@ from fastapi.security import HTTPBasic, HTTPBasicCredentials from secrets import compare_digest import modules.shared as shared -from modules import sd_samplers, deepbooru, sd_hijack +from modules import sd_samplers, deepbooru, sd_hijack, images from modules.api.models import * from modules.processing import StableDiffusionProcessingTxt2Img, StableDiffusionProcessingImg2Img, process_images -from modules.extras import run_extras, run_pnginfo +from modules.extras import run_extras from modules.textual_inversion.textual_inversion import create_embedding, train_embedding from modules.textual_inversion.preprocess import preprocess from modules.hypernetworks.hypernetwork import create_hypernetwork, train_hypernetwork @@ -233,9 +233,17 @@ class Api: if(not req.image.strip()): return PNGInfoResponse(info="") - result = run_pnginfo(decode_base64_to_image(req.image.strip())) + image = decode_base64_to_image(req.image.strip()) + if image is None: + return PNGInfoResponse(info="") + + geninfo, items = images.read_info_from_image(image) + if geninfo is None: + geninfo = "" + + items = {**{'parameters': geninfo}, **items} - return PNGInfoResponse(info=result[1]) + return PNGInfoResponse(info=geninfo, items=items) def progressapi(self, req: ProgressRequest = Depends()): # copy from check_progress_call of ui.py diff --git a/modules/api/models.py b/modules/api/models.py index 4a632c68..d8198a27 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -157,7 +157,8 @@ class PNGInfoRequest(BaseModel): image: str = Field(title="Image", description="The base64 encoded PNG image") class PNGInfoResponse(BaseModel): - info: str = Field(title="Image info", description="A string with all the info the image had") + info: str = Field(title="Image info", description="A string with the parameters used to generate the image") + items: dict = Field(title="Items", description="An object containing all the info the image had") class ProgressRequest(BaseModel): skip_current_image: bool = Field(default=False, title="Skip current image", description="Skip current image serialization") @@ -258,4 +259,4 @@ class EmbeddingItem(BaseModel): class EmbeddingsResponse(BaseModel): loaded: Dict[str, EmbeddingItem] = Field(title="Loaded", description="Embeddings loaded for the current model") - skipped: Dict[str, EmbeddingItem] = Field(title="Skipped", description="Embeddings skipped for the current model (likely due to architecture incompatibility)") \ No newline at end of file + skipped: Dict[str, EmbeddingItem] = Field(title="Skipped", description="Embeddings skipped for the current model (likely due to architecture incompatibility)") -- cgit v1.2.3 From bc43293c640aef65df3136de9e5bd8b7e79eb3e0 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 23:56:43 +0300 Subject: fix incorrect display/calculation for number of steps for hires fix in progress bars --- modules/processing.py | 9 ++++++--- modules/sd_samplers.py | 5 +++-- modules/shared.py | 4 +++- 3 files changed, 12 insertions(+), 6 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 9cad05f2..f28e7212 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -685,10 +685,13 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): def init(self, all_prompts, all_seeds, all_subseeds): if self.enable_hr: - if state.job_count == -1: - state.job_count = self.n_iter * 2 - else: + if not state.processing_has_refined_job_count: + if state.job_count == -1: + state.job_count = self.n_iter + + shared.total_tqdm.updateTotal((self.steps + (self.hr_second_pass_steps or self.steps)) * state.job_count) state.job_count = state.job_count * 2 + state.processing_has_refined_job_count = True if self.hr_resize_x == 0 and self.hr_resize_y == 0: self.extra_generation_params["Hires upscale"] = self.hr_scale diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py index e904d860..3851a77f 100644 --- a/modules/sd_samplers.py +++ b/modules/sd_samplers.py @@ -97,8 +97,9 @@ sampler_extra_params = { def setup_img2img_steps(p, steps=None): if opts.img2img_fix_steps or steps is not None: - steps = int((steps or p.steps) / min(p.denoising_strength, 0.999)) if p.denoising_strength > 0 else 0 - t_enc = p.steps - 1 + requested_steps = (steps or p.steps) + steps = int(requested_steps / min(p.denoising_strength, 0.999)) if p.denoising_strength > 0 else 0 + t_enc = requested_steps - 1 else: steps = p.steps t_enc = int(min(p.denoising_strength, 0.999) * steps) diff --git a/modules/shared.py b/modules/shared.py index 54a6ba23..04c545ee 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -153,6 +153,7 @@ class State: job = "" job_no = 0 job_count = 0 + processing_has_refined_job_count = False job_timestamp = '0' sampling_step = 0 sampling_steps = 0 @@ -194,6 +195,7 @@ class State: def begin(self): self.sampling_step = 0 self.job_count = -1 + self.processing_has_refined_job_count = False self.job_no = 0 self.job_timestamp = datetime.datetime.now().strftime("%Y%m%d%H%M%S") self.current_latent = None @@ -608,7 +610,7 @@ class TotalTQDM: return if self._tqdm is None: self.reset() - self._tqdm.total=new_total + self._tqdm.total = new_total def clear(self): if self._tqdm is not None: -- cgit v1.2.3 From 5851bc839b6f639cda59e84eb1ee8c706986633d Mon Sep 17 00:00:00 2001 From: me <25877290+Kryptortio@users.noreply.github.com> Date: Wed, 4 Jan 2023 22:03:32 +0100 Subject: Add element ids for script components and a few more in ui.py --- modules/ui.py | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 04091e67..bb64fe20 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -560,7 +560,7 @@ Requested path was: {f} generation_info = None with gr.Column(): with gr.Row(elem_id=f"image_buttons_{tabname}"): - open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else 'open_folder') + open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') if tabname != "extras": save = gr.Button('Save', elem_id=f'save_{tabname}') @@ -576,13 +576,13 @@ Requested path was: {f} if tabname != "extras": with gr.Row(): - download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False) + download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') with gr.Group(): - html_info = gr.HTML() - html_log = gr.HTML() + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') - generation_info = gr.Textbox(visible=False) + generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') if tabname == 'txt2img' or tabname == 'img2img': generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") generation_info_button.click( @@ -624,9 +624,9 @@ Requested path was: {f} ) else: - html_info_x = gr.HTML() - html_info = gr.HTML() - html_log = gr.HTML() + html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log -- cgit v1.2.3 From 99b67cff0b48c4a1ad6e14d9cc591b11db6e293c Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 01:25:52 +0300 Subject: make hires fix not do anything if the user chooses the second pass resolution to be the same as first pass resolution --- modules/processing.py | 25 +++++++++++++++++-------- 1 file changed, 17 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index f28e7212..7e853287 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -683,16 +683,9 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): self.truncate_x = 0 self.truncate_y = 0 + def init(self, all_prompts, all_seeds, all_subseeds): if self.enable_hr: - if not state.processing_has_refined_job_count: - if state.job_count == -1: - state.job_count = self.n_iter - - shared.total_tqdm.updateTotal((self.steps + (self.hr_second_pass_steps or self.steps)) * state.job_count) - state.job_count = state.job_count * 2 - state.processing_has_refined_job_count = True - if self.hr_resize_x == 0 and self.hr_resize_y == 0: self.extra_generation_params["Hires upscale"] = self.hr_scale self.hr_upscale_to_x = int(self.width * self.hr_scale) @@ -722,6 +715,22 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): self.truncate_x = (self.hr_upscale_to_x - target_w) // opt_f self.truncate_y = (self.hr_upscale_to_y - target_h) // opt_f + # special case: the user has chosen to do nothing + if self.hr_upscale_to_x == self.width and self.hr_upscale_to_y == self.height: + self.enable_hr = False + self.denoising_strength = None + self.extra_generation_params.pop("Hires upscale", None) + self.extra_generation_params.pop("Hires resize", None) + return + + if not state.processing_has_refined_job_count: + if state.job_count == -1: + state.job_count = self.n_iter + + shared.total_tqdm.updateTotal((self.steps + (self.hr_second_pass_steps or self.steps)) * state.job_count) + state.job_count = state.job_count * 2 + state.processing_has_refined_job_count = True + if self.hr_second_pass_steps: self.extra_generation_params["Hires steps"] = self.hr_second_pass_steps -- cgit v1.2.3 From 2e30997450835ed8f80ab5e8f02f7d4c7f26dd3f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 10:21:17 +0300 Subject: move sd_model assignment to the place where we change the sd_model --- modules/processing.py | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index a12bd9e8..61e97077 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -466,12 +466,16 @@ def process_images(p: StableDiffusionProcessing) -> Processed: try: for k, v in p.override_settings.items(): setattr(opts, k, v) - if k == 'sd_hypernetwork': shared.reload_hypernetworks() # make onchange call for changing hypernet - if k == 'sd_model_checkpoint': sd_models.reload_model_weights() # make onchange call for changing SD model - if k == 'sd_vae': sd_vae.reload_vae_weights() # make onchange call for changing VAE + if k == 'sd_hypernetwork': + shared.reload_hypernetworks() # make onchange call for changing hypernet + + if k == 'sd_model_checkpoint': + sd_models.reload_model_weights() # make onchange call for changing SD model + p.sd_model = shared.sd_model + + if k == 'sd_vae': + sd_vae.reload_vae_weights() # make onchange call for changing VAE - # Assign sd_model here to ensure that it reflects the model after any changes - p.sd_model = shared.sd_model res = process_images_inner(p) finally: -- cgit v1.2.3 From 42fcc79bd31e5e5485f1cf115ad505cc623d0ac9 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 10:43:21 +0300 Subject: add Discard penultimate sigma to infotext --- modules/sd_samplers.py | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py index 31b255a3..01221b89 100644 --- a/modules/sd_samplers.py +++ b/modules/sd_samplers.py @@ -463,8 +463,12 @@ class KDiffusionSampler: return extra_params_kwargs def get_sigmas(self, p, steps): - disc = opts.always_discard_next_to_last_sigma or (self.config is not None and self.config.options.get('discard_next_to_last_sigma', False)) - steps += 1 if disc else 0 + discard_next_to_last_sigma = self.config is not None and self.config.options.get('discard_next_to_last_sigma', False) + if opts.always_discard_next_to_last_sigma and not discard_next_to_last_sigma: + discard_next_to_last_sigma = True + p.extra_generation_params["Discard penultimate sigma"] = True + + steps += 1 if discard_next_to_last_sigma else 0 if p.sampler_noise_scheduler_override: sigmas = p.sampler_noise_scheduler_override(steps) @@ -475,7 +479,7 @@ class KDiffusionSampler: else: sigmas = self.model_wrap.get_sigmas(steps) - if disc: + if discard_next_to_last_sigma: sigmas = torch.cat([sigmas[:-2], sigmas[-1:]]) return sigmas -- cgit v1.2.3 From 997461d3dd86f51c06ea0c2eff17ce8b8b48c0af Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 11:57:01 +0300 Subject: add footer with versions --- modules/ui.py | 31 ++++++++++++++++++++++++++++++- 1 file changed, 30 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index bb64fe20..81d96c5b 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1696,7 +1696,9 @@ def create_ui(): if os.path.exists("html/footer.html"): with open("html/footer.html", encoding="utf8") as file: - gr.HTML(file.read(), elem_id="footer") + footer = file.read() + footer = footer.format(versions=versions_html()) + gr.HTML(footer, elem_id="footer") text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) settings_submit.click( @@ -1857,3 +1859,30 @@ def reload_javascript(): if not hasattr(shared, 'GradioTemplateResponseOriginal'): shared.GradioTemplateResponseOriginal = gradio.routes.templates.TemplateResponse + + +def versions_html(): + import torch + import launch + + python_version = ".".join([str(x) for x in sys.version_info[0:3]]) + commit = launch.commit_hash() + short_commit = commit[0:8] + + if shared.xformers_available: + import xformers + xformers_version = xformers.__version__ + else: + xformers_version = "N/A" + + return f""" +python: {python_version} + •  +torch: {torch.__version__} + •  +xformers: {xformers_version} + •  +gradio: {gr.__version__} + •  +commit: {short_commit} +""" -- cgit v1.2.3 From f8d0cf6a6ec4911559cfecb9a9d1d46b547b38e8 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 12:08:11 +0300 Subject: rework #6329 to remove duplicate code and add prevent tab names for showing in ids for scripts that only exist on one tab --- modules/scripts.py | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'modules') diff --git a/modules/scripts.py b/modules/scripts.py index 722f8685..0c44f191 100644 --- a/modules/scripts.py +++ b/modules/scripts.py @@ -1,4 +1,5 @@ import os +import re import sys import traceback from collections import namedtuple @@ -128,6 +129,15 @@ class Script: """unused""" return "" + def elem_id(self, item_id): + """helper function to generate id for a HTML element, constructs final id out of script name, tab and user-supplied item_id""" + + need_tabname = self.show(True) == self.show(False) + tabname = ('img2img' if self.is_img2img else 'txt2txt') + "_" if need_tabname else "" + title = re.sub(r'[^a-z_0-9]', '', re.sub(r'\s', '_', self.title().lower())) + + return f'script_{tabname}{title}_{item_id}' + current_basedir = paths.script_path -- cgit v1.2.3 From eea8fc40e16664ddc8a9aec77206da704a35dde0 Mon Sep 17 00:00:00 2001 From: timntorres Date: Thu, 5 Jan 2023 07:24:22 -0800 Subject: Add option to save ti settings to file. --- modules/shared.py | 1 + modules/textual_inversion/textual_inversion.py | 30 +++++++++++++++++++++++--- 2 files changed, 28 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index e0f44c6d..933cd738 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -362,6 +362,7 @@ options_templates.update(options_section(('training', "Training"), { "unload_models_when_training": OptionInfo(False, "Move VAE and CLIP to RAM when training if possible. Saves VRAM."), "pin_memory": OptionInfo(False, "Turn on pin_memory for DataLoader. Makes training slightly faster but can increase memory usage."), "save_optimizer_state": OptionInfo(False, "Saves Optimizer state as separate *.optim file. Training of embedding or HN can be resumed with the matching optim file."), + "save_train_settings_to_txt": OptionInfo(False, "Save textual inversion and hypernet settings to a text file when training starts."), "dataset_filename_word_regex": OptionInfo("", "Filename word regex"), "dataset_filename_join_string": OptionInfo(" ", "Filename join string"), "training_image_repeats_per_epoch": OptionInfo(1, "Number of repeats for a single input image per epoch; used only for displaying epoch number", gr.Number, {"precision": 0}), diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 71e07bcc..2bed2ecb 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -1,6 +1,7 @@ import os import sys import traceback +import inspect import torch import tqdm @@ -229,6 +230,28 @@ def write_loss(log_directory, filename, step, epoch_len, values): **values, }) +def save_settings_to_file(initial_step, num_of_dataset_images, embedding_name, vectors_per_token, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): + checkpoint = sd_models.select_checkpoint() + model_name = checkpoint.model_name + model_hash = '[{}]'.format(checkpoint.hash) + + # Get a list of the argument names. + arg_names = inspect.getfullargspec(save_settings_to_file).args + + # Create a list of the argument names to include in the settings string. + names = arg_names[:16] # Include all arguments up until the preview-related ones. + if preview_from_txt2img: + names.extend(arg_names[16:]) # Include all remaining arguments if `preview_from_txt2img` is True. + + # Build the settings string. + settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n" + for name in names: + value = locals()[name] + settings_str += f"{name}: {value}\n" + + with open(os.path.join(log_directory, 'settings.txt'), "a+") as fout: + fout.write(settings_str + "\n\n") + def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, data_root, template_file, steps, save_model_every, create_image_every, log_directory, name="embedding"): assert model_name, f"{name} not selected" assert learn_rate, "Learning rate is empty or 0" @@ -292,13 +315,13 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ if initial_step >= steps: shared.state.textinfo = "Model has already been trained beyond specified max steps" return embedding, filename + scheduler = LearnRateScheduler(learn_rate, steps, initial_step) - clip_grad = torch.nn.utils.clip_grad_value_ if clip_grad_mode == "value" else \ torch.nn.utils.clip_grad_norm_ if clip_grad_mode == "norm" else \ None if clip_grad: - clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, ititial_step, verbose=False) + clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, initial_step, verbose=False) # dataset loading may take a while, so input validations and early returns should be done before this shared.state.textinfo = f"Preparing dataset from {html.escape(data_root)}..." old_parallel_processing_allowed = shared.parallel_processing_allowed @@ -306,7 +329,8 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ pin_memory = shared.opts.pin_memory ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=embedding_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) - + if shared.opts.save_train_settings_to_txt: + save_settings_to_file(initial_step , len(ds) , embedding_name, len(embedding.vec) , learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) latent_sampling_method = ds.latent_sampling_method dl = modules.textual_inversion.dataset.PersonalizedDataLoader(ds, latent_sampling_method=latent_sampling_method, batch_size=ds.batch_size, pin_memory=pin_memory) -- cgit v1.2.3 From 19a81ac2871ec900fc8b7955bbc2554b6c5ac6b1 Mon Sep 17 00:00:00 2001 From: cat Date: Thu, 5 Jan 2023 20:17:39 +0500 Subject: hires-fix: add "nearest-exact" latent upscale mode. --- modules/shared.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index e0f44c6d..b7a3ce5c 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -576,6 +576,7 @@ latent_upscale_modes = { "Latent (bicubic)": {"mode": "bicubic", "antialias": False}, "Latent (bicubic antialiased)": {"mode": "bicubic", "antialias": True}, "Latent (nearest)": {"mode": "nearest", "antialias": False}, + "Latent (nearest-exact)": {"mode": "nearest-exact", "antialias": False}, } sd_upscalers = [] -- cgit v1.2.3 From b85c2b5cf4a6809bc871718cf4680d49c3e95e94 Mon Sep 17 00:00:00 2001 From: timntorres Date: Thu, 5 Jan 2023 08:14:38 -0800 Subject: Clean up ti, add same behavior to hypernetwork. --- modules/hypernetworks/hypernetwork.py | 31 +++++++++++++++++++++++++- modules/shared.py | 2 +- modules/textual_inversion/textual_inversion.py | 14 +++++++----- 3 files changed, 40 insertions(+), 7 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 6a9b1398..d5985263 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -401,7 +401,33 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, hypernet.save(fn) shared.reload_hypernetworks() +# Note: textual_inversion.py has a nearly identical function of the same name. +def save_settings_to_file(initial_step, num_of_dataset_images, hypernetwork_name, layer_structure, activation_func, weight_init, add_layer_norm, use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): + checkpoint = sd_models.select_checkpoint() + model_name = checkpoint.model_name + model_hash = '[{}]'.format(checkpoint.hash) + # Starting index of preview-related arguments. + border_index = 19 + + # Get a list of the argument names, excluding default argument. + sig = inspect.signature(save_settings_to_file) + arg_names = [p.name for p in sig.parameters.values() if p.default == p.empty] + + # Create a list of the argument names to include in the settings string. + names = arg_names[:border_index] # Include all arguments up until the preview-related ones. + + # Include preview-related arguments if applicable. + if preview_from_txt2img: + names.extend(arg_names[border_index:]) + + # Build the settings string. + settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n" + for name in names: + value = locals()[name] + settings_str += f"{name}: {value}\n" + with open(os.path.join(log_directory, 'settings.txt'), "a+") as fout: + fout.write(settings_str + "\n\n") def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): # images allows training previews to have infotext. Importing it at the top causes a circular import problem. @@ -457,7 +483,10 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, pin_memory = shared.opts.pin_memory ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=hypernetwork_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, include_cond=True, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) - + + if shared.opts.save_training_settings_to_txt: + save_settings_to_file(initial_step, len(ds), hypernetwork_name, hypernetwork.layer_structure, hypernetwork.activation_func, hypernetwork.weight_init, hypernetwork.add_layer_norm, hypernetwork.use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) + latent_sampling_method = ds.latent_sampling_method dl = modules.textual_inversion.dataset.PersonalizedDataLoader(ds, latent_sampling_method=latent_sampling_method, batch_size=ds.batch_size, pin_memory=pin_memory) diff --git a/modules/shared.py b/modules/shared.py index 933cd738..10231a75 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -362,7 +362,7 @@ options_templates.update(options_section(('training', "Training"), { "unload_models_when_training": OptionInfo(False, "Move VAE and CLIP to RAM when training if possible. Saves VRAM."), "pin_memory": OptionInfo(False, "Turn on pin_memory for DataLoader. Makes training slightly faster but can increase memory usage."), "save_optimizer_state": OptionInfo(False, "Saves Optimizer state as separate *.optim file. Training of embedding or HN can be resumed with the matching optim file."), - "save_train_settings_to_txt": OptionInfo(False, "Save textual inversion and hypernet settings to a text file when training starts."), + "save_training_settings_to_txt": OptionInfo(False, "Save textual inversion and hypernet settings to a text file whenever training starts."), "dataset_filename_word_regex": OptionInfo("", "Filename word regex"), "dataset_filename_join_string": OptionInfo(" ", "Filename join string"), "training_image_repeats_per_epoch": OptionInfo(1, "Number of repeats for a single input image per epoch; used only for displaying epoch number", gr.Number, {"precision": 0}), diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 2bed2ecb..68648550 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -230,18 +230,20 @@ def write_loss(log_directory, filename, step, epoch_len, values): **values, }) +# Note: hypernetwork.py has a nearly identical function of the same name. def save_settings_to_file(initial_step, num_of_dataset_images, embedding_name, vectors_per_token, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): checkpoint = sd_models.select_checkpoint() model_name = checkpoint.model_name model_hash = '[{}]'.format(checkpoint.hash) - + # Starting index of preview-related arguments. + border_index = 16 # Get a list of the argument names. arg_names = inspect.getfullargspec(save_settings_to_file).args # Create a list of the argument names to include in the settings string. - names = arg_names[:16] # Include all arguments up until the preview-related ones. + names = arg_names[:border_index] # Include all arguments up until the preview-related ones. if preview_from_txt2img: - names.extend(arg_names[16:]) # Include all remaining arguments if `preview_from_txt2img` is True. + names.extend(arg_names[border_index:]) # Include all remaining arguments if `preview_from_txt2img` is True. # Build the settings string. settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n" @@ -329,8 +331,10 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ pin_memory = shared.opts.pin_memory ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=embedding_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) - if shared.opts.save_train_settings_to_txt: - save_settings_to_file(initial_step , len(ds) , embedding_name, len(embedding.vec) , learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) + + if shared.opts.save_training_settings_to_txt: + save_settings_to_file(initial_step, len(ds), embedding_name, len(embedding.vec), learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) + latent_sampling_method = ds.latent_sampling_method dl = modules.textual_inversion.dataset.PersonalizedDataLoader(ds, latent_sampling_method=latent_sampling_method, batch_size=ds.batch_size, pin_memory=pin_memory) -- cgit v1.2.3 From b6bab2f052b32c0ffebe6aecc1819ccf20cf8c5d Mon Sep 17 00:00:00 2001 From: timntorres Date: Thu, 5 Jan 2023 09:14:56 -0800 Subject: Include model in log file. Exclude directory. --- modules/hypernetworks/hypernetwork.py | 28 +++++++++----------------- modules/textual_inversion/textual_inversion.py | 22 +++++++++----------- 2 files changed, 19 insertions(+), 31 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index d5985263..3237c37a 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -402,30 +402,22 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, shared.reload_hypernetworks() # Note: textual_inversion.py has a nearly identical function of the same name. -def save_settings_to_file(initial_step, num_of_dataset_images, hypernetwork_name, layer_structure, activation_func, weight_init, add_layer_norm, use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): - checkpoint = sd_models.select_checkpoint() - model_name = checkpoint.model_name - model_hash = '[{}]'.format(checkpoint.hash) +def save_settings_to_file(model_name, model_hash, initial_step, num_of_dataset_images, hypernetwork_name, layer_structure, activation_func, weight_init, add_layer_norm, use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): # Starting index of preview-related arguments. - border_index = 19 - - # Get a list of the argument names, excluding default argument. - sig = inspect.signature(save_settings_to_file) - arg_names = [p.name for p in sig.parameters.values() if p.default == p.empty] - + border_index = 21 + # Get a list of the argument names. + arg_names = inspect.getfullargspec(save_settings_to_file).args # Create a list of the argument names to include in the settings string. names = arg_names[:border_index] # Include all arguments up until the preview-related ones. - - # Include preview-related arguments if applicable. if preview_from_txt2img: - names.extend(arg_names[border_index:]) - + names.extend(arg_names[border_index:]) # Include preview-related arguments if applicable. # Build the settings string. settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n" for name in names: - value = locals()[name] - settings_str += f"{name}: {value}\n" - + if name != 'log_directory': # It's useless and redundant to save log_directory. + value = locals()[name] + settings_str += f"{name}: {value}\n" + # Create or append to the file. with open(os.path.join(log_directory, 'settings.txt'), "a+") as fout: fout.write(settings_str + "\n\n") @@ -485,7 +477,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=hypernetwork_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, include_cond=True, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) if shared.opts.save_training_settings_to_txt: - save_settings_to_file(initial_step, len(ds), hypernetwork_name, hypernetwork.layer_structure, hypernetwork.activation_func, hypernetwork.weight_init, hypernetwork.add_layer_norm, hypernetwork.use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) + save_settings_to_file(checkpoint.model_name, '[{}]'.format(checkpoint.hash), initial_step, len(ds), hypernetwork_name, hypernetwork.layer_structure, hypernetwork.activation_func, hypernetwork.weight_init, hypernetwork.add_layer_norm, hypernetwork.use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) latent_sampling_method = ds.latent_sampling_method diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 68648550..ce7e4f5d 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -231,26 +231,22 @@ def write_loss(log_directory, filename, step, epoch_len, values): }) # Note: hypernetwork.py has a nearly identical function of the same name. -def save_settings_to_file(initial_step, num_of_dataset_images, embedding_name, vectors_per_token, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): - checkpoint = sd_models.select_checkpoint() - model_name = checkpoint.model_name - model_hash = '[{}]'.format(checkpoint.hash) +def save_settings_to_file(model_name, model_hash, initial_step, num_of_dataset_images, embedding_name, vectors_per_token, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): # Starting index of preview-related arguments. - border_index = 16 + border_index = 18 # Get a list of the argument names. - arg_names = inspect.getfullargspec(save_settings_to_file).args - + arg_names = inspect.getfullargspec(save_settings_to_file).args # Create a list of the argument names to include in the settings string. names = arg_names[:border_index] # Include all arguments up until the preview-related ones. if preview_from_txt2img: - names.extend(arg_names[border_index:]) # Include all remaining arguments if `preview_from_txt2img` is True. - + names.extend(arg_names[border_index:]) # Include preview-related arguments if applicable. # Build the settings string. settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n" for name in names: - value = locals()[name] - settings_str += f"{name}: {value}\n" - + if name != 'log_directory': # It's useless and redundant to save log_directory. + value = locals()[name] + settings_str += f"{name}: {value}\n" + # Create or append to the file. with open(os.path.join(log_directory, 'settings.txt'), "a+") as fout: fout.write(settings_str + "\n\n") @@ -333,7 +329,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=embedding_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) if shared.opts.save_training_settings_to_txt: - save_settings_to_file(initial_step, len(ds), embedding_name, len(embedding.vec), learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) + save_settings_to_file(checkpoint.model_name, '[{}]'.format(checkpoint.hash), initial_step, len(ds), embedding_name, len(embedding.vec), learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) latent_sampling_method = ds.latent_sampling_method -- cgit v1.2.3 From fda04e620d529031e2134520e74756d0efa30464 Mon Sep 17 00:00:00 2001 From: Kuma <36082288+KumiIT@users.noreply.github.com> Date: Thu, 5 Jan 2023 18:44:19 +0100 Subject: typo in TI --- modules/textual_inversion/textual_inversion.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 71e07bcc..24b43045 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -298,7 +298,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ torch.nn.utils.clip_grad_norm_ if clip_grad_mode == "norm" else \ None if clip_grad: - clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, ititial_step, verbose=False) + clip_grad_sched = LearnRateScheduler(clip_grad_value, steps, initial_step, verbose=False) # dataset loading may take a while, so input validations and early returns should be done before this shared.state.textinfo = f"Preparing dataset from {html.escape(data_root)}..." old_parallel_processing_allowed = shared.parallel_processing_allowed -- cgit v1.2.3 From 847f869c67c7108e3e792fc193331d0e6acca29c Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 21:00:52 +0300 Subject: experimental optimization --- modules/processing.py | 28 +++++++++++++++++++++++++--- 1 file changed, 25 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 61e97077..a408d622 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -544,6 +544,29 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: infotexts = [] output_images = [] + cached_uc = [None, None] + cached_c = [None, None] + + def get_conds_with_caching(function, required_prompts, steps, cache): + """ + Returns the result of calling function(shared.sd_model, required_prompts, steps) + using a cache to store the result if the same arguments have been used before. + + cache is an array containing two elements. The first element is a tuple + representing the previously used arguments, or None if no arguments + have been used before. The second element is where the previously + computed result is stored. + """ + + if cache[0] is not None and (required_prompts, steps) == cache[0]: + return cache[1] + + with devices.autocast(): + cache[1] = function(shared.sd_model, required_prompts, steps) + + cache[0] = (required_prompts, steps) + return cache[1] + with torch.no_grad(), p.sd_model.ema_scope(): with devices.autocast(): p.init(p.all_prompts, p.all_seeds, p.all_subseeds) @@ -571,9 +594,8 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if p.scripts is not None: p.scripts.process_batch(p, batch_number=n, prompts=prompts, seeds=seeds, subseeds=subseeds) - with devices.autocast(): - uc = prompt_parser.get_learned_conditioning(shared.sd_model, negative_prompts, p.steps) - c = prompt_parser.get_multicond_learned_conditioning(shared.sd_model, prompts, p.steps) + uc = get_conds_with_caching(prompt_parser.get_learned_conditioning, negative_prompts, p.steps, cached_uc) + c = get_conds_with_caching(prompt_parser.get_multicond_learned_conditioning, prompts, p.steps, cached_c) if len(model_hijack.comments) > 0: for comment in model_hijack.comments: -- cgit v1.2.3 From 81133d4168ae0bae9bf8bf1a1d4983319a589112 Mon Sep 17 00:00:00 2001 From: Faber Date: Fri, 6 Jan 2023 03:38:37 +0700 Subject: allow loading embeddings from subdirectories --- modules/textual_inversion/textual_inversion.py | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 24b43045..0a059044 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -149,19 +149,20 @@ class EmbeddingDatabase: else: self.skipped_embeddings[name] = embedding - for fn in os.listdir(self.embeddings_dir): - try: - fullfn = os.path.join(self.embeddings_dir, fn) - - if os.stat(fullfn).st_size == 0: + for root, dirs, fns in os.walk(self.embeddings_dir): + for fn in fns: + try: + fullfn = os.path.join(root, fn) + + if os.stat(fullfn).st_size == 0: + continue + + process_file(fullfn, fn) + except Exception: + print(f"Error loading embedding {fn}:", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) continue - process_file(fullfn, fn) - except Exception: - print(f"Error loading embedding {fn}:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - continue - print(f"Textual inversion embeddings loaded({len(self.word_embeddings)}): {', '.join(self.word_embeddings.keys())}") if len(self.skipped_embeddings) > 0: print(f"Textual inversion embeddings skipped({len(self.skipped_embeddings)}): {', '.join(self.skipped_embeddings.keys())}") -- cgit v1.2.3 From b5253f0dab529707f1fe2e11211a10ce2f264617 Mon Sep 17 00:00:00 2001 From: noodleanon <122053346+noodleanon@users.noreply.github.com> Date: Thu, 5 Jan 2023 21:21:48 +0000 Subject: allow img2img api to run scripts --- modules/api/api.py | 27 ++++++++++++++++++++++++--- modules/api/models.py | 2 +- modules/processing.py | 4 ++-- 3 files changed, 27 insertions(+), 6 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 2103709b..aa62a42e 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -11,7 +11,7 @@ from fastapi.security import HTTPBasic, HTTPBasicCredentials from secrets import compare_digest import modules.shared as shared -from modules import sd_samplers, deepbooru, sd_hijack, images +from modules import sd_samplers, deepbooru, sd_hijack, images, scripts, ui from modules.api.models import * from modules.processing import StableDiffusionProcessingTxt2Img, StableDiffusionProcessingImg2Img, process_images from modules.extras import run_extras @@ -28,8 +28,13 @@ def upscaler_to_index(name: str): try: return [x.name.lower() for x in shared.sd_upscalers].index(name.lower()) except: - raise HTTPException(status_code=400, detail=f"Invalid upscaler, needs to be on of these: {' , '.join([x.name for x in sd_upscalers])}") + raise HTTPException(status_code=400, detail=f"Invalid upscaler, needs to be one of these: {' , '.join([x.name for x in sd_upscalers])}") +def script_name_to_index(name, scripts): + try: + return [script.title().lower() for script in scripts].index(name.lower()) + except: + raise HTTPException(status_code=422, detail=f"Script '{name}' not found") def validate_sampler_name(name): config = sd_samplers.all_samplers_map.get(name, None) @@ -170,6 +175,14 @@ class Api: if init_images is None: raise HTTPException(status_code=404, detail="Init image not found") + if img2imgreq.script_name is not None: + if scripts.scripts_img2img.scripts == []: + scripts.scripts_img2img.initialize_scripts(True) + ui.create_ui() + + script_idx = script_name_to_index(img2imgreq.script_name, scripts.scripts_img2img.selectable_scripts) + script = scripts.scripts_img2img.selectable_scripts[script_idx] + mask = img2imgreq.mask if mask: mask = decode_base64_to_image(mask) @@ -186,13 +199,21 @@ class Api: args = vars(populate) args.pop('include_init_images', None) # this is meant to be done by "exclude": True in model, but it's for a reason that I cannot determine. + args.pop('script_name', None) with self.queue_lock: p = StableDiffusionProcessingImg2Img(sd_model=shared.sd_model, **args) p.init_images = [decode_base64_to_image(x) for x in init_images] shared.state.begin() - processed = process_images(p) + if 'script' in locals(): + p.outpath_grids = opts.outdir_img2img_grids + p.outpath_samples = opts.outdir_img2img_samples + p.script_args = [script_idx + 1] + [None] * (script.args_from - 1) + p.script_args + processed = scripts.scripts_img2img.run(p, *p.script_args) + else: + processed = process_images(p) + shared.state.end() b64images = list(map(encode_pil_to_base64, processed.images)) diff --git a/modules/api/models.py b/modules/api/models.py index d8198a27..862477e7 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -106,7 +106,7 @@ StableDiffusionTxt2ImgProcessingAPI = PydanticModelGenerator( StableDiffusionImg2ImgProcessingAPI = PydanticModelGenerator( "StableDiffusionProcessingImg2Img", StableDiffusionProcessingImg2Img, - [{"key": "sampler_index", "type": str, "default": "Euler"}, {"key": "init_images", "type": list, "default": None}, {"key": "denoising_strength", "type": float, "default": 0.75}, {"key": "mask", "type": str, "default": None}, {"key": "include_init_images", "type": bool, "default": False, "exclude" : True}] + [{"key": "sampler_index", "type": str, "default": "Euler"}, {"key": "init_images", "type": list, "default": None}, {"key": "denoising_strength", "type": float, "default": 0.75}, {"key": "mask", "type": str, "default": None}, {"key": "include_init_images", "type": bool, "default": False, "exclude" : True}, {"key": "script_name", "type": str, "default": None}, {"key": "script_args", "type": list, "default": []}] ).generate_model() class TextToImageResponse(BaseModel): diff --git a/modules/processing.py b/modules/processing.py index a408d622..d5ac7eb1 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -98,7 +98,7 @@ class StableDiffusionProcessing(): """ The first set of paramaters: sd_models -> do_not_reload_embeddings represent the minimum required to create a StableDiffusionProcessing """ - def __init__(self, sd_model=None, outpath_samples=None, outpath_grids=None, prompt: str = "", styles: List[str] = None, seed: int = -1, subseed: int = -1, subseed_strength: float = 0, seed_resize_from_h: int = -1, seed_resize_from_w: int = -1, seed_enable_extras: bool = True, sampler_name: str = None, batch_size: int = 1, n_iter: int = 1, steps: int = 50, cfg_scale: float = 7.0, width: int = 512, height: int = 512, restore_faces: bool = False, tiling: bool = False, do_not_save_samples: bool = False, do_not_save_grid: bool = False, extra_generation_params: Dict[Any, Any] = None, overlay_images: Any = None, negative_prompt: str = None, eta: float = None, do_not_reload_embeddings: bool = False, denoising_strength: float = 0, ddim_discretize: str = None, s_churn: float = 0.0, s_tmax: float = None, s_tmin: float = 0.0, s_noise: float = 1.0, override_settings: Dict[str, Any] = None, override_settings_restore_afterwards: bool = True, sampler_index: int = None): + def __init__(self, sd_model=None, outpath_samples=None, outpath_grids=None, prompt: str = "", styles: List[str] = None, seed: int = -1, subseed: int = -1, subseed_strength: float = 0, seed_resize_from_h: int = -1, seed_resize_from_w: int = -1, seed_enable_extras: bool = True, sampler_name: str = None, batch_size: int = 1, n_iter: int = 1, steps: int = 50, cfg_scale: float = 7.0, width: int = 512, height: int = 512, restore_faces: bool = False, tiling: bool = False, do_not_save_samples: bool = False, do_not_save_grid: bool = False, extra_generation_params: Dict[Any, Any] = None, overlay_images: Any = None, negative_prompt: str = None, eta: float = None, do_not_reload_embeddings: bool = False, denoising_strength: float = 0, ddim_discretize: str = None, s_churn: float = 0.0, s_tmax: float = None, s_tmin: float = 0.0, s_noise: float = 1.0, override_settings: Dict[str, Any] = None, override_settings_restore_afterwards: bool = True, sampler_index: int = None, script_args: list = None): if sampler_index is not None: print("sampler_index argument for StableDiffusionProcessing does not do anything; use sampler_name", file=sys.stderr) @@ -149,7 +149,7 @@ class StableDiffusionProcessing(): self.seed_resize_from_w = 0 self.scripts = None - self.script_args = None + self.script_args = script_args self.all_prompts = None self.all_negative_prompts = None self.all_seeds = None -- cgit v1.2.3 From 8111b5569d07c7ac3b695e28171aede728b4ae56 Mon Sep 17 00:00:00 2001 From: brkirch Date: Tue, 3 Jan 2023 20:43:05 -0500 Subject: Add support for PyTorch nightly and local builds --- modules/devices.py | 28 +++++++++++++++++++++++----- 1 file changed, 23 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/devices.py b/modules/devices.py index 800510b7..caeb0276 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -133,8 +133,26 @@ def numpy_fix(self, *args, **kwargs): return orig_tensor_numpy(self, *args, **kwargs) -# PyTorch 1.13 doesn't need these fixes but unfortunately is slower and has regressions that prevent training from working -if has_mps() and version.parse(torch.__version__) < version.parse("1.13"): - torch.Tensor.to = tensor_to_fix - torch.nn.functional.layer_norm = layer_norm_fix - torch.Tensor.numpy = numpy_fix +# MPS workaround for https://github.com/pytorch/pytorch/issues/89784 +orig_cumsum = torch.cumsum +orig_Tensor_cumsum = torch.Tensor.cumsum +def cumsum_fix(input, cumsum_func, *args, **kwargs): + if input.device.type == 'mps': + output_dtype = kwargs.get('dtype', input.dtype) + if any(output_dtype == broken_dtype for broken_dtype in [torch.bool, torch.int8, torch.int16, torch.int64]): + return cumsum_func(input.cpu(), *args, **kwargs).to(input.device) + return cumsum_func(input, *args, **kwargs) + + +if has_mps(): + if version.parse(torch.__version__) < version.parse("1.13"): + # PyTorch 1.13 doesn't need these fixes but unfortunately is slower and has regressions that prevent training from working + torch.Tensor.to = tensor_to_fix + torch.nn.functional.layer_norm = layer_norm_fix + torch.Tensor.numpy = numpy_fix + elif version.parse(torch.__version__) > version.parse("1.13.1"): + if not torch.Tensor([1,2]).to(torch.device("mps")).equal(torch.Tensor([1,1]).to(torch.device("mps")).cumsum(0, dtype=torch.int16)): + torch.cumsum = lambda input, *args, **kwargs: ( cumsum_fix(input, orig_cumsum, *args, **kwargs) ) + torch.Tensor.cumsum = lambda self, *args, **kwargs: ( cumsum_fix(self, orig_Tensor_cumsum, *args, **kwargs) ) + orig_narrow = torch.narrow + torch.narrow = lambda *args, **kwargs: ( orig_narrow(*args, **kwargs).clone() ) -- cgit v1.2.3 From d61a5aa4f623f6630670241aca8fc5c2a6381769 Mon Sep 17 00:00:00 2001 From: acncagua Date: Fri, 6 Jan 2023 10:58:22 +0900 Subject: Add files via upload --- modules/ui.py | 2 ++ 1 file changed, 2 insertions(+) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 81d96c5b..030f0685 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -550,6 +550,8 @@ Requested path was: {f} os.startfile(path) elif platform.system() == "Darwin": sp.Popen(["open", path]) + elif "microsoft-standard-WSL2" in platform.uname().release: + sp.Popen(["wsl-open", path]) else: sp.Popen(["xdg-open", path]) -- cgit v1.2.3 From d782a95967c9eea753df3333cd1954b6ec73eba0 Mon Sep 17 00:00:00 2001 From: brkirch Date: Tue, 27 Dec 2022 08:50:55 -0500 Subject: Add Birch-san's sub-quadratic attention implementation --- modules/sd_hijack.py | 15 ++- modules/sd_hijack_optimizations.py | 124 ++++++++++++++++++----- modules/shared.py | 4 + modules/sub_quadratic_attention.py | 201 +++++++++++++++++++++++++++++++++++++ 4 files changed, 310 insertions(+), 34 deletions(-) create mode 100644 modules/sub_quadratic_attention.py (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index 690a9ec2..019a6f3f 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -7,8 +7,6 @@ from modules.hypernetworks import hypernetwork from modules.shared import cmd_opts from modules import sd_hijack_clip, sd_hijack_open_clip, sd_hijack_unet -from modules.sd_hijack_optimizations import invokeAI_mps_available - import ldm.modules.attention import ldm.modules.diffusionmodules.model import ldm.modules.diffusionmodules.openaimodel @@ -40,17 +38,16 @@ def apply_optimizations(): print("Applying xformers cross attention optimization.") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.xformers_attention_forward ldm.modules.diffusionmodules.model.AttnBlock.forward = sd_hijack_optimizations.xformers_attnblock_forward + elif cmd_opts.opt_sub_quad_attention: + print("Applying sub-quadratic cross attention optimization.") + ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.sub_quad_attention_forward + ldm.modules.diffusionmodules.model.AttnBlock.forward = sd_hijack_optimizations.sub_quad_attnblock_forward elif cmd_opts.opt_split_attention_v1: print("Applying v1 cross attention optimization.") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_v1 elif not cmd_opts.disable_opt_split_attention and (cmd_opts.opt_split_attention_invokeai or not torch.cuda.is_available()): - if not invokeAI_mps_available and shared.device.type == 'mps': - print("The InvokeAI cross attention optimization for MPS requires the psutil package which is not installed.") - print("Applying v1 cross attention optimization.") - ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_v1 - else: - print("Applying cross attention optimization (InvokeAI).") - ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_invokeAI + print("Applying cross attention optimization (InvokeAI).") + ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_invokeAI elif not cmd_opts.disable_opt_split_attention and (cmd_opts.opt_split_attention or torch.cuda.is_available()): print("Applying cross attention optimization (Doggettx).") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward diff --git a/modules/sd_hijack_optimizations.py b/modules/sd_hijack_optimizations.py index 02c87f40..f5c153e8 100644 --- a/modules/sd_hijack_optimizations.py +++ b/modules/sd_hijack_optimizations.py @@ -1,7 +1,7 @@ import math import sys import traceback -import importlib +import psutil import torch from torch import einsum @@ -12,6 +12,8 @@ from einops import rearrange from modules import shared from modules.hypernetworks import hypernetwork +from .sub_quadratic_attention import efficient_dot_product_attention + if shared.cmd_opts.xformers or shared.cmd_opts.force_enable_xformers: try: @@ -22,6 +24,19 @@ if shared.cmd_opts.xformers or shared.cmd_opts.force_enable_xformers: print(traceback.format_exc(), file=sys.stderr) +def get_available_vram(): + if shared.device.type == 'cuda': + stats = torch.cuda.memory_stats(shared.device) + mem_active = stats['active_bytes.all.current'] + mem_reserved = stats['reserved_bytes.all.current'] + mem_free_cuda, _ = torch.cuda.mem_get_info(torch.cuda.current_device()) + mem_free_torch = mem_reserved - mem_active + mem_free_total = mem_free_cuda + mem_free_torch + return mem_free_total + else: + return psutil.virtual_memory().available + + # see https://github.com/basujindal/stable-diffusion/pull/117 for discussion def split_cross_attention_forward_v1(self, x, context=None, mask=None): h = self.heads @@ -76,12 +91,7 @@ def split_cross_attention_forward(self, x, context=None, mask=None): r1 = torch.zeros(q.shape[0], q.shape[1], v.shape[2], device=q.device, dtype=q.dtype) - stats = torch.cuda.memory_stats(q.device) - mem_active = stats['active_bytes.all.current'] - mem_reserved = stats['reserved_bytes.all.current'] - mem_free_cuda, _ = torch.cuda.mem_get_info(torch.cuda.current_device()) - mem_free_torch = mem_reserved - mem_active - mem_free_total = mem_free_cuda + mem_free_torch + mem_free_total = get_available_vram() gb = 1024 ** 3 tensor_size = q.shape[0] * q.shape[1] * k.shape[1] * q.element_size() @@ -118,19 +128,8 @@ def split_cross_attention_forward(self, x, context=None, mask=None): return self.to_out(r2) -def check_for_psutil(): - try: - spec = importlib.util.find_spec('psutil') - return spec is not None - except ModuleNotFoundError: - return False - -invokeAI_mps_available = check_for_psutil() - # -- Taken from https://github.com/invoke-ai/InvokeAI and modified -- -if invokeAI_mps_available: - import psutil - mem_total_gb = psutil.virtual_memory().total // (1 << 30) +mem_total_gb = psutil.virtual_memory().total // (1 << 30) def einsum_op_compvis(q, k, v): s = einsum('b i d, b j d -> b i j', q, k) @@ -215,6 +214,70 @@ def split_cross_attention_forward_invokeAI(self, x, context=None, mask=None): # -- End of code from https://github.com/invoke-ai/InvokeAI -- + +# Based on Birch-san's modified implementation of sub-quadratic attention from https://github.com/Birch-san/diffusers/pull/1 +def sub_quad_attention_forward(self, x, context=None, mask=None): + assert mask is None, "attention-mask not currently implemented for SubQuadraticCrossAttnProcessor." + + h = self.heads + + q = self.to_q(x) + context = default(context, x) + + context_k, context_v = hypernetwork.apply_hypernetwork(shared.loaded_hypernetwork, context) + k = self.to_k(context_k) + v = self.to_v(context_v) + del context, context_k, context_v, x + + q = q.unflatten(-1, (h, -1)).transpose(1,2).flatten(end_dim=1) + k = k.unflatten(-1, (h, -1)).transpose(1,2).flatten(end_dim=1) + v = v.unflatten(-1, (h, -1)).transpose(1,2).flatten(end_dim=1) + + x = sub_quad_attention(q, k, v, q_chunk_size=shared.cmd_opts.sub_quad_q_chunk_size, kv_chunk_size=shared.cmd_opts.sub_quad_kv_chunk_size, chunk_threshold_bytes=shared.cmd_opts.sub_quad_chunk_threshold, use_checkpoint=self.training) + + x = x.unflatten(0, (-1, h)).transpose(1,2).flatten(start_dim=2) + + out_proj, dropout = self.to_out + x = out_proj(x) + x = dropout(x) + + return x + +def sub_quad_attention(q, k, v, q_chunk_size=1024, kv_chunk_size=None, kv_chunk_size_min=None, chunk_threshold_bytes=None, use_checkpoint=True): + bytes_per_token = torch.finfo(q.dtype).bits//8 + batch_x_heads, q_tokens, _ = q.shape + _, k_tokens, _ = k.shape + qk_matmul_size_bytes = batch_x_heads * bytes_per_token * q_tokens * k_tokens + + available_vram = int(get_available_vram() * 0.9) if q.device.type == 'mps' else int(get_available_vram() * 0.7) + + if chunk_threshold_bytes is None: + chunk_threshold_bytes = available_vram + elif chunk_threshold_bytes == 0: + chunk_threshold_bytes = None + + if kv_chunk_size_min is None: + kv_chunk_size_min = chunk_threshold_bytes // (batch_x_heads * bytes_per_token * (k.shape[2] + v.shape[2])) + elif kv_chunk_size_min == 0: + kv_chunk_size_min = None + + if chunk_threshold_bytes is not None and qk_matmul_size_bytes <= chunk_threshold_bytes: + # the big matmul fits into our memory limit; do everything in 1 chunk, + # i.e. send it down the unchunked fast-path + query_chunk_size = q_tokens + kv_chunk_size = k_tokens + + return efficient_dot_product_attention( + q, + k, + v, + query_chunk_size=q_chunk_size, + kv_chunk_size=kv_chunk_size, + kv_chunk_size_min = kv_chunk_size_min, + use_checkpoint=use_checkpoint, + ) + + def xformers_attention_forward(self, x, context=None, mask=None): h = self.heads q_in = self.to_q(x) @@ -252,12 +315,7 @@ def cross_attention_attnblock_forward(self, x): h_ = torch.zeros_like(k, device=q.device) - stats = torch.cuda.memory_stats(q.device) - mem_active = stats['active_bytes.all.current'] - mem_reserved = stats['reserved_bytes.all.current'] - mem_free_cuda, _ = torch.cuda.mem_get_info(torch.cuda.current_device()) - mem_free_torch = mem_reserved - mem_active - mem_free_total = mem_free_cuda + mem_free_torch + mem_free_total = get_available_vram() tensor_size = q.shape[0] * q.shape[1] * k.shape[2] * q.element_size() mem_required = tensor_size * 2.5 @@ -312,3 +370,19 @@ def xformers_attnblock_forward(self, x): return x + out except NotImplementedError: return cross_attention_attnblock_forward(self, x) + +def sub_quad_attnblock_forward(self, x): + h_ = x + h_ = self.norm(h_) + q = self.q(h_) + k = self.k(h_) + v = self.v(h_) + b, c, h, w = q.shape + q, k, v = map(lambda t: rearrange(t, 'b c h w -> b (h w) c'), (q, k, v)) + q = q.contiguous() + k = k.contiguous() + v = v.contiguous() + out = sub_quad_attention(q, k, v, q_chunk_size=shared.cmd_opts.sub_quad_q_chunk_size, kv_chunk_size=shared.cmd_opts.sub_quad_kv_chunk_size, chunk_threshold_bytes=shared.cmd_opts.sub_quad_chunk_threshold, use_checkpoint=self.training) + out = rearrange(out, 'b (h w) c -> b c h w', h=h) + out = self.proj_out(out) + return x + out diff --git a/modules/shared.py b/modules/shared.py index d4ddeea0..487a7792 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -56,6 +56,10 @@ parser.add_argument("--xformers", action='store_true', help="enable xformers for parser.add_argument("--force-enable-xformers", action='store_true', help="enable xformers for cross attention layers regardless of whether the checking code thinks you can run it; do not make bug reports if this fails to work") parser.add_argument("--deepdanbooru", action='store_true', help="does not do anything") parser.add_argument("--opt-split-attention", action='store_true', help="force-enables Doggettx's cross-attention layer optimization. By default, it's on for torch cuda.") +parser.add_argument("--opt-sub-quad-attention", action='store_true', help="enable memory efficient sub-quadratic cross-attention layer optimization") +parser.add_argument("--sub-quad-q-chunk-size", type=int, help="query chunk size for the sub-quadratic cross-attention layer optimization to use", default=1024) +parser.add_argument("--sub-quad-kv-chunk-size", type=int, help="kv chunk size for the sub-quadratic cross-attention layer optimization to use", default=None) +parser.add_argument("--sub-quad-chunk-threshold", type=int, help="the size threshold in bytes for the sub-quadratic cross-attention layer optimization to use chunking", default=None) parser.add_argument("--opt-split-attention-invokeai", action='store_true', help="force-enables InvokeAI's cross-attention layer optimization. By default, it's on when cuda is unavailable.") parser.add_argument("--opt-split-attention-v1", action='store_true', help="enable older version of split attention optimization that does not consume all the VRAM it can find") parser.add_argument("--disable-opt-split-attention", action='store_true', help="force-disables cross-attention layer optimization") diff --git a/modules/sub_quadratic_attention.py b/modules/sub_quadratic_attention.py new file mode 100644 index 00000000..b11dc1c7 --- /dev/null +++ b/modules/sub_quadratic_attention.py @@ -0,0 +1,201 @@ +# original source: +# https://github.com/AminRezaei0x443/memory-efficient-attention/blob/1bc0d9e6ac5f82ea43a375135c4e1d3896ee1694/memory_efficient_attention/attention_torch.py +# license: +# unspecified +# credit: +# Amin Rezaei (original author) +# Alex Birch (optimized algorithm for 3D tensors, at the expense of removing bias, masking and callbacks) +# implementation of: +# Self-attention Does Not Need O(n2) Memory": +# https://arxiv.org/abs/2112.05682v2 + +from functools import partial +import torch +from torch import Tensor +from torch.utils.checkpoint import checkpoint +import math +from typing import Optional, NamedTuple, Protocol, List + +def dynamic_slice( + x: Tensor, + starts: List[int], + sizes: List[int], +) -> Tensor: + slicing = [slice(start, start + size) for start, size in zip(starts, sizes)] + return x[slicing] + +class AttnChunk(NamedTuple): + exp_values: Tensor + exp_weights_sum: Tensor + max_score: Tensor + +class SummarizeChunk(Protocol): + @staticmethod + def __call__( + query: Tensor, + key: Tensor, + value: Tensor, + ) -> AttnChunk: ... + +class ComputeQueryChunkAttn(Protocol): + @staticmethod + def __call__( + query: Tensor, + key: Tensor, + value: Tensor, + ) -> Tensor: ... + +def _summarize_chunk( + query: Tensor, + key: Tensor, + value: Tensor, + scale: float, +) -> AttnChunk: + attn_weights = torch.baddbmm( + torch.empty(1, 1, 1, device=query.device, dtype=query.dtype), + query, + key.transpose(1,2), + alpha=scale, + beta=0, + ) + max_score, _ = torch.max(attn_weights, -1, keepdim=True) + max_score = max_score.detach() + exp_weights = torch.exp(attn_weights - max_score) + exp_values = torch.bmm(exp_weights, value) + max_score = max_score.squeeze(-1) + return AttnChunk(exp_values, exp_weights.sum(dim=-1), max_score) + +def _query_chunk_attention( + query: Tensor, + key: Tensor, + value: Tensor, + summarize_chunk: SummarizeChunk, + kv_chunk_size: int, +) -> Tensor: + batch_x_heads, k_tokens, k_channels_per_head = key.shape + _, _, v_channels_per_head = value.shape + + def chunk_scanner(chunk_idx: int) -> AttnChunk: + key_chunk = dynamic_slice( + key, + (0, chunk_idx, 0), + (batch_x_heads, kv_chunk_size, k_channels_per_head) + ) + value_chunk = dynamic_slice( + value, + (0, chunk_idx, 0), + (batch_x_heads, kv_chunk_size, v_channels_per_head) + ) + return summarize_chunk(query, key_chunk, value_chunk) + + chunks: List[AttnChunk] = [ + chunk_scanner(chunk) for chunk in torch.arange(0, k_tokens, kv_chunk_size) + ] + acc_chunk = AttnChunk(*map(torch.stack, zip(*chunks))) + chunk_values, chunk_weights, chunk_max = acc_chunk + + global_max, _ = torch.max(chunk_max, 0, keepdim=True) + max_diffs = torch.exp(chunk_max - global_max) + chunk_values *= torch.unsqueeze(max_diffs, -1) + chunk_weights *= max_diffs + + all_values = chunk_values.sum(dim=0) + all_weights = torch.unsqueeze(chunk_weights, -1).sum(dim=0) + return all_values / all_weights + +# TODO: refactor CrossAttention#get_attention_scores to share code with this +def _get_attention_scores_no_kv_chunking( + query: Tensor, + key: Tensor, + value: Tensor, + scale: float, +) -> Tensor: + attn_scores = torch.baddbmm( + torch.empty(1, 1, 1, device=query.device, dtype=query.dtype), + query, + key.transpose(1,2), + alpha=scale, + beta=0, + ) + attn_probs = attn_scores.softmax(dim=-1) + del attn_scores + hidden_states_slice = torch.bmm(attn_probs, value) + return hidden_states_slice + +class ScannedChunk(NamedTuple): + chunk_idx: int + attn_chunk: AttnChunk + +def efficient_dot_product_attention( + query: Tensor, + key: Tensor, + value: Tensor, + query_chunk_size=1024, + kv_chunk_size: Optional[int] = None, + kv_chunk_size_min: Optional[int] = None, + use_checkpoint=True, +): + """Computes efficient dot-product attention given query, key, and value. + This is efficient version of attention presented in + https://arxiv.org/abs/2112.05682v2 which comes with O(sqrt(n)) memory requirements. + Args: + query: queries for calculating attention with shape of + `[batch * num_heads, tokens, channels_per_head]`. + key: keys for calculating attention with shape of + `[batch * num_heads, tokens, channels_per_head]`. + value: values to be used in attention with shape of + `[batch * num_heads, tokens, channels_per_head]`. + query_chunk_size: int: query chunks size + kv_chunk_size: Optional[int]: key/value chunks size. if None: defaults to sqrt(key_tokens) + kv_chunk_size_min: Optional[int]: key/value minimum chunk size. only considered when kv_chunk_size is None. changes `sqrt(key_tokens)` into `max(sqrt(key_tokens), kv_chunk_size_min)`, to ensure our chunk sizes don't get too small (smaller chunks = more chunks = less concurrent work done). + use_checkpoint: bool: whether to use checkpointing (recommended True for training, False for inference) + Returns: + Output of shape `[batch * num_heads, query_tokens, channels_per_head]`. + """ + batch_x_heads, q_tokens, q_channels_per_head = query.shape + _, k_tokens, _ = key.shape + scale = q_channels_per_head ** -0.5 + + kv_chunk_size = min(kv_chunk_size or int(math.sqrt(k_tokens)), k_tokens) + if kv_chunk_size_min is not None: + kv_chunk_size = max(kv_chunk_size, kv_chunk_size_min) + + def get_query_chunk(chunk_idx: int) -> Tensor: + return dynamic_slice( + query, + (0, chunk_idx, 0), + (batch_x_heads, min(query_chunk_size, q_tokens), q_channels_per_head) + ) + + summarize_chunk: SummarizeChunk = partial(_summarize_chunk, scale=scale) + summarize_chunk: SummarizeChunk = partial(checkpoint, summarize_chunk) if use_checkpoint else summarize_chunk + compute_query_chunk_attn: ComputeQueryChunkAttn = partial( + _get_attention_scores_no_kv_chunking, + scale=scale + ) if k_tokens <= kv_chunk_size else ( + # fast-path for when there's just 1 key-value chunk per query chunk (this is just sliced attention btw) + partial( + _query_chunk_attention, + kv_chunk_size=kv_chunk_size, + summarize_chunk=summarize_chunk, + ) + ) + + if q_tokens <= query_chunk_size: + # fast-path for when there's just 1 query chunk + return compute_query_chunk_attn( + query=query, + key=key, + value=value, + ) + + # TODO: maybe we should use torch.empty_like(query) to allocate storage in-advance, + # and pass slices to be mutated, instead of torch.cat()ing the returned slices + res = torch.cat([ + compute_query_chunk_attn( + query=get_query_chunk(i * query_chunk_size), + key=key, + value=value, + ) for i in range(math.ceil(q_tokens / query_chunk_size)) + ], dim=1) + return res -- cgit v1.2.3 From b119815333026164f2bd7d1ca71f3e4f7a9afd0d Mon Sep 17 00:00:00 2001 From: brkirch Date: Thu, 5 Jan 2023 04:37:17 -0500 Subject: Use narrow instead of dynamic_slice --- modules/sub_quadratic_attention.py | 34 +++++++++++++++++++--------------- 1 file changed, 19 insertions(+), 15 deletions(-) (limited to 'modules') diff --git a/modules/sub_quadratic_attention.py b/modules/sub_quadratic_attention.py index b11dc1c7..95924d24 100644 --- a/modules/sub_quadratic_attention.py +++ b/modules/sub_quadratic_attention.py @@ -5,6 +5,7 @@ # credit: # Amin Rezaei (original author) # Alex Birch (optimized algorithm for 3D tensors, at the expense of removing bias, masking and callbacks) +# brkirch (modified to use torch.narrow instead of dynamic_slice implementation) # implementation of: # Self-attention Does Not Need O(n2) Memory": # https://arxiv.org/abs/2112.05682v2 @@ -16,13 +17,13 @@ from torch.utils.checkpoint import checkpoint import math from typing import Optional, NamedTuple, Protocol, List -def dynamic_slice( - x: Tensor, - starts: List[int], - sizes: List[int], +def narrow_trunc( + input: Tensor, + dim: int, + start: int, + length: int ) -> Tensor: - slicing = [slice(start, start + size) for start, size in zip(starts, sizes)] - return x[slicing] + return torch.narrow(input, dim, start, length if input.shape[dim] >= start + length else input.shape[dim] - start) class AttnChunk(NamedTuple): exp_values: Tensor @@ -76,15 +77,17 @@ def _query_chunk_attention( _, _, v_channels_per_head = value.shape def chunk_scanner(chunk_idx: int) -> AttnChunk: - key_chunk = dynamic_slice( + key_chunk = narrow_trunc( key, - (0, chunk_idx, 0), - (batch_x_heads, kv_chunk_size, k_channels_per_head) + 1, + chunk_idx, + kv_chunk_size ) - value_chunk = dynamic_slice( + value_chunk = narrow_trunc( value, - (0, chunk_idx, 0), - (batch_x_heads, kv_chunk_size, v_channels_per_head) + 1, + chunk_idx, + kv_chunk_size ) return summarize_chunk(query, key_chunk, value_chunk) @@ -161,10 +164,11 @@ def efficient_dot_product_attention( kv_chunk_size = max(kv_chunk_size, kv_chunk_size_min) def get_query_chunk(chunk_idx: int) -> Tensor: - return dynamic_slice( + return narrow_trunc( query, - (0, chunk_idx, 0), - (batch_x_heads, min(query_chunk_size, q_tokens), q_channels_per_head) + 1, + chunk_idx, + min(query_chunk_size, q_tokens) ) summarize_chunk: SummarizeChunk = partial(_summarize_chunk, scale=scale) -- cgit v1.2.3 From 683287d87f6401083a8d63eedc00ca7410214ca1 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 6 Jan 2023 08:52:06 +0300 Subject: rework saving training params to file #6372 --- modules/hypernetworks/hypernetwork.py | 28 +++++++------------------- modules/shared.py | 2 +- modules/textual_inversion/logging.py | 24 ++++++++++++++++++++++ modules/textual_inversion/textual_inversion.py | 23 +++------------------ 4 files changed, 35 insertions(+), 42 deletions(-) create mode 100644 modules/textual_inversion/logging.py (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 3237c37a..b0cfbe71 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -13,7 +13,7 @@ import tqdm from einops import rearrange, repeat from ldm.util import default from modules import devices, processing, sd_models, shared, sd_samplers -from modules.textual_inversion import textual_inversion +from modules.textual_inversion import textual_inversion, logging from modules.textual_inversion.learn_schedule import LearnRateScheduler from torch import einsum from torch.nn.init import normal_, xavier_normal_, xavier_uniform_, kaiming_normal_, kaiming_uniform_, zeros_ @@ -401,25 +401,7 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, hypernet.save(fn) shared.reload_hypernetworks() -# Note: textual_inversion.py has a nearly identical function of the same name. -def save_settings_to_file(model_name, model_hash, initial_step, num_of_dataset_images, hypernetwork_name, layer_structure, activation_func, weight_init, add_layer_norm, use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): - # Starting index of preview-related arguments. - border_index = 21 - # Get a list of the argument names. - arg_names = inspect.getfullargspec(save_settings_to_file).args - # Create a list of the argument names to include in the settings string. - names = arg_names[:border_index] # Include all arguments up until the preview-related ones. - if preview_from_txt2img: - names.extend(arg_names[border_index:]) # Include preview-related arguments if applicable. - # Build the settings string. - settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n" - for name in names: - if name != 'log_directory': # It's useless and redundant to save log_directory. - value = locals()[name] - settings_str += f"{name}: {value}\n" - # Create or append to the file. - with open(os.path.join(log_directory, 'settings.txt'), "a+") as fout: - fout.write(settings_str + "\n\n") + def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): # images allows training previews to have infotext. Importing it at the top causes a circular import problem. @@ -477,7 +459,11 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=hypernetwork_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, include_cond=True, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) if shared.opts.save_training_settings_to_txt: - save_settings_to_file(checkpoint.model_name, '[{}]'.format(checkpoint.hash), initial_step, len(ds), hypernetwork_name, hypernetwork.layer_structure, hypernetwork.activation_func, hypernetwork.weight_init, hypernetwork.add_layer_norm, hypernetwork.use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) + saved_params = dict( + model_name=checkpoint.model_name, model_hash=checkpoint.hash, num_of_dataset_images=len(ds), + **{field: getattr(hypernetwork, field) for field in ['layer_structure', 'activation_func', 'weight_init', 'add_layer_norm', 'use_dropout', ]} + ) + logging.save_settings_to_file(log_directory, {**saved_params, **locals()}) latent_sampling_method = ds.latent_sampling_method diff --git a/modules/shared.py b/modules/shared.py index f0e10b35..57e489d0 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -362,7 +362,7 @@ options_templates.update(options_section(('training', "Training"), { "unload_models_when_training": OptionInfo(False, "Move VAE and CLIP to RAM when training if possible. Saves VRAM."), "pin_memory": OptionInfo(False, "Turn on pin_memory for DataLoader. Makes training slightly faster but can increase memory usage."), "save_optimizer_state": OptionInfo(False, "Saves Optimizer state as separate *.optim file. Training of embedding or HN can be resumed with the matching optim file."), - "save_training_settings_to_txt": OptionInfo(False, "Save textual inversion and hypernet settings to a text file whenever training starts."), + "save_training_settings_to_txt": OptionInfo(True, "Save textual inversion and hypernet settings to a text file whenever training starts."), "dataset_filename_word_regex": OptionInfo("", "Filename word regex"), "dataset_filename_join_string": OptionInfo(" ", "Filename join string"), "training_image_repeats_per_epoch": OptionInfo(1, "Number of repeats for a single input image per epoch; used only for displaying epoch number", gr.Number, {"precision": 0}), diff --git a/modules/textual_inversion/logging.py b/modules/textual_inversion/logging.py new file mode 100644 index 00000000..8b1981d5 --- /dev/null +++ b/modules/textual_inversion/logging.py @@ -0,0 +1,24 @@ +import datetime +import json +import os + +saved_params_shared = {"model_name", "model_hash", "initial_step", "num_of_dataset_images", "learn_rate", "batch_size", "data_root", "log_directory", "training_width", "training_height", "steps", "create_image_every", "template_file"} +saved_params_ti = {"embedding_name", "num_vectors_per_token", "save_embedding_every", "save_image_with_stored_embedding"} +saved_params_hypernet = {"hypernetwork_name", "layer_structure", "activation_func", "weight_init", "add_layer_norm", "use_dropout", "save_hypernetwork_every"} +saved_params_all = saved_params_shared | saved_params_ti | saved_params_hypernet +saved_params_previews = {"preview_prompt", "preview_negative_prompt", "preview_steps", "preview_sampler_index", "preview_cfg_scale", "preview_seed", "preview_width", "preview_height"} + + +def save_settings_to_file(log_directory, all_params): + now = datetime.datetime.now() + params = {"datetime": now.strftime("%Y-%m-%d %H:%M:%S")} + + keys = saved_params_all + if all_params.get('preview_from_txt2img'): + keys = keys | saved_params_previews + + params.update({k: v for k, v in all_params.items() if k in keys}) + + filename = f'settings-{now.strftime("%Y-%m-%d-%H-%M-%S")}.json' + with open(os.path.join(log_directory, filename), "w") as file: + json.dump(params, file, indent=4) diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index e9cf432f..f9f5e8cd 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -18,6 +18,8 @@ from modules.textual_inversion.learn_schedule import LearnRateScheduler from modules.textual_inversion.image_embedding import (embedding_to_b64, embedding_from_b64, insert_image_data_embed, extract_image_data_embed, caption_image_overlay) +from modules.textual_inversion.logging import save_settings_to_file + class Embedding: def __init__(self, vec, name, step=None): @@ -231,25 +233,6 @@ def write_loss(log_directory, filename, step, epoch_len, values): **values, }) -# Note: hypernetwork.py has a nearly identical function of the same name. -def save_settings_to_file(model_name, model_hash, initial_step, num_of_dataset_images, embedding_name, vectors_per_token, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): - # Starting index of preview-related arguments. - border_index = 18 - # Get a list of the argument names. - arg_names = inspect.getfullargspec(save_settings_to_file).args - # Create a list of the argument names to include in the settings string. - names = arg_names[:border_index] # Include all arguments up until the preview-related ones. - if preview_from_txt2img: - names.extend(arg_names[border_index:]) # Include preview-related arguments if applicable. - # Build the settings string. - settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n" - for name in names: - if name != 'log_directory': # It's useless and redundant to save log_directory. - value = locals()[name] - settings_str += f"{name}: {value}\n" - # Create or append to the file. - with open(os.path.join(log_directory, 'settings.txt'), "a+") as fout: - fout.write(settings_str + "\n\n") def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, data_root, template_file, steps, save_model_every, create_image_every, log_directory, name="embedding"): assert model_name, f"{name} not selected" @@ -330,7 +313,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=embedding_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) if shared.opts.save_training_settings_to_txt: - save_settings_to_file(checkpoint.model_name, '[{}]'.format(checkpoint.hash), initial_step, len(ds), embedding_name, len(embedding.vec), learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height) + save_settings_to_file(log_directory, {**dict(model_name=checkpoint.model_name, model_hash=checkpoint.hash, num_of_dataset_images=len(ds), num_vectors_per_token=len(embedding.vec)), **locals()}) latent_sampling_method = ds.latent_sampling_method -- cgit v1.2.3 From b95a4c0ce5ab9c414e0494193bfff665f45e9e65 Mon Sep 17 00:00:00 2001 From: brkirch Date: Fri, 6 Jan 2023 01:01:51 -0500 Subject: Change sub-quad chunk threshold to use percentage --- modules/sd_hijack_optimizations.py | 18 +++++++++--------- modules/shared.py | 2 +- 2 files changed, 10 insertions(+), 10 deletions(-) (limited to 'modules') diff --git a/modules/sd_hijack_optimizations.py b/modules/sd_hijack_optimizations.py index f5c153e8..b416e9ac 100644 --- a/modules/sd_hijack_optimizations.py +++ b/modules/sd_hijack_optimizations.py @@ -233,7 +233,7 @@ def sub_quad_attention_forward(self, x, context=None, mask=None): k = k.unflatten(-1, (h, -1)).transpose(1,2).flatten(end_dim=1) v = v.unflatten(-1, (h, -1)).transpose(1,2).flatten(end_dim=1) - x = sub_quad_attention(q, k, v, q_chunk_size=shared.cmd_opts.sub_quad_q_chunk_size, kv_chunk_size=shared.cmd_opts.sub_quad_kv_chunk_size, chunk_threshold_bytes=shared.cmd_opts.sub_quad_chunk_threshold, use_checkpoint=self.training) + x = sub_quad_attention(q, k, v, q_chunk_size=shared.cmd_opts.sub_quad_q_chunk_size, kv_chunk_size=shared.cmd_opts.sub_quad_kv_chunk_size, chunk_threshold=shared.cmd_opts.sub_quad_chunk_threshold, use_checkpoint=self.training) x = x.unflatten(0, (-1, h)).transpose(1,2).flatten(start_dim=2) @@ -243,20 +243,20 @@ def sub_quad_attention_forward(self, x, context=None, mask=None): return x -def sub_quad_attention(q, k, v, q_chunk_size=1024, kv_chunk_size=None, kv_chunk_size_min=None, chunk_threshold_bytes=None, use_checkpoint=True): +def sub_quad_attention(q, k, v, q_chunk_size=1024, kv_chunk_size=None, kv_chunk_size_min=None, chunk_threshold=None, use_checkpoint=True): bytes_per_token = torch.finfo(q.dtype).bits//8 batch_x_heads, q_tokens, _ = q.shape _, k_tokens, _ = k.shape qk_matmul_size_bytes = batch_x_heads * bytes_per_token * q_tokens * k_tokens - available_vram = int(get_available_vram() * 0.9) if q.device.type == 'mps' else int(get_available_vram() * 0.7) - - if chunk_threshold_bytes is None: - chunk_threshold_bytes = available_vram - elif chunk_threshold_bytes == 0: + if chunk_threshold is None: + chunk_threshold_bytes = int(get_available_vram() * 0.9) if q.device.type == 'mps' else int(get_available_vram() * 0.7) + elif chunk_threshold == 0: chunk_threshold_bytes = None + else: + chunk_threshold_bytes = int(0.01 * chunk_threshold * get_available_vram()) - if kv_chunk_size_min is None: + if kv_chunk_size_min is None and chunk_threshold_bytes is not None: kv_chunk_size_min = chunk_threshold_bytes // (batch_x_heads * bytes_per_token * (k.shape[2] + v.shape[2])) elif kv_chunk_size_min == 0: kv_chunk_size_min = None @@ -382,7 +382,7 @@ def sub_quad_attnblock_forward(self, x): q = q.contiguous() k = k.contiguous() v = v.contiguous() - out = sub_quad_attention(q, k, v, q_chunk_size=shared.cmd_opts.sub_quad_q_chunk_size, kv_chunk_size=shared.cmd_opts.sub_quad_kv_chunk_size, chunk_threshold_bytes=shared.cmd_opts.sub_quad_chunk_threshold, use_checkpoint=self.training) + out = sub_quad_attention(q, k, v, q_chunk_size=shared.cmd_opts.sub_quad_q_chunk_size, kv_chunk_size=shared.cmd_opts.sub_quad_kv_chunk_size, chunk_threshold=shared.cmd_opts.sub_quad_chunk_threshold, use_checkpoint=self.training) out = rearrange(out, 'b (h w) c -> b c h w', h=h) out = self.proj_out(out) return x + out diff --git a/modules/shared.py b/modules/shared.py index cb1dc312..d7a81db1 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -59,7 +59,7 @@ parser.add_argument("--opt-split-attention", action='store_true', help="force-en parser.add_argument("--opt-sub-quad-attention", action='store_true', help="enable memory efficient sub-quadratic cross-attention layer optimization") parser.add_argument("--sub-quad-q-chunk-size", type=int, help="query chunk size for the sub-quadratic cross-attention layer optimization to use", default=1024) parser.add_argument("--sub-quad-kv-chunk-size", type=int, help="kv chunk size for the sub-quadratic cross-attention layer optimization to use", default=None) -parser.add_argument("--sub-quad-chunk-threshold", type=int, help="the size threshold in bytes for the sub-quadratic cross-attention layer optimization to use chunking", default=None) +parser.add_argument("--sub-quad-chunk-threshold", type=int, help="the percentage of VRAM threshold for the sub-quadratic cross-attention layer optimization to use chunking", default=None) parser.add_argument("--opt-split-attention-invokeai", action='store_true', help="force-enables InvokeAI's cross-attention layer optimization. By default, it's on when cuda is unavailable.") parser.add_argument("--opt-split-attention-v1", action='store_true', help="enable older version of split attention optimization that does not consume all the VRAM it can find") parser.add_argument("--disable-opt-split-attention", action='store_true', help="force-disables cross-attention layer optimization") -- cgit v1.2.3 From 5deb2a19ccea57a50252e8fcb07b4d17c6599def Mon Sep 17 00:00:00 2001 From: brkirch Date: Fri, 6 Jan 2023 01:33:15 -0500 Subject: Allow Doggettx's cross attention opt without CUDA --- modules/sd_hijack.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index ef25dadb..bd101e5b 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -50,7 +50,7 @@ def apply_optimizations(): print("Applying v1 cross attention optimization.") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_v1 optimization_method = 'V1' - elif not cmd_opts.disable_opt_split_attention and (cmd_opts.opt_split_attention_invokeai or not torch.cuda.is_available()): + elif not cmd_opts.disable_opt_split_attention and (cmd_opts.opt_split_attention_invokeai or not cmd_opts.opt_split_attention and not torch.cuda.is_available()): print("Applying cross attention optimization (InvokeAI).") ldm.modules.attention.CrossAttention.forward = sd_hijack_optimizations.split_cross_attention_forward_invokeAI optimization_method = 'InvokeAI' -- cgit v1.2.3 From c9bded39ee05bd0507ccd27d2b674d86d6c0c8e8 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 6 Jan 2023 12:32:44 +0300 Subject: sort extensions by date and add an option to sort by other columns --- modules/ui_extensions.py | 44 ++++++++++++++++++++++++++++++++------------ 1 file changed, 32 insertions(+), 12 deletions(-) (limited to 'modules') diff --git a/modules/ui_extensions.py b/modules/ui_extensions.py index eec9586f..742e745e 100644 --- a/modules/ui_extensions.py +++ b/modules/ui_extensions.py @@ -162,15 +162,15 @@ def install_extension_from_url(dirname, url): shutil.rmtree(tmpdir, True) -def install_extension_from_index(url, hide_tags): +def install_extension_from_index(url, hide_tags, sort_column): ext_table, message = install_extension_from_url(None, url) - code, _ = refresh_available_extensions_from_data(hide_tags) + code, _ = refresh_available_extensions_from_data(hide_tags, sort_column) return code, ext_table, message -def refresh_available_extensions(url, hide_tags): +def refresh_available_extensions(url, hide_tags, sort_column): global available_extensions import urllib.request @@ -179,18 +179,28 @@ def refresh_available_extensions(url, hide_tags): available_extensions = json.loads(text) - code, tags = refresh_available_extensions_from_data(hide_tags) + code, tags = refresh_available_extensions_from_data(hide_tags, sort_column) return url, code, gr.CheckboxGroup.update(choices=tags), '' -def refresh_available_extensions_for_tags(hide_tags): - code, _ = refresh_available_extensions_from_data(hide_tags) +def refresh_available_extensions_for_tags(hide_tags, sort_column): + code, _ = refresh_available_extensions_from_data(hide_tags, sort_column) return code, '' -def refresh_available_extensions_from_data(hide_tags): +sort_ordering = [ + # (reverse, order_by_function) + (True, lambda x: x.get('added', 'z')), + (False, lambda x: x.get('added', 'z')), + (False, lambda x: x.get('name', 'z')), + (True, lambda x: x.get('name', 'z')), + (False, lambda x: 'z'), +] + + +def refresh_available_extensions_from_data(hide_tags, sort_column): extlist = available_extensions["extensions"] installed_extension_urls = {normalize_git_url(extension.remote): extension.name for extension in extensions.extensions} @@ -210,8 +220,11 @@ def refresh_available_extensions_from_data(hide_tags): """ - for ext in extlist: + sort_reverse, sort_function = sort_ordering[sort_column if 0 <= sort_column < len(sort_ordering) else 0] + + for ext in sorted(extlist, key=sort_function, reverse=sort_reverse): name = ext.get("name", "noname") + added = ext.get('added', 'unknown') url = ext.get("url", None) description = ext.get("description", "") extension_tags = ext.get("tags", []) @@ -233,7 +246,7 @@ def refresh_available_extensions_from_data(hide_tags): code += f""" {html.escape(name)}
{tags_text} - {html.escape(description)} + {html.escape(description)}

Added: {html.escape(added)}

{install_code} @@ -291,25 +304,32 @@ def create_ui(): with gr.Row(): hide_tags = gr.CheckboxGroup(value=["ads", "localization", "installed"], label="Hide extensions with tags", choices=["script", "ads", "localization", "installed"]) + sort_column = gr.Radio(value="newest first", label="Order", choices=["newest first", "oldest first", "a-z", "z-a", "internal order", ], type="index") install_result = gr.HTML() available_extensions_table = gr.HTML() refresh_available_extensions_button.click( fn=modules.ui.wrap_gradio_call(refresh_available_extensions, extra_outputs=[gr.update(), gr.update(), gr.update()]), - inputs=[available_extensions_index, hide_tags], + inputs=[available_extensions_index, hide_tags, sort_column], outputs=[available_extensions_index, available_extensions_table, hide_tags, install_result], ) install_extension_button.click( fn=modules.ui.wrap_gradio_call(install_extension_from_index, extra_outputs=[gr.update(), gr.update()]), - inputs=[extension_to_install, hide_tags], + inputs=[extension_to_install, hide_tags, sort_column], outputs=[available_extensions_table, extensions_table, install_result], ) hide_tags.change( fn=modules.ui.wrap_gradio_call(refresh_available_extensions_for_tags, extra_outputs=[gr.update()]), - inputs=[hide_tags], + inputs=[hide_tags, sort_column], + outputs=[available_extensions_table, install_result] + ) + + sort_column.change( + fn=modules.ui.wrap_gradio_call(refresh_available_extensions_for_tags, extra_outputs=[gr.update()]), + inputs=[hide_tags, sort_column], outputs=[available_extensions_table, install_result] ) -- cgit v1.2.3 From 65ed4421e609dda3112f236c13e4db14caa71364 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 6 Jan 2023 13:55:50 +0300 Subject: add callback for when the script is unloaded --- modules/script_callbacks.py | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/script_callbacks.py b/modules/script_callbacks.py index de69fd9f..608c5300 100644 --- a/modules/script_callbacks.py +++ b/modules/script_callbacks.py @@ -71,6 +71,7 @@ callback_map = dict( callbacks_before_component=[], callbacks_after_component=[], callbacks_image_grid=[], + callbacks_script_unloaded=[], ) @@ -171,6 +172,14 @@ def image_grid_callback(params: ImageGridLoopParams): report_exception(c, 'image_grid') +def script_unloaded_callback(): + for c in reversed(callback_map['callbacks_script_unloaded']): + try: + c.callback() + except Exception: + report_exception(c, 'script_unloaded') + + def add_callback(callbacks, fun): stack = [x for x in inspect.stack() if x.filename != __file__] filename = stack[0].filename if len(stack) > 0 else 'unknown file' @@ -202,7 +211,7 @@ def on_app_started(callback): def on_model_loaded(callback): """register a function to be called when the stable diffusion model is created; the model is - passed as an argument""" + passed as an argument; this function is also called when the script is reloaded. """ add_callback(callback_map['callbacks_model_loaded'], callback) @@ -279,3 +288,10 @@ def on_image_grid(callback): - params: ImageGridLoopParams - parameters to be used for grid creation. Can be modified. """ add_callback(callback_map['callbacks_image_grid'], callback) + + +def on_script_unloaded(callback): + """register a function to be called before the script is unloaded. Any hooks/hijacks/monkeying about that + the script did should be reverted here""" + + add_callback(callback_map['callbacks_script_unloaded'], callback) -- cgit v1.2.3 From 3246a2d6b898da6a98fe9df4dc67944635a41bd3 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 6 Jan 2023 16:03:43 +0300 Subject: remove restriction for saving dropdowns to ui-config.json --- modules/scripts.py | 1 - modules/ui.py | 10 ++-------- 2 files changed, 2 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/scripts.py b/modules/scripts.py index 0c44f191..35164093 100644 --- a/modules/scripts.py +++ b/modules/scripts.py @@ -290,7 +290,6 @@ class ScriptRunner: script.group = group dropdown = gr.Dropdown(label="Script", elem_id="script_list", choices=["None"] + self.titles, value="None", type="index") - dropdown.save_to_config = True inputs[0] = dropdown for script in self.selectable_scripts: diff --git a/modules/ui.py b/modules/ui.py index 030f0685..b79d24ee 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -435,11 +435,9 @@ def create_toprow(is_img2img): with gr.Row(): with gr.Column(scale=1, elem_id="style_pos_col"): prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - prompt_style.save_to_config = True with gr.Column(scale=1, elem_id="style_neg_col"): prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - prompt_style2.save_to_config = True return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button @@ -638,7 +636,6 @@ def create_sampler_and_steps_selection(choices, tabname): if opts.samplers_in_dropdown: with FormRow(elem_id=f"sampler_selection_{tabname}"): sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - sampler_index.save_to_config = True steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) else: with FormGroup(elem_id=f"sampler_selection_{tabname}"): @@ -1794,7 +1791,7 @@ def create_ui(): if init_field is not None: init_field(saved_value) - if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number] and x.visible: + if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number, gr.Dropdown] and x.visible: apply_field(x, 'visible') if type(x) == gr.Slider: @@ -1815,11 +1812,8 @@ def create_ui(): if type(x) == gr.Number: apply_field(x, 'value') - # Since there are many dropdowns that shouldn't be saved, - # we only mark dropdowns that should be saved. - if type(x) == gr.Dropdown and getattr(x, 'save_to_config', False): + if type(x) == gr.Dropdown: apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) - apply_field(x, 'visible') visit(txt2img_interface, loadsave, "txt2img") visit(img2img_interface, loadsave, "img2img") -- cgit v1.2.3 From 50194de93ffc9db763d9b08fcc9c3bde1aa86151 Mon Sep 17 00:00:00 2001 From: Kuma <36082288+KumiIT@users.noreply.github.com> Date: Fri, 6 Jan 2023 16:12:45 +0100 Subject: typo UI fixes #6391 --- modules/shared.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 57e489d0..865c3c07 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -430,7 +430,7 @@ options_templates.update(options_section(('ui', "User interface"), { "samplers_in_dropdown": OptionInfo(True, "Use dropdown for sampler selection instead of radio group"), "dimensions_and_batch_together": OptionInfo(True, "Show Witdth/Height and Batch sliders in same row"), 'quicksettings': OptionInfo("sd_model_checkpoint", "Quicksettings list"), - 'ui_reorder': OptionInfo(", ".join(ui_reorder_categories), "txt2img/ing2img UI item order"), + 'ui_reorder': OptionInfo(", ".join(ui_reorder_categories), "txt2img/img2img UI item order"), 'localization': OptionInfo("None", "Localization (requires restart)", gr.Dropdown, lambda: {"choices": ["None"] + list(localization.localizations.keys())}, refresh=lambda: localization.list_localizations(cmd_opts.localizations_dir)), })) -- cgit v1.2.3 From 3992ecbe6e46a465062508c677964534e7397f72 Mon Sep 17 00:00:00 2001 From: Mitchell Boot <47387831+Mitchell1711@users.noreply.github.com> Date: Fri, 6 Jan 2023 18:02:46 +0100 Subject: Added UI elements Added a new row to hires fix that shows the new resolution after scaling --- modules/ui.py | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index b79d24ee..20f7d2a2 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -255,6 +255,12 @@ def add_style(name: str, prompt: str, negative_prompt: str): return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] +def calc_resolution_hires(x, y, scale): + #final res can only be a multiple of 8 + scaled_x = int(x * scale // 8) * 8 + scaled_y = int(y * scale // 8) * 8 + + return "

Upscaled Resolution: "+str(scaled_x)+"x"+str(scaled_y)+"

" def apply_styles(prompt, prompt_neg, style1_name, style2_name): prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) @@ -718,6 +724,12 @@ def create_ui(): hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") + + with FormRow(elem_id="txt2img_hires_fix_row3"): + hr_final_resolution = gr.HTML(value=calc_resolution_hires(width.value, height.value, hr_scale.value), elem_id="txtimg_hr_finalres") + hr_scale.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) + width.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) + height.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) elif category == "batch": if not opts.dimensions_and_batch_together: -- cgit v1.2.3 From 991368c8d54404d8e13d4c6e76a0f32644e65ad4 Mon Sep 17 00:00:00 2001 From: Mitchell Boot <47387831+Mitchell1711@users.noreply.github.com> Date: Fri, 6 Jan 2023 18:24:29 +0100 Subject: remove camelcase --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 20f7d2a2..6fc8b7d7 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -260,7 +260,7 @@ def calc_resolution_hires(x, y, scale): scaled_x = int(x * scale // 8) * 8 scaled_y = int(y * scale // 8) * 8 - return "

Upscaled Resolution: "+str(scaled_x)+"x"+str(scaled_y)+"

" + return "

Upscaled resolution: "+str(scaled_x)+"x"+str(scaled_y)+"

" def apply_styles(prompt, prompt_neg, style1_name, style2_name): prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) -- cgit v1.2.3 From c18add68ef7d2de3617cbbaff864b0c74cfdf6c0 Mon Sep 17 00:00:00 2001 From: brkirch Date: Fri, 6 Jan 2023 16:42:47 -0500 Subject: Added license --- modules/sd_hijack_optimizations.py | 1 + modules/sub_quadratic_attention.py | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_hijack_optimizations.py b/modules/sd_hijack_optimizations.py index b416e9ac..cdc63ed7 100644 --- a/modules/sd_hijack_optimizations.py +++ b/modules/sd_hijack_optimizations.py @@ -216,6 +216,7 @@ def split_cross_attention_forward_invokeAI(self, x, context=None, mask=None): # Based on Birch-san's modified implementation of sub-quadratic attention from https://github.com/Birch-san/diffusers/pull/1 +# The sub_quad_attention_forward function is under the MIT License listed under Memory Efficient Attention in the Licenses section of the web UI interface def sub_quad_attention_forward(self, x, context=None, mask=None): assert mask is None, "attention-mask not currently implemented for SubQuadraticCrossAttnProcessor." diff --git a/modules/sub_quadratic_attention.py b/modules/sub_quadratic_attention.py index 95924d24..fea7aaac 100644 --- a/modules/sub_quadratic_attention.py +++ b/modules/sub_quadratic_attention.py @@ -1,7 +1,7 @@ # original source: # https://github.com/AminRezaei0x443/memory-efficient-attention/blob/1bc0d9e6ac5f82ea43a375135c4e1d3896ee1694/memory_efficient_attention/attention_torch.py # license: -# unspecified +# MIT License (see Memory Efficient Attention under the Licenses section in the web UI interface for the full license) # credit: # Amin Rezaei (original author) # Alex Birch (optimized algorithm for 3D tensors, at the expense of removing bias, masking and callbacks) -- cgit v1.2.3 From 82c1f10b144f733460feead0bdc37a861489dc57 Mon Sep 17 00:00:00 2001 From: Dean Hopkins Date: Fri, 6 Jan 2023 22:00:12 +0000 Subject: increase upscale api validation limit --- modules/api/models.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/api/models.py b/modules/api/models.py index f77951fc..22b88c59 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -125,7 +125,7 @@ class ExtrasBaseRequest(BaseModel): gfpgan_visibility: float = Field(default=0, title="GFPGAN Visibility", ge=0, le=1, allow_inf_nan=False, description="Sets the visibility of GFPGAN, values should be between 0 and 1.") codeformer_visibility: float = Field(default=0, title="CodeFormer Visibility", ge=0, le=1, allow_inf_nan=False, description="Sets the visibility of CodeFormer, values should be between 0 and 1.") codeformer_weight: float = Field(default=0, title="CodeFormer Weight", ge=0, le=1, allow_inf_nan=False, description="Sets the weight of CodeFormer, values should be between 0 and 1.") - upscaling_resize: float = Field(default=2, title="Upscaling Factor", ge=1, le=4, description="By how much to upscale the image, only used when resize_mode=0.") + upscaling_resize: float = Field(default=2, title="Upscaling Factor", ge=1, le=8, description="By how much to upscale the image, only used when resize_mode=0.") upscaling_resize_w: int = Field(default=512, title="Target Width", ge=1, description="Target width for the upscaler to hit. Only used when resize_mode=1.") upscaling_resize_h: int = Field(default=512, title="Target Height", ge=1, description="Target height for the upscaler to hit. Only used when resize_mode=1.") upscaling_crop: bool = Field(default=True, title="Crop to fit", description="Should the upscaler crop the image to fit in the choosen size?") -- cgit v1.2.3 From 79e39fae6110c20a3ee6255e2841c877f65e8cbd Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 7 Jan 2023 01:45:28 +0300 Subject: CLIP hijack rework --- modules/sd_hijack.py | 6 +- modules/sd_hijack_clip.py | 348 ++++++++++++------------- modules/sd_hijack_clip_old.py | 81 ++++++ modules/textual_inversion/textual_inversion.py | 1 - modules/ui.py | 2 +- 5 files changed, 256 insertions(+), 182 deletions(-) create mode 100644 modules/sd_hijack_clip_old.py (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index fa2cd4bb..71cc145a 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -150,10 +150,10 @@ class StableDiffusionModelHijack: def clear_comments(self): self.comments = [] - def tokenize(self, text): - _, remade_batch_tokens, _, _, _, token_count = self.clip.process_text([text]) + def get_prompt_lengths(self, text): + _, token_count = self.clip.process_texts([text]) - return remade_batch_tokens[0], token_count, sd_hijack_clip.get_target_prompt_token_count(token_count) + return token_count, self.clip.get_target_prompt_token_count(token_count) class EmbeddingsWithFixes(torch.nn.Module): diff --git a/modules/sd_hijack_clip.py b/modules/sd_hijack_clip.py index ca92b142..ac3020d7 100644 --- a/modules/sd_hijack_clip.py +++ b/modules/sd_hijack_clip.py @@ -1,12 +1,28 @@ import math +from collections import namedtuple import torch from modules import prompt_parser, devices from modules.shared import opts -def get_target_prompt_token_count(token_count): - return math.ceil(max(token_count, 1) / 75) * 75 + +class PromptChunk: + """ + This object contains token ids, weight (multipliers:1.4) and textual inversion embedding info for a chunk of prompt. + If a prompt is short, it is represented by one PromptChunk, otherwise, multiple are necessary. + Each PromptChunk contains an exact amount of tokens - 77, which includes one for start and end token, + so just 75 tokens from prompt. + """ + + def __init__(self): + self.tokens = [] + self.multipliers = [] + self.fixes = [] + + +PromptChunkFix = namedtuple('PromptChunkFix', ['offset', 'embedding']) +"""This is a marker showing that textual inversion embedding's vectors have to placed at offset in the prompt chunk""" class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): @@ -14,17 +30,49 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): super().__init__() self.wrapped = wrapped self.hijack = hijack + self.chunk_length = 75 + + def empty_chunk(self): + """creates an empty PromptChunk and returns it""" + + chunk = PromptChunk() + chunk.tokens = [self.id_start] + [self.id_end] * (self.chunk_length + 1) + chunk.multipliers = [1.0] * (self.chunk_length + 2) + return chunk + + def get_target_prompt_token_count(self, token_count): + """returns the maximum number of tokens a prompt of a known length can have before it requires one more PromptChunk to be represented""" + + return math.ceil(max(token_count, 1) / self.chunk_length) * self.chunk_length def tokenize(self, texts): + """Converts a batch of texts into a batch of token ids""" + raise NotImplementedError def encode_with_transformers(self, tokens): + """ + converts a batch of token ids (in python lists) into a single tensor with numeric respresentation of those tokens; + All python lists with tokens are assumed to have same length, usually 77. + if input is a list with B elements and each element has T tokens, expected output shape is (B, T, C), where C depends on + model - can be 768 and 1024 + """ + raise NotImplementedError def encode_embedding_init_text(self, init_text, nvpt): + """Converts text into a tensor with this text's tokens' embeddings. Note that those are embeddings before they are passed through + transformers. nvpt is used as a maximum length in tokens. If text produces less teokens than nvpt, only this many is returned.""" + raise NotImplementedError - def tokenize_line(self, line, used_custom_terms, hijack_comments): + def tokenize_line(self, line): + """ + this transforms a single prompt into a list of PromptChunk objects - as many as needed to + represent the prompt. + Returns the list and the total number of tokens in the prompt. + """ + if opts.enable_emphasis: parsed = prompt_parser.parse_prompt_attention(line) else: @@ -32,205 +80,152 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): tokenized = self.tokenize([text for text, _ in parsed]) - fixes = [] - remade_tokens = [] - multipliers = [] + chunks = [] + chunk = PromptChunk() + token_count = 0 last_comma = -1 - for tokens, (text, weight) in zip(tokenized, parsed): - i = 0 - while i < len(tokens): - token = tokens[i] + def next_chunk(): + """puts current chunk into the list of results and produces the next one - empty""" + nonlocal token_count + nonlocal last_comma + nonlocal chunk + + token_count += len(chunk.tokens) + to_add = self.chunk_length - len(chunk.tokens) + if to_add > 0: + chunk.tokens += [self.id_end] * to_add + chunk.multipliers += [1.0] * to_add - embedding, embedding_length_in_tokens = self.hijack.embedding_db.find_embedding_at_position(tokens, i) + chunk.tokens = [self.id_start] + chunk.tokens + [self.id_end] + chunk.multipliers = [1.0] + chunk.multipliers + [1.0] + + last_comma = -1 + chunks.append(chunk) + chunk = PromptChunk() + + for tokens, (text, weight) in zip(tokenized, parsed): + position = 0 + while position < len(tokens): + token = tokens[position] if token == self.comma_token: - last_comma = len(remade_tokens) - elif opts.comma_padding_backtrack != 0 and max(len(remade_tokens), 1) % 75 == 0 and last_comma != -1 and len(remade_tokens) - last_comma <= opts.comma_padding_backtrack: - last_comma += 1 - reloc_tokens = remade_tokens[last_comma:] - reloc_mults = multipliers[last_comma:] + last_comma = len(chunk.tokens) + + # this is when we are at the end of alloted 75 tokens for the current chunk, and the current token is not a comma. opts.comma_padding_backtrack + # is a setting that specifies that is there is a comma nearby, the text after comma should be moved out of this chunk and into the next. + elif opts.comma_padding_backtrack != 0 and len(chunk.tokens) == self.chunk_length and last_comma != -1 and len(chunk.tokens) - last_comma <= opts.comma_padding_backtrack: + break_location = last_comma + 1 + + reloc_tokens = chunk.tokens[break_location:] + reloc_mults = chunk.multipliers[break_location:] - remade_tokens = remade_tokens[:last_comma] - length = len(remade_tokens) + chunk.tokens = chunk.tokens[:break_location] + chunk.multipliers = chunk.multipliers[:break_location] - rem = int(math.ceil(length / 75)) * 75 - length - remade_tokens += [self.id_end] * rem + reloc_tokens - multipliers = multipliers[:last_comma] + [1.0] * rem + reloc_mults + next_chunk() + chunk.tokens = reloc_tokens + chunk.multipliers = reloc_mults + if len(chunk.tokens) == self.chunk_length: + next_chunk() + + embedding, embedding_length_in_tokens = self.hijack.embedding_db.find_embedding_at_position(tokens, position) if embedding is None: - remade_tokens.append(token) - multipliers.append(weight) - i += 1 - else: - emb_len = int(embedding.vec.shape[0]) - iteration = len(remade_tokens) // 75 - if (len(remade_tokens) + emb_len) // 75 != iteration: - rem = (75 * (iteration + 1) - len(remade_tokens)) - remade_tokens += [self.id_end] * rem - multipliers += [1.0] * rem - iteration += 1 - fixes.append((iteration, (len(remade_tokens) % 75, embedding))) - remade_tokens += [0] * emb_len - multipliers += [weight] * emb_len - used_custom_terms.append((embedding.name, embedding.checksum())) - i += embedding_length_in_tokens - - token_count = len(remade_tokens) - prompt_target_length = get_target_prompt_token_count(token_count) - tokens_to_add = prompt_target_length - len(remade_tokens) - - remade_tokens = remade_tokens + [self.id_end] * tokens_to_add - multipliers = multipliers + [1.0] * tokens_to_add - - return remade_tokens, fixes, multipliers, token_count - - def process_text(self, texts): - used_custom_terms = [] - remade_batch_tokens = [] - hijack_comments = [] - hijack_fixes = [] + chunk.tokens.append(token) + chunk.multipliers.append(weight) + position += 1 + continue + + emb_len = int(embedding.vec.shape[0]) + if len(chunk.tokens) + emb_len > self.chunk_length: + next_chunk() + + chunk.fixes.append(PromptChunkFix(len(chunk.tokens), embedding)) + + chunk.tokens += [0] * emb_len + chunk.multipliers += [weight] * emb_len + position += embedding_length_in_tokens + + if len(chunk.tokens) > 0: + next_chunk() + + return chunks, token_count + + def process_texts(self, texts): + """ + Accepts a list of texts and calls tokenize_line() on each, with cache. Returns the list of results and maximum + length, in tokens, of all texts. + """ + token_count = 0 cache = {} - batch_multipliers = [] + batch_chunks = [] for line in texts: if line in cache: - remade_tokens, fixes, multipliers = cache[line] + chunks = cache[line] else: - remade_tokens, fixes, multipliers, current_token_count = self.tokenize_line(line, used_custom_terms, hijack_comments) + chunks, current_token_count = self.tokenize_line(line) token_count = max(current_token_count, token_count) - cache[line] = (remade_tokens, fixes, multipliers) + cache[line] = chunks - remade_batch_tokens.append(remade_tokens) - hijack_fixes.append(fixes) - batch_multipliers.append(multipliers) + batch_chunks.append(chunks) - return batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count + return batch_chunks, token_count - def process_text_old(self, texts): - id_start = self.id_start - id_end = self.id_end - maxlen = self.wrapped.max_length # you get to stay at 77 - used_custom_terms = [] - remade_batch_tokens = [] - hijack_comments = [] - hijack_fixes = [] - token_count = 0 + def forward(self, texts): + """ + Accepts an array of texts; Passes texts through transformers network to create a tensor with numerical representation of those texts. + Returns a tensor with shape of (B, T, C), where B is length of the array; T is length, in tokens, of texts (including padding) - T will + be a multiple of 77; and C is dimensionality of each token - for SD1 it's 768, and for SD2 it's 1024. + An example shape returned by this function can be: (2, 77, 768). + Webui usually sends just one text at a time through this function - the only time when texts is an array with more than one elemenet + is when you do prompt editing: "a picture of a [cat:dog:0.4] eating ice cream" + """ - cache = {} - batch_tokens = self.tokenize(texts) - batch_multipliers = [] - for tokens in batch_tokens: - tuple_tokens = tuple(tokens) + if opts.use_old_emphasis_implementation: + import modules.sd_hijack_clip_old + return modules.sd_hijack_clip_old.forward_old(self, texts) - if tuple_tokens in cache: - remade_tokens, fixes, multipliers = cache[tuple_tokens] - else: - fixes = [] - remade_tokens = [] - multipliers = [] - mult = 1.0 - - i = 0 - while i < len(tokens): - token = tokens[i] - - embedding, embedding_length_in_tokens = self.hijack.embedding_db.find_embedding_at_position(tokens, i) - - mult_change = self.token_mults.get(token) if opts.enable_emphasis else None - if mult_change is not None: - mult *= mult_change - i += 1 - elif embedding is None: - remade_tokens.append(token) - multipliers.append(mult) - i += 1 - else: - emb_len = int(embedding.vec.shape[0]) - fixes.append((len(remade_tokens), embedding)) - remade_tokens += [0] * emb_len - multipliers += [mult] * emb_len - used_custom_terms.append((embedding.name, embedding.checksum())) - i += embedding_length_in_tokens - - if len(remade_tokens) > maxlen - 2: - vocab = {v: k for k, v in self.wrapped.tokenizer.get_vocab().items()} - ovf = remade_tokens[maxlen - 2:] - overflowing_words = [vocab.get(int(x), "") for x in ovf] - overflowing_text = self.wrapped.tokenizer.convert_tokens_to_string(''.join(overflowing_words)) - hijack_comments.append(f"Warning: too many input tokens; some ({len(overflowing_words)}) have been truncated:\n{overflowing_text}\n") - - token_count = len(remade_tokens) - remade_tokens = remade_tokens + [id_end] * (maxlen - 2 - len(remade_tokens)) - remade_tokens = [id_start] + remade_tokens[0:maxlen - 2] + [id_end] - cache[tuple_tokens] = (remade_tokens, fixes, multipliers) - - multipliers = multipliers + [1.0] * (maxlen - 2 - len(multipliers)) - multipliers = [1.0] + multipliers[0:maxlen - 2] + [1.0] - - remade_batch_tokens.append(remade_tokens) - hijack_fixes.append(fixes) - batch_multipliers.append(multipliers) - return batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count - - def forward(self, text): - use_old = opts.use_old_emphasis_implementation - if use_old: - batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count = self.process_text_old(text) - else: - batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count = self.process_text(text) - - self.hijack.comments += hijack_comments - - if len(used_custom_terms) > 0: - self.hijack.comments.append("Used embeddings: " + ", ".join([f'{word} [{checksum}]' for word, checksum in used_custom_terms])) - - if use_old: - self.hijack.fixes = hijack_fixes - return self.process_tokens(remade_batch_tokens, batch_multipliers) - - z = None - i = 0 - while max(map(len, remade_batch_tokens)) != 0: - rem_tokens = [x[75:] for x in remade_batch_tokens] - rem_multipliers = [x[75:] for x in batch_multipliers] - - self.hijack.fixes = [] - for unfiltered in hijack_fixes: - fixes = [] - for fix in unfiltered: - if fix[0] == i: - fixes.append(fix[1]) - self.hijack.fixes.append(fixes) - - tokens = [] - multipliers = [] - for j in range(len(remade_batch_tokens)): - if len(remade_batch_tokens[j]) > 0: - tokens.append(remade_batch_tokens[j][:75]) - multipliers.append(batch_multipliers[j][:75]) - else: - tokens.append([self.id_end] * 75) - multipliers.append([1.0] * 75) - - z1 = self.process_tokens(tokens, multipliers) - z = z1 if z is None else torch.cat((z, z1), axis=-2) - - remade_batch_tokens = rem_tokens - batch_multipliers = rem_multipliers - i += 1 + batch_chunks, token_count = self.process_texts(texts) - return z + used_embeddings = {} + chunk_count = max([len(x) for x in batch_chunks]) - def process_tokens(self, remade_batch_tokens, batch_multipliers): - if not opts.use_old_emphasis_implementation: - remade_batch_tokens = [[self.id_start] + x[:75] + [self.id_end] for x in remade_batch_tokens] - batch_multipliers = [[1.0] + x[:75] + [1.0] for x in batch_multipliers] + zs = [] + for i in range(chunk_count): + batch_chunk = [chunks[i] if i < len(chunks) else self.empty_chunk() for chunks in batch_chunks] + + tokens = [x.tokens for x in batch_chunk] + multipliers = [x.multipliers for x in batch_chunk] + self.hijack.fixes = [x.fixes for x in batch_chunk] + for fixes in self.hijack.fixes: + for position, embedding in fixes: + used_embeddings[embedding.name] = embedding + + z = self.process_tokens(tokens, multipliers) + zs.append(z) + + if len(used_embeddings) > 0: + embeddings_list = ", ".join([f'{name} [{embedding.checksum()}]' for name, embedding in used_embeddings.items()]) + self.hijack.comments.append(f"Used embeddings: {embeddings_list}") + + return torch.hstack(zs) + + def process_tokens(self, remade_batch_tokens, batch_multipliers): + """ + sends one single prompt chunk to be encoded by transformers neural network. + remade_batch_tokens is a batch of tokens - a list, where every element is a list of tokens; usually + there are exactly 77 tokens in the list. batch_multipliers is the same but for multipliers instead of tokens. + Multipliers are used to give more or less weight to the outputs of transformers network. Each multiplier + corresponds to one token. + """ tokens = torch.asarray(remade_batch_tokens).to(devices.device) + # this is for SD2: SD1 uses the same token for padding and end of text, while SD2 uses different ones. if self.id_end != self.id_pad: for batch_pos in range(len(remade_batch_tokens)): index = remade_batch_tokens[batch_pos].index(self.id_end) @@ -239,8 +234,7 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): z = self.encode_with_transformers(tokens) # restoring original mean is likely not correct, but it seems to work well to prevent artifacts that happen otherwise - batch_multipliers_of_same_length = [x + [1.0] * (75 - len(x)) for x in batch_multipliers] - batch_multipliers = torch.asarray(batch_multipliers_of_same_length).to(devices.device) + batch_multipliers = torch.asarray(batch_multipliers).to(devices.device) original_mean = z.mean() z *= batch_multipliers.reshape(batch_multipliers.shape + (1,)).expand(z.shape) new_mean = z.mean() diff --git a/modules/sd_hijack_clip_old.py b/modules/sd_hijack_clip_old.py new file mode 100644 index 00000000..6d9fbbe6 --- /dev/null +++ b/modules/sd_hijack_clip_old.py @@ -0,0 +1,81 @@ +from modules import sd_hijack_clip +from modules import shared + + +def process_text_old(self: sd_hijack_clip.FrozenCLIPEmbedderWithCustomWordsBase, texts): + id_start = self.id_start + id_end = self.id_end + maxlen = self.wrapped.max_length # you get to stay at 77 + used_custom_terms = [] + remade_batch_tokens = [] + hijack_comments = [] + hijack_fixes = [] + token_count = 0 + + cache = {} + batch_tokens = self.tokenize(texts) + batch_multipliers = [] + for tokens in batch_tokens: + tuple_tokens = tuple(tokens) + + if tuple_tokens in cache: + remade_tokens, fixes, multipliers = cache[tuple_tokens] + else: + fixes = [] + remade_tokens = [] + multipliers = [] + mult = 1.0 + + i = 0 + while i < len(tokens): + token = tokens[i] + + embedding, embedding_length_in_tokens = self.hijack.embedding_db.find_embedding_at_position(tokens, i) + + mult_change = self.token_mults.get(token) if shared.opts.enable_emphasis else None + if mult_change is not None: + mult *= mult_change + i += 1 + elif embedding is None: + remade_tokens.append(token) + multipliers.append(mult) + i += 1 + else: + emb_len = int(embedding.vec.shape[0]) + fixes.append((len(remade_tokens), embedding)) + remade_tokens += [0] * emb_len + multipliers += [mult] * emb_len + used_custom_terms.append((embedding.name, embedding.checksum())) + i += embedding_length_in_tokens + + if len(remade_tokens) > maxlen - 2: + vocab = {v: k for k, v in self.wrapped.tokenizer.get_vocab().items()} + ovf = remade_tokens[maxlen - 2:] + overflowing_words = [vocab.get(int(x), "") for x in ovf] + overflowing_text = self.wrapped.tokenizer.convert_tokens_to_string(''.join(overflowing_words)) + hijack_comments.append(f"Warning: too many input tokens; some ({len(overflowing_words)}) have been truncated:\n{overflowing_text}\n") + + token_count = len(remade_tokens) + remade_tokens = remade_tokens + [id_end] * (maxlen - 2 - len(remade_tokens)) + remade_tokens = [id_start] + remade_tokens[0:maxlen - 2] + [id_end] + cache[tuple_tokens] = (remade_tokens, fixes, multipliers) + + multipliers = multipliers + [1.0] * (maxlen - 2 - len(multipliers)) + multipliers = [1.0] + multipliers[0:maxlen - 2] + [1.0] + + remade_batch_tokens.append(remade_tokens) + hijack_fixes.append(fixes) + batch_multipliers.append(multipliers) + return batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count + + +def forward_old(self: sd_hijack_clip.FrozenCLIPEmbedderWithCustomWordsBase, texts): + batch_multipliers, remade_batch_tokens, used_custom_terms, hijack_comments, hijack_fixes, token_count = process_text_old(self, texts) + + self.hijack.comments += hijack_comments + + if len(used_custom_terms) > 0: + self.hijack.comments.append("Used embeddings: " + ", ".join([f'{word} [{checksum}]' for word, checksum in used_custom_terms])) + + self.hijack.fixes = hijack_fixes + return self.process_tokens(remade_batch_tokens, batch_multipliers) diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index f9f5e8cd..45882ed6 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -79,7 +79,6 @@ class EmbeddingDatabase: self.word_embeddings[embedding.name] = embedding - # TODO changing between clip and open clip changes tokenization, which will cause embeddings to stop working ids = model.cond_stage_model.tokenize([embedding.name])[0] first_id = ids[0] diff --git a/modules/ui.py b/modules/ui.py index b79d24ee..5d2f5bad 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -368,7 +368,7 @@ def update_token_counter(text, steps): flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) prompts = [prompt_text for step, prompt_text in flat_prompts] - tokens, token_count, max_length = max([model_hijack.tokenize(prompt) for prompt in prompts], key=lambda args: args[1]) + token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) style_class = ' class="red"' if (token_count > max_length) else "" return f"{token_count}/{max_length}" -- cgit v1.2.3 From f94cfc563bbedd923d5e95563a5e8d93c8516ac3 Mon Sep 17 00:00:00 2001 From: Mitchell Boot <47387831+Mitchell1711@users.noreply.github.com> Date: Sat, 7 Jan 2023 01:15:22 +0100 Subject: Changed HTML to textbox instead Using HTML caused an issue where the row would expand for a frame when changing the sliders because of the loading animation. This solution also doesn't use any additional HTML padding --- modules/ui.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 6fc8b7d7..6ea1b5d7 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -260,7 +260,7 @@ def calc_resolution_hires(x, y, scale): scaled_x = int(x * scale // 8) * 8 scaled_y = int(y * scale // 8) * 8 - return "

Upscaled resolution: "+str(scaled_x)+"x"+str(scaled_y)+"

" + return str(scaled_x)+"x"+str(scaled_y) def apply_styles(prompt, prompt_neg, style1_name, style2_name): prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) @@ -726,7 +726,10 @@ def create_ui(): hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") with FormRow(elem_id="txt2img_hires_fix_row3"): - hr_final_resolution = gr.HTML(value=calc_resolution_hires(width.value, height.value, hr_scale.value), elem_id="txtimg_hr_finalres") + hr_final_resolution = gr.Textbox(value=calc_resolution_hires(width.value, height.value, hr_scale.value), + elem_id="txtimg_hr_finalres", + label="Upscaled resolution", + interactive=False) hr_scale.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) width.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) height.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) -- cgit v1.2.3 From 08066676a47b560235d4c085dd3cfcb470b80997 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 7 Jan 2023 07:22:07 +0300 Subject: make it not break on empty inputs; thank you tarded, we are --- modules/sd_hijack_clip.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_hijack_clip.py b/modules/sd_hijack_clip.py index ac3020d7..16aef76a 100644 --- a/modules/sd_hijack_clip.py +++ b/modules/sd_hijack_clip.py @@ -147,7 +147,7 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): chunk.multipliers += [weight] * emb_len position += embedding_length_in_tokens - if len(chunk.tokens) > 0: + if len(chunk.tokens) > 0 or len(chunks) == 0: next_chunk() return chunks, token_count -- cgit v1.2.3 From 1740c33547b62f692834c95914a2b295d51684c7 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 7 Jan 2023 07:48:44 +0300 Subject: more comments --- modules/sd_hijack_clip.py | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/sd_hijack_clip.py b/modules/sd_hijack_clip.py index 16aef76a..5520c9b2 100644 --- a/modules/sd_hijack_clip.py +++ b/modules/sd_hijack_clip.py @@ -3,7 +3,7 @@ from collections import namedtuple import torch -from modules import prompt_parser, devices +from modules import prompt_parser, devices, sd_hijack from modules.shared import opts @@ -22,14 +22,24 @@ class PromptChunk: PromptChunkFix = namedtuple('PromptChunkFix', ['offset', 'embedding']) -"""This is a marker showing that textual inversion embedding's vectors have to placed at offset in the prompt chunk""" +"""An object of this type is a marker showing that textual inversion embedding's vectors have to placed at offset in the prompt +chunk. Thos objects are found in PromptChunk.fixes and, are placed into FrozenCLIPEmbedderWithCustomWordsBase.hijack.fixes, and finally +are applied by sd_hijack.EmbeddingsWithFixes's forward function.""" class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): + """A pytorch module that is a wrapper for FrozenCLIPEmbedder module. it enhances FrozenCLIPEmbedder, making it possible to + have unlimited prompt length and assign weights to tokens in prompt. + """ + def __init__(self, wrapped, hijack): super().__init__() + self.wrapped = wrapped - self.hijack = hijack + """Original FrozenCLIPEmbedder module; can also be FrozenOpenCLIPEmbedder or xlmr.BertSeriesModelWithTransformation, + depending on model.""" + + self.hijack: sd_hijack.StableDiffusionModelHijack = hijack self.chunk_length = 75 def empty_chunk(self): @@ -55,7 +65,8 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): converts a batch of token ids (in python lists) into a single tensor with numeric respresentation of those tokens; All python lists with tokens are assumed to have same length, usually 77. if input is a list with B elements and each element has T tokens, expected output shape is (B, T, C), where C depends on - model - can be 768 and 1024 + model - can be 768 and 1024. + Among other things, this call will read self.hijack.fixes, apply it to its inputs, and clear it (setting it to None). """ raise NotImplementedError @@ -113,7 +124,7 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): last_comma = len(chunk.tokens) # this is when we are at the end of alloted 75 tokens for the current chunk, and the current token is not a comma. opts.comma_padding_backtrack - # is a setting that specifies that is there is a comma nearby, the text after comma should be moved out of this chunk and into the next. + # is a setting that specifies that if there is a comma nearby, the text after the comma should be moved out of this chunk and into the next. elif opts.comma_padding_backtrack != 0 and len(chunk.tokens) == self.chunk_length and last_comma != -1 and len(chunk.tokens) - last_comma <= opts.comma_padding_backtrack: break_location = last_comma + 1 -- cgit v1.2.3 From de9738044571877450d1038e18f1ecce93d24af3 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 7 Jan 2023 08:53:53 +0300 Subject: this breaks on default config because width, height, hr_scale are None at that point. --- modules/ui.py | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index f946382d..a18b9007 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -725,14 +725,8 @@ def create_ui(): hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") - with FormRow(elem_id="txt2img_hires_fix_row3"): - hr_final_resolution = gr.Textbox(value=calc_resolution_hires(width.value, height.value, hr_scale.value), - elem_id="txtimg_hr_finalres", - label="Upscaled resolution", - interactive=False) - hr_scale.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) - width.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) - height.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) + with FormRow(elem_id="txt2img_hires_fix_row3"): + hr_final_resolution = gr.Textbox(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) elif category == "batch": if not opts.dimensions_and_batch_together: @@ -744,6 +738,10 @@ def create_ui(): with FormGroup(elem_id="txt2img_script_container"): custom_inputs = modules.scripts.scripts_txt2img.setup_ui() + hr_scale.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) + width.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) + height.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) + txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) -- cgit v1.2.3 From 1a5b86ad65fd738eadea1ad72f4abad3a4aabf17 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 7 Jan 2023 09:56:37 +0300 Subject: rework hires fix preview for #6437: movie it to where it takes less place, make it actually account for all relevant sliders and calculate dimensions correctly --- modules/processing.py | 1 - modules/ui.py | 40 +++++++++++++++++++++++++++------------- modules/ui_components.py | 8 ++++++++ 3 files changed, 35 insertions(+), 14 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index a408d622..82157bc9 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -711,7 +711,6 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): self.truncate_x = 0 self.truncate_y = 0 - def init(self, all_prompts, all_seeds, all_subseeds): if self.enable_hr: if self.hr_resize_x == 0 and self.hr_resize_y == 0: diff --git a/modules/ui.py b/modules/ui.py index a18b9007..6c765262 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -20,7 +20,7 @@ from PIL import Image, PngImagePlugin from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru -from modules.ui_components import FormRow, FormGroup, ToolButton +from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML from modules.paths import script_path from modules.shared import opts, cmd_opts, restricted_opts @@ -255,12 +255,20 @@ def add_style(name: str, prompt: str, negative_prompt: str): return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] -def calc_resolution_hires(x, y, scale): - #final res can only be a multiple of 8 - scaled_x = int(x * scale // 8) * 8 - scaled_y = int(y * scale // 8) * 8 - - return str(scaled_x)+"x"+str(scaled_y) + +def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): + from modules import processing, devices + + if not enable: + return "" + + p = processing.StableDiffusionProcessingTxt2Img(width=width, height=height, enable_hr=True, hr_scale=hr_scale, hr_resize_x=hr_resize_x, hr_resize_y=hr_resize_y) + + with devices.autocast(): + p.init([""], [0], [0]) + + return f"resize to: {p.hr_upscale_to_x}x{p.hr_upscale_to_y}" + def apply_styles(prompt, prompt_neg, style1_name, style2_name): prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) @@ -712,6 +720,7 @@ def create_ui(): restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") + hr_final_resolution = FormHTML(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) elif category == "hires_fix": with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: @@ -724,9 +733,6 @@ def create_ui(): hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") - - with FormRow(elem_id="txt2img_hires_fix_row3"): - hr_final_resolution = gr.Textbox(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) elif category == "batch": if not opts.dimensions_and_batch_together: @@ -738,9 +744,16 @@ def create_ui(): with FormGroup(elem_id="txt2img_script_container"): custom_inputs = modules.scripts.scripts_txt2img.setup_ui() - hr_scale.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) - width.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) - height.change(fn=calc_resolution_hires, inputs=[width, height, hr_scale], outputs=hr_final_resolution, show_progress=False) + hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] + hr_resolution_preview_args = dict( + fn=calc_resolution_hires, + inputs=hr_resolution_preview_inputs, + outputs=[hr_final_resolution], + show_progress=False + ) + + for input in hr_resolution_preview_inputs: + input.change(**hr_resolution_preview_args) txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) @@ -803,6 +816,7 @@ def create_ui(): fn=lambda x: gr_show(x), inputs=[enable_hr], outputs=[hr_options], + show_progress = False, ) txt2img_paste_fields = [ diff --git a/modules/ui_components.py b/modules/ui_components.py index 91eb0e3d..cac001dc 100644 --- a/modules/ui_components.py +++ b/modules/ui_components.py @@ -23,3 +23,11 @@ class FormGroup(gr.Group, gr.components.FormComponent): def get_block_name(self): return "group" + + +class FormHTML(gr.HTML, gr.components.FormComponent): + """Same as gr.HTML but fits inside gradio forms""" + + def get_block_name(self): + return "html" + -- cgit v1.2.3 From fdfce4711076c2ebac1089bac8169d043eb7978f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 7 Jan 2023 13:29:47 +0300 Subject: add "from" resolution for hires fix to be less confusing. --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 6c765262..99483130 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -267,7 +267,7 @@ def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resiz with devices.autocast(): p.init([""], [0], [0]) - return f"resize to: {p.hr_upscale_to_x}x{p.hr_upscale_to_y}" + return f"resize: from {width}x{height} to {p.hr_upscale_to_x}x{p.hr_upscale_to_y}" def apply_styles(prompt, prompt_neg, style1_name, style2_name): -- cgit v1.2.3 From df3b31eb559ab9fabf7e513bdeddd5282c16f124 Mon Sep 17 00:00:00 2001 From: brkirch Date: Sat, 7 Jan 2023 07:04:59 -0500 Subject: In-place operations can break gradient calculation --- modules/sd_hijack_clip.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/sd_hijack_clip.py b/modules/sd_hijack_clip.py index 5520c9b2..852afc66 100644 --- a/modules/sd_hijack_clip.py +++ b/modules/sd_hijack_clip.py @@ -247,9 +247,9 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): # restoring original mean is likely not correct, but it seems to work well to prevent artifacts that happen otherwise batch_multipliers = torch.asarray(batch_multipliers).to(devices.device) original_mean = z.mean() - z *= batch_multipliers.reshape(batch_multipliers.shape + (1,)).expand(z.shape) + z = z * batch_multipliers.reshape(batch_multipliers.shape + (1,)).expand(z.shape) new_mean = z.mean() - z *= original_mean / new_mean + z = z * (original_mean / new_mean) return z -- cgit v1.2.3 From 47534577eda63b0db1eeb8921c2a161773ec434c Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Sat, 7 Jan 2023 07:51:35 -0500 Subject: api-get-memory --- modules/api/api.py | 37 +++++++++++++++++++++++++++++++++++++ modules/api/models.py | 4 ++++ 2 files changed, 41 insertions(+) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 2103709b..d2222b18 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -130,6 +130,7 @@ class Api: self.add_api_route("/sdapi/v1/preprocess", self.preprocess, methods=["POST"], response_model=PreprocessResponse) self.add_api_route("/sdapi/v1/train/embedding", self.train_embedding, methods=["POST"], response_model=TrainResponse) self.add_api_route("/sdapi/v1/train/hypernetwork", self.train_hypernetwork, methods=["POST"], response_model=TrainResponse) + self.add_api_route("/sdapi/v1/memory", self.get_memory, methods=["GET"], response_model=MemoryResponse) def add_api_route(self, path: str, endpoint, **kwargs): if shared.cmd_opts.api_auth: @@ -465,6 +466,42 @@ class Api: shared.state.end() return TrainResponse(info = "train embedding error: {error}".format(error = error)) + def get_memory(self): + def gb(val: float): + return round(val / 1024 / 1024 / 1024, 2) + try: + import os, psutil + process = psutil.Process(os.getpid()) + res = process.memory_info() + ram_total = 100 * res.rss / process.memory_percent() + ram = { 'free': gb(ram_total - res.rss), 'used': gb(res.rss), 'total': gb(ram_total) } + except Exception as err: + ram = { 'error': f'{err}' } + try: + import torch + if torch.cuda.is_available(): + s = torch.cuda.mem_get_info() + system = { 'free': gb(s[0]), 'used': gb(s[1] - s[0]), 'total': gb(s[1]) } + s = dict(torch.cuda.memory_stats(shared.device)) + allocated = { 'current': gb(s['allocated_bytes.all.current']), 'peak': gb(s['allocated_bytes.all.peak']) } + reserved = { 'current': gb(s['reserved_bytes.all.current']), 'peak': gb(s['reserved_bytes.all.peak']) } + active = { 'current': gb(s['active_bytes.all.current']), 'peak': gb(s['active_bytes.all.peak']) } + inactive = { 'current': gb(s['inactive_split_bytes.all.current']), 'peak': gb(s['inactive_split_bytes.all.peak']) } + warnings = { 'retries': s['num_alloc_retries'], 'oom': s['num_ooms'] } + cuda = { + 'system': system, + 'active': active, + 'allocated': allocated, + 'reserved': reserved, + 'inactive': inactive, + 'events': warnings, + } + else: + cuda = { 'error': 'unavailable' } + except Exception as err: + cuda = { 'error': f'{err}' } + return MemoryResponse(ram = ram, cuda = cuda) + def launch(self, server_name, port): self.app.include_router(self.router) uvicorn.run(self.app, host=server_name, port=port) diff --git a/modules/api/models.py b/modules/api/models.py index 5fa63774..49bf1e7a 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -260,3 +260,7 @@ class EmbeddingItem(BaseModel): class EmbeddingsResponse(BaseModel): loaded: Dict[str, EmbeddingItem] = Field(title="Loaded", description="Embeddings loaded for the current model") skipped: Dict[str, EmbeddingItem] = Field(title="Skipped", description="Embeddings skipped for the current model (likely due to architecture incompatibility)") + +class MemoryResponse(BaseModel): + ram: dict[str, str] | dict[str, float] = Field(title="RAM", description="System memory stats") + cuda: dict[str, str] | dict[str, dict] = Field(title="CUDA", description="nVidia CUDA memory stats") -- cgit v1.2.3 From d38ede71d5330958f4bbac5f99c1be3c146b506a Mon Sep 17 00:00:00 2001 From: noodleanon <122053346+noodleanon@users.noreply.github.com> Date: Sat, 7 Jan 2023 14:21:31 +0000 Subject: Added script support in txt2img endpoint --- modules/api/api.py | 22 +++++++++++++++++++--- modules/api/models.py | 2 +- 2 files changed, 20 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index aa62a42e..0e8ea263 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -149,6 +149,14 @@ class Api: raise HTTPException(status_code=401, detail="Incorrect username or password", headers={"WWW-Authenticate": "Basic"}) def text2imgapi(self, txt2imgreq: StableDiffusionTxt2ImgProcessingAPI): + if txt2imgreq.script_name is not None: + if scripts.scripts_txt2img.scripts == []: + scripts.scripts_txt2img.initialize_scripts(True) + ui.create_ui() + + script_idx = script_name_to_index(txt2imgreq.script_name, scripts.scripts_txt2img.selectable_scripts) + script = scripts.scripts_txt2img.selectable_scripts[script_idx] + populate = txt2imgreq.copy(update={ # Override __init__ params "sampler_name": validate_sampler_name(txt2imgreq.sampler_name or txt2imgreq.sampler_index), "do_not_save_samples": True, @@ -158,11 +166,20 @@ class Api: if populate.sampler_name: populate.sampler_index = None # prevent a warning later on + args = vars(populate) + args.pop('script_name', None) + with self.queue_lock: - p = StableDiffusionProcessingTxt2Img(sd_model=shared.sd_model, **vars(populate)) + p = StableDiffusionProcessingTxt2Img(sd_model=shared.sd_model, **args) shared.state.begin() - processed = process_images(p) + if 'script' in locals(): + p.outpath_grids = opts.outdir_txt2img_grids + p.outpath_samples = opts.outdir_txt2img_samples + p.script_args = [script_idx + 1] + [None] * (script.args_from - 1) + p.script_args + processed = scripts.scripts_txt2img.run(p, *p.script_args) + else: + processed = process_images(p) shared.state.end() @@ -213,7 +230,6 @@ class Api: processed = scripts.scripts_img2img.run(p, *p.script_args) else: processed = process_images(p) - shared.state.end() b64images = list(map(encode_pil_to_base64, processed.images)) diff --git a/modules/api/models.py b/modules/api/models.py index c85eb94d..ce43c858 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -100,7 +100,7 @@ class PydanticModelGenerator: StableDiffusionTxt2ImgProcessingAPI = PydanticModelGenerator( "StableDiffusionProcessingTxt2Img", StableDiffusionProcessingTxt2Img, - [{"key": "sampler_index", "type": str, "default": "Euler"}] + [{"key": "sampler_index", "type": str, "default": "Euler"}, {"key": "script_name", "type": str, "default": None}, {"key": "script_args", "type": list, "default": []}] ).generate_model() StableDiffusionImg2ImgProcessingAPI = PydanticModelGenerator( -- cgit v1.2.3 From 448b9cedab66e05b5b2800513ca334a769b42aa7 Mon Sep 17 00:00:00 2001 From: dan Date: Sat, 7 Jan 2023 21:07:27 +0800 Subject: Allow variable img size --- modules/textual_inversion/dataset.py | 18 +++++++++++------- modules/textual_inversion/textual_inversion.py | 4 ++-- 2 files changed, 13 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py index 88d68c76..375178ed 100644 --- a/modules/textual_inversion/dataset.py +++ b/modules/textual_inversion/dataset.py @@ -17,7 +17,7 @@ re_numbers_at_start = re.compile(r"^[-\d]+\s*") class DatasetEntry: - def __init__(self, filename=None, filename_text=None, latent_dist=None, latent_sample=None, cond=None, cond_text=None, pixel_values=None): + def __init__(self, filename=None, filename_text=None, latent_dist=None, latent_sample=None, cond=None, cond_text=None, pixel_values=None, img_shape=None): self.filename = filename self.filename_text = filename_text self.latent_dist = latent_dist @@ -25,6 +25,7 @@ class DatasetEntry: self.cond = cond self.cond_text = cond_text self.pixel_values = pixel_values + self.img_shape = img_shape class PersonalizedBase(Dataset): @@ -33,8 +34,6 @@ class PersonalizedBase(Dataset): self.placeholder_token = placeholder_token - self.width = width - self.height = height self.flip = transforms.RandomHorizontalFlip(p=flip_p) self.dataset = [] @@ -59,7 +58,11 @@ class PersonalizedBase(Dataset): if shared.state.interrupted: raise Exception("interrupted") try: - image = Image.open(path).convert('RGB').resize((self.width, self.height), PIL.Image.BICUBIC) + image = Image.open(path).convert('RGB') + if width < 2000: + image = image.resize((width, height), PIL.Image.BICUBIC) + else: + assert batch_size == 1, 'variable img size must have batch size 1' except Exception: continue @@ -88,14 +91,14 @@ class PersonalizedBase(Dataset): if latent_sampling_method == "once" or (latent_sampling_method == "deterministic" and not isinstance(latent_dist, DiagonalGaussianDistribution)): latent_sample = model.get_first_stage_encoding(latent_dist).squeeze().to(devices.cpu) latent_sampling_method = "once" - entry = DatasetEntry(filename=path, filename_text=filename_text, latent_sample=latent_sample) + entry = DatasetEntry(filename=path, filename_text=filename_text, latent_sample=latent_sample, img_shape=image.size) elif latent_sampling_method == "deterministic": # Works only for DiagonalGaussianDistribution latent_dist.std = 0 latent_sample = model.get_first_stage_encoding(latent_dist).squeeze().to(devices.cpu) - entry = DatasetEntry(filename=path, filename_text=filename_text, latent_sample=latent_sample) + entry = DatasetEntry(filename=path, filename_text=filename_text, latent_sample=latent_sample, img_shape=image.size) elif latent_sampling_method == "random": - entry = DatasetEntry(filename=path, filename_text=filename_text, latent_dist=latent_dist) + entry = DatasetEntry(filename=path, filename_text=filename_text, latent_dist=latent_dist, img_shape=image.size) if not (self.tag_drop_out != 0 or self.shuffle_tags): entry.cond_text = self.create_text(filename_text) @@ -151,6 +154,7 @@ class BatchLoader: self.cond_text = [entry.cond_text for entry in data] self.cond = [entry.cond for entry in data] self.latent_sample = torch.stack([entry.latent_sample for entry in data]).squeeze(1) + self.img_shape = [entry.img_shape for entry in data] #self.emb_index = [entry.emb_index for entry in data] #print(self.latent_sample.device) diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 45882ed6..9f96d0fd 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -451,8 +451,8 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ else: p.prompt = batch.cond_text[0] p.steps = 20 - p.width = training_width - p.height = training_height + p.width = batch.img_shape[0][0] + p.height = batch.img_shape[0][1] preview_text = p.prompt -- cgit v1.2.3 From 669fb18d5222f53ae48abe0f30393d846c50ad91 Mon Sep 17 00:00:00 2001 From: dan Date: Sun, 8 Jan 2023 01:34:52 +0800 Subject: Add checkbox for variable training dims --- modules/hypernetworks/hypernetwork.py | 2 +- modules/textual_inversion/dataset.py | 4 ++-- modules/textual_inversion/textual_inversion.py | 4 ++-- modules/ui.py | 3 +++ 4 files changed, 8 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index b0cfbe71..dba52841 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -403,7 +403,7 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, shared.reload_hypernetworks() -def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): +def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): # images allows training previews to have infotext. Importing it at the top causes a circular import problem. from modules import images diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py index 375178ed..7f8a314f 100644 --- a/modules/textual_inversion/dataset.py +++ b/modules/textual_inversion/dataset.py @@ -29,7 +29,7 @@ class DatasetEntry: class PersonalizedBase(Dataset): - def __init__(self, data_root, width, height, repeats, flip_p=0.5, placeholder_token="*", model=None, cond_model=None, device=None, template_file=None, include_cond=False, batch_size=1, gradient_step=1, shuffle_tags=False, tag_drop_out=0, latent_sampling_method='once'): + def __init__(self, data_root, width, height, repeats, flip_p=0.5, placeholder_token="*", model=None, cond_model=None, device=None, template_file=None, include_cond=False, batch_size=1, gradient_step=1, shuffle_tags=False, tag_drop_out=0, latent_sampling_method='once', varsize=False): re_word = re.compile(shared.opts.dataset_filename_word_regex) if len(shared.opts.dataset_filename_word_regex) > 0 else None self.placeholder_token = placeholder_token @@ -59,7 +59,7 @@ class PersonalizedBase(Dataset): raise Exception("interrupted") try: image = Image.open(path).convert('RGB') - if width < 2000: + if not varsize: image = image.resize((width, height), PIL.Image.BICUBIC) else: assert batch_size == 1, 'variable img size must have batch size 1' diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 9f96d0fd..110efd19 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -255,7 +255,7 @@ def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, dat if save_model_every or create_image_every: assert log_directory, "Log directory is empty" -def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): +def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): save_embedding_every = save_embedding_every or 0 create_image_every = create_image_every or 0 validate_train_inputs(embedding_name, learn_rate, batch_size, gradient_step, data_root, template_file, steps, save_embedding_every, create_image_every, log_directory, name="embedding") @@ -309,7 +309,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ pin_memory = shared.opts.pin_memory - ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=embedding_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) + ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=embedding_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method, varsize=varsize) if shared.opts.save_training_settings_to_txt: save_settings_to_file(log_directory, {**dict(model_name=checkpoint.model_name, model_hash=checkpoint.hash, num_of_dataset_images=len(ds), num_vectors_per_token=len(embedding.vec)), **locals()}) diff --git a/modules/ui.py b/modules/ui.py index 99483130..4e709a71 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1343,6 +1343,7 @@ def create_ui(): template_file = gr.Textbox(label='Prompt template file', value=os.path.join(script_path, "textual_inversion_templates", "style_filewords.txt"), elem_id="train_template_file") training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") + varsize = gr.Checkbox(label="Ignore dimension settings and do not resize images", value=False, elem_id="train_varsize") steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") with FormRow(): @@ -1449,6 +1450,7 @@ def create_ui(): log_directory, training_width, training_height, + varsize, steps, clip_grad_mode, clip_grad_value, @@ -1480,6 +1482,7 @@ def create_ui(): log_directory, training_width, training_height, + varsize, steps, clip_grad_mode, clip_grad_value, -- cgit v1.2.3 From 72497895b9b1948f86d9309fe897cbb70c20ba7e Mon Sep 17 00:00:00 2001 From: dan Date: Sun, 8 Jan 2023 01:36:00 +0800 Subject: Move batchsize check --- modules/hypernetworks/hypernetwork.py | 2 +- modules/textual_inversion/dataset.py | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index dba52841..32c67ccc 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -456,7 +456,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, pin_memory = shared.opts.pin_memory - ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=hypernetwork_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, include_cond=True, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method) + ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=hypernetwork_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, include_cond=True, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method, varsize=varsize) if shared.opts.save_training_settings_to_txt: saved_params = dict( diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py index 7f8a314f..bcad6848 100644 --- a/modules/textual_inversion/dataset.py +++ b/modules/textual_inversion/dataset.py @@ -46,6 +46,8 @@ class PersonalizedBase(Dataset): assert data_root, 'dataset directory not specified' assert os.path.isdir(data_root), "Dataset directory doesn't exist" assert os.listdir(data_root), "Dataset directory is empty" + if varsize: + assert batch_size == 1, 'variable img size must have batch size 1' self.image_paths = [os.path.join(data_root, file_path) for file_path in os.listdir(data_root)] @@ -61,8 +63,6 @@ class PersonalizedBase(Dataset): image = Image.open(path).convert('RGB') if not varsize: image = image.resize((width, height), PIL.Image.BICUBIC) - else: - assert batch_size == 1, 'variable img size must have batch size 1' except Exception: continue -- cgit v1.2.3 From 984b86dd0abf0da7f6b116864c791a2bfe8859ef Mon Sep 17 00:00:00 2001 From: ProGamerGov Date: Sat, 7 Jan 2023 13:08:21 -0700 Subject: Add fallback for Protocol import --- modules/sub_quadratic_attention.py | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sub_quadratic_attention.py b/modules/sub_quadratic_attention.py index fea7aaac..93381bae 100644 --- a/modules/sub_quadratic_attention.py +++ b/modules/sub_quadratic_attention.py @@ -15,7 +15,13 @@ import torch from torch import Tensor from torch.utils.checkpoint import checkpoint import math -from typing import Optional, NamedTuple, Protocol, List + +try: + from typing import Protocol +except: + from typing_extensions import Protocol + +from typing import Optional, NamedTuple, List def narrow_trunc( input: Tensor, -- cgit v1.2.3 From a0c87f1fdf2b76b2ae4ef6c4b01ddaede3afab06 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 8 Jan 2023 08:52:26 +0300 Subject: skip images in embeddings dir if they have a second .preview extension --- modules/textual_inversion/textual_inversion.py | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 45882ed6..e85dd549 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -109,6 +109,10 @@ class EmbeddingDatabase: ext = ext.upper() if ext in ['.PNG', '.WEBP', '.JXL', '.AVIF']: + _, second_ext = os.path.splitext(name) + if second_ext.upper() == '.PREVIEW': + return + embed_image = Image.open(path) if hasattr(embed_image, 'text') and 'sd-ti-embedding' in embed_image.text: data = embedding_from_b64(embed_image.text['sd-ti-embedding']) -- cgit v1.2.3 From 085427de0efc9e9e7a6e9a5aebc6b5a69f0365e7 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 8 Jan 2023 09:37:33 +0300 Subject: make it possible for extensions/scripts to add their own embedding directories --- modules/sd_hijack.py | 7 +- modules/textual_inversion/textual_inversion.py | 170 +++++++++++++++---------- 2 files changed, 108 insertions(+), 69 deletions(-) (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index cfdb09d6..6b0d95af 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -83,10 +83,12 @@ class StableDiffusionModelHijack: clip = None optimization_method = None - embedding_db = modules.textual_inversion.textual_inversion.EmbeddingDatabase(cmd_opts.embeddings_dir) + embedding_db = modules.textual_inversion.textual_inversion.EmbeddingDatabase() - def hijack(self, m): + def __init__(self): + self.embedding_db.add_embedding_dir(cmd_opts.embeddings_dir) + def hijack(self, m): if type(m.cond_stage_model) == xlmr.BertSeriesModelWithTransformation: model_embeddings = m.cond_stage_model.roberta.embeddings model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.word_embeddings, self) @@ -117,7 +119,6 @@ class StableDiffusionModelHijack: self.layers = flatten(m) def undo_hijack(self, m): - if type(m.cond_stage_model) == xlmr.BertSeriesModelWithTransformation: m.cond_stage_model = m.cond_stage_model.wrapped diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index e85dd549..217fe9eb 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -66,17 +66,41 @@ class Embedding: return self.cached_checksum +class DirWithTextualInversionEmbeddings: + def __init__(self, path): + self.path = path + self.mtime = None + + def has_changed(self): + if not os.path.isdir(self.path): + return False + + mt = os.path.getmtime(self.path) + if self.mtime is None or mt > self.mtime: + return True + + def update(self): + if not os.path.isdir(self.path): + return + + self.mtime = os.path.getmtime(self.path) + + class EmbeddingDatabase: - def __init__(self, embeddings_dir): + def __init__(self): self.ids_lookup = {} self.word_embeddings = {} self.skipped_embeddings = {} - self.dir_mtime = None - self.embeddings_dir = embeddings_dir self.expected_shape = -1 + self.embedding_dirs = {} - def register_embedding(self, embedding, model): + def add_embedding_dir(self, path): + self.embedding_dirs[path] = DirWithTextualInversionEmbeddings(path) + + def clear_embedding_dirs(self): + self.embedding_dirs.clear() + def register_embedding(self, embedding, model): self.word_embeddings[embedding.name] = embedding ids = model.cond_stage_model.tokenize([embedding.name])[0] @@ -93,69 +117,62 @@ class EmbeddingDatabase: vec = shared.sd_model.cond_stage_model.encode_embedding_init_text(",", 1) return vec.shape[1] - def load_textual_inversion_embeddings(self, force_reload = False): - mt = os.path.getmtime(self.embeddings_dir) - if not force_reload and self.dir_mtime is not None and mt <= self.dir_mtime: - return + def load_from_file(self, path, filename): + name, ext = os.path.splitext(filename) + ext = ext.upper() - self.dir_mtime = mt - self.ids_lookup.clear() - self.word_embeddings.clear() - self.skipped_embeddings.clear() - self.expected_shape = self.get_expected_shape() - - def process_file(path, filename): - name, ext = os.path.splitext(filename) - ext = ext.upper() - - if ext in ['.PNG', '.WEBP', '.JXL', '.AVIF']: - _, second_ext = os.path.splitext(name) - if second_ext.upper() == '.PREVIEW': - return - - embed_image = Image.open(path) - if hasattr(embed_image, 'text') and 'sd-ti-embedding' in embed_image.text: - data = embedding_from_b64(embed_image.text['sd-ti-embedding']) - name = data.get('name', name) - else: - data = extract_image_data_embed(embed_image) - name = data.get('name', name) - elif ext in ['.BIN', '.PT']: - data = torch.load(path, map_location="cpu") - else: + if ext in ['.PNG', '.WEBP', '.JXL', '.AVIF']: + _, second_ext = os.path.splitext(name) + if second_ext.upper() == '.PREVIEW': return - # textual inversion embeddings - if 'string_to_param' in data: - param_dict = data['string_to_param'] - if hasattr(param_dict, '_parameters'): - param_dict = getattr(param_dict, '_parameters') # fix for torch 1.12.1 loading saved file from torch 1.11 - assert len(param_dict) == 1, 'embedding file has multiple terms in it' - emb = next(iter(param_dict.items()))[1] - # diffuser concepts - elif type(data) == dict and type(next(iter(data.values()))) == torch.Tensor: - assert len(data.keys()) == 1, 'embedding file has multiple terms in it' - - emb = next(iter(data.values())) - if len(emb.shape) == 1: - emb = emb.unsqueeze(0) - else: - raise Exception(f"Couldn't identify {filename} as neither textual inversion embedding nor diffuser concept.") - - vec = emb.detach().to(devices.device, dtype=torch.float32) - embedding = Embedding(vec, name) - embedding.step = data.get('step', None) - embedding.sd_checkpoint = data.get('sd_checkpoint', None) - embedding.sd_checkpoint_name = data.get('sd_checkpoint_name', None) - embedding.vectors = vec.shape[0] - embedding.shape = vec.shape[-1] - - if self.expected_shape == -1 or self.expected_shape == embedding.shape: - self.register_embedding(embedding, shared.sd_model) + embed_image = Image.open(path) + if hasattr(embed_image, 'text') and 'sd-ti-embedding' in embed_image.text: + data = embedding_from_b64(embed_image.text['sd-ti-embedding']) + name = data.get('name', name) else: - self.skipped_embeddings[name] = embedding + data = extract_image_data_embed(embed_image) + name = data.get('name', name) + elif ext in ['.BIN', '.PT']: + data = torch.load(path, map_location="cpu") + else: + return + + # textual inversion embeddings + if 'string_to_param' in data: + param_dict = data['string_to_param'] + if hasattr(param_dict, '_parameters'): + param_dict = getattr(param_dict, '_parameters') # fix for torch 1.12.1 loading saved file from torch 1.11 + assert len(param_dict) == 1, 'embedding file has multiple terms in it' + emb = next(iter(param_dict.items()))[1] + # diffuser concepts + elif type(data) == dict and type(next(iter(data.values()))) == torch.Tensor: + assert len(data.keys()) == 1, 'embedding file has multiple terms in it' + + emb = next(iter(data.values())) + if len(emb.shape) == 1: + emb = emb.unsqueeze(0) + else: + raise Exception(f"Couldn't identify {filename} as neither textual inversion embedding nor diffuser concept.") + + vec = emb.detach().to(devices.device, dtype=torch.float32) + embedding = Embedding(vec, name) + embedding.step = data.get('step', None) + embedding.sd_checkpoint = data.get('sd_checkpoint', None) + embedding.sd_checkpoint_name = data.get('sd_checkpoint_name', None) + embedding.vectors = vec.shape[0] + embedding.shape = vec.shape[-1] + + if self.expected_shape == -1 or self.expected_shape == embedding.shape: + self.register_embedding(embedding, shared.sd_model) + else: + self.skipped_embeddings[name] = embedding - for root, dirs, fns in os.walk(self.embeddings_dir): + def load_from_dir(self, embdir): + if not os.path.isdir(embdir.path): + return + + for root, dirs, fns in os.walk(embdir.path): for fn in fns: try: fullfn = os.path.join(root, fn) @@ -163,12 +180,32 @@ class EmbeddingDatabase: if os.stat(fullfn).st_size == 0: continue - process_file(fullfn, fn) + self.load_from_file(fullfn, fn) except Exception: print(f"Error loading embedding {fn}:", file=sys.stderr) print(traceback.format_exc(), file=sys.stderr) continue + def load_textual_inversion_embeddings(self, force_reload=False): + if not force_reload: + need_reload = False + for path, embdir in self.embedding_dirs.items(): + if embdir.has_changed(): + need_reload = True + break + + if not need_reload: + return + + self.ids_lookup.clear() + self.word_embeddings.clear() + self.skipped_embeddings.clear() + self.expected_shape = self.get_expected_shape() + + for path, embdir in self.embedding_dirs.items(): + self.load_from_dir(embdir) + embdir.update() + print(f"Textual inversion embeddings loaded({len(self.word_embeddings)}): {', '.join(self.word_embeddings.keys())}") if len(self.skipped_embeddings) > 0: print(f"Textual inversion embeddings skipped({len(self.skipped_embeddings)}): {', '.join(self.skipped_embeddings.keys())}") @@ -251,14 +288,15 @@ def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, dat assert os.path.isfile(template_file), "Prompt template file doesn't exist" assert steps, "Max steps is empty or 0" assert isinstance(steps, int), "Max steps must be integer" - assert steps > 0 , "Max steps must be positive" + assert steps > 0, "Max steps must be positive" assert isinstance(save_model_every, int), "Save {name} must be integer" - assert save_model_every >= 0 , "Save {name} must be positive or 0" + assert save_model_every >= 0, "Save {name} must be positive or 0" assert isinstance(create_image_every, int), "Create image must be integer" - assert create_image_every >= 0 , "Create image must be positive or 0" + assert create_image_every >= 0, "Create image must be positive or 0" if save_model_every or create_image_every: assert log_directory, "Log directory is empty" + def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): save_embedding_every = save_embedding_every or 0 create_image_every = create_image_every or 0 -- cgit v1.2.3 From 6d0cc1e239e0a43a2e6d696eae20c66fad0819bb Mon Sep 17 00:00:00 2001 From: noodleanon <122053346+noodleanon@users.noreply.github.com> Date: Sun, 8 Jan 2023 11:03:48 +0000 Subject: Corrected is_img2img param --- modules/api/api.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 0e8ea263..1785a6b4 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -151,7 +151,7 @@ class Api: def text2imgapi(self, txt2imgreq: StableDiffusionTxt2ImgProcessingAPI): if txt2imgreq.script_name is not None: if scripts.scripts_txt2img.scripts == []: - scripts.scripts_txt2img.initialize_scripts(True) + scripts.scripts_txt2img.initialize_scripts(False) ui.create_ui() script_idx = script_name_to_index(txt2imgreq.script_name, scripts.scripts_txt2img.selectable_scripts) -- cgit v1.2.3 From 137ce534b2355a527cd1a50c192909161258b442 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 8 Jan 2023 16:14:38 +0300 Subject: remove some code duplication remove calls to locals() add a test for img2img with script --- modules/api/api.py | 33 ++++++++++++++++----------------- 1 file changed, 16 insertions(+), 17 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 1785a6b4..5b6125f8 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -148,14 +148,20 @@ class Api: raise HTTPException(status_code=401, detail="Incorrect username or password", headers={"WWW-Authenticate": "Basic"}) - def text2imgapi(self, txt2imgreq: StableDiffusionTxt2ImgProcessingAPI): - if txt2imgreq.script_name is not None: - if scripts.scripts_txt2img.scripts == []: - scripts.scripts_txt2img.initialize_scripts(False) - ui.create_ui() + def get_script(self, script_name, script_runner): + if script_name is None: + return None, None + + if not script_runner.scripts: + script_runner.initialize_scripts(False) + ui.create_ui() + + script_idx = script_name_to_index(script_name, script_runner.selectable_scripts) + script = script_runner.selectable_scripts[script_idx] + return script, script_idx - script_idx = script_name_to_index(txt2imgreq.script_name, scripts.scripts_txt2img.selectable_scripts) - script = scripts.scripts_txt2img.selectable_scripts[script_idx] + def text2imgapi(self, txt2imgreq: StableDiffusionTxt2ImgProcessingAPI): + script, script_idx = self.get_script(txt2imgreq.script_name, scripts.scripts_txt2img) populate = txt2imgreq.copy(update={ # Override __init__ params "sampler_name": validate_sampler_name(txt2imgreq.sampler_name or txt2imgreq.sampler_index), @@ -173,7 +179,7 @@ class Api: p = StableDiffusionProcessingTxt2Img(sd_model=shared.sd_model, **args) shared.state.begin() - if 'script' in locals(): + if script is not None: p.outpath_grids = opts.outdir_txt2img_grids p.outpath_samples = opts.outdir_txt2img_samples p.script_args = [script_idx + 1] + [None] * (script.args_from - 1) + p.script_args @@ -182,7 +188,6 @@ class Api: processed = process_images(p) shared.state.end() - b64images = list(map(encode_pil_to_base64, processed.images)) return TextToImageResponse(images=b64images, parameters=vars(txt2imgreq), info=processed.js()) @@ -192,13 +197,7 @@ class Api: if init_images is None: raise HTTPException(status_code=404, detail="Init image not found") - if img2imgreq.script_name is not None: - if scripts.scripts_img2img.scripts == []: - scripts.scripts_img2img.initialize_scripts(True) - ui.create_ui() - - script_idx = script_name_to_index(img2imgreq.script_name, scripts.scripts_img2img.selectable_scripts) - script = scripts.scripts_img2img.selectable_scripts[script_idx] + script, script_idx = self.get_script(img2imgreq.script_name, scripts.scripts_img2img) mask = img2imgreq.mask if mask: @@ -223,7 +222,7 @@ class Api: p.init_images = [decode_base64_to_image(x) for x in init_images] shared.state.begin() - if 'script' in locals(): + if script is not None: p.outpath_grids = opts.outdir_img2img_grids p.outpath_samples = opts.outdir_img2img_samples p.script_args = [script_idx + 1] + [None] * (script.args_from - 1) + p.script_args -- cgit v1.2.3 From cb255faec6e5f6b47b7632e6b7d450b9e2f6678b Mon Sep 17 00:00:00 2001 From: Lee Bousfield Date: Sun, 8 Jan 2023 10:17:50 -0700 Subject: Add support for loading VAEs from safetensor files --- modules/sd_vae.py | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/sd_vae.py b/modules/sd_vae.py index ac71d62d..9fcfd9db 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -1,4 +1,5 @@ import torch +import safetensors.torch import os import collections from collections import namedtuple @@ -72,8 +73,10 @@ def refresh_vae_list(vae_path=vae_path, model_path=model_path): candidates = [ *glob.iglob(os.path.join(model_path, '**/*.vae.ckpt'), recursive=True), *glob.iglob(os.path.join(model_path, '**/*.vae.pt'), recursive=True), + *glob.iglob(os.path.join(model_path, '**/*.vae.safetensors'), recursive=True), *glob.iglob(os.path.join(vae_path, '**/*.ckpt'), recursive=True), - *glob.iglob(os.path.join(vae_path, '**/*.pt'), recursive=True) + *glob.iglob(os.path.join(vae_path, '**/*.pt'), recursive=True), + *glob.iglob(os.path.join(vae_path, '**/*.safetensors'), recursive=True), ] if shared.cmd_opts.vae_path is not None and os.path.isfile(shared.cmd_opts.vae_path): candidates.append(shared.cmd_opts.vae_path) @@ -137,6 +140,12 @@ def resolve_vae(checkpoint_file=None, vae_file="auto"): if os.path.isfile(vae_file_try): vae_file = vae_file_try print(f"Using VAE found similar to selected model: {vae_file}") + # if still not found, try look for ".vae.safetensors" beside model + if vae_file == "auto": + vae_file_try = model_path + ".vae.safetensors" + if os.path.isfile(vae_file_try): + vae_file = vae_file_try + print(f"Using VAE found similar to selected model: {vae_file}") # No more fallbacks for auto if vae_file == "auto": vae_file = None @@ -163,8 +172,14 @@ def load_vae(model, vae_file=None): assert os.path.isfile(vae_file), f"VAE file doesn't exist: {vae_file}" print(f"Loading VAE weights from: {vae_file}") store_base_vae(model) - vae_ckpt = torch.load(vae_file, map_location=shared.weight_load_location) - vae_dict_1 = {k: v for k, v in vae_ckpt["state_dict"].items() if k[0:4] != "loss" and k not in vae_ignore_keys} + _, extension = os.path.splitext(vae_file) + if extension.lower() == ".safetensors": + vae_ckpt = safetensors.torch.load_file(vae_file, device=shared.weight_load_location) + else: + vae_ckpt = torch.load(vae_file, map_location=shared.weight_load_location) + if "state_dict" in vae_ckpt: + vae_ckpt = vae_ckpt["state_dict"] + vae_dict_1 = {k: v for k, v in vae_ckpt.items() if k[0:4] != "loss" and k not in vae_ignore_keys} _load_vae_dict(model, vae_dict_1) if cache_enabled: -- cgit v1.2.3 From d4fd2418efb0986a8226add0b800fb5c73ffb58c Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 9 Jan 2023 14:57:47 +0300 Subject: add an option to use old hiresfix width/height behavior add a visual effect to inactive hires fix elements --- modules/generation_parameters_copypaste.py | 17 +++++++++++------ modules/processing.py | 26 ++++++++++++++++++++++++-- modules/shared.py | 1 + modules/ui.py | 23 ++++++++++++++--------- 4 files changed, 50 insertions(+), 17 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 12a9de3d..f7f68b67 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -197,6 +197,15 @@ def restore_old_hires_fix_params(res): firstpass_width = res.get('First pass size-1', None) firstpass_height = res.get('First pass size-2', None) + if shared.opts.use_old_hires_fix_width_height: + hires_width = int(res.get("Hires resize-1", None)) + hires_height = int(res.get("Hires resize-2", None)) + + if hires_width is not None and hires_height is not None: + res['Size-1'] = hires_width + res['Size-2'] = hires_height + return + if firstpass_width is None or firstpass_height is None: return @@ -205,12 +214,8 @@ def restore_old_hires_fix_params(res): height = int(res.get("Size-2", 512)) if firstpass_width == 0 or firstpass_height == 0: - # old algorithm for auto-calculating first pass size - desired_pixel_count = 512 * 512 - actual_pixel_count = width * height - scale = math.sqrt(desired_pixel_count / actual_pixel_count) - firstpass_width = math.ceil(scale * width / 64) * 64 - firstpass_height = math.ceil(scale * height / 64) * 64 + from modules import processing + firstpass_width, firstpass_height = processing.old_hires_fix_first_pass_dimensions(width, height) res['Size-1'] = firstpass_width res['Size-2'] = firstpass_height diff --git a/modules/processing.py b/modules/processing.py index 1d23b15f..f04a0e1e 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -687,6 +687,18 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: return res +def old_hires_fix_first_pass_dimensions(width, height): + """old algorithm for auto-calculating first pass size""" + + desired_pixel_count = 512 * 512 + actual_pixel_count = width * height + scale = math.sqrt(desired_pixel_count / actual_pixel_count) + width = math.ceil(scale * width / 64) * 64 + height = math.ceil(scale * height / 64) * 64 + + return width, height + + class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): sampler = None @@ -703,16 +715,26 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): self.hr_upscale_to_y = hr_resize_y if firstphase_width != 0 or firstphase_height != 0: - print("firstphase_width/firstphase_height no longer supported; use hr_scale", file=sys.stderr) - self.hr_scale = self.width / firstphase_width + self.hr_upscale_to_x = self.width + self.hr_upscale_to_y = self.height self.width = firstphase_width self.height = firstphase_height self.truncate_x = 0 self.truncate_y = 0 + self.applied_old_hires_behavior_to = None def init(self, all_prompts, all_seeds, all_subseeds): if self.enable_hr: + if opts.use_old_hires_fix_width_height and self.applied_old_hires_behavior_to != (self.width, self.height): + self.hr_resize_x = self.width + self.hr_resize_y = self.height + self.hr_upscale_to_x = self.width + self.hr_upscale_to_y = self.height + + self.width, self.height = old_hires_fix_first_pass_dimensions(self.width, self.height) + self.applied_old_hires_behavior_to = (self.width, self.height) + if self.hr_resize_x == 0 and self.hr_resize_y == 0: self.extra_generation_params["Hires upscale"] = self.hr_scale self.hr_upscale_to_x = int(self.width * self.hr_scale) diff --git a/modules/shared.py b/modules/shared.py index a6712dae..a1e10201 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -398,6 +398,7 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { options_templates.update(options_section(('compatibility', "Compatibility"), { "use_old_emphasis_implementation": OptionInfo(False, "Use old emphasis implementation. Can be useful to reproduce old seeds."), "use_old_karras_scheduler_sigmas": OptionInfo(False, "Use old karras scheduler sigmas (0.1 to 10)."), + "use_old_hires_fix_width_height": OptionInfo(False, "For hires fix, use width/height sliders to set final resolution rather than first pass (disables Upscale by, Resize width/height to)."), })) options_templates.update(options_section(('interrogate', "Interrogate Options"), { diff --git a/modules/ui.py b/modules/ui.py index 99483130..719c26b3 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -267,7 +267,7 @@ def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resiz with devices.autocast(): p.init([""], [0], [0]) - return f"resize: from {width}x{height} to {p.hr_upscale_to_x}x{p.hr_upscale_to_y}" + return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" def apply_styles(prompt, prompt_neg, style1_name, style2_name): @@ -745,15 +745,20 @@ def create_ui(): custom_inputs = modules.scripts.scripts_txt2img.setup_ui() hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] - hr_resolution_preview_args = dict( - fn=calc_resolution_hires, - inputs=hr_resolution_preview_inputs, - outputs=[hr_final_resolution], - show_progress=False - ) - for input in hr_resolution_preview_inputs: - input.change(**hr_resolution_preview_args) + input.change( + fn=calc_resolution_hires, + inputs=hr_resolution_preview_inputs, + outputs=[hr_final_resolution], + show_progress=False, + ) + input.change( + None, + _js="onCalcResolutionHires", + inputs=hr_resolution_preview_inputs, + outputs=[], + show_progress=False, + ) txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) -- cgit v1.2.3 From 49c4509ce2302350210ff650fd26373518c46a79 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 9 Jan 2023 19:58:35 +0300 Subject: use existing function for loading VAE weights from file --- modules/sd_vae.py | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/sd_vae.py b/modules/sd_vae.py index 9fcfd9db..0a49daa1 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -3,7 +3,7 @@ import safetensors.torch import os import collections from collections import namedtuple -from modules import shared, devices, script_callbacks +from modules import shared, devices, script_callbacks, sd_models from modules.paths import models_path import glob from copy import deepcopy @@ -172,13 +172,8 @@ def load_vae(model, vae_file=None): assert os.path.isfile(vae_file), f"VAE file doesn't exist: {vae_file}" print(f"Loading VAE weights from: {vae_file}") store_base_vae(model) - _, extension = os.path.splitext(vae_file) - if extension.lower() == ".safetensors": - vae_ckpt = safetensors.torch.load_file(vae_file, device=shared.weight_load_location) - else: - vae_ckpt = torch.load(vae_file, map_location=shared.weight_load_location) - if "state_dict" in vae_ckpt: - vae_ckpt = vae_ckpt["state_dict"] + + vae_ckpt = sd_models.read_state_dict(vae_file, map_location=shared.weight_load_location) vae_dict_1 = {k: v for k, v in vae_ckpt.items() if k[0:4] != "loss" and k not in vae_ignore_keys} _load_vae_dict(model, vae_dict_1) @@ -210,10 +205,12 @@ def _load_vae_dict(model, vae_dict_1): model.first_stage_model.load_state_dict(vae_dict_1) model.first_stage_model.to(devices.dtype_vae) + def clear_loaded_vae(): global loaded_vae_file loaded_vae_file = None + def reload_vae_weights(sd_model=None, vae_file="auto"): from modules import lowvram, devices, sd_hijack -- cgit v1.2.3 From cdfcbd995932ffa728db0cc00a5f97665c752103 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 9 Jan 2023 20:08:48 +0300 Subject: Remove fallback for Protocol import and remove Protocol import and remove instances of Protocol in code add some whitespace between functions to be in line with other code in the repo --- modules/sub_quadratic_attention.py | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/sub_quadratic_attention.py b/modules/sub_quadratic_attention.py index 93381bae..55052815 100644 --- a/modules/sub_quadratic_attention.py +++ b/modules/sub_quadratic_attention.py @@ -15,14 +15,9 @@ import torch from torch import Tensor from torch.utils.checkpoint import checkpoint import math - -try: - from typing import Protocol -except: - from typing_extensions import Protocol - from typing import Optional, NamedTuple, List + def narrow_trunc( input: Tensor, dim: int, @@ -31,12 +26,14 @@ def narrow_trunc( ) -> Tensor: return torch.narrow(input, dim, start, length if input.shape[dim] >= start + length else input.shape[dim] - start) + class AttnChunk(NamedTuple): exp_values: Tensor exp_weights_sum: Tensor max_score: Tensor -class SummarizeChunk(Protocol): + +class SummarizeChunk: @staticmethod def __call__( query: Tensor, @@ -44,7 +41,8 @@ class SummarizeChunk(Protocol): value: Tensor, ) -> AttnChunk: ... -class ComputeQueryChunkAttn(Protocol): + +class ComputeQueryChunkAttn: @staticmethod def __call__( query: Tensor, @@ -52,6 +50,7 @@ class ComputeQueryChunkAttn(Protocol): value: Tensor, ) -> Tensor: ... + def _summarize_chunk( query: Tensor, key: Tensor, @@ -72,6 +71,7 @@ def _summarize_chunk( max_score = max_score.squeeze(-1) return AttnChunk(exp_values, exp_weights.sum(dim=-1), max_score) + def _query_chunk_attention( query: Tensor, key: Tensor, @@ -112,6 +112,7 @@ def _query_chunk_attention( all_weights = torch.unsqueeze(chunk_weights, -1).sum(dim=0) return all_values / all_weights + # TODO: refactor CrossAttention#get_attention_scores to share code with this def _get_attention_scores_no_kv_chunking( query: Tensor, @@ -131,10 +132,12 @@ def _get_attention_scores_no_kv_chunking( hidden_states_slice = torch.bmm(attn_probs, value) return hidden_states_slice + class ScannedChunk(NamedTuple): chunk_idx: int attn_chunk: AttnChunk + def efficient_dot_product_attention( query: Tensor, key: Tensor, -- cgit v1.2.3 From 43bb5190fc9e7ae479a5dc6640be202c9a71e464 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 9 Jan 2023 22:52:23 +0300 Subject: remove/simplify some changes from #6481 --- modules/textual_inversion/dataset.py | 14 +++++--------- modules/textual_inversion/textual_inversion.py | 4 ++-- modules/ui.py | 2 +- 3 files changed, 8 insertions(+), 12 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py index bcad6848..fa48708e 100644 --- a/modules/textual_inversion/dataset.py +++ b/modules/textual_inversion/dataset.py @@ -17,7 +17,7 @@ re_numbers_at_start = re.compile(r"^[-\d]+\s*") class DatasetEntry: - def __init__(self, filename=None, filename_text=None, latent_dist=None, latent_sample=None, cond=None, cond_text=None, pixel_values=None, img_shape=None): + def __init__(self, filename=None, filename_text=None, latent_dist=None, latent_sample=None, cond=None, cond_text=None, pixel_values=None): self.filename = filename self.filename_text = filename_text self.latent_dist = latent_dist @@ -25,7 +25,6 @@ class DatasetEntry: self.cond = cond self.cond_text = cond_text self.pixel_values = pixel_values - self.img_shape = img_shape class PersonalizedBase(Dataset): @@ -46,12 +45,10 @@ class PersonalizedBase(Dataset): assert data_root, 'dataset directory not specified' assert os.path.isdir(data_root), "Dataset directory doesn't exist" assert os.listdir(data_root), "Dataset directory is empty" - if varsize: - assert batch_size == 1, 'variable img size must have batch size 1' + assert batch_size == 1 or not varsize, 'variable img size must have batch size 1' self.image_paths = [os.path.join(data_root, file_path) for file_path in os.listdir(data_root)] - self.shuffle_tags = shuffle_tags self.tag_drop_out = tag_drop_out @@ -91,14 +88,14 @@ class PersonalizedBase(Dataset): if latent_sampling_method == "once" or (latent_sampling_method == "deterministic" and not isinstance(latent_dist, DiagonalGaussianDistribution)): latent_sample = model.get_first_stage_encoding(latent_dist).squeeze().to(devices.cpu) latent_sampling_method = "once" - entry = DatasetEntry(filename=path, filename_text=filename_text, latent_sample=latent_sample, img_shape=image.size) + entry = DatasetEntry(filename=path, filename_text=filename_text, latent_sample=latent_sample) elif latent_sampling_method == "deterministic": # Works only for DiagonalGaussianDistribution latent_dist.std = 0 latent_sample = model.get_first_stage_encoding(latent_dist).squeeze().to(devices.cpu) - entry = DatasetEntry(filename=path, filename_text=filename_text, latent_sample=latent_sample, img_shape=image.size) + entry = DatasetEntry(filename=path, filename_text=filename_text, latent_sample=latent_sample) elif latent_sampling_method == "random": - entry = DatasetEntry(filename=path, filename_text=filename_text, latent_dist=latent_dist, img_shape=image.size) + entry = DatasetEntry(filename=path, filename_text=filename_text, latent_dist=latent_dist) if not (self.tag_drop_out != 0 or self.shuffle_tags): entry.cond_text = self.create_text(filename_text) @@ -154,7 +151,6 @@ class BatchLoader: self.cond_text = [entry.cond_text for entry in data] self.cond = [entry.cond for entry in data] self.latent_sample = torch.stack([entry.latent_sample for entry in data]).squeeze(1) - self.img_shape = [entry.img_shape for entry in data] #self.emb_index = [entry.emb_index for entry in data] #print(self.latent_sample.device) diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index ad76297e..14be2c96 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -492,8 +492,8 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ else: p.prompt = batch.cond_text[0] p.steps = 20 - p.width = batch.img_shape[0][0] - p.height = batch.img_shape[0][1] + p.width = training_width + p.height = training_height preview_text = p.prompt diff --git a/modules/ui.py b/modules/ui.py index 9d6b141e..ddfe1b1a 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1348,7 +1348,7 @@ def create_ui(): template_file = gr.Textbox(label='Prompt template file', value=os.path.join(script_path, "textual_inversion_templates", "style_filewords.txt"), elem_id="train_template_file") training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") - varsize = gr.Checkbox(label="Ignore dimension settings and do not resize images", value=False, elem_id="train_varsize") + varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") with FormRow(): -- cgit v1.2.3 From 1fbb6f9ebe48326a3b12ecf611105dbc4a46891e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 9 Jan 2023 23:35:40 +0300 Subject: make a dropdown for prompt template selection --- modules/hypernetworks/hypernetwork.py | 7 ++++-- modules/shared.py | 1 + modules/textual_inversion/textual_inversion.py | 35 ++++++++++++++++++++------ modules/ui.py | 11 ++++++-- 4 files changed, 42 insertions(+), 12 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 32c67ccc..ea3f1db9 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -24,6 +24,7 @@ from statistics import stdev, mean optimizer_dict = {optim_name : cls_obj for optim_name, cls_obj in inspect.getmembers(torch.optim, inspect.isclass) if optim_name != "Optimizer"} + class HypernetworkModule(torch.nn.Module): multiplier = 1.0 activation_dict = { @@ -403,13 +404,15 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, shared.reload_hypernetworks() -def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): +def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_filename, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): # images allows training previews to have infotext. Importing it at the top causes a circular import problem. from modules import images save_hypernetwork_every = save_hypernetwork_every or 0 create_image_every = create_image_every or 0 - textual_inversion.validate_train_inputs(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, template_file, steps, save_hypernetwork_every, create_image_every, log_directory, name="hypernetwork") + template_file = textual_inversion.textual_inversion_templates.get(template_filename, None) + textual_inversion.validate_train_inputs(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, template_file, template_filename, steps, save_hypernetwork_every, create_image_every, log_directory, name="hypernetwork") + template_file = template_file.path path = shared.hypernetworks.get(hypernetwork_name, None) shared.loaded_hypernetwork = Hypernetwork() diff --git a/modules/shared.py b/modules/shared.py index a1e10201..aa37c8ce 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -33,6 +33,7 @@ parser.add_argument("--no-half-vae", action='store_true', help="do not switch th parser.add_argument("--no-progressbar-hiding", action='store_true', help="do not hide progressbar in gradio UI (we hide it because it slows down ML if you have hardware acceleration in browser)") parser.add_argument("--max-batch-count", type=int, default=16, help="maximum batch count value for the UI") parser.add_argument("--embeddings-dir", type=str, default=os.path.join(script_path, 'embeddings'), help="embeddings directory for textual inversion (default: embeddings)") +parser.add_argument("--textual-inversion-templates-dir", type=str, default=os.path.join(script_path, 'textual_inversion_templates'), help="directory with textual inversion templates") parser.add_argument("--hypernetwork-dir", type=str, default=os.path.join(models_path, 'hypernetworks'), help="hypernetwork directory") parser.add_argument("--localizations-dir", type=str, default=os.path.join(script_path, 'localizations'), help="localizations directory") parser.add_argument("--allow-code", action='store_true', help="allow custom script execution from webui") diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 14be2c96..5420903f 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -2,6 +2,7 @@ import os import sys import traceback import inspect +from collections import namedtuple import torch import tqdm @@ -15,12 +16,26 @@ from modules import shared, devices, sd_hijack, processing, sd_models, images, s import modules.textual_inversion.dataset from modules.textual_inversion.learn_schedule import LearnRateScheduler -from modules.textual_inversion.image_embedding import (embedding_to_b64, embedding_from_b64, - insert_image_data_embed, extract_image_data_embed, - caption_image_overlay) +from modules.textual_inversion.image_embedding import embedding_to_b64, embedding_from_b64, insert_image_data_embed, extract_image_data_embed, caption_image_overlay from modules.textual_inversion.logging import save_settings_to_file +TextualInversionTemplate = namedtuple("TextualInversionTemplate", ["name", "path"]) +textual_inversion_templates = {} + + +def list_textual_inversion_templates(): + textual_inversion_templates.clear() + + for root, dirs, fns in os.walk(shared.cmd_opts.textual_inversion_templates_dir): + for fn in fns: + path = os.path.join(root, fn) + + textual_inversion_templates[fn] = TextualInversionTemplate(fn, path) + + return textual_inversion_templates + + class Embedding: def __init__(self, vec, name, step=None): self.vec = vec @@ -274,7 +289,7 @@ def write_loss(log_directory, filename, step, epoch_len, values): }) -def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, data_root, template_file, steps, save_model_every, create_image_every, log_directory, name="embedding"): +def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, data_root, template_file, template_filename, steps, save_model_every, create_image_every, log_directory, name="embedding"): assert model_name, f"{name} not selected" assert learn_rate, "Learning rate is empty or 0" assert isinstance(batch_size, int), "Batch size must be integer" @@ -284,8 +299,9 @@ def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, dat assert data_root, "Dataset directory is empty" assert os.path.isdir(data_root), "Dataset directory doesn't exist" assert os.listdir(data_root), "Dataset directory is empty" - assert template_file, "Prompt template file is empty" - assert os.path.isfile(template_file), "Prompt template file doesn't exist" + assert template_filename, "Prompt template file not selected" + assert template_file, f"Prompt template file {template_filename} not found" + assert os.path.isfile(template_file.path), f"Prompt template file {template_filename} doesn't exist" assert steps, "Max steps is empty or 0" assert isinstance(steps, int), "Max steps must be integer" assert steps > 0, "Max steps must be positive" @@ -296,10 +312,13 @@ def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, dat if save_model_every or create_image_every: assert log_directory, "Log directory is empty" -def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): + +def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_embedding_every, template_filename, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): save_embedding_every = save_embedding_every or 0 create_image_every = create_image_every or 0 - validate_train_inputs(embedding_name, learn_rate, batch_size, gradient_step, data_root, template_file, steps, save_embedding_every, create_image_every, log_directory, name="embedding") + template_file = textual_inversion_templates.get(template_filename, None) + validate_train_inputs(embedding_name, learn_rate, batch_size, gradient_step, data_root, template_file, template_filename, steps, save_embedding_every, create_image_every, log_directory, name="embedding") + template_file = template_file.path shared.state.job = "train-embedding" shared.state.textinfo = "Initializing textual inversion training..." diff --git a/modules/ui.py b/modules/ui.py index ddfe1b1a..b6079aec 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -37,7 +37,7 @@ from modules import prompt_parser from modules.images import save_image from modules.sd_hijack import model_hijack from modules.sd_samplers import samplers, samplers_for_img2img -import modules.textual_inversion.ui +from modules.textual_inversion import textual_inversion import modules.hypernetworks.ui from modules.generation_parameters_copypaste import image_from_url_text @@ -1322,6 +1322,9 @@ def create_ui(): outputs=[process_focal_crop_row], ) + def get_textual_inversion_template_names(): + return sorted([x for x in textual_inversion.textual_inversion_templates]) + with gr.Tab(label="Train"): gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") with FormRow(): @@ -1345,7 +1348,11 @@ def create_ui(): dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") - template_file = gr.Textbox(label='Prompt template file', value=os.path.join(script_path, "textual_inversion_templates", "style_filewords.txt"), elem_id="train_template_file") + + with FormRow(): + template_file = gr.Dropdown(label='Prompt template', value="style_filewords.txt", elem_id="train_template_file", choices=get_textual_inversion_template_names()) + create_refresh_button(template_file, textual_inversion.list_textual_inversion_templates, lambda: {"choices": get_textual_inversion_template_names()}, "refrsh_train_template_file") + training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") -- cgit v1.2.3 From 95727312ca5913876aa1c74f47d1ff6d93bb6b1f Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Mon, 9 Jan 2023 16:54:12 -0500 Subject: remove bytes -> gb conversion --- modules/api/api.py | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index d2222b18..1c121ff0 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -467,26 +467,24 @@ class Api: return TrainResponse(info = "train embedding error: {error}".format(error = error)) def get_memory(self): - def gb(val: float): - return round(val / 1024 / 1024 / 1024, 2) try: import os, psutil process = psutil.Process(os.getpid()) - res = process.memory_info() - ram_total = 100 * res.rss / process.memory_percent() - ram = { 'free': gb(ram_total - res.rss), 'used': gb(res.rss), 'total': gb(ram_total) } + res = process.memory_info() # only rss is cross-platform guaranteed so we dont rely on other values + ram_total = 100 * res.rss / process.memory_percent() # and total memory is calculated as actual value is not cross-platform safe + ram = { 'free': ram_total - res.rss, 'used': res.rss, 'total': ram_total } except Exception as err: ram = { 'error': f'{err}' } try: import torch if torch.cuda.is_available(): s = torch.cuda.mem_get_info() - system = { 'free': gb(s[0]), 'used': gb(s[1] - s[0]), 'total': gb(s[1]) } + system = { 'free': s[0], 'used': s[1] - s[0], 'total': s[1] } s = dict(torch.cuda.memory_stats(shared.device)) - allocated = { 'current': gb(s['allocated_bytes.all.current']), 'peak': gb(s['allocated_bytes.all.peak']) } - reserved = { 'current': gb(s['reserved_bytes.all.current']), 'peak': gb(s['reserved_bytes.all.peak']) } - active = { 'current': gb(s['active_bytes.all.current']), 'peak': gb(s['active_bytes.all.peak']) } - inactive = { 'current': gb(s['inactive_split_bytes.all.current']), 'peak': gb(s['inactive_split_bytes.all.peak']) } + allocated = { 'current': s['allocated_bytes.all.current'], 'peak': s['allocated_bytes.all.peak'] } + reserved = { 'current': s['reserved_bytes.all.current'], 'peak': s['reserved_bytes.all.peak'] } + active = { 'current': s['active_bytes.all.current'], 'peak': s['active_bytes.all.peak'] } + inactive = { 'current': s['inactive_split_bytes.all.current'], 'peak': s['inactive_split_bytes.all.peak'] } warnings = { 'retries': s['num_alloc_retries'], 'oom': s['num_ooms'] } cuda = { 'system': system, -- cgit v1.2.3 From 3fe9e9e54dcfc41d7c5ee6976f83b0de29fd3dda Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 10 Jan 2023 02:17:33 +0300 Subject: fix broken resolution detection when pasting parameters with old hires fix enabled --- modules/generation_parameters_copypaste.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index f7f68b67..620aa606 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -198,10 +198,10 @@ def restore_old_hires_fix_params(res): firstpass_height = res.get('First pass size-2', None) if shared.opts.use_old_hires_fix_width_height: - hires_width = int(res.get("Hires resize-1", None)) - hires_height = int(res.get("Hires resize-2", None)) + hires_width = int(res.get("Hires resize-1", 0)) + hires_height = int(res.get("Hires resize-2", 0)) - if hires_width is not None and hires_height is not None: + if hires_width and hires_height: res['Size-1'] = hires_width res['Size-2'] = hires_height return -- cgit v1.2.3 From 552d7b90bf483c160cd20740f7acd7fccbc02e6f Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Mon, 9 Jan 2023 18:34:26 -0500 Subject: allow model load if previous model failed --- modules/sd_models.py | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index 76a89e88..0a6d55ca 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -49,6 +49,9 @@ def checkpoint_tiles(): def find_checkpoint_config(info): + if info is None: + return shared.cmd_opts.config + config = os.path.splitext(info.filename)[0] + ".yaml" if os.path.exists(config): return config @@ -345,14 +348,16 @@ def reload_model_weights(sd_model=None, info=None): if not sd_model: sd_model = shared.sd_model + if sd_model is None: # previous model load failed + current_checkpoint_info = None + else: + current_checkpoint_info = sd_model.sd_checkpoint_info + if sd_model.sd_model_checkpoint == checkpoint_info.filename: + return - current_checkpoint_info = sd_model.sd_checkpoint_info checkpoint_config = find_checkpoint_config(current_checkpoint_info) - if sd_model.sd_model_checkpoint == checkpoint_info.filename: - return - - if checkpoint_config != find_checkpoint_config(checkpoint_info) or should_hijack_inpainting(checkpoint_info) != should_hijack_inpainting(sd_model.sd_checkpoint_info): + if current_checkpoint_info is None or checkpoint_config != find_checkpoint_config(checkpoint_info) or should_hijack_inpainting(checkpoint_info) != should_hijack_inpainting(sd_model.sd_checkpoint_info): del sd_model checkpoints_loaded.clear() load_model(checkpoint_info) -- cgit v1.2.3 From 2275f130bfe437c3245a66559f92af94d0e4d8ff Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Mon, 9 Jan 2023 21:23:58 -0500 Subject: relax reponse type check enforcement --- modules/api/models.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/api/models.py b/modules/api/models.py index 880edde6..034b4aa0 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -262,5 +262,5 @@ class EmbeddingsResponse(BaseModel): skipped: Dict[str, EmbeddingItem] = Field(title="Skipped", description="Embeddings skipped for the current model (likely due to architecture incompatibility)") class MemoryResponse(BaseModel): - ram: dict[str, str] | dict[str, float] = Field(title="RAM", description="System memory stats") - cuda: dict[str, str] | dict[str, dict] = Field(title="CUDA", description="nVidia CUDA memory stats") + ram: dict = Field(title="RAM", description="System memory stats") + cuda: dict = Field(title="CUDA", description="nVidia CUDA memory stats") -- cgit v1.2.3 From a4a5475cfa3c68af6cb046081002a72f862ce4be Mon Sep 17 00:00:00 2001 From: aria1th <35677394+aria1th@users.noreply.github.com> Date: Tue, 10 Jan 2023 14:56:57 +0900 Subject: Variable dropout rate Implements variable dropout rate from #4549 Fixes hypernetwork multiplier being able to modified during training, also fixes user-errors by setting multiplier value to lower values for training. Changes function name to match torch.nn.module standard Fixes RNG reset issue when generating previews by restoring RNG state --- modules/hypernetworks/hypernetwork.py | 101 +++++++++++++++++++++++++--------- modules/hypernetworks/ui.py | 4 +- modules/ui.py | 4 +- 3 files changed, 81 insertions(+), 28 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index ea3f1db9..300d3975 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -39,7 +39,7 @@ class HypernetworkModule(torch.nn.Module): activation_dict.update({cls_name.lower(): cls_obj for cls_name, cls_obj in inspect.getmembers(torch.nn.modules.activation) if inspect.isclass(cls_obj) and cls_obj.__module__ == 'torch.nn.modules.activation'}) def __init__(self, dim, state_dict=None, layer_structure=None, activation_func=None, weight_init='Normal', - add_layer_norm=False, use_dropout=False, activate_output=False, last_layer_dropout=False): + add_layer_norm=False, activate_output=False, dropout_structure=None): super().__init__() assert layer_structure is not None, "layer_structure must not be None" @@ -64,9 +64,12 @@ class HypernetworkModule(torch.nn.Module): if add_layer_norm: linears.append(torch.nn.LayerNorm(int(dim * layer_structure[i+1]))) - # Add dropout except last layer - if use_dropout and (i < len(layer_structure) - 3 or last_layer_dropout and i < len(layer_structure) - 2): - linears.append(torch.nn.Dropout(p=0.3)) + # Everything should be now parsed into dropout structure, and applied here. + # Since we only have dropouts after layers, dropout structure should start with 0 and end with 0. + if dropout_structure is not None and dropout_structure[i+1] > 0: + assert 0 < dropout_structure[i+1] < 1, "Dropout probability should be 0 or float between 0 and 1!" + linears.append(torch.nn.Dropout(p=dropout_structure[i+1])) + # Code explanation : [1, 2, 1] -> dropout is missing when last_layer_dropout is false. [1, 2, 2, 1] -> [0, 0.3, 0, 0], when its True, [0, 0.3, 0.3, 0]. self.linear = torch.nn.Sequential(*linears) @@ -113,7 +116,7 @@ class HypernetworkModule(torch.nn.Module): state_dict[to] = x def forward(self, x): - return x + self.linear(x) * self.multiplier + return x + self.linear(x) * (HypernetworkModule.multiplier if not self.training else 1) def trainables(self): layer_structure = [] @@ -126,6 +129,21 @@ class HypernetworkModule(torch.nn.Module): def apply_strength(value=None): HypernetworkModule.multiplier = value if value is not None else shared.opts.sd_hypernetwork_strength +#param layer_structure : sequence used for length, use_dropout : controlling boolean, last_layer_dropout : for compatibility check. +def parse_dropout_structure(layer_structure, use_dropout, last_layer_dropout): + if layer_structure is None: + layer_structure = [1, 2, 1] + if not use_dropout: + return [0] * len(layer_structure) + dropout_values = [0] + dropout_values.extend([0.3] * (len(layer_structure) - 3)) + if last_layer_dropout: + dropout_values.append(0.3) + else: + dropout_values.append(0) + dropout_values.append(0) + return dropout_values + class Hypernetwork: filename = None @@ -144,18 +162,22 @@ class Hypernetwork: self.add_layer_norm = add_layer_norm self.use_dropout = use_dropout self.activate_output = activate_output - self.last_layer_dropout = kwargs['last_layer_dropout'] if 'last_layer_dropout' in kwargs else True + self.last_layer_dropout = kwargs.get('last_layer_dropout', True) + self.dropout_structure = kwargs.get('dropout_structure', None) + if self.dropout_structure is None: + self.dropout_structure = parse_dropout_structure(self.layer_structure, self.use_dropout, self.last_layer_dropout) self.optimizer_name = None self.optimizer_state_dict = None + self.optional_info = None for size in enable_sizes or []: self.layers[size] = ( HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.weight_init, - self.add_layer_norm, self.use_dropout, self.activate_output, last_layer_dropout=self.last_layer_dropout), + self.add_layer_norm, self.activate_output, dropout_structure=self.dropout_structure), HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.weight_init, - self.add_layer_norm, self.use_dropout, self.activate_output, last_layer_dropout=self.last_layer_dropout), + self.add_layer_norm, self.activate_output, dropout_structure=self.dropout_structure), ) - self.eval_mode() + self.eval() def weights(self): res = [] @@ -164,14 +186,14 @@ class Hypernetwork: res += layer.parameters() return res - def train_mode(self): + def train(self, mode=True): for k, layers in self.layers.items(): for layer in layers: - layer.train() + layer.train(mode=mode) for param in layer.parameters(): - param.requires_grad = True + param.requires_grad = mode - def eval_mode(self): + def eval(self): for k, layers in self.layers.items(): for layer in layers: layer.eval() @@ -191,11 +213,13 @@ class Hypernetwork: state_dict['activation_func'] = self.activation_func state_dict['is_layer_norm'] = self.add_layer_norm state_dict['weight_initialization'] = self.weight_init - state_dict['use_dropout'] = self.use_dropout state_dict['sd_checkpoint'] = self.sd_checkpoint state_dict['sd_checkpoint_name'] = self.sd_checkpoint_name state_dict['activate_output'] = self.activate_output - state_dict['last_layer_dropout'] = self.last_layer_dropout + state_dict['use_dropout'] = self.use_dropout + state_dict['dropout_structure'] = self.dropout_structure + state_dict['last_layer_dropout'] = (self.dropout_structure[-2] != 0) if self.dropout_structure is not None else self.last_layer_dropout + state_dict['optional_info'] = self.optional_info if self.optional_info else None if self.optimizer_name is not None: optimizer_saved_dict['optimizer_name'] = self.optimizer_name @@ -215,43 +239,56 @@ class Hypernetwork: self.layer_structure = state_dict.get('layer_structure', [1, 2, 1]) print(self.layer_structure) + optional_info = state_dict.get('optional_info', None) + if optional_info is not None: + print(f"INFO:\n {optional_info}\n") + self.optional_info = optional_info self.activation_func = state_dict.get('activation_func', None) print(f"Activation function is {self.activation_func}") self.weight_init = state_dict.get('weight_initialization', 'Normal') print(f"Weight initialization is {self.weight_init}") self.add_layer_norm = state_dict.get('is_layer_norm', False) print(f"Layer norm is set to {self.add_layer_norm}") - self.use_dropout = state_dict.get('use_dropout', False) + self.dropout_structure = state_dict.get('dropout_structure', None) + self.use_dropout = True if self.dropout_structure is not None and any(self.dropout_structure) else state_dict.get('use_dropout', False) print(f"Dropout usage is set to {self.use_dropout}" ) self.activate_output = state_dict.get('activate_output', True) print(f"Activate last layer is set to {self.activate_output}") self.last_layer_dropout = state_dict.get('last_layer_dropout', False) + # Dropout structure should have same length as layer structure, Every digits should be in [0,1), and last digit must be 0. + if self.dropout_structure is None: + print("Using previous dropout structure") + self.dropout_structure = parse_dropout_structure(self.layer_structure, self.use_dropout, self.last_layer_dropout) + print(f"Dropout structure is set to {self.dropout_structure}") optimizer_saved_dict = torch.load(self.filename + '.optim', map_location = 'cpu') if os.path.exists(self.filename + '.optim') else {} - self.optimizer_name = optimizer_saved_dict.get('optimizer_name', 'AdamW') - print(f"Optimizer name is {self.optimizer_name}") + if sd_models.model_hash(filename) == optimizer_saved_dict.get('hash', None): self.optimizer_state_dict = optimizer_saved_dict.get('optimizer_state_dict', None) else: self.optimizer_state_dict = None if self.optimizer_state_dict: + self.optimizer_name = optimizer_saved_dict.get('optimizer_name', 'AdamW') print("Loaded existing optimizer from checkpoint") + print(f"Optimizer name is {self.optimizer_name}") else: + self.optimizer_name = "AdamW" print("No saved optimizer exists in checkpoint") for size, sd in state_dict.items(): if type(size) == int: self.layers[size] = ( HypernetworkModule(size, sd[0], self.layer_structure, self.activation_func, self.weight_init, - self.add_layer_norm, self.use_dropout, self.activate_output, last_layer_dropout=self.last_layer_dropout), + self.add_layer_norm, self.activate_output, self.dropout_structure), HypernetworkModule(size, sd[1], self.layer_structure, self.activation_func, self.weight_init, - self.add_layer_norm, self.use_dropout, self.activate_output, last_layer_dropout=self.last_layer_dropout), + self.add_layer_norm, self.activate_output, self.dropout_structure), ) self.name = state_dict.get('name', self.name) self.step = state_dict.get('step', 0) self.sd_checkpoint = state_dict.get('sd_checkpoint', None) self.sd_checkpoint_name = state_dict.get('sd_checkpoint_name', None) + self.eval() def list_hypernetworks(path): @@ -379,9 +416,10 @@ def report_statistics(loss_info:dict): print(e) -def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False): +def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False, dropout_structure=None): # Remove illegal characters from name. name = "".join( x for x in name if (x.isalnum() or x in "._- ")) + assert name, "Name cannot be empty!" fn = os.path.join(shared.cmd_opts.hypernetwork_dir, f"{name}.pt") if not overwrite_old: @@ -390,6 +428,11 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, if type(layer_structure) == str: layer_structure = [float(x.strip()) for x in layer_structure.split(",")] + if use_dropout and dropout_structure and type(dropout_structure) == str: + dropout_structure = [float(x.strip()) for x in dropout_structure.split(",")] + else: + dropout_structure = [0] * len(layer_structure) + hypernet = modules.hypernetworks.hypernetwork.Hypernetwork( name=name, enable_sizes=[int(x) for x in enable_sizes], @@ -398,6 +441,7 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, weight_init=weight_init, add_layer_norm=add_layer_norm, use_dropout=use_dropout, + dropout_structure=dropout_structure ) hypernet.save(fn) @@ -480,7 +524,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, shared.sd_model.first_stage_model.to(devices.cpu) weights = hypernetwork.weights() - hypernetwork.train_mode() + hypernetwork.train() # Here we use optimizer from saved HN, or we can specify as UI option. if hypernetwork.optimizer_name in optimizer_dict: @@ -594,7 +638,11 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, if images_dir is not None and steps_done % create_image_every == 0: forced_filename = f'{hypernetwork_name}-{steps_done}' last_saved_image = os.path.join(images_dir, forced_filename) - hypernetwork.eval_mode() + hypernetwork.eval() + rng_state = torch.get_rng_state() + cuda_rng_state = None + if torch.cuda.is_available(): + cuda_rng_state = torch.cuda.get_rng_state_all() shared.sd_model.cond_stage_model.to(devices.device) shared.sd_model.first_stage_model.to(devices.device) @@ -627,7 +675,10 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, if unload: shared.sd_model.cond_stage_model.to(devices.cpu) shared.sd_model.first_stage_model.to(devices.cpu) - hypernetwork.train_mode() + torch.set_rng_state(rng_state) + if torch.cuda.is_available(): + torch.cuda.set_rng_state_all(cuda_rng_state) + hypernetwork.train() if image is not None: shared.state.current_image = image last_saved_image, last_text_info = images.save_image(image, images_dir, "", p.seed, p.prompt, shared.opts.samples_format, processed.infotexts[0], p=p, forced_filename=forced_filename, save_to_dirs=False) @@ -649,7 +700,7 @@ Last saved image: {html.escape(last_saved_image)}
finally: pbar.leave = False pbar.close() - hypernetwork.eval_mode() + hypernetwork.eval() #report_statistics(loss_dict) filename = os.path.join(shared.cmd_opts.hypernetwork_dir, f'{hypernetwork_name}.pt') diff --git a/modules/hypernetworks/ui.py b/modules/hypernetworks/ui.py index e7f9e593..81e3f519 100644 --- a/modules/hypernetworks/ui.py +++ b/modules/hypernetworks/ui.py @@ -9,8 +9,8 @@ from modules import devices, sd_hijack, shared not_available = ["hardswish", "multiheadattention"] keys = list(x for x in modules.hypernetworks.hypernetwork.HypernetworkModule.activation_dict.keys() if x not in not_available) -def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False): - filename = modules.hypernetworks.hypernetwork.create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure, activation_func, weight_init, add_layer_norm, use_dropout) +def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False, dropout_structure=None): + filename = modules.hypernetworks.hypernetwork.create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure, activation_func, weight_init, add_layer_norm, use_dropout, dropout_structure) return gr.Dropdown.update(choices=sorted([x for x in shared.hypernetworks.keys()])), f"Created: {filename}", "" diff --git a/modules/ui.py b/modules/ui.py index b6079aec..9b9081b5 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1268,6 +1268,7 @@ def create_ui(): new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") + new_hypernetwork_dropout_structure = gr.Textbox("0, 0, 0", label="Enter hypernetwork Dropout structure (or empty). Recommended : 0~0.35 incrementing sequence: 0, 0.05, 0.15", placeholder="1st and last digit must be 0 and values should be between 0 and 1. ex:'0, 0.01, 0'") overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") with gr.Row(): @@ -1414,7 +1415,8 @@ def create_ui(): new_hypernetwork_activation_func, new_hypernetwork_initialization_option, new_hypernetwork_add_layer_norm, - new_hypernetwork_use_dropout + new_hypernetwork_use_dropout, + new_hypernetwork_dropout_structure ], outputs=[ train_hypernetwork_name, -- cgit v1.2.3 From e9f8292a3a6792b722696fcf8e32b3fcb43ba436 Mon Sep 17 00:00:00 2001 From: Andrey <16777216c@gmail.com> Date: Tue, 10 Jan 2023 11:54:48 +0300 Subject: Split history ui.py to ui_progress.py --- modules/ui.py | 1928 ------------------------------------------------ modules/ui_progress.py | 1928 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 1928 insertions(+), 1928 deletions(-) delete mode 100644 modules/ui.py create mode 100644 modules/ui_progress.py (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py deleted file mode 100644 index 9b9081b5..00000000 --- a/modules/ui.py +++ /dev/null @@ -1,1928 +0,0 @@ -import html -import json -import math -import mimetypes -import os -import platform -import random -import subprocess as sp -import sys -import tempfile -import time -import traceback -from functools import partial, reduce - -import gradio as gr -import gradio.routes -import gradio.utils -import numpy as np -from PIL import Image, PngImagePlugin -from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call - -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru -from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML -from modules.paths import script_path - -from modules.shared import opts, cmd_opts, restricted_opts - -import modules.codeformer_model -import modules.generation_parameters_copypaste as parameters_copypaste -import modules.gfpgan_model -import modules.hypernetworks.ui -import modules.scripts -import modules.shared as shared -import modules.styles -import modules.textual_inversion.ui -from modules import prompt_parser -from modules.images import save_image -from modules.sd_hijack import model_hijack -from modules.sd_samplers import samplers, samplers_for_img2img -from modules.textual_inversion import textual_inversion -import modules.hypernetworks.ui -from modules.generation_parameters_copypaste import image_from_url_text - -# this is a fix for Windows users. Without it, javascript files will be served with text/html content-type and the browser will not show any UI -mimetypes.init() -mimetypes.add_type('application/javascript', '.js') - -if not cmd_opts.share and not cmd_opts.listen: - # fix gradio phoning home - gradio.utils.version_check = lambda: None - gradio.utils.get_local_ip_address = lambda: '127.0.0.1' - -if cmd_opts.ngrok is not None: - import modules.ngrok as ngrok - print('ngrok authtoken detected, trying to connect...') - ngrok.connect( - cmd_opts.ngrok, - cmd_opts.port if cmd_opts.port is not None else 7860, - cmd_opts.ngrok_region - ) - - -def gr_show(visible=True): - return {"visible": visible, "__type__": "update"} - - -sample_img2img = "assets/stable-samples/img2img/sketch-mountains-input.jpg" -sample_img2img = sample_img2img if os.path.exists(sample_img2img) else None - -css_hide_progressbar = """ -.wrap .m-12 svg { display:none!important; } -.wrap .m-12::before { content:"Loading..." } -.wrap .z-20 svg { display:none!important; } -.wrap .z-20::before { content:"Loading..." } -.progress-bar { display:none!important; } -.meta-text { display:none!important; } -.meta-text-center { display:none!important; } -""" - -# Using constants for these since the variation selector isn't visible. -# Important that they exactly match script.js for tooltip to work. -random_symbol = '\U0001f3b2\ufe0f' # 🎲️ -reuse_symbol = '\u267b\ufe0f' # ♻️ -paste_symbol = '\u2199\ufe0f' # ↙ -folder_symbol = '\U0001f4c2' # 📂 -refresh_symbol = '\U0001f504' # 🔄 -save_style_symbol = '\U0001f4be' # 💾 -apply_style_symbol = '\U0001f4cb' # 📋 -clear_prompt_symbol = '\U0001F5D1' # 🗑️ - - -def plaintext_to_html(text): - text = "

" + "
\n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

" - return text - -def send_gradio_gallery_to_image(x): - if len(x) == 0: - return None - return image_from_url_text(x[0]) - -def save_files(js_data, images, do_make_zip, index): - import csv - filenames = [] - fullfns = [] - - #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it - class MyObject: - def __init__(self, d=None): - if d is not None: - for key, value in d.items(): - setattr(self, key, value) - - data = json.loads(js_data) - - p = MyObject(data) - path = opts.outdir_save - save_to_dirs = opts.use_save_to_dirs_for_ui - extension: str = opts.samples_format - start_index = 0 - - if index > -1 and opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only - - images = [images[index]] - start_index = index - - os.makedirs(opts.outdir_save, exist_ok=True) - - with open(os.path.join(opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: - at_start = file.tell() == 0 - writer = csv.writer(file) - if at_start: - writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) - - for image_index, filedata in enumerate(images, start_index): - image = image_from_url_text(filedata) - - is_grid = image_index < p.index_of_first_image - i = 0 if is_grid else (image_index - p.index_of_first_image) - - fullfn, txt_fullfn = save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) - - filename = os.path.relpath(fullfn, path) - filenames.append(filename) - fullfns.append(fullfn) - if txt_fullfn: - filenames.append(os.path.basename(txt_fullfn)) - fullfns.append(txt_fullfn) - - writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) - - # Make Zip - if do_make_zip: - zip_filepath = os.path.join(path, "images.zip") - - from zipfile import ZipFile - with ZipFile(zip_filepath, "w") as zip_file: - for i in range(len(fullfns)): - with open(fullfns[i], mode="rb") as f: - zip_file.writestr(filenames[i], f.read()) - fullfns.insert(0, zip_filepath) - - return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") - - -def calc_time_left(progress, threshold, label, force_display, show_eta): - if progress == 0: - return "" - else: - time_since_start = time.time() - shared.state.time_start - eta = (time_since_start/progress) - eta_relative = eta-time_since_start - if (eta_relative > threshold and show_eta) or force_display: - if eta_relative > 3600: - return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) - elif eta_relative > 60: - return label + time.strftime('%M:%S', time.gmtime(eta_relative)) - else: - return label + time.strftime('%Ss', time.gmtime(eta_relative)) - else: - return "" - - -def check_progress_call(id_part): - if shared.state.job_count == 0: - return "", gr_show(False), gr_show(False), gr_show(False) - - progress = 0 - - if shared.state.job_count > 0: - progress += shared.state.job_no / shared.state.job_count - if shared.state.sampling_steps > 0: - progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps - - # Show progress percentage and time left at the same moment, and base it also on steps done - show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 - - time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) - if time_left != "": - shared.state.time_left_force_display = True - - progress = min(progress, 1) - - progressbar = "" - if opts.show_progressbar: - progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" - - image = gr_show(False) - preview_visibility = gr_show(False) - - if opts.show_progress_every_n_steps != 0: - shared.state.set_current_image() - image = shared.state.current_image - - if image is None: - image = gr.update(value=None) - else: - preview_visibility = gr_show(True) - - if shared.state.textinfo is not None: - textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) - else: - textinfo_result = gr_show(False) - - return f"

{progressbar}

", preview_visibility, image, textinfo_result - - -def check_progress_call_initial(id_part): - shared.state.job_count = -1 - shared.state.current_latent = None - shared.state.current_image = None - shared.state.textinfo = None - shared.state.time_start = time.time() - shared.state.time_left_force_display = False - - return check_progress_call(id_part) - - -def visit(x, func, path=""): - if hasattr(x, 'children'): - for c in x.children: - visit(c, func, path) - elif x.label is not None: - func(path + "/" + str(x.label), x) - - -def add_style(name: str, prompt: str, negative_prompt: str): - if name is None: - return [gr_show() for x in range(4)] - - style = modules.styles.PromptStyle(name, prompt, negative_prompt) - shared.prompt_styles.styles[style.name] = style - # Save all loaded prompt styles: this allows us to update the storage format in the future more easily, because we - # reserialize all styles every time we save them - shared.prompt_styles.save_styles(shared.styles_filename) - - return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] - - -def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): - from modules import processing, devices - - if not enable: - return "" - - p = processing.StableDiffusionProcessingTxt2Img(width=width, height=height, enable_hr=True, hr_scale=hr_scale, hr_resize_x=hr_resize_x, hr_resize_y=hr_resize_y) - - with devices.autocast(): - p.init([""], [0], [0]) - - return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" - - -def apply_styles(prompt, prompt_neg, style1_name, style2_name): - prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) - prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, [style1_name, style2_name]) - - return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] - - -def interrogate(image): - prompt = shared.interrogator.interrogate(image.convert("RGB")) - - return gr_show(True) if prompt is None else prompt - - -def interrogate_deepbooru(image): - prompt = deepbooru.model.tag(image) - return gr_show(True) if prompt is None else prompt - - -def create_seed_inputs(target_interface): - with FormRow(elem_id=target_interface + '_seed_row'): - seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') - seed.style(container=False) - random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') - reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') - - with gr.Group(elem_id=target_interface + '_subseed_show_box'): - seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) - - # Components to show/hide based on the 'Extra' checkbox - seed_extras = [] - - with FormRow(visible=False, elem_id=target_interface + '_subseed_row') as seed_extra_row_1: - seed_extras.append(seed_extra_row_1) - subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') - subseed.style(container=False) - random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') - reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') - subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') - - with FormRow(visible=False) as seed_extra_row_2: - seed_extras.append(seed_extra_row_2) - seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') - seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') - - random_seed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[seed]) - random_subseed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[subseed]) - - def change_visibility(show): - return {comp: gr_show(show) for comp in seed_extras} - - seed_checkbox.change(change_visibility, show_progress=False, inputs=[seed_checkbox], outputs=seed_extras) - - return seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox - - - -def connect_clear_prompt(button): - """Given clear button, prompt, and token_counter objects, setup clear prompt button click event""" - button.click( - _js="clear_prompt", - fn=None, - inputs=[], - outputs=[], - ) - - -def connect_reuse_seed(seed: gr.Number, reuse_seed: gr.Button, generation_info: gr.Textbox, dummy_component, is_subseed): - """ Connects a 'reuse (sub)seed' button's click event so that it copies last used - (sub)seed value from generation info the to the seed field. If copying subseed and subseed strength - was 0, i.e. no variation seed was used, it copies the normal seed value instead.""" - def copy_seed(gen_info_string: str, index): - res = -1 - - try: - gen_info = json.loads(gen_info_string) - index -= gen_info.get('index_of_first_image', 0) - - if is_subseed and gen_info.get('subseed_strength', 0) > 0: - all_subseeds = gen_info.get('all_subseeds', [-1]) - res = all_subseeds[index if 0 <= index < len(all_subseeds) else 0] - else: - all_seeds = gen_info.get('all_seeds', [-1]) - res = all_seeds[index if 0 <= index < len(all_seeds) else 0] - - except json.decoder.JSONDecodeError as e: - if gen_info_string != '': - print("Error parsing JSON generation info:", file=sys.stderr) - print(gen_info_string, file=sys.stderr) - - return [res, gr_show(False)] - - reuse_seed.click( - fn=copy_seed, - _js="(x, y) => [x, selected_gallery_index()]", - show_progress=False, - inputs=[generation_info, dummy_component], - outputs=[seed, dummy_component] - ) - - -def update_token_counter(text, steps): - try: - _, prompt_flat_list, _ = prompt_parser.get_multicond_prompt_list([text]) - prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules(prompt_flat_list, steps) - - except Exception: - # a parsing error can happen here during typing, and we don't want to bother the user with - # messages related to it in console - prompt_schedules = [[[steps, text]]] - - flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) - prompts = [prompt_text for step, prompt_text in flat_prompts] - token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) - style_class = ' class="red"' if (token_count > max_length) else "" - return f"{token_count}/{max_length}" - - -def create_toprow(is_img2img): - id_part = "img2img" if is_img2img else "txt2img" - - with gr.Row(elem_id="toprow"): - with gr.Column(scale=6): - with gr.Row(): - with gr.Column(scale=80): - with gr.Row(): - prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, - placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) - - with gr.Row(): - with gr.Column(scale=80): - with gr.Row(): - negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, - placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) - - with gr.Column(scale=1, elem_id="roll_col"): - paste = gr.Button(value=paste_symbol, elem_id="paste") - save_style = gr.Button(value=save_style_symbol, elem_id="style_create") - prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") - clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") - token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") - token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") - - clear_prompt_button.click( - fn=lambda *x: x, - _js="confirm_clear_prompt", - inputs=[prompt, negative_prompt], - outputs=[prompt, negative_prompt], - ) - - button_interrogate = None - button_deepbooru = None - if is_img2img: - with gr.Column(scale=1, elem_id="interrogate_col"): - button_interrogate = gr.Button('Interrogate\nCLIP', elem_id="interrogate") - button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") - - with gr.Column(scale=1): - with gr.Row(): - skip = gr.Button('Skip', elem_id=f"{id_part}_skip") - interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") - submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') - - skip.click( - fn=lambda: shared.state.skip(), - inputs=[], - outputs=[], - ) - - interrupt.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - with gr.Row(): - with gr.Column(scale=1, elem_id="style_pos_col"): - prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - with gr.Column(scale=1, elem_id="style_neg_col"): - prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button - - -def setup_progressbar(progressbar, preview, id_part, textinfo=None): - if textinfo is None: - textinfo = gr.HTML(visible=False) - - check_progress = gr.Button('Check progress', elem_id=f"{id_part}_check_progress", visible=False) - check_progress.click( - fn=lambda: check_progress_call(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) - - check_progress_initial = gr.Button('Check progress (first)', elem_id=f"{id_part}_check_progress_initial", visible=False) - check_progress_initial.click( - fn=lambda: check_progress_call_initial(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) - - -def apply_setting(key, value): - if value is None: - return gr.update() - - if shared.cmd_opts.freeze_settings: - return gr.update() - - # dont allow model to be swapped when model hash exists in prompt - if key == "sd_model_checkpoint" and opts.disable_weights_auto_swap: - return gr.update() - - if key == "sd_model_checkpoint": - ckpt_info = sd_models.get_closet_checkpoint_match(value) - - if ckpt_info is not None: - value = ckpt_info.title - else: - return gr.update() - - comp_args = opts.data_labels[key].component_args - if comp_args and isinstance(comp_args, dict) and comp_args.get('visible') is False: - return - - valtype = type(opts.data_labels[key].default) - oldval = opts.data.get(key, None) - opts.data[key] = valtype(value) if valtype != type(None) else value - if oldval != value and opts.data_labels[key].onchange is not None: - opts.data_labels[key].onchange() - - opts.save(shared.config_filename) - return value - - -def update_generation_info(args): - generation_info, html_info, img_index = args - try: - generation_info = json.loads(generation_info) - if img_index < 0 or img_index >= len(generation_info["infotexts"]): - return html_info - return plaintext_to_html(generation_info["infotexts"][img_index]) - except Exception: - pass - # if the json parse or anything else fails, just return the old html_info - return html_info - - -def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): - def refresh(): - refresh_method() - args = refreshed_args() if callable(refreshed_args) else refreshed_args - - for k, v in args.items(): - setattr(refresh_component, k, v) - - return gr.update(**(args or {})) - - refresh_button = ToolButton(value=refresh_symbol, elem_id=elem_id) - refresh_button.click( - fn=refresh, - inputs=[], - outputs=[refresh_component] - ) - return refresh_button - - -def create_output_panel(tabname, outdir): - def open_folder(f): - if not os.path.exists(f): - print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') - return - elif not os.path.isdir(f): - print(f""" -WARNING -An open_folder request was made with an argument that is not a folder. -This could be an error or a malicious attempt to run code on your computer. -Requested path was: {f} -""", file=sys.stderr) - return - - if not shared.cmd_opts.hide_ui_dir_config: - path = os.path.normpath(f) - if platform.system() == "Windows": - os.startfile(path) - elif platform.system() == "Darwin": - sp.Popen(["open", path]) - elif "microsoft-standard-WSL2" in platform.uname().release: - sp.Popen(["wsl-open", path]) - else: - sp.Popen(["xdg-open", path]) - - with gr.Column(variant='panel'): - with gr.Group(): - result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) - - generation_info = None - with gr.Column(): - with gr.Row(elem_id=f"image_buttons_{tabname}"): - open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') - - if tabname != "extras": - save = gr.Button('Save', elem_id=f'save_{tabname}') - save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') - - buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) - - open_folder_button.click( - fn=lambda: open_folder(opts.outdir_samples or outdir), - inputs=[], - outputs=[], - ) - - if tabname != "extras": - with gr.Row(): - download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') - - with gr.Group(): - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') - if tabname == 'txt2img' or tabname == 'img2img': - generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") - generation_info_button.click( - fn=update_generation_info, - _js="(x, y) => [x, y, selected_gallery_index()]", - inputs=[generation_info, html_info], - outputs=[html_info], - preprocess=False - ) - - save.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - save_zip.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - else: - html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) - return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log - - -def create_sampler_and_steps_selection(choices, tabname): - if opts.samplers_in_dropdown: - with FormRow(elem_id=f"sampler_selection_{tabname}"): - sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) - else: - with FormGroup(elem_id=f"sampler_selection_{tabname}"): - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) - sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - - return steps, sampler_index - - -def ordered_ui_categories(): - user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} - - for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): - yield category - - -def create_ui(): - import modules.img2img - import modules.txt2img - - reload_javascript() - - parameters_copypaste.reset() - - modules.scripts.scripts_current = modules.scripts.scripts_txt2img - modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) - - with gr.Blocks(analytics_enabled=False) as txt2img_interface: - txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) - - dummy_component = gr.Label(visible=False) - txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Row(elem_id='txt2img_progress_row'): - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="txt2img_progressbar") - txt2img_preview = gr.Image(elem_id='txt2img_preview', visible=False) - setup_progressbar(progressbar, txt2img_preview, 'txt2img') - - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel', elem_id="txt2img_settings"): - for category in ordered_ui_categories(): - if category == "sampler": - steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") - - elif category == "dimensions": - with FormRow(): - with gr.Column(elem_id="txt2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") - - if opts.dimensions_and_batch_together: - with gr.Column(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - elif category == "cfg": - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") - - elif category == "seed": - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') - - elif category == "checkboxes": - with FormRow(elem_id="txt2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") - enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") - hr_final_resolution = FormHTML(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) - - elif category == "hires_fix": - with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: - with FormRow(elem_id="txt2img_hires_fix_row1"): - hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) - hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") - - with FormRow(elem_id="txt2img_hires_fix_row2"): - hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") - hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") - hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") - - elif category == "batch": - if not opts.dimensions_and_batch_together: - with FormRow(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - elif category == "scripts": - with FormGroup(elem_id="txt2img_script_container"): - custom_inputs = modules.scripts.scripts_txt2img.setup_ui() - - hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] - for input in hr_resolution_preview_inputs: - input.change( - fn=calc_resolution_hires, - inputs=hr_resolution_preview_inputs, - outputs=[hr_final_resolution], - show_progress=False, - ) - input.change( - None, - _js="onCalcResolutionHires", - inputs=hr_resolution_preview_inputs, - outputs=[], - show_progress=False, - ) - - txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) - parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) - - connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) - connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) - - txt2img_args = dict( - fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), - _js="submit", - inputs=[ - txt2img_prompt, - txt2img_negative_prompt, - txt2img_prompt_style, - txt2img_prompt_style2, - steps, - sampler_index, - restore_faces, - tiling, - batch_count, - batch_size, - cfg_scale, - seed, - subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, - height, - width, - enable_hr, - denoising_strength, - hr_scale, - hr_upscaler, - hr_second_pass_steps, - hr_resize_x, - hr_resize_y, - ] + custom_inputs, - - outputs=[ - txt2img_gallery, - generation_info, - html_info, - html_log, - ], - show_progress=False, - ) - - txt2img_prompt.submit(**txt2img_args) - submit.click(**txt2img_args) - - txt_prompt_img.change( - fn=modules.images.image_data, - inputs=[ - txt_prompt_img - ], - outputs=[ - txt2img_prompt, - txt_prompt_img - ] - ) - - enable_hr.change( - fn=lambda x: gr_show(x), - inputs=[enable_hr], - outputs=[hr_options], - show_progress = False, - ) - - txt2img_paste_fields = [ - (txt2img_prompt, "Prompt"), - (txt2img_negative_prompt, "Negative prompt"), - (steps, "Steps"), - (sampler_index, "Sampler"), - (restore_faces, "Face restoration"), - (cfg_scale, "CFG scale"), - (seed, "Seed"), - (width, "Size-1"), - (height, "Size-2"), - (batch_size, "Batch size"), - (subseed, "Variation seed"), - (subseed_strength, "Variation seed strength"), - (seed_resize_from_w, "Seed resize from-1"), - (seed_resize_from_h, "Seed resize from-2"), - (denoising_strength, "Denoising strength"), - (enable_hr, lambda d: "Denoising strength" in d), - (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), - (hr_scale, "Hires upscale"), - (hr_upscaler, "Hires upscaler"), - (hr_second_pass_steps, "Hires steps"), - (hr_resize_x, "Hires resize-1"), - (hr_resize_y, "Hires resize-2"), - *modules.scripts.scripts_txt2img.infotext_fields - ] - parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) - - txt2img_preview_params = [ - txt2img_prompt, - txt2img_negative_prompt, - steps, - sampler_index, - cfg_scale, - seed, - width, - height, - ] - - token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) - - modules.scripts.scripts_current = modules.scripts.scripts_img2img - modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) - - with gr.Blocks(analytics_enabled=False) as img2img_interface: - img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) - - with gr.Row(elem_id='img2img_progress_row'): - img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="img2img_progressbar") - img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) - setup_progressbar(progressbar, img2img_preview, 'img2img') - - with FormRow().style(equal_height=False): - with gr.Column(variant='panel', elem_id="img2img_settings"): - - with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: - with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): - init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) - - with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): - init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) - init_img_with_mask_orig = gr.State(None) - - use_color_sketch = cmd_opts.gradio_inpaint_tool == "color-sketch" - if use_color_sketch: - def update_orig(image, state): - if image is not None: - same_size = state is not None and state.size == image.size - has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) - edited = same_size and has_exact_match - return image if not edited or state is None else state - - init_img_with_mask.change(update_orig, [init_img_with_mask, init_img_with_mask_orig], init_img_with_mask_orig) - - init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") - init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") - - with FormRow(): - mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") - mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") - - with FormRow(): - mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") - inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") - - with FormRow(): - inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") - - with FormRow(): - with gr.Column(): - inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") - - with gr.Column(scale=4): - inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") - - with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): - hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' - gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") - img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") - img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") - - with FormRow(): - resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") - - for category in ordered_ui_categories(): - if category == "sampler": - steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") - - elif category == "dimensions": - with FormRow(): - with gr.Column(elem_id="img2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") - - if opts.dimensions_and_batch_together: - with gr.Column(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - elif category == "cfg": - with FormGroup(): - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") - - elif category == "seed": - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') - - elif category == "checkboxes": - with FormRow(elem_id="img2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") - - elif category == "batch": - if not opts.dimensions_and_batch_together: - with FormRow(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - elif category == "scripts": - with FormGroup(elem_id="img2img_script_container"): - custom_inputs = modules.scripts.scripts_img2img.setup_ui() - - img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) - parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) - - connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) - connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) - - img2img_prompt_img.change( - fn=modules.images.image_data, - inputs=[ - img2img_prompt_img - ], - outputs=[ - img2img_prompt, - img2img_prompt_img - ] - ) - - mask_mode.change( - lambda mode, img: { - init_img_with_mask: gr_show(mode == 0), - init_img_inpaint: gr_show(mode == 1), - init_mask_inpaint: gr_show(mode == 1), - }, - inputs=[mask_mode, init_img_with_mask], - outputs=[ - init_img_with_mask, - init_img_inpaint, - init_mask_inpaint, - ], - ) - - img2img_args = dict( - fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), - _js="submit_img2img", - inputs=[ - dummy_component, - img2img_prompt, - img2img_negative_prompt, - img2img_prompt_style, - img2img_prompt_style2, - init_img, - init_img_with_mask, - init_img_with_mask_orig, - init_img_inpaint, - init_mask_inpaint, - mask_mode, - steps, - sampler_index, - mask_blur, - mask_alpha, - inpainting_fill, - restore_faces, - tiling, - batch_count, - batch_size, - cfg_scale, - denoising_strength, - seed, - subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, - height, - width, - resize_mode, - inpaint_full_res, - inpaint_full_res_padding, - inpainting_mask_invert, - img2img_batch_input_dir, - img2img_batch_output_dir, - ] + custom_inputs, - outputs=[ - img2img_gallery, - generation_info, - html_info, - html_log, - ], - show_progress=False, - ) - - img2img_prompt.submit(**img2img_args) - submit.click(**img2img_args) - - img2img_interrogate.click( - fn=interrogate, - inputs=[init_img], - outputs=[img2img_prompt], - ) - - img2img_deepbooru.click( - fn=interrogate_deepbooru, - inputs=[init_img], - outputs=[img2img_prompt], - ) - - prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] - style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] - style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] - - for button, (prompt, negative_prompt) in zip([txt2img_save_style, img2img_save_style], prompts): - button.click( - fn=add_style, - _js="ask_for_style_name", - # Have to pass empty dummy component here, because the JavaScript and Python function have to accept - # the same number of parameters, but we only know the style-name after the JavaScript prompt - inputs=[dummy_component, prompt, negative_prompt], - outputs=[txt2img_prompt_style, img2img_prompt_style, txt2img_prompt_style2, img2img_prompt_style2], - ) - - for button, (prompt, negative_prompt), (style1, style2), js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): - button.click( - fn=apply_styles, - _js=js_func, - inputs=[prompt, negative_prompt, style1, style2], - outputs=[prompt, negative_prompt, style1, style2], - ) - - token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) - - img2img_paste_fields = [ - (img2img_prompt, "Prompt"), - (img2img_negative_prompt, "Negative prompt"), - (steps, "Steps"), - (sampler_index, "Sampler"), - (restore_faces, "Face restoration"), - (cfg_scale, "CFG scale"), - (seed, "Seed"), - (width, "Size-1"), - (height, "Size-2"), - (batch_size, "Batch size"), - (subseed, "Variation seed"), - (subseed_strength, "Variation seed strength"), - (seed_resize_from_w, "Seed resize from-1"), - (seed_resize_from_h, "Seed resize from-2"), - (denoising_strength, "Denoising strength"), - (mask_blur, "Mask blur"), - *modules.scripts.scripts_img2img.infotext_fields - ] - parameters_copypaste.add_paste_fields("img2img", init_img, img2img_paste_fields) - parameters_copypaste.add_paste_fields("inpaint", init_img_with_mask, img2img_paste_fields) - - modules.scripts.scripts_current = None - - with gr.Blocks(analytics_enabled=False) as extras_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - with gr.Tabs(elem_id="mode_extras"): - with gr.TabItem('Single Image', elem_id="extras_single_tab"): - extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") - - with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): - image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") - - with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): - extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") - extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") - show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") - - submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') - - with gr.Tabs(elem_id="extras_resize_mode"): - with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): - upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") - with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): - with gr.Group(): - with gr.Row(): - upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") - upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") - upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") - - with gr.Group(): - extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - - with gr.Group(): - extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") - - with gr.Group(): - gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") - - with gr.Group(): - codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") - codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") - - with gr.Group(): - upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") - - result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) - - submit.click( - fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), - _js="get_extras_tab_index", - inputs=[ - dummy_component, - dummy_component, - extras_image, - image_batch, - extras_batch_input_dir, - extras_batch_output_dir, - show_extras_results, - gfpgan_visibility, - codeformer_visibility, - codeformer_weight, - upscaling_resize, - upscaling_resize_w, - upscaling_resize_h, - upscaling_crop, - extras_upscaler_1, - extras_upscaler_2, - extras_upscaler_2_visibility, - upscale_before_face_fix, - ], - outputs=[ - result_images, - html_info_x, - html_info, - ] - ) - parameters_copypaste.add_paste_fields("extras", extras_image, None) - - extras_image.change( - fn=modules.extras.clear_cache, - inputs=[], outputs=[] - ) - - with gr.Blocks(analytics_enabled=False) as pnginfo_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - image = gr.Image(elem_id="pnginfo_image", label="Source", source="upload", interactive=True, type="pil") - - with gr.Column(variant='panel'): - html = gr.HTML() - generation_info = gr.Textbox(visible=False, elem_id="pnginfo_generation_info") - html2 = gr.HTML() - with gr.Row(): - buttons = parameters_copypaste.create_buttons(["txt2img", "img2img", "inpaint", "extras"]) - parameters_copypaste.bind_buttons(buttons, image, generation_info) - - image.change( - fn=wrap_gradio_call(modules.extras.run_pnginfo), - inputs=[image], - outputs=[html, generation_info, html2], - ) - - with gr.Blocks(analytics_enabled=False) as modelmerger_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") - - with gr.Row(): - primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") - create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") - - secondary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_secondary_model_name", label="Secondary model (B)") - create_refresh_button(secondary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_B") - - tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") - create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") - - custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") - interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") - interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") - - with gr.Row(): - checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") - save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") - - modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') - - with gr.Column(variant='panel'): - submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) - - with gr.Blocks(analytics_enabled=False) as train_interface: - with gr.Row().style(equal_height=False): - gr.HTML(value="

See wiki for detailed explanation.

") - - with gr.Row().style(equal_height=False): - with gr.Tabs(elem_id="train_tabs"): - - with gr.Tab(label="Create embedding"): - new_embedding_name = gr.Textbox(label="Name", elem_id="train_new_embedding_name") - initialization_text = gr.Textbox(label="Initialization text", value="*", elem_id="train_initialization_text") - nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1, elem_id="train_nvpt") - overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding", elem_id="train_overwrite_old_embedding") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - create_embedding = gr.Button(value="Create embedding", variant='primary', elem_id="train_create_embedding") - - with gr.Tab(label="Create hypernetwork"): - new_hypernetwork_name = gr.Textbox(label="Name", elem_id="train_new_hypernetwork_name") - new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"], elem_id="train_new_hypernetwork_sizes") - new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'", elem_id="train_new_hypernetwork_layer_structure") - new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys, elem_id="train_new_hypernetwork_activation_func") - new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") - new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") - new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") - new_hypernetwork_dropout_structure = gr.Textbox("0, 0, 0", label="Enter hypernetwork Dropout structure (or empty). Recommended : 0~0.35 incrementing sequence: 0, 0.05, 0.15", placeholder="1st and last digit must be 0 and values should be between 0 and 1. ex:'0, 0.01, 0'") - overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary', elem_id="train_create_hypernetwork") - - with gr.Tab(label="Preprocess images"): - process_src = gr.Textbox(label='Source directory', elem_id="train_process_src") - process_dst = gr.Textbox(label='Destination directory', elem_id="train_process_dst") - process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_process_width") - process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_process_height") - preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"], elem_id="train_preprocess_txt_action") - - with gr.Row(): - process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") - process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") - process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") - process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") - process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") - - with gr.Row(visible=False) as process_split_extra_row: - process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_split_threshold") - process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05, elem_id="train_process_overlap_ratio") - - with gr.Row(visible=False) as process_focal_crop_row: - process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_face_weight") - process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") - process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") - process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - with gr.Row(): - interrupt_preprocessing = gr.Button("Interrupt", elem_id="train_interrupt_preprocessing") - run_preprocess = gr.Button(value="Preprocess", variant='primary', elem_id="train_run_preprocess") - - process_split.change( - fn=lambda show: gr_show(show), - inputs=[process_split], - outputs=[process_split_extra_row], - ) - - process_focal_crop.change( - fn=lambda show: gr_show(show), - inputs=[process_focal_crop], - outputs=[process_focal_crop_row], - ) - - def get_textual_inversion_template_names(): - return sorted([x for x in textual_inversion.textual_inversion_templates]) - - with gr.Tab(label="Train"): - gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") - with FormRow(): - train_embedding_name = gr.Dropdown(label='Embedding', elem_id="train_embedding", choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())) - create_refresh_button(train_embedding_name, sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings, lambda: {"choices": sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())}, "refresh_train_embedding_name") - - train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) - create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") - - with FormRow(): - embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") - hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") - - with FormRow(): - clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) - clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) - - with FormRow(): - batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") - gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") - - dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") - log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") - - with FormRow(): - template_file = gr.Dropdown(label='Prompt template', value="style_filewords.txt", elem_id="train_template_file", choices=get_textual_inversion_template_names()) - create_refresh_button(template_file, textual_inversion.list_textual_inversion_templates, lambda: {"choices": get_textual_inversion_template_names()}, "refrsh_train_template_file") - - training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") - training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") - varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") - steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") - - with FormRow(): - create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") - save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") - - save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") - preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") - - shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") - tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") - - latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") - - with gr.Row(): - train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") - interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") - train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") - - params = script_callbacks.UiTrainTabParams(txt2img_preview_params) - - script_callbacks.ui_train_tabs_callback(params) - - with gr.Column(): - progressbar = gr.HTML(elem_id="ti_progressbar") - ti_output = gr.Text(elem_id="ti_output", value="", show_label=False) - - ti_gallery = gr.Gallery(label='Output', show_label=False, elem_id='ti_gallery').style(grid=4) - ti_preview = gr.Image(elem_id='ti_preview', visible=False) - ti_progress = gr.HTML(elem_id="ti_progress", value="") - ti_outcome = gr.HTML(elem_id="ti_error", value="") - setup_progressbar(progressbar, ti_preview, 'ti', textinfo=ti_progress) - - create_embedding.click( - fn=modules.textual_inversion.ui.create_embedding, - inputs=[ - new_embedding_name, - initialization_text, - nvpt, - overwrite_old_embedding, - ], - outputs=[ - train_embedding_name, - ti_output, - ti_outcome, - ] - ) - - create_hypernetwork.click( - fn=modules.hypernetworks.ui.create_hypernetwork, - inputs=[ - new_hypernetwork_name, - new_hypernetwork_sizes, - overwrite_old_hypernetwork, - new_hypernetwork_layer_structure, - new_hypernetwork_activation_func, - new_hypernetwork_initialization_option, - new_hypernetwork_add_layer_norm, - new_hypernetwork_use_dropout, - new_hypernetwork_dropout_structure - ], - outputs=[ - train_hypernetwork_name, - ti_output, - ti_outcome, - ] - ) - - run_preprocess.click( - fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.preprocess, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - process_src, - process_dst, - process_width, - process_height, - preprocess_txt_action, - process_flip, - process_split, - process_caption, - process_caption_deepbooru, - process_split_threshold, - process_overlap_ratio, - process_focal_crop, - process_focal_crop_face_weight, - process_focal_crop_entropy_weight, - process_focal_crop_edges_weight, - process_focal_crop_debug, - ], - outputs=[ - ti_output, - ti_outcome, - ], - ) - - train_embedding.click( - fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.train_embedding, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - train_embedding_name, - embedding_learn_rate, - batch_size, - gradient_step, - dataset_directory, - log_directory, - training_width, - training_height, - varsize, - steps, - clip_grad_mode, - clip_grad_value, - shuffle_tags, - tag_drop_out, - latent_sampling_method, - create_image_every, - save_embedding_every, - template_file, - save_image_with_stored_embedding, - preview_from_txt2img, - *txt2img_preview_params, - ], - outputs=[ - ti_output, - ti_outcome, - ] - ) - - train_hypernetwork.click( - fn=wrap_gradio_gpu_call(modules.hypernetworks.ui.train_hypernetwork, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - train_hypernetwork_name, - hypernetwork_learn_rate, - batch_size, - gradient_step, - dataset_directory, - log_directory, - training_width, - training_height, - varsize, - steps, - clip_grad_mode, - clip_grad_value, - shuffle_tags, - tag_drop_out, - latent_sampling_method, - create_image_every, - save_embedding_every, - template_file, - preview_from_txt2img, - *txt2img_preview_params, - ], - outputs=[ - ti_output, - ti_outcome, - ] - ) - - interrupt_training.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - interrupt_preprocessing.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - def create_setting_component(key, is_quicksettings=False): - def fun(): - return opts.data[key] if key in opts.data else opts.data_labels[key].default - - info = opts.data_labels[key] - t = type(info.default) - - args = info.component_args() if callable(info.component_args) else info.component_args - - if info.component is not None: - comp = info.component - elif t == str: - comp = gr.Textbox - elif t == int: - comp = gr.Number - elif t == bool: - comp = gr.Checkbox - else: - raise Exception(f'bad options item type: {str(t)} for key {key}') - - elem_id = "setting_"+key - - if info.refresh is not None: - if is_quicksettings: - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) - else: - with FormRow(): - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) - else: - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - - return res - - components = [] - component_dict = {} - - script_callbacks.ui_settings_callback() - opts.reorder() - - def run_settings(*args): - changed = [] - - for key, value, comp in zip(opts.data_labels.keys(), args, components): - assert comp == dummy_component or opts.same_type(value, opts.data_labels[key].default), f"Bad value for setting {key}: {value}; expecting {type(opts.data_labels[key].default).__name__}" - - for key, value, comp in zip(opts.data_labels.keys(), args, components): - if comp == dummy_component: - continue - - if opts.set(key, value): - changed.append(key) - - try: - opts.save(shared.config_filename) - except RuntimeError: - return opts.dumpjson(), f'{len(changed)} settings changed without save: {", ".join(changed)}.' - return opts.dumpjson(), f'{len(changed)} settings changed{": " if len(changed) > 0 else ""}{", ".join(changed)}.' - - def run_settings_single(value, key): - if not opts.same_type(value, opts.data_labels[key].default): - return gr.update(visible=True), opts.dumpjson() - - if not opts.set(key, value): - return gr.update(value=getattr(opts, key)), opts.dumpjson() - - opts.save(shared.config_filename) - - return gr.update(value=value), opts.dumpjson() - - with gr.Blocks(analytics_enabled=False) as settings_interface: - with gr.Row(): - with gr.Column(scale=6): - settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") - with gr.Column(): - restart_gradio = gr.Button(value='Reload UI', variant='primary', elem_id="settings_restart_gradio") - - result = gr.HTML(elem_id="settings_result") - - quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] - quicksettings_names = {x: i for i, x in enumerate(quicksettings_names) if x != 'quicksettings'} - - quicksettings_list = [] - - previous_section = None - current_tab = None - with gr.Tabs(elem_id="settings"): - for i, (k, item) in enumerate(opts.data_labels.items()): - section_must_be_skipped = item.section[0] is None - - if previous_section != item.section and not section_must_be_skipped: - elem_id, text = item.section - - if current_tab is not None: - current_tab.__exit__() - - current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) - current_tab.__enter__() - - previous_section = item.section - - if k in quicksettings_names and not shared.cmd_opts.freeze_settings: - quicksettings_list.append((i, k, item)) - components.append(dummy_component) - elif section_must_be_skipped: - components.append(dummy_component) - else: - component = create_setting_component(k) - component_dict[k] = component - components.append(component) - - if current_tab is not None: - current_tab.__exit__() - - with gr.TabItem("Actions"): - request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") - download_localization = gr.Button(value='Download localization template', elem_id="download_localization") - reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") - - if os.path.exists("html/licenses.html"): - with open("html/licenses.html", encoding="utf8") as file: - with gr.TabItem("Licenses"): - gr.HTML(file.read(), elem_id="licenses") - - gr.Button(value="Show all pages", elem_id="settings_show_all_pages") - - request_notifications.click( - fn=lambda: None, - inputs=[], - outputs=[], - _js='function(){}' - ) - - download_localization.click( - fn=lambda: None, - inputs=[], - outputs=[], - _js='download_localization' - ) - - def reload_scripts(): - modules.scripts.reload_script_body_only() - reload_javascript() # need to refresh the html page - - reload_script_bodies.click( - fn=reload_scripts, - inputs=[], - outputs=[] - ) - - def request_restart(): - shared.state.interrupt() - shared.state.need_restart = True - - restart_gradio.click( - fn=request_restart, - _js='restart_reload', - inputs=[], - outputs=[], - ) - - interfaces = [ - (txt2img_interface, "txt2img", "txt2img"), - (img2img_interface, "img2img", "img2img"), - (extras_interface, "Extras", "extras"), - (pnginfo_interface, "PNG Info", "pnginfo"), - (modelmerger_interface, "Checkpoint Merger", "modelmerger"), - (train_interface, "Train", "ti"), - ] - - css = "" - - for cssfile in modules.scripts.list_files_with_name("style.css"): - if not os.path.isfile(cssfile): - continue - - with open(cssfile, "r", encoding="utf8") as file: - css += file.read() + "\n" - - if os.path.exists(os.path.join(script_path, "user.css")): - with open(os.path.join(script_path, "user.css"), "r", encoding="utf8") as file: - css += file.read() + "\n" - - if not cmd_opts.no_progressbar_hiding: - css += css_hide_progressbar - - interfaces += script_callbacks.ui_tabs_callback() - interfaces += [(settings_interface, "Settings", "settings")] - - extensions_interface = ui_extensions.create_ui() - interfaces += [(extensions_interface, "Extensions", "extensions")] - - with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: - with gr.Row(elem_id="quicksettings"): - for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): - component = create_setting_component(k, is_quicksettings=True) - component_dict[k] = component - - parameters_copypaste.integrate_settings_paste_fields(component_dict) - parameters_copypaste.run_bind() - - with gr.Tabs(elem_id="tabs") as tabs: - for interface, label, ifid in interfaces: - with gr.TabItem(label, id=ifid, elem_id='tab_' + ifid): - interface.render() - - if os.path.exists(os.path.join(script_path, "notification.mp3")): - audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) - - if os.path.exists("html/footer.html"): - with open("html/footer.html", encoding="utf8") as file: - footer = file.read() - footer = footer.format(versions=versions_html()) - gr.HTML(footer, elem_id="footer") - - text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) - settings_submit.click( - fn=wrap_gradio_call(run_settings, extra_outputs=[gr.update()]), - inputs=components, - outputs=[text_settings, result], - ) - - for i, k, item in quicksettings_list: - component = component_dict[k] - - component.change( - fn=lambda value, k=k: run_settings_single(value, key=k), - inputs=[component], - outputs=[component, text_settings], - ) - - component_keys = [k for k in opts.data_labels.keys() if k in component_dict] - - def get_settings_values(): - return [getattr(opts, key) for key in component_keys] - - demo.load( - fn=get_settings_values, - inputs=[], - outputs=[component_dict[k] for k in component_keys], - ) - - def modelmerger(*args): - try: - results = modules.extras.run_modelmerger(*args) - except Exception as e: - print("Error loading/saving model file:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - modules.sd_models.list_models() # to remove the potentially missing models from the list - return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] - return results - - modelmerger_merge.click( - fn=modelmerger, - inputs=[ - primary_model_name, - secondary_model_name, - tertiary_model_name, - interp_method, - interp_amount, - save_as_half, - custom_name, - checkpoint_format, - ], - outputs=[ - submit_result, - primary_model_name, - secondary_model_name, - tertiary_model_name, - component_dict['sd_model_checkpoint'], - ] - ) - - ui_config_file = cmd_opts.ui_config_file - ui_settings = {} - settings_count = len(ui_settings) - error_loading = False - - try: - if os.path.exists(ui_config_file): - with open(ui_config_file, "r", encoding="utf8") as file: - ui_settings = json.load(file) - except Exception: - error_loading = True - print("Error loading settings:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - - def loadsave(path, x): - def apply_field(obj, field, condition=None, init_field=None): - key = path + "/" + field - - if getattr(obj, 'custom_script_source', None) is not None: - key = 'customscript/' + obj.custom_script_source + '/' + key - - if getattr(obj, 'do_not_save_to_config', False): - return - - saved_value = ui_settings.get(key, None) - if saved_value is None: - ui_settings[key] = getattr(obj, field) - elif condition and not condition(saved_value): - print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') - else: - setattr(obj, field, saved_value) - if init_field is not None: - init_field(saved_value) - - if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number, gr.Dropdown] and x.visible: - apply_field(x, 'visible') - - if type(x) == gr.Slider: - apply_field(x, 'value') - apply_field(x, 'minimum') - apply_field(x, 'maximum') - apply_field(x, 'step') - - if type(x) == gr.Radio: - apply_field(x, 'value', lambda val: val in x.choices) - - if type(x) == gr.Checkbox: - apply_field(x, 'value') - - if type(x) == gr.Textbox: - apply_field(x, 'value') - - if type(x) == gr.Number: - apply_field(x, 'value') - - if type(x) == gr.Dropdown: - apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) - - visit(txt2img_interface, loadsave, "txt2img") - visit(img2img_interface, loadsave, "img2img") - visit(extras_interface, loadsave, "extras") - visit(modelmerger_interface, loadsave, "modelmerger") - visit(train_interface, loadsave, "train") - - if not error_loading and (not os.path.exists(ui_config_file) or settings_count != len(ui_settings)): - with open(ui_config_file, "w", encoding="utf8") as file: - json.dump(ui_settings, file, indent=4) - - return demo - - -def reload_javascript(): - with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: - javascript = f'' - - scripts_list = modules.scripts.list_scripts("javascript", ".js") - - for basedir, filename, path in scripts_list: - with open(path, "r", encoding="utf8") as jsfile: - javascript += f"\n" - - if cmd_opts.theme is not None: - javascript += f"\n\n" - - javascript += f"\n" - - def template_response(*args, **kwargs): - res = shared.GradioTemplateResponseOriginal(*args, **kwargs) - res.body = res.body.replace( - b'', f'{javascript}'.encode("utf8")) - res.init_headers() - return res - - gradio.routes.templates.TemplateResponse = template_response - - -if not hasattr(shared, 'GradioTemplateResponseOriginal'): - shared.GradioTemplateResponseOriginal = gradio.routes.templates.TemplateResponse - - -def versions_html(): - import torch - import launch - - python_version = ".".join([str(x) for x in sys.version_info[0:3]]) - commit = launch.commit_hash() - short_commit = commit[0:8] - - if shared.xformers_available: - import xformers - xformers_version = xformers.__version__ - else: - xformers_version = "N/A" - - return f""" -python: {python_version} - •  -torch: {torch.__version__} - •  -xformers: {xformers_version} - •  -gradio: {gr.__version__} - •  -commit: {short_commit} -""" diff --git a/modules/ui_progress.py b/modules/ui_progress.py new file mode 100644 index 00000000..9b9081b5 --- /dev/null +++ b/modules/ui_progress.py @@ -0,0 +1,1928 @@ +import html +import json +import math +import mimetypes +import os +import platform +import random +import subprocess as sp +import sys +import tempfile +import time +import traceback +from functools import partial, reduce + +import gradio as gr +import gradio.routes +import gradio.utils +import numpy as np +from PIL import Image, PngImagePlugin +from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call + +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru +from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML +from modules.paths import script_path + +from modules.shared import opts, cmd_opts, restricted_opts + +import modules.codeformer_model +import modules.generation_parameters_copypaste as parameters_copypaste +import modules.gfpgan_model +import modules.hypernetworks.ui +import modules.scripts +import modules.shared as shared +import modules.styles +import modules.textual_inversion.ui +from modules import prompt_parser +from modules.images import save_image +from modules.sd_hijack import model_hijack +from modules.sd_samplers import samplers, samplers_for_img2img +from modules.textual_inversion import textual_inversion +import modules.hypernetworks.ui +from modules.generation_parameters_copypaste import image_from_url_text + +# this is a fix for Windows users. Without it, javascript files will be served with text/html content-type and the browser will not show any UI +mimetypes.init() +mimetypes.add_type('application/javascript', '.js') + +if not cmd_opts.share and not cmd_opts.listen: + # fix gradio phoning home + gradio.utils.version_check = lambda: None + gradio.utils.get_local_ip_address = lambda: '127.0.0.1' + +if cmd_opts.ngrok is not None: + import modules.ngrok as ngrok + print('ngrok authtoken detected, trying to connect...') + ngrok.connect( + cmd_opts.ngrok, + cmd_opts.port if cmd_opts.port is not None else 7860, + cmd_opts.ngrok_region + ) + + +def gr_show(visible=True): + return {"visible": visible, "__type__": "update"} + + +sample_img2img = "assets/stable-samples/img2img/sketch-mountains-input.jpg" +sample_img2img = sample_img2img if os.path.exists(sample_img2img) else None + +css_hide_progressbar = """ +.wrap .m-12 svg { display:none!important; } +.wrap .m-12::before { content:"Loading..." } +.wrap .z-20 svg { display:none!important; } +.wrap .z-20::before { content:"Loading..." } +.progress-bar { display:none!important; } +.meta-text { display:none!important; } +.meta-text-center { display:none!important; } +""" + +# Using constants for these since the variation selector isn't visible. +# Important that they exactly match script.js for tooltip to work. +random_symbol = '\U0001f3b2\ufe0f' # 🎲️ +reuse_symbol = '\u267b\ufe0f' # ♻️ +paste_symbol = '\u2199\ufe0f' # ↙ +folder_symbol = '\U0001f4c2' # 📂 +refresh_symbol = '\U0001f504' # 🔄 +save_style_symbol = '\U0001f4be' # 💾 +apply_style_symbol = '\U0001f4cb' # 📋 +clear_prompt_symbol = '\U0001F5D1' # 🗑️ + + +def plaintext_to_html(text): + text = "

" + "
\n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

" + return text + +def send_gradio_gallery_to_image(x): + if len(x) == 0: + return None + return image_from_url_text(x[0]) + +def save_files(js_data, images, do_make_zip, index): + import csv + filenames = [] + fullfns = [] + + #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it + class MyObject: + def __init__(self, d=None): + if d is not None: + for key, value in d.items(): + setattr(self, key, value) + + data = json.loads(js_data) + + p = MyObject(data) + path = opts.outdir_save + save_to_dirs = opts.use_save_to_dirs_for_ui + extension: str = opts.samples_format + start_index = 0 + + if index > -1 and opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only + + images = [images[index]] + start_index = index + + os.makedirs(opts.outdir_save, exist_ok=True) + + with open(os.path.join(opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: + at_start = file.tell() == 0 + writer = csv.writer(file) + if at_start: + writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) + + for image_index, filedata in enumerate(images, start_index): + image = image_from_url_text(filedata) + + is_grid = image_index < p.index_of_first_image + i = 0 if is_grid else (image_index - p.index_of_first_image) + + fullfn, txt_fullfn = save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) + + filename = os.path.relpath(fullfn, path) + filenames.append(filename) + fullfns.append(fullfn) + if txt_fullfn: + filenames.append(os.path.basename(txt_fullfn)) + fullfns.append(txt_fullfn) + + writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) + + # Make Zip + if do_make_zip: + zip_filepath = os.path.join(path, "images.zip") + + from zipfile import ZipFile + with ZipFile(zip_filepath, "w") as zip_file: + for i in range(len(fullfns)): + with open(fullfns[i], mode="rb") as f: + zip_file.writestr(filenames[i], f.read()) + fullfns.insert(0, zip_filepath) + + return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") + + +def calc_time_left(progress, threshold, label, force_display, show_eta): + if progress == 0: + return "" + else: + time_since_start = time.time() - shared.state.time_start + eta = (time_since_start/progress) + eta_relative = eta-time_since_start + if (eta_relative > threshold and show_eta) or force_display: + if eta_relative > 3600: + return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) + elif eta_relative > 60: + return label + time.strftime('%M:%S', time.gmtime(eta_relative)) + else: + return label + time.strftime('%Ss', time.gmtime(eta_relative)) + else: + return "" + + +def check_progress_call(id_part): + if shared.state.job_count == 0: + return "", gr_show(False), gr_show(False), gr_show(False) + + progress = 0 + + if shared.state.job_count > 0: + progress += shared.state.job_no / shared.state.job_count + if shared.state.sampling_steps > 0: + progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps + + # Show progress percentage and time left at the same moment, and base it also on steps done + show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 + + time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) + if time_left != "": + shared.state.time_left_force_display = True + + progress = min(progress, 1) + + progressbar = "" + if opts.show_progressbar: + progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" + + image = gr_show(False) + preview_visibility = gr_show(False) + + if opts.show_progress_every_n_steps != 0: + shared.state.set_current_image() + image = shared.state.current_image + + if image is None: + image = gr.update(value=None) + else: + preview_visibility = gr_show(True) + + if shared.state.textinfo is not None: + textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) + else: + textinfo_result = gr_show(False) + + return f"

{progressbar}

", preview_visibility, image, textinfo_result + + +def check_progress_call_initial(id_part): + shared.state.job_count = -1 + shared.state.current_latent = None + shared.state.current_image = None + shared.state.textinfo = None + shared.state.time_start = time.time() + shared.state.time_left_force_display = False + + return check_progress_call(id_part) + + +def visit(x, func, path=""): + if hasattr(x, 'children'): + for c in x.children: + visit(c, func, path) + elif x.label is not None: + func(path + "/" + str(x.label), x) + + +def add_style(name: str, prompt: str, negative_prompt: str): + if name is None: + return [gr_show() for x in range(4)] + + style = modules.styles.PromptStyle(name, prompt, negative_prompt) + shared.prompt_styles.styles[style.name] = style + # Save all loaded prompt styles: this allows us to update the storage format in the future more easily, because we + # reserialize all styles every time we save them + shared.prompt_styles.save_styles(shared.styles_filename) + + return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] + + +def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): + from modules import processing, devices + + if not enable: + return "" + + p = processing.StableDiffusionProcessingTxt2Img(width=width, height=height, enable_hr=True, hr_scale=hr_scale, hr_resize_x=hr_resize_x, hr_resize_y=hr_resize_y) + + with devices.autocast(): + p.init([""], [0], [0]) + + return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" + + +def apply_styles(prompt, prompt_neg, style1_name, style2_name): + prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) + prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, [style1_name, style2_name]) + + return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] + + +def interrogate(image): + prompt = shared.interrogator.interrogate(image.convert("RGB")) + + return gr_show(True) if prompt is None else prompt + + +def interrogate_deepbooru(image): + prompt = deepbooru.model.tag(image) + return gr_show(True) if prompt is None else prompt + + +def create_seed_inputs(target_interface): + with FormRow(elem_id=target_interface + '_seed_row'): + seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') + seed.style(container=False) + random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') + reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') + + with gr.Group(elem_id=target_interface + '_subseed_show_box'): + seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) + + # Components to show/hide based on the 'Extra' checkbox + seed_extras = [] + + with FormRow(visible=False, elem_id=target_interface + '_subseed_row') as seed_extra_row_1: + seed_extras.append(seed_extra_row_1) + subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') + subseed.style(container=False) + random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') + reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') + subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') + + with FormRow(visible=False) as seed_extra_row_2: + seed_extras.append(seed_extra_row_2) + seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') + seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') + + random_seed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[seed]) + random_subseed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[subseed]) + + def change_visibility(show): + return {comp: gr_show(show) for comp in seed_extras} + + seed_checkbox.change(change_visibility, show_progress=False, inputs=[seed_checkbox], outputs=seed_extras) + + return seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox + + + +def connect_clear_prompt(button): + """Given clear button, prompt, and token_counter objects, setup clear prompt button click event""" + button.click( + _js="clear_prompt", + fn=None, + inputs=[], + outputs=[], + ) + + +def connect_reuse_seed(seed: gr.Number, reuse_seed: gr.Button, generation_info: gr.Textbox, dummy_component, is_subseed): + """ Connects a 'reuse (sub)seed' button's click event so that it copies last used + (sub)seed value from generation info the to the seed field. If copying subseed and subseed strength + was 0, i.e. no variation seed was used, it copies the normal seed value instead.""" + def copy_seed(gen_info_string: str, index): + res = -1 + + try: + gen_info = json.loads(gen_info_string) + index -= gen_info.get('index_of_first_image', 0) + + if is_subseed and gen_info.get('subseed_strength', 0) > 0: + all_subseeds = gen_info.get('all_subseeds', [-1]) + res = all_subseeds[index if 0 <= index < len(all_subseeds) else 0] + else: + all_seeds = gen_info.get('all_seeds', [-1]) + res = all_seeds[index if 0 <= index < len(all_seeds) else 0] + + except json.decoder.JSONDecodeError as e: + if gen_info_string != '': + print("Error parsing JSON generation info:", file=sys.stderr) + print(gen_info_string, file=sys.stderr) + + return [res, gr_show(False)] + + reuse_seed.click( + fn=copy_seed, + _js="(x, y) => [x, selected_gallery_index()]", + show_progress=False, + inputs=[generation_info, dummy_component], + outputs=[seed, dummy_component] + ) + + +def update_token_counter(text, steps): + try: + _, prompt_flat_list, _ = prompt_parser.get_multicond_prompt_list([text]) + prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules(prompt_flat_list, steps) + + except Exception: + # a parsing error can happen here during typing, and we don't want to bother the user with + # messages related to it in console + prompt_schedules = [[[steps, text]]] + + flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) + prompts = [prompt_text for step, prompt_text in flat_prompts] + token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) + style_class = ' class="red"' if (token_count > max_length) else "" + return f"{token_count}/{max_length}" + + +def create_toprow(is_img2img): + id_part = "img2img" if is_img2img else "txt2img" + + with gr.Row(elem_id="toprow"): + with gr.Column(scale=6): + with gr.Row(): + with gr.Column(scale=80): + with gr.Row(): + prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, + placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)" + ) + + with gr.Row(): + with gr.Column(scale=80): + with gr.Row(): + negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, + placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)" + ) + + with gr.Column(scale=1, elem_id="roll_col"): + paste = gr.Button(value=paste_symbol, elem_id="paste") + save_style = gr.Button(value=save_style_symbol, elem_id="style_create") + prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") + clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") + token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") + token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") + + clear_prompt_button.click( + fn=lambda *x: x, + _js="confirm_clear_prompt", + inputs=[prompt, negative_prompt], + outputs=[prompt, negative_prompt], + ) + + button_interrogate = None + button_deepbooru = None + if is_img2img: + with gr.Column(scale=1, elem_id="interrogate_col"): + button_interrogate = gr.Button('Interrogate\nCLIP', elem_id="interrogate") + button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") + + with gr.Column(scale=1): + with gr.Row(): + skip = gr.Button('Skip', elem_id=f"{id_part}_skip") + interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") + submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') + + skip.click( + fn=lambda: shared.state.skip(), + inputs=[], + outputs=[], + ) + + interrupt.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + with gr.Row(): + with gr.Column(scale=1, elem_id="style_pos_col"): + prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) + + with gr.Column(scale=1, elem_id="style_neg_col"): + prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) + + return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button + + +def setup_progressbar(progressbar, preview, id_part, textinfo=None): + if textinfo is None: + textinfo = gr.HTML(visible=False) + + check_progress = gr.Button('Check progress', elem_id=f"{id_part}_check_progress", visible=False) + check_progress.click( + fn=lambda: check_progress_call(id_part), + show_progress=False, + inputs=[], + outputs=[progressbar, preview, preview, textinfo], + ) + + check_progress_initial = gr.Button('Check progress (first)', elem_id=f"{id_part}_check_progress_initial", visible=False) + check_progress_initial.click( + fn=lambda: check_progress_call_initial(id_part), + show_progress=False, + inputs=[], + outputs=[progressbar, preview, preview, textinfo], + ) + + +def apply_setting(key, value): + if value is None: + return gr.update() + + if shared.cmd_opts.freeze_settings: + return gr.update() + + # dont allow model to be swapped when model hash exists in prompt + if key == "sd_model_checkpoint" and opts.disable_weights_auto_swap: + return gr.update() + + if key == "sd_model_checkpoint": + ckpt_info = sd_models.get_closet_checkpoint_match(value) + + if ckpt_info is not None: + value = ckpt_info.title + else: + return gr.update() + + comp_args = opts.data_labels[key].component_args + if comp_args and isinstance(comp_args, dict) and comp_args.get('visible') is False: + return + + valtype = type(opts.data_labels[key].default) + oldval = opts.data.get(key, None) + opts.data[key] = valtype(value) if valtype != type(None) else value + if oldval != value and opts.data_labels[key].onchange is not None: + opts.data_labels[key].onchange() + + opts.save(shared.config_filename) + return value + + +def update_generation_info(args): + generation_info, html_info, img_index = args + try: + generation_info = json.loads(generation_info) + if img_index < 0 or img_index >= len(generation_info["infotexts"]): + return html_info + return plaintext_to_html(generation_info["infotexts"][img_index]) + except Exception: + pass + # if the json parse or anything else fails, just return the old html_info + return html_info + + +def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): + def refresh(): + refresh_method() + args = refreshed_args() if callable(refreshed_args) else refreshed_args + + for k, v in args.items(): + setattr(refresh_component, k, v) + + return gr.update(**(args or {})) + + refresh_button = ToolButton(value=refresh_symbol, elem_id=elem_id) + refresh_button.click( + fn=refresh, + inputs=[], + outputs=[refresh_component] + ) + return refresh_button + + +def create_output_panel(tabname, outdir): + def open_folder(f): + if not os.path.exists(f): + print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') + return + elif not os.path.isdir(f): + print(f""" +WARNING +An open_folder request was made with an argument that is not a folder. +This could be an error or a malicious attempt to run code on your computer. +Requested path was: {f} +""", file=sys.stderr) + return + + if not shared.cmd_opts.hide_ui_dir_config: + path = os.path.normpath(f) + if platform.system() == "Windows": + os.startfile(path) + elif platform.system() == "Darwin": + sp.Popen(["open", path]) + elif "microsoft-standard-WSL2" in platform.uname().release: + sp.Popen(["wsl-open", path]) + else: + sp.Popen(["xdg-open", path]) + + with gr.Column(variant='panel'): + with gr.Group(): + result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) + + generation_info = None + with gr.Column(): + with gr.Row(elem_id=f"image_buttons_{tabname}"): + open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') + + if tabname != "extras": + save = gr.Button('Save', elem_id=f'save_{tabname}') + save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') + + buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) + + open_folder_button.click( + fn=lambda: open_folder(opts.outdir_samples or outdir), + inputs=[], + outputs=[], + ) + + if tabname != "extras": + with gr.Row(): + download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') + + with gr.Group(): + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') + + generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') + if tabname == 'txt2img' or tabname == 'img2img': + generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") + generation_info_button.click( + fn=update_generation_info, + _js="(x, y) => [x, y, selected_gallery_index()]", + inputs=[generation_info, html_info], + outputs=[html_info], + preprocess=False + ) + + save.click( + fn=wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ] + ) + + save_zip.click( + fn=wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ] + ) + + else: + html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') + + parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) + return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log + + +def create_sampler_and_steps_selection(choices, tabname): + if opts.samplers_in_dropdown: + with FormRow(elem_id=f"sampler_selection_{tabname}"): + sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) + else: + with FormGroup(elem_id=f"sampler_selection_{tabname}"): + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) + sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + + return steps, sampler_index + + +def ordered_ui_categories(): + user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} + + for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): + yield category + + +def create_ui(): + import modules.img2img + import modules.txt2img + + reload_javascript() + + parameters_copypaste.reset() + + modules.scripts.scripts_current = modules.scripts.scripts_txt2img + modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) + + with gr.Blocks(analytics_enabled=False) as txt2img_interface: + txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) + + dummy_component = gr.Label(visible=False) + txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) + + with gr.Row(elem_id='txt2img_progress_row'): + with gr.Column(scale=1): + pass + + with gr.Column(scale=1): + progressbar = gr.HTML(elem_id="txt2img_progressbar") + txt2img_preview = gr.Image(elem_id='txt2img_preview', visible=False) + setup_progressbar(progressbar, txt2img_preview, 'txt2img') + + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel', elem_id="txt2img_settings"): + for category in ordered_ui_categories(): + if category == "sampler": + steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") + + elif category == "dimensions": + with FormRow(): + with gr.Column(elem_id="txt2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") + + if opts.dimensions_and_batch_together: + with gr.Column(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") + + elif category == "cfg": + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") + + elif category == "seed": + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') + + elif category == "checkboxes": + with FormRow(elem_id="txt2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") + enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") + hr_final_resolution = FormHTML(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) + + elif category == "hires_fix": + with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: + with FormRow(elem_id="txt2img_hires_fix_row1"): + hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) + hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") + + with FormRow(elem_id="txt2img_hires_fix_row2"): + hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") + hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") + hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") + + elif category == "batch": + if not opts.dimensions_and_batch_together: + with FormRow(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") + + elif category == "scripts": + with FormGroup(elem_id="txt2img_script_container"): + custom_inputs = modules.scripts.scripts_txt2img.setup_ui() + + hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] + for input in hr_resolution_preview_inputs: + input.change( + fn=calc_resolution_hires, + inputs=hr_resolution_preview_inputs, + outputs=[hr_final_resolution], + show_progress=False, + ) + input.change( + None, + _js="onCalcResolutionHires", + inputs=hr_resolution_preview_inputs, + outputs=[], + show_progress=False, + ) + + txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) + parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) + + connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) + connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) + + txt2img_args = dict( + fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), + _js="submit", + inputs=[ + txt2img_prompt, + txt2img_negative_prompt, + txt2img_prompt_style, + txt2img_prompt_style2, + steps, + sampler_index, + restore_faces, + tiling, + batch_count, + batch_size, + cfg_scale, + seed, + subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, + height, + width, + enable_hr, + denoising_strength, + hr_scale, + hr_upscaler, + hr_second_pass_steps, + hr_resize_x, + hr_resize_y, + ] + custom_inputs, + + outputs=[ + txt2img_gallery, + generation_info, + html_info, + html_log, + ], + show_progress=False, + ) + + txt2img_prompt.submit(**txt2img_args) + submit.click(**txt2img_args) + + txt_prompt_img.change( + fn=modules.images.image_data, + inputs=[ + txt_prompt_img + ], + outputs=[ + txt2img_prompt, + txt_prompt_img + ] + ) + + enable_hr.change( + fn=lambda x: gr_show(x), + inputs=[enable_hr], + outputs=[hr_options], + show_progress = False, + ) + + txt2img_paste_fields = [ + (txt2img_prompt, "Prompt"), + (txt2img_negative_prompt, "Negative prompt"), + (steps, "Steps"), + (sampler_index, "Sampler"), + (restore_faces, "Face restoration"), + (cfg_scale, "CFG scale"), + (seed, "Seed"), + (width, "Size-1"), + (height, "Size-2"), + (batch_size, "Batch size"), + (subseed, "Variation seed"), + (subseed_strength, "Variation seed strength"), + (seed_resize_from_w, "Seed resize from-1"), + (seed_resize_from_h, "Seed resize from-2"), + (denoising_strength, "Denoising strength"), + (enable_hr, lambda d: "Denoising strength" in d), + (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), + (hr_scale, "Hires upscale"), + (hr_upscaler, "Hires upscaler"), + (hr_second_pass_steps, "Hires steps"), + (hr_resize_x, "Hires resize-1"), + (hr_resize_y, "Hires resize-2"), + *modules.scripts.scripts_txt2img.infotext_fields + ] + parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) + + txt2img_preview_params = [ + txt2img_prompt, + txt2img_negative_prompt, + steps, + sampler_index, + cfg_scale, + seed, + width, + height, + ] + + token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) + + modules.scripts.scripts_current = modules.scripts.scripts_img2img + modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) + + with gr.Blocks(analytics_enabled=False) as img2img_interface: + img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) + + with gr.Row(elem_id='img2img_progress_row'): + img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) + + with gr.Column(scale=1): + pass + + with gr.Column(scale=1): + progressbar = gr.HTML(elem_id="img2img_progressbar") + img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) + setup_progressbar(progressbar, img2img_preview, 'img2img') + + with FormRow().style(equal_height=False): + with gr.Column(variant='panel', elem_id="img2img_settings"): + + with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: + with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): + init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) + + with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): + init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) + init_img_with_mask_orig = gr.State(None) + + use_color_sketch = cmd_opts.gradio_inpaint_tool == "color-sketch" + if use_color_sketch: + def update_orig(image, state): + if image is not None: + same_size = state is not None and state.size == image.size + has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) + edited = same_size and has_exact_match + return image if not edited or state is None else state + + init_img_with_mask.change(update_orig, [init_img_with_mask, init_img_with_mask_orig], init_img_with_mask_orig) + + init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") + init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") + + with FormRow(): + mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") + mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") + + with FormRow(): + mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") + inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") + + with FormRow(): + inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") + + with FormRow(): + with gr.Column(): + inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") + + with gr.Column(scale=4): + inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") + + with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): + hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' + gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") + img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") + img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") + + with FormRow(): + resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") + + for category in ordered_ui_categories(): + if category == "sampler": + steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") + + elif category == "dimensions": + with FormRow(): + with gr.Column(elem_id="img2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") + + if opts.dimensions_and_batch_together: + with gr.Column(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") + + elif category == "cfg": + with FormGroup(): + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") + + elif category == "seed": + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') + + elif category == "checkboxes": + with FormRow(elem_id="img2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") + + elif category == "batch": + if not opts.dimensions_and_batch_together: + with FormRow(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") + + elif category == "scripts": + with FormGroup(elem_id="img2img_script_container"): + custom_inputs = modules.scripts.scripts_img2img.setup_ui() + + img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) + parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) + + connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) + connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) + + img2img_prompt_img.change( + fn=modules.images.image_data, + inputs=[ + img2img_prompt_img + ], + outputs=[ + img2img_prompt, + img2img_prompt_img + ] + ) + + mask_mode.change( + lambda mode, img: { + init_img_with_mask: gr_show(mode == 0), + init_img_inpaint: gr_show(mode == 1), + init_mask_inpaint: gr_show(mode == 1), + }, + inputs=[mask_mode, init_img_with_mask], + outputs=[ + init_img_with_mask, + init_img_inpaint, + init_mask_inpaint, + ], + ) + + img2img_args = dict( + fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), + _js="submit_img2img", + inputs=[ + dummy_component, + img2img_prompt, + img2img_negative_prompt, + img2img_prompt_style, + img2img_prompt_style2, + init_img, + init_img_with_mask, + init_img_with_mask_orig, + init_img_inpaint, + init_mask_inpaint, + mask_mode, + steps, + sampler_index, + mask_blur, + mask_alpha, + inpainting_fill, + restore_faces, + tiling, + batch_count, + batch_size, + cfg_scale, + denoising_strength, + seed, + subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, + height, + width, + resize_mode, + inpaint_full_res, + inpaint_full_res_padding, + inpainting_mask_invert, + img2img_batch_input_dir, + img2img_batch_output_dir, + ] + custom_inputs, + outputs=[ + img2img_gallery, + generation_info, + html_info, + html_log, + ], + show_progress=False, + ) + + img2img_prompt.submit(**img2img_args) + submit.click(**img2img_args) + + img2img_interrogate.click( + fn=interrogate, + inputs=[init_img], + outputs=[img2img_prompt], + ) + + img2img_deepbooru.click( + fn=interrogate_deepbooru, + inputs=[init_img], + outputs=[img2img_prompt], + ) + + prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] + style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] + style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] + + for button, (prompt, negative_prompt) in zip([txt2img_save_style, img2img_save_style], prompts): + button.click( + fn=add_style, + _js="ask_for_style_name", + # Have to pass empty dummy component here, because the JavaScript and Python function have to accept + # the same number of parameters, but we only know the style-name after the JavaScript prompt + inputs=[dummy_component, prompt, negative_prompt], + outputs=[txt2img_prompt_style, img2img_prompt_style, txt2img_prompt_style2, img2img_prompt_style2], + ) + + for button, (prompt, negative_prompt), (style1, style2), js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): + button.click( + fn=apply_styles, + _js=js_func, + inputs=[prompt, negative_prompt, style1, style2], + outputs=[prompt, negative_prompt, style1, style2], + ) + + token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) + + img2img_paste_fields = [ + (img2img_prompt, "Prompt"), + (img2img_negative_prompt, "Negative prompt"), + (steps, "Steps"), + (sampler_index, "Sampler"), + (restore_faces, "Face restoration"), + (cfg_scale, "CFG scale"), + (seed, "Seed"), + (width, "Size-1"), + (height, "Size-2"), + (batch_size, "Batch size"), + (subseed, "Variation seed"), + (subseed_strength, "Variation seed strength"), + (seed_resize_from_w, "Seed resize from-1"), + (seed_resize_from_h, "Seed resize from-2"), + (denoising_strength, "Denoising strength"), + (mask_blur, "Mask blur"), + *modules.scripts.scripts_img2img.infotext_fields + ] + parameters_copypaste.add_paste_fields("img2img", init_img, img2img_paste_fields) + parameters_copypaste.add_paste_fields("inpaint", init_img_with_mask, img2img_paste_fields) + + modules.scripts.scripts_current = None + + with gr.Blocks(analytics_enabled=False) as extras_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + with gr.Tabs(elem_id="mode_extras"): + with gr.TabItem('Single Image', elem_id="extras_single_tab"): + extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") + + with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): + image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") + + with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): + extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") + extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") + show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") + + submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') + + with gr.Tabs(elem_id="extras_resize_mode"): + with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): + upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") + with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): + with gr.Group(): + with gr.Row(): + upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") + upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") + upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") + + with gr.Group(): + extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") + + with gr.Group(): + extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") + extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") + + with gr.Group(): + gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") + + with gr.Group(): + codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") + codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") + + with gr.Group(): + upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") + + result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) + + submit.click( + fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), + _js="get_extras_tab_index", + inputs=[ + dummy_component, + dummy_component, + extras_image, + image_batch, + extras_batch_input_dir, + extras_batch_output_dir, + show_extras_results, + gfpgan_visibility, + codeformer_visibility, + codeformer_weight, + upscaling_resize, + upscaling_resize_w, + upscaling_resize_h, + upscaling_crop, + extras_upscaler_1, + extras_upscaler_2, + extras_upscaler_2_visibility, + upscale_before_face_fix, + ], + outputs=[ + result_images, + html_info_x, + html_info, + ] + ) + parameters_copypaste.add_paste_fields("extras", extras_image, None) + + extras_image.change( + fn=modules.extras.clear_cache, + inputs=[], outputs=[] + ) + + with gr.Blocks(analytics_enabled=False) as pnginfo_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + image = gr.Image(elem_id="pnginfo_image", label="Source", source="upload", interactive=True, type="pil") + + with gr.Column(variant='panel'): + html = gr.HTML() + generation_info = gr.Textbox(visible=False, elem_id="pnginfo_generation_info") + html2 = gr.HTML() + with gr.Row(): + buttons = parameters_copypaste.create_buttons(["txt2img", "img2img", "inpaint", "extras"]) + parameters_copypaste.bind_buttons(buttons, image, generation_info) + + image.change( + fn=wrap_gradio_call(modules.extras.run_pnginfo), + inputs=[image], + outputs=[html, generation_info, html2], + ) + + with gr.Blocks(analytics_enabled=False) as modelmerger_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") + + with gr.Row(): + primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") + create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") + + secondary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_secondary_model_name", label="Secondary model (B)") + create_refresh_button(secondary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_B") + + tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") + create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") + + custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") + interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") + interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") + + with gr.Row(): + checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") + save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") + + modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') + + with gr.Column(variant='panel'): + submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) + + with gr.Blocks(analytics_enabled=False) as train_interface: + with gr.Row().style(equal_height=False): + gr.HTML(value="

See wiki for detailed explanation.

") + + with gr.Row().style(equal_height=False): + with gr.Tabs(elem_id="train_tabs"): + + with gr.Tab(label="Create embedding"): + new_embedding_name = gr.Textbox(label="Name", elem_id="train_new_embedding_name") + initialization_text = gr.Textbox(label="Initialization text", value="*", elem_id="train_initialization_text") + nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1, elem_id="train_nvpt") + overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding", elem_id="train_overwrite_old_embedding") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + create_embedding = gr.Button(value="Create embedding", variant='primary', elem_id="train_create_embedding") + + with gr.Tab(label="Create hypernetwork"): + new_hypernetwork_name = gr.Textbox(label="Name", elem_id="train_new_hypernetwork_name") + new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"], elem_id="train_new_hypernetwork_sizes") + new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'", elem_id="train_new_hypernetwork_layer_structure") + new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys, elem_id="train_new_hypernetwork_activation_func") + new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") + new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") + new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") + new_hypernetwork_dropout_structure = gr.Textbox("0, 0, 0", label="Enter hypernetwork Dropout structure (or empty). Recommended : 0~0.35 incrementing sequence: 0, 0.05, 0.15", placeholder="1st and last digit must be 0 and values should be between 0 and 1. ex:'0, 0.01, 0'") + overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary', elem_id="train_create_hypernetwork") + + with gr.Tab(label="Preprocess images"): + process_src = gr.Textbox(label='Source directory', elem_id="train_process_src") + process_dst = gr.Textbox(label='Destination directory', elem_id="train_process_dst") + process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_process_width") + process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_process_height") + preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"], elem_id="train_preprocess_txt_action") + + with gr.Row(): + process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") + process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") + process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") + process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") + process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") + + with gr.Row(visible=False) as process_split_extra_row: + process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_split_threshold") + process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05, elem_id="train_process_overlap_ratio") + + with gr.Row(visible=False) as process_focal_crop_row: + process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_face_weight") + process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") + process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") + process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + with gr.Row(): + interrupt_preprocessing = gr.Button("Interrupt", elem_id="train_interrupt_preprocessing") + run_preprocess = gr.Button(value="Preprocess", variant='primary', elem_id="train_run_preprocess") + + process_split.change( + fn=lambda show: gr_show(show), + inputs=[process_split], + outputs=[process_split_extra_row], + ) + + process_focal_crop.change( + fn=lambda show: gr_show(show), + inputs=[process_focal_crop], + outputs=[process_focal_crop_row], + ) + + def get_textual_inversion_template_names(): + return sorted([x for x in textual_inversion.textual_inversion_templates]) + + with gr.Tab(label="Train"): + gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") + with FormRow(): + train_embedding_name = gr.Dropdown(label='Embedding', elem_id="train_embedding", choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())) + create_refresh_button(train_embedding_name, sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings, lambda: {"choices": sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())}, "refresh_train_embedding_name") + + train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) + create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") + + with FormRow(): + embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") + hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") + + with FormRow(): + clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) + clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) + + with FormRow(): + batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") + gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") + + dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") + log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") + + with FormRow(): + template_file = gr.Dropdown(label='Prompt template', value="style_filewords.txt", elem_id="train_template_file", choices=get_textual_inversion_template_names()) + create_refresh_button(template_file, textual_inversion.list_textual_inversion_templates, lambda: {"choices": get_textual_inversion_template_names()}, "refrsh_train_template_file") + + training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") + training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") + varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") + steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") + + with FormRow(): + create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") + save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") + + save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") + preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") + + shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") + tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") + + latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") + + with gr.Row(): + train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") + interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") + train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") + + params = script_callbacks.UiTrainTabParams(txt2img_preview_params) + + script_callbacks.ui_train_tabs_callback(params) + + with gr.Column(): + progressbar = gr.HTML(elem_id="ti_progressbar") + ti_output = gr.Text(elem_id="ti_output", value="", show_label=False) + + ti_gallery = gr.Gallery(label='Output', show_label=False, elem_id='ti_gallery').style(grid=4) + ti_preview = gr.Image(elem_id='ti_preview', visible=False) + ti_progress = gr.HTML(elem_id="ti_progress", value="") + ti_outcome = gr.HTML(elem_id="ti_error", value="") + setup_progressbar(progressbar, ti_preview, 'ti', textinfo=ti_progress) + + create_embedding.click( + fn=modules.textual_inversion.ui.create_embedding, + inputs=[ + new_embedding_name, + initialization_text, + nvpt, + overwrite_old_embedding, + ], + outputs=[ + train_embedding_name, + ti_output, + ti_outcome, + ] + ) + + create_hypernetwork.click( + fn=modules.hypernetworks.ui.create_hypernetwork, + inputs=[ + new_hypernetwork_name, + new_hypernetwork_sizes, + overwrite_old_hypernetwork, + new_hypernetwork_layer_structure, + new_hypernetwork_activation_func, + new_hypernetwork_initialization_option, + new_hypernetwork_add_layer_norm, + new_hypernetwork_use_dropout, + new_hypernetwork_dropout_structure + ], + outputs=[ + train_hypernetwork_name, + ti_output, + ti_outcome, + ] + ) + + run_preprocess.click( + fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.preprocess, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + process_src, + process_dst, + process_width, + process_height, + preprocess_txt_action, + process_flip, + process_split, + process_caption, + process_caption_deepbooru, + process_split_threshold, + process_overlap_ratio, + process_focal_crop, + process_focal_crop_face_weight, + process_focal_crop_entropy_weight, + process_focal_crop_edges_weight, + process_focal_crop_debug, + ], + outputs=[ + ti_output, + ti_outcome, + ], + ) + + train_embedding.click( + fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.train_embedding, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + train_embedding_name, + embedding_learn_rate, + batch_size, + gradient_step, + dataset_directory, + log_directory, + training_width, + training_height, + varsize, + steps, + clip_grad_mode, + clip_grad_value, + shuffle_tags, + tag_drop_out, + latent_sampling_method, + create_image_every, + save_embedding_every, + template_file, + save_image_with_stored_embedding, + preview_from_txt2img, + *txt2img_preview_params, + ], + outputs=[ + ti_output, + ti_outcome, + ] + ) + + train_hypernetwork.click( + fn=wrap_gradio_gpu_call(modules.hypernetworks.ui.train_hypernetwork, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + train_hypernetwork_name, + hypernetwork_learn_rate, + batch_size, + gradient_step, + dataset_directory, + log_directory, + training_width, + training_height, + varsize, + steps, + clip_grad_mode, + clip_grad_value, + shuffle_tags, + tag_drop_out, + latent_sampling_method, + create_image_every, + save_embedding_every, + template_file, + preview_from_txt2img, + *txt2img_preview_params, + ], + outputs=[ + ti_output, + ti_outcome, + ] + ) + + interrupt_training.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + interrupt_preprocessing.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + def create_setting_component(key, is_quicksettings=False): + def fun(): + return opts.data[key] if key in opts.data else opts.data_labels[key].default + + info = opts.data_labels[key] + t = type(info.default) + + args = info.component_args() if callable(info.component_args) else info.component_args + + if info.component is not None: + comp = info.component + elif t == str: + comp = gr.Textbox + elif t == int: + comp = gr.Number + elif t == bool: + comp = gr.Checkbox + else: + raise Exception(f'bad options item type: {str(t)} for key {key}') + + elem_id = "setting_"+key + + if info.refresh is not None: + if is_quicksettings: + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) + else: + with FormRow(): + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) + else: + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + + return res + + components = [] + component_dict = {} + + script_callbacks.ui_settings_callback() + opts.reorder() + + def run_settings(*args): + changed = [] + + for key, value, comp in zip(opts.data_labels.keys(), args, components): + assert comp == dummy_component or opts.same_type(value, opts.data_labels[key].default), f"Bad value for setting {key}: {value}; expecting {type(opts.data_labels[key].default).__name__}" + + for key, value, comp in zip(opts.data_labels.keys(), args, components): + if comp == dummy_component: + continue + + if opts.set(key, value): + changed.append(key) + + try: + opts.save(shared.config_filename) + except RuntimeError: + return opts.dumpjson(), f'{len(changed)} settings changed without save: {", ".join(changed)}.' + return opts.dumpjson(), f'{len(changed)} settings changed{": " if len(changed) > 0 else ""}{", ".join(changed)}.' + + def run_settings_single(value, key): + if not opts.same_type(value, opts.data_labels[key].default): + return gr.update(visible=True), opts.dumpjson() + + if not opts.set(key, value): + return gr.update(value=getattr(opts, key)), opts.dumpjson() + + opts.save(shared.config_filename) + + return gr.update(value=value), opts.dumpjson() + + with gr.Blocks(analytics_enabled=False) as settings_interface: + with gr.Row(): + with gr.Column(scale=6): + settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") + with gr.Column(): + restart_gradio = gr.Button(value='Reload UI', variant='primary', elem_id="settings_restart_gradio") + + result = gr.HTML(elem_id="settings_result") + + quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] + quicksettings_names = {x: i for i, x in enumerate(quicksettings_names) if x != 'quicksettings'} + + quicksettings_list = [] + + previous_section = None + current_tab = None + with gr.Tabs(elem_id="settings"): + for i, (k, item) in enumerate(opts.data_labels.items()): + section_must_be_skipped = item.section[0] is None + + if previous_section != item.section and not section_must_be_skipped: + elem_id, text = item.section + + if current_tab is not None: + current_tab.__exit__() + + current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) + current_tab.__enter__() + + previous_section = item.section + + if k in quicksettings_names and not shared.cmd_opts.freeze_settings: + quicksettings_list.append((i, k, item)) + components.append(dummy_component) + elif section_must_be_skipped: + components.append(dummy_component) + else: + component = create_setting_component(k) + component_dict[k] = component + components.append(component) + + if current_tab is not None: + current_tab.__exit__() + + with gr.TabItem("Actions"): + request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") + download_localization = gr.Button(value='Download localization template', elem_id="download_localization") + reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") + + if os.path.exists("html/licenses.html"): + with open("html/licenses.html", encoding="utf8") as file: + with gr.TabItem("Licenses"): + gr.HTML(file.read(), elem_id="licenses") + + gr.Button(value="Show all pages", elem_id="settings_show_all_pages") + + request_notifications.click( + fn=lambda: None, + inputs=[], + outputs=[], + _js='function(){}' + ) + + download_localization.click( + fn=lambda: None, + inputs=[], + outputs=[], + _js='download_localization' + ) + + def reload_scripts(): + modules.scripts.reload_script_body_only() + reload_javascript() # need to refresh the html page + + reload_script_bodies.click( + fn=reload_scripts, + inputs=[], + outputs=[] + ) + + def request_restart(): + shared.state.interrupt() + shared.state.need_restart = True + + restart_gradio.click( + fn=request_restart, + _js='restart_reload', + inputs=[], + outputs=[], + ) + + interfaces = [ + (txt2img_interface, "txt2img", "txt2img"), + (img2img_interface, "img2img", "img2img"), + (extras_interface, "Extras", "extras"), + (pnginfo_interface, "PNG Info", "pnginfo"), + (modelmerger_interface, "Checkpoint Merger", "modelmerger"), + (train_interface, "Train", "ti"), + ] + + css = "" + + for cssfile in modules.scripts.list_files_with_name("style.css"): + if not os.path.isfile(cssfile): + continue + + with open(cssfile, "r", encoding="utf8") as file: + css += file.read() + "\n" + + if os.path.exists(os.path.join(script_path, "user.css")): + with open(os.path.join(script_path, "user.css"), "r", encoding="utf8") as file: + css += file.read() + "\n" + + if not cmd_opts.no_progressbar_hiding: + css += css_hide_progressbar + + interfaces += script_callbacks.ui_tabs_callback() + interfaces += [(settings_interface, "Settings", "settings")] + + extensions_interface = ui_extensions.create_ui() + interfaces += [(extensions_interface, "Extensions", "extensions")] + + with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: + with gr.Row(elem_id="quicksettings"): + for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): + component = create_setting_component(k, is_quicksettings=True) + component_dict[k] = component + + parameters_copypaste.integrate_settings_paste_fields(component_dict) + parameters_copypaste.run_bind() + + with gr.Tabs(elem_id="tabs") as tabs: + for interface, label, ifid in interfaces: + with gr.TabItem(label, id=ifid, elem_id='tab_' + ifid): + interface.render() + + if os.path.exists(os.path.join(script_path, "notification.mp3")): + audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) + + if os.path.exists("html/footer.html"): + with open("html/footer.html", encoding="utf8") as file: + footer = file.read() + footer = footer.format(versions=versions_html()) + gr.HTML(footer, elem_id="footer") + + text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) + settings_submit.click( + fn=wrap_gradio_call(run_settings, extra_outputs=[gr.update()]), + inputs=components, + outputs=[text_settings, result], + ) + + for i, k, item in quicksettings_list: + component = component_dict[k] + + component.change( + fn=lambda value, k=k: run_settings_single(value, key=k), + inputs=[component], + outputs=[component, text_settings], + ) + + component_keys = [k for k in opts.data_labels.keys() if k in component_dict] + + def get_settings_values(): + return [getattr(opts, key) for key in component_keys] + + demo.load( + fn=get_settings_values, + inputs=[], + outputs=[component_dict[k] for k in component_keys], + ) + + def modelmerger(*args): + try: + results = modules.extras.run_modelmerger(*args) + except Exception as e: + print("Error loading/saving model file:", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + modules.sd_models.list_models() # to remove the potentially missing models from the list + return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] + return results + + modelmerger_merge.click( + fn=modelmerger, + inputs=[ + primary_model_name, + secondary_model_name, + tertiary_model_name, + interp_method, + interp_amount, + save_as_half, + custom_name, + checkpoint_format, + ], + outputs=[ + submit_result, + primary_model_name, + secondary_model_name, + tertiary_model_name, + component_dict['sd_model_checkpoint'], + ] + ) + + ui_config_file = cmd_opts.ui_config_file + ui_settings = {} + settings_count = len(ui_settings) + error_loading = False + + try: + if os.path.exists(ui_config_file): + with open(ui_config_file, "r", encoding="utf8") as file: + ui_settings = json.load(file) + except Exception: + error_loading = True + print("Error loading settings:", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + + def loadsave(path, x): + def apply_field(obj, field, condition=None, init_field=None): + key = path + "/" + field + + if getattr(obj, 'custom_script_source', None) is not None: + key = 'customscript/' + obj.custom_script_source + '/' + key + + if getattr(obj, 'do_not_save_to_config', False): + return + + saved_value = ui_settings.get(key, None) + if saved_value is None: + ui_settings[key] = getattr(obj, field) + elif condition and not condition(saved_value): + print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') + else: + setattr(obj, field, saved_value) + if init_field is not None: + init_field(saved_value) + + if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number, gr.Dropdown] and x.visible: + apply_field(x, 'visible') + + if type(x) == gr.Slider: + apply_field(x, 'value') + apply_field(x, 'minimum') + apply_field(x, 'maximum') + apply_field(x, 'step') + + if type(x) == gr.Radio: + apply_field(x, 'value', lambda val: val in x.choices) + + if type(x) == gr.Checkbox: + apply_field(x, 'value') + + if type(x) == gr.Textbox: + apply_field(x, 'value') + + if type(x) == gr.Number: + apply_field(x, 'value') + + if type(x) == gr.Dropdown: + apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) + + visit(txt2img_interface, loadsave, "txt2img") + visit(img2img_interface, loadsave, "img2img") + visit(extras_interface, loadsave, "extras") + visit(modelmerger_interface, loadsave, "modelmerger") + visit(train_interface, loadsave, "train") + + if not error_loading and (not os.path.exists(ui_config_file) or settings_count != len(ui_settings)): + with open(ui_config_file, "w", encoding="utf8") as file: + json.dump(ui_settings, file, indent=4) + + return demo + + +def reload_javascript(): + with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: + javascript = f'' + + scripts_list = modules.scripts.list_scripts("javascript", ".js") + + for basedir, filename, path in scripts_list: + with open(path, "r", encoding="utf8") as jsfile: + javascript += f"\n" + + if cmd_opts.theme is not None: + javascript += f"\n\n" + + javascript += f"\n" + + def template_response(*args, **kwargs): + res = shared.GradioTemplateResponseOriginal(*args, **kwargs) + res.body = res.body.replace( + b'', f'{javascript}'.encode("utf8")) + res.init_headers() + return res + + gradio.routes.templates.TemplateResponse = template_response + + +if not hasattr(shared, 'GradioTemplateResponseOriginal'): + shared.GradioTemplateResponseOriginal = gradio.routes.templates.TemplateResponse + + +def versions_html(): + import torch + import launch + + python_version = ".".join([str(x) for x in sys.version_info[0:3]]) + commit = launch.commit_hash() + short_commit = commit[0:8] + + if shared.xformers_available: + import xformers + xformers_version = xformers.__version__ + else: + xformers_version = "N/A" + + return f""" +python: {python_version} + •  +torch: {torch.__version__} + •  +xformers: {xformers_version} + •  +gradio: {gr.__version__} + •  +commit: {short_commit} +""" -- cgit v1.2.3 From 27ea6949d3206c9a52fa77db587bac0012cb0b52 Mon Sep 17 00:00:00 2001 From: Andrey <16777216c@gmail.com> Date: Tue, 10 Jan 2023 11:54:48 +0300 Subject: Split history ui.py to ui_progress.py --- modules/temp | 1928 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ modules/ui.py | 1928 --------------------------------------------------------- 2 files changed, 1928 insertions(+), 1928 deletions(-) create mode 100644 modules/temp delete mode 100644 modules/ui.py (limited to 'modules') diff --git a/modules/temp b/modules/temp new file mode 100644 index 00000000..9b9081b5 --- /dev/null +++ b/modules/temp @@ -0,0 +1,1928 @@ +import html +import json +import math +import mimetypes +import os +import platform +import random +import subprocess as sp +import sys +import tempfile +import time +import traceback +from functools import partial, reduce + +import gradio as gr +import gradio.routes +import gradio.utils +import numpy as np +from PIL import Image, PngImagePlugin +from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call + +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru +from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML +from modules.paths import script_path + +from modules.shared import opts, cmd_opts, restricted_opts + +import modules.codeformer_model +import modules.generation_parameters_copypaste as parameters_copypaste +import modules.gfpgan_model +import modules.hypernetworks.ui +import modules.scripts +import modules.shared as shared +import modules.styles +import modules.textual_inversion.ui +from modules import prompt_parser +from modules.images import save_image +from modules.sd_hijack import model_hijack +from modules.sd_samplers import samplers, samplers_for_img2img +from modules.textual_inversion import textual_inversion +import modules.hypernetworks.ui +from modules.generation_parameters_copypaste import image_from_url_text + +# this is a fix for Windows users. Without it, javascript files will be served with text/html content-type and the browser will not show any UI +mimetypes.init() +mimetypes.add_type('application/javascript', '.js') + +if not cmd_opts.share and not cmd_opts.listen: + # fix gradio phoning home + gradio.utils.version_check = lambda: None + gradio.utils.get_local_ip_address = lambda: '127.0.0.1' + +if cmd_opts.ngrok is not None: + import modules.ngrok as ngrok + print('ngrok authtoken detected, trying to connect...') + ngrok.connect( + cmd_opts.ngrok, + cmd_opts.port if cmd_opts.port is not None else 7860, + cmd_opts.ngrok_region + ) + + +def gr_show(visible=True): + return {"visible": visible, "__type__": "update"} + + +sample_img2img = "assets/stable-samples/img2img/sketch-mountains-input.jpg" +sample_img2img = sample_img2img if os.path.exists(sample_img2img) else None + +css_hide_progressbar = """ +.wrap .m-12 svg { display:none!important; } +.wrap .m-12::before { content:"Loading..." } +.wrap .z-20 svg { display:none!important; } +.wrap .z-20::before { content:"Loading..." } +.progress-bar { display:none!important; } +.meta-text { display:none!important; } +.meta-text-center { display:none!important; } +""" + +# Using constants for these since the variation selector isn't visible. +# Important that they exactly match script.js for tooltip to work. +random_symbol = '\U0001f3b2\ufe0f' # 🎲️ +reuse_symbol = '\u267b\ufe0f' # ♻️ +paste_symbol = '\u2199\ufe0f' # ↙ +folder_symbol = '\U0001f4c2' # 📂 +refresh_symbol = '\U0001f504' # 🔄 +save_style_symbol = '\U0001f4be' # 💾 +apply_style_symbol = '\U0001f4cb' # 📋 +clear_prompt_symbol = '\U0001F5D1' # 🗑️ + + +def plaintext_to_html(text): + text = "

" + "
\n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

" + return text + +def send_gradio_gallery_to_image(x): + if len(x) == 0: + return None + return image_from_url_text(x[0]) + +def save_files(js_data, images, do_make_zip, index): + import csv + filenames = [] + fullfns = [] + + #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it + class MyObject: + def __init__(self, d=None): + if d is not None: + for key, value in d.items(): + setattr(self, key, value) + + data = json.loads(js_data) + + p = MyObject(data) + path = opts.outdir_save + save_to_dirs = opts.use_save_to_dirs_for_ui + extension: str = opts.samples_format + start_index = 0 + + if index > -1 and opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only + + images = [images[index]] + start_index = index + + os.makedirs(opts.outdir_save, exist_ok=True) + + with open(os.path.join(opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: + at_start = file.tell() == 0 + writer = csv.writer(file) + if at_start: + writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) + + for image_index, filedata in enumerate(images, start_index): + image = image_from_url_text(filedata) + + is_grid = image_index < p.index_of_first_image + i = 0 if is_grid else (image_index - p.index_of_first_image) + + fullfn, txt_fullfn = save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) + + filename = os.path.relpath(fullfn, path) + filenames.append(filename) + fullfns.append(fullfn) + if txt_fullfn: + filenames.append(os.path.basename(txt_fullfn)) + fullfns.append(txt_fullfn) + + writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) + + # Make Zip + if do_make_zip: + zip_filepath = os.path.join(path, "images.zip") + + from zipfile import ZipFile + with ZipFile(zip_filepath, "w") as zip_file: + for i in range(len(fullfns)): + with open(fullfns[i], mode="rb") as f: + zip_file.writestr(filenames[i], f.read()) + fullfns.insert(0, zip_filepath) + + return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") + + +def calc_time_left(progress, threshold, label, force_display, show_eta): + if progress == 0: + return "" + else: + time_since_start = time.time() - shared.state.time_start + eta = (time_since_start/progress) + eta_relative = eta-time_since_start + if (eta_relative > threshold and show_eta) or force_display: + if eta_relative > 3600: + return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) + elif eta_relative > 60: + return label + time.strftime('%M:%S', time.gmtime(eta_relative)) + else: + return label + time.strftime('%Ss', time.gmtime(eta_relative)) + else: + return "" + + +def check_progress_call(id_part): + if shared.state.job_count == 0: + return "", gr_show(False), gr_show(False), gr_show(False) + + progress = 0 + + if shared.state.job_count > 0: + progress += shared.state.job_no / shared.state.job_count + if shared.state.sampling_steps > 0: + progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps + + # Show progress percentage and time left at the same moment, and base it also on steps done + show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 + + time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) + if time_left != "": + shared.state.time_left_force_display = True + + progress = min(progress, 1) + + progressbar = "" + if opts.show_progressbar: + progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" + + image = gr_show(False) + preview_visibility = gr_show(False) + + if opts.show_progress_every_n_steps != 0: + shared.state.set_current_image() + image = shared.state.current_image + + if image is None: + image = gr.update(value=None) + else: + preview_visibility = gr_show(True) + + if shared.state.textinfo is not None: + textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) + else: + textinfo_result = gr_show(False) + + return f"

{progressbar}

", preview_visibility, image, textinfo_result + + +def check_progress_call_initial(id_part): + shared.state.job_count = -1 + shared.state.current_latent = None + shared.state.current_image = None + shared.state.textinfo = None + shared.state.time_start = time.time() + shared.state.time_left_force_display = False + + return check_progress_call(id_part) + + +def visit(x, func, path=""): + if hasattr(x, 'children'): + for c in x.children: + visit(c, func, path) + elif x.label is not None: + func(path + "/" + str(x.label), x) + + +def add_style(name: str, prompt: str, negative_prompt: str): + if name is None: + return [gr_show() for x in range(4)] + + style = modules.styles.PromptStyle(name, prompt, negative_prompt) + shared.prompt_styles.styles[style.name] = style + # Save all loaded prompt styles: this allows us to update the storage format in the future more easily, because we + # reserialize all styles every time we save them + shared.prompt_styles.save_styles(shared.styles_filename) + + return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] + + +def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): + from modules import processing, devices + + if not enable: + return "" + + p = processing.StableDiffusionProcessingTxt2Img(width=width, height=height, enable_hr=True, hr_scale=hr_scale, hr_resize_x=hr_resize_x, hr_resize_y=hr_resize_y) + + with devices.autocast(): + p.init([""], [0], [0]) + + return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" + + +def apply_styles(prompt, prompt_neg, style1_name, style2_name): + prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) + prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, [style1_name, style2_name]) + + return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] + + +def interrogate(image): + prompt = shared.interrogator.interrogate(image.convert("RGB")) + + return gr_show(True) if prompt is None else prompt + + +def interrogate_deepbooru(image): + prompt = deepbooru.model.tag(image) + return gr_show(True) if prompt is None else prompt + + +def create_seed_inputs(target_interface): + with FormRow(elem_id=target_interface + '_seed_row'): + seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') + seed.style(container=False) + random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') + reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') + + with gr.Group(elem_id=target_interface + '_subseed_show_box'): + seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) + + # Components to show/hide based on the 'Extra' checkbox + seed_extras = [] + + with FormRow(visible=False, elem_id=target_interface + '_subseed_row') as seed_extra_row_1: + seed_extras.append(seed_extra_row_1) + subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') + subseed.style(container=False) + random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') + reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') + subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') + + with FormRow(visible=False) as seed_extra_row_2: + seed_extras.append(seed_extra_row_2) + seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') + seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') + + random_seed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[seed]) + random_subseed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[subseed]) + + def change_visibility(show): + return {comp: gr_show(show) for comp in seed_extras} + + seed_checkbox.change(change_visibility, show_progress=False, inputs=[seed_checkbox], outputs=seed_extras) + + return seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox + + + +def connect_clear_prompt(button): + """Given clear button, prompt, and token_counter objects, setup clear prompt button click event""" + button.click( + _js="clear_prompt", + fn=None, + inputs=[], + outputs=[], + ) + + +def connect_reuse_seed(seed: gr.Number, reuse_seed: gr.Button, generation_info: gr.Textbox, dummy_component, is_subseed): + """ Connects a 'reuse (sub)seed' button's click event so that it copies last used + (sub)seed value from generation info the to the seed field. If copying subseed and subseed strength + was 0, i.e. no variation seed was used, it copies the normal seed value instead.""" + def copy_seed(gen_info_string: str, index): + res = -1 + + try: + gen_info = json.loads(gen_info_string) + index -= gen_info.get('index_of_first_image', 0) + + if is_subseed and gen_info.get('subseed_strength', 0) > 0: + all_subseeds = gen_info.get('all_subseeds', [-1]) + res = all_subseeds[index if 0 <= index < len(all_subseeds) else 0] + else: + all_seeds = gen_info.get('all_seeds', [-1]) + res = all_seeds[index if 0 <= index < len(all_seeds) else 0] + + except json.decoder.JSONDecodeError as e: + if gen_info_string != '': + print("Error parsing JSON generation info:", file=sys.stderr) + print(gen_info_string, file=sys.stderr) + + return [res, gr_show(False)] + + reuse_seed.click( + fn=copy_seed, + _js="(x, y) => [x, selected_gallery_index()]", + show_progress=False, + inputs=[generation_info, dummy_component], + outputs=[seed, dummy_component] + ) + + +def update_token_counter(text, steps): + try: + _, prompt_flat_list, _ = prompt_parser.get_multicond_prompt_list([text]) + prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules(prompt_flat_list, steps) + + except Exception: + # a parsing error can happen here during typing, and we don't want to bother the user with + # messages related to it in console + prompt_schedules = [[[steps, text]]] + + flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) + prompts = [prompt_text for step, prompt_text in flat_prompts] + token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) + style_class = ' class="red"' if (token_count > max_length) else "" + return f"{token_count}/{max_length}" + + +def create_toprow(is_img2img): + id_part = "img2img" if is_img2img else "txt2img" + + with gr.Row(elem_id="toprow"): + with gr.Column(scale=6): + with gr.Row(): + with gr.Column(scale=80): + with gr.Row(): + prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, + placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)" + ) + + with gr.Row(): + with gr.Column(scale=80): + with gr.Row(): + negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, + placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)" + ) + + with gr.Column(scale=1, elem_id="roll_col"): + paste = gr.Button(value=paste_symbol, elem_id="paste") + save_style = gr.Button(value=save_style_symbol, elem_id="style_create") + prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") + clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") + token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") + token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") + + clear_prompt_button.click( + fn=lambda *x: x, + _js="confirm_clear_prompt", + inputs=[prompt, negative_prompt], + outputs=[prompt, negative_prompt], + ) + + button_interrogate = None + button_deepbooru = None + if is_img2img: + with gr.Column(scale=1, elem_id="interrogate_col"): + button_interrogate = gr.Button('Interrogate\nCLIP', elem_id="interrogate") + button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") + + with gr.Column(scale=1): + with gr.Row(): + skip = gr.Button('Skip', elem_id=f"{id_part}_skip") + interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") + submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') + + skip.click( + fn=lambda: shared.state.skip(), + inputs=[], + outputs=[], + ) + + interrupt.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + with gr.Row(): + with gr.Column(scale=1, elem_id="style_pos_col"): + prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) + + with gr.Column(scale=1, elem_id="style_neg_col"): + prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) + + return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button + + +def setup_progressbar(progressbar, preview, id_part, textinfo=None): + if textinfo is None: + textinfo = gr.HTML(visible=False) + + check_progress = gr.Button('Check progress', elem_id=f"{id_part}_check_progress", visible=False) + check_progress.click( + fn=lambda: check_progress_call(id_part), + show_progress=False, + inputs=[], + outputs=[progressbar, preview, preview, textinfo], + ) + + check_progress_initial = gr.Button('Check progress (first)', elem_id=f"{id_part}_check_progress_initial", visible=False) + check_progress_initial.click( + fn=lambda: check_progress_call_initial(id_part), + show_progress=False, + inputs=[], + outputs=[progressbar, preview, preview, textinfo], + ) + + +def apply_setting(key, value): + if value is None: + return gr.update() + + if shared.cmd_opts.freeze_settings: + return gr.update() + + # dont allow model to be swapped when model hash exists in prompt + if key == "sd_model_checkpoint" and opts.disable_weights_auto_swap: + return gr.update() + + if key == "sd_model_checkpoint": + ckpt_info = sd_models.get_closet_checkpoint_match(value) + + if ckpt_info is not None: + value = ckpt_info.title + else: + return gr.update() + + comp_args = opts.data_labels[key].component_args + if comp_args and isinstance(comp_args, dict) and comp_args.get('visible') is False: + return + + valtype = type(opts.data_labels[key].default) + oldval = opts.data.get(key, None) + opts.data[key] = valtype(value) if valtype != type(None) else value + if oldval != value and opts.data_labels[key].onchange is not None: + opts.data_labels[key].onchange() + + opts.save(shared.config_filename) + return value + + +def update_generation_info(args): + generation_info, html_info, img_index = args + try: + generation_info = json.loads(generation_info) + if img_index < 0 or img_index >= len(generation_info["infotexts"]): + return html_info + return plaintext_to_html(generation_info["infotexts"][img_index]) + except Exception: + pass + # if the json parse or anything else fails, just return the old html_info + return html_info + + +def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): + def refresh(): + refresh_method() + args = refreshed_args() if callable(refreshed_args) else refreshed_args + + for k, v in args.items(): + setattr(refresh_component, k, v) + + return gr.update(**(args or {})) + + refresh_button = ToolButton(value=refresh_symbol, elem_id=elem_id) + refresh_button.click( + fn=refresh, + inputs=[], + outputs=[refresh_component] + ) + return refresh_button + + +def create_output_panel(tabname, outdir): + def open_folder(f): + if not os.path.exists(f): + print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') + return + elif not os.path.isdir(f): + print(f""" +WARNING +An open_folder request was made with an argument that is not a folder. +This could be an error or a malicious attempt to run code on your computer. +Requested path was: {f} +""", file=sys.stderr) + return + + if not shared.cmd_opts.hide_ui_dir_config: + path = os.path.normpath(f) + if platform.system() == "Windows": + os.startfile(path) + elif platform.system() == "Darwin": + sp.Popen(["open", path]) + elif "microsoft-standard-WSL2" in platform.uname().release: + sp.Popen(["wsl-open", path]) + else: + sp.Popen(["xdg-open", path]) + + with gr.Column(variant='panel'): + with gr.Group(): + result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) + + generation_info = None + with gr.Column(): + with gr.Row(elem_id=f"image_buttons_{tabname}"): + open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') + + if tabname != "extras": + save = gr.Button('Save', elem_id=f'save_{tabname}') + save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') + + buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) + + open_folder_button.click( + fn=lambda: open_folder(opts.outdir_samples or outdir), + inputs=[], + outputs=[], + ) + + if tabname != "extras": + with gr.Row(): + download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') + + with gr.Group(): + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') + + generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') + if tabname == 'txt2img' or tabname == 'img2img': + generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") + generation_info_button.click( + fn=update_generation_info, + _js="(x, y) => [x, y, selected_gallery_index()]", + inputs=[generation_info, html_info], + outputs=[html_info], + preprocess=False + ) + + save.click( + fn=wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ] + ) + + save_zip.click( + fn=wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ] + ) + + else: + html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') + + parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) + return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log + + +def create_sampler_and_steps_selection(choices, tabname): + if opts.samplers_in_dropdown: + with FormRow(elem_id=f"sampler_selection_{tabname}"): + sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) + else: + with FormGroup(elem_id=f"sampler_selection_{tabname}"): + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) + sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + + return steps, sampler_index + + +def ordered_ui_categories(): + user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} + + for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): + yield category + + +def create_ui(): + import modules.img2img + import modules.txt2img + + reload_javascript() + + parameters_copypaste.reset() + + modules.scripts.scripts_current = modules.scripts.scripts_txt2img + modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) + + with gr.Blocks(analytics_enabled=False) as txt2img_interface: + txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) + + dummy_component = gr.Label(visible=False) + txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) + + with gr.Row(elem_id='txt2img_progress_row'): + with gr.Column(scale=1): + pass + + with gr.Column(scale=1): + progressbar = gr.HTML(elem_id="txt2img_progressbar") + txt2img_preview = gr.Image(elem_id='txt2img_preview', visible=False) + setup_progressbar(progressbar, txt2img_preview, 'txt2img') + + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel', elem_id="txt2img_settings"): + for category in ordered_ui_categories(): + if category == "sampler": + steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") + + elif category == "dimensions": + with FormRow(): + with gr.Column(elem_id="txt2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") + + if opts.dimensions_and_batch_together: + with gr.Column(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") + + elif category == "cfg": + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") + + elif category == "seed": + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') + + elif category == "checkboxes": + with FormRow(elem_id="txt2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") + enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") + hr_final_resolution = FormHTML(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) + + elif category == "hires_fix": + with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: + with FormRow(elem_id="txt2img_hires_fix_row1"): + hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) + hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") + + with FormRow(elem_id="txt2img_hires_fix_row2"): + hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") + hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") + hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") + + elif category == "batch": + if not opts.dimensions_and_batch_together: + with FormRow(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") + + elif category == "scripts": + with FormGroup(elem_id="txt2img_script_container"): + custom_inputs = modules.scripts.scripts_txt2img.setup_ui() + + hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] + for input in hr_resolution_preview_inputs: + input.change( + fn=calc_resolution_hires, + inputs=hr_resolution_preview_inputs, + outputs=[hr_final_resolution], + show_progress=False, + ) + input.change( + None, + _js="onCalcResolutionHires", + inputs=hr_resolution_preview_inputs, + outputs=[], + show_progress=False, + ) + + txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) + parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) + + connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) + connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) + + txt2img_args = dict( + fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), + _js="submit", + inputs=[ + txt2img_prompt, + txt2img_negative_prompt, + txt2img_prompt_style, + txt2img_prompt_style2, + steps, + sampler_index, + restore_faces, + tiling, + batch_count, + batch_size, + cfg_scale, + seed, + subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, + height, + width, + enable_hr, + denoising_strength, + hr_scale, + hr_upscaler, + hr_second_pass_steps, + hr_resize_x, + hr_resize_y, + ] + custom_inputs, + + outputs=[ + txt2img_gallery, + generation_info, + html_info, + html_log, + ], + show_progress=False, + ) + + txt2img_prompt.submit(**txt2img_args) + submit.click(**txt2img_args) + + txt_prompt_img.change( + fn=modules.images.image_data, + inputs=[ + txt_prompt_img + ], + outputs=[ + txt2img_prompt, + txt_prompt_img + ] + ) + + enable_hr.change( + fn=lambda x: gr_show(x), + inputs=[enable_hr], + outputs=[hr_options], + show_progress = False, + ) + + txt2img_paste_fields = [ + (txt2img_prompt, "Prompt"), + (txt2img_negative_prompt, "Negative prompt"), + (steps, "Steps"), + (sampler_index, "Sampler"), + (restore_faces, "Face restoration"), + (cfg_scale, "CFG scale"), + (seed, "Seed"), + (width, "Size-1"), + (height, "Size-2"), + (batch_size, "Batch size"), + (subseed, "Variation seed"), + (subseed_strength, "Variation seed strength"), + (seed_resize_from_w, "Seed resize from-1"), + (seed_resize_from_h, "Seed resize from-2"), + (denoising_strength, "Denoising strength"), + (enable_hr, lambda d: "Denoising strength" in d), + (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), + (hr_scale, "Hires upscale"), + (hr_upscaler, "Hires upscaler"), + (hr_second_pass_steps, "Hires steps"), + (hr_resize_x, "Hires resize-1"), + (hr_resize_y, "Hires resize-2"), + *modules.scripts.scripts_txt2img.infotext_fields + ] + parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) + + txt2img_preview_params = [ + txt2img_prompt, + txt2img_negative_prompt, + steps, + sampler_index, + cfg_scale, + seed, + width, + height, + ] + + token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) + + modules.scripts.scripts_current = modules.scripts.scripts_img2img + modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) + + with gr.Blocks(analytics_enabled=False) as img2img_interface: + img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) + + with gr.Row(elem_id='img2img_progress_row'): + img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) + + with gr.Column(scale=1): + pass + + with gr.Column(scale=1): + progressbar = gr.HTML(elem_id="img2img_progressbar") + img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) + setup_progressbar(progressbar, img2img_preview, 'img2img') + + with FormRow().style(equal_height=False): + with gr.Column(variant='panel', elem_id="img2img_settings"): + + with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: + with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): + init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) + + with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): + init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) + init_img_with_mask_orig = gr.State(None) + + use_color_sketch = cmd_opts.gradio_inpaint_tool == "color-sketch" + if use_color_sketch: + def update_orig(image, state): + if image is not None: + same_size = state is not None and state.size == image.size + has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) + edited = same_size and has_exact_match + return image if not edited or state is None else state + + init_img_with_mask.change(update_orig, [init_img_with_mask, init_img_with_mask_orig], init_img_with_mask_orig) + + init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") + init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") + + with FormRow(): + mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") + mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") + + with FormRow(): + mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") + inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") + + with FormRow(): + inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") + + with FormRow(): + with gr.Column(): + inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") + + with gr.Column(scale=4): + inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") + + with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): + hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' + gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") + img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") + img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") + + with FormRow(): + resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") + + for category in ordered_ui_categories(): + if category == "sampler": + steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") + + elif category == "dimensions": + with FormRow(): + with gr.Column(elem_id="img2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") + + if opts.dimensions_and_batch_together: + with gr.Column(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") + + elif category == "cfg": + with FormGroup(): + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") + + elif category == "seed": + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') + + elif category == "checkboxes": + with FormRow(elem_id="img2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") + + elif category == "batch": + if not opts.dimensions_and_batch_together: + with FormRow(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") + + elif category == "scripts": + with FormGroup(elem_id="img2img_script_container"): + custom_inputs = modules.scripts.scripts_img2img.setup_ui() + + img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) + parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) + + connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) + connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) + + img2img_prompt_img.change( + fn=modules.images.image_data, + inputs=[ + img2img_prompt_img + ], + outputs=[ + img2img_prompt, + img2img_prompt_img + ] + ) + + mask_mode.change( + lambda mode, img: { + init_img_with_mask: gr_show(mode == 0), + init_img_inpaint: gr_show(mode == 1), + init_mask_inpaint: gr_show(mode == 1), + }, + inputs=[mask_mode, init_img_with_mask], + outputs=[ + init_img_with_mask, + init_img_inpaint, + init_mask_inpaint, + ], + ) + + img2img_args = dict( + fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), + _js="submit_img2img", + inputs=[ + dummy_component, + img2img_prompt, + img2img_negative_prompt, + img2img_prompt_style, + img2img_prompt_style2, + init_img, + init_img_with_mask, + init_img_with_mask_orig, + init_img_inpaint, + init_mask_inpaint, + mask_mode, + steps, + sampler_index, + mask_blur, + mask_alpha, + inpainting_fill, + restore_faces, + tiling, + batch_count, + batch_size, + cfg_scale, + denoising_strength, + seed, + subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, + height, + width, + resize_mode, + inpaint_full_res, + inpaint_full_res_padding, + inpainting_mask_invert, + img2img_batch_input_dir, + img2img_batch_output_dir, + ] + custom_inputs, + outputs=[ + img2img_gallery, + generation_info, + html_info, + html_log, + ], + show_progress=False, + ) + + img2img_prompt.submit(**img2img_args) + submit.click(**img2img_args) + + img2img_interrogate.click( + fn=interrogate, + inputs=[init_img], + outputs=[img2img_prompt], + ) + + img2img_deepbooru.click( + fn=interrogate_deepbooru, + inputs=[init_img], + outputs=[img2img_prompt], + ) + + prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] + style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] + style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] + + for button, (prompt, negative_prompt) in zip([txt2img_save_style, img2img_save_style], prompts): + button.click( + fn=add_style, + _js="ask_for_style_name", + # Have to pass empty dummy component here, because the JavaScript and Python function have to accept + # the same number of parameters, but we only know the style-name after the JavaScript prompt + inputs=[dummy_component, prompt, negative_prompt], + outputs=[txt2img_prompt_style, img2img_prompt_style, txt2img_prompt_style2, img2img_prompt_style2], + ) + + for button, (prompt, negative_prompt), (style1, style2), js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): + button.click( + fn=apply_styles, + _js=js_func, + inputs=[prompt, negative_prompt, style1, style2], + outputs=[prompt, negative_prompt, style1, style2], + ) + + token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) + + img2img_paste_fields = [ + (img2img_prompt, "Prompt"), + (img2img_negative_prompt, "Negative prompt"), + (steps, "Steps"), + (sampler_index, "Sampler"), + (restore_faces, "Face restoration"), + (cfg_scale, "CFG scale"), + (seed, "Seed"), + (width, "Size-1"), + (height, "Size-2"), + (batch_size, "Batch size"), + (subseed, "Variation seed"), + (subseed_strength, "Variation seed strength"), + (seed_resize_from_w, "Seed resize from-1"), + (seed_resize_from_h, "Seed resize from-2"), + (denoising_strength, "Denoising strength"), + (mask_blur, "Mask blur"), + *modules.scripts.scripts_img2img.infotext_fields + ] + parameters_copypaste.add_paste_fields("img2img", init_img, img2img_paste_fields) + parameters_copypaste.add_paste_fields("inpaint", init_img_with_mask, img2img_paste_fields) + + modules.scripts.scripts_current = None + + with gr.Blocks(analytics_enabled=False) as extras_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + with gr.Tabs(elem_id="mode_extras"): + with gr.TabItem('Single Image', elem_id="extras_single_tab"): + extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") + + with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): + image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") + + with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): + extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") + extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") + show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") + + submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') + + with gr.Tabs(elem_id="extras_resize_mode"): + with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): + upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") + with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): + with gr.Group(): + with gr.Row(): + upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") + upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") + upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") + + with gr.Group(): + extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") + + with gr.Group(): + extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") + extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") + + with gr.Group(): + gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") + + with gr.Group(): + codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") + codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") + + with gr.Group(): + upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") + + result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) + + submit.click( + fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), + _js="get_extras_tab_index", + inputs=[ + dummy_component, + dummy_component, + extras_image, + image_batch, + extras_batch_input_dir, + extras_batch_output_dir, + show_extras_results, + gfpgan_visibility, + codeformer_visibility, + codeformer_weight, + upscaling_resize, + upscaling_resize_w, + upscaling_resize_h, + upscaling_crop, + extras_upscaler_1, + extras_upscaler_2, + extras_upscaler_2_visibility, + upscale_before_face_fix, + ], + outputs=[ + result_images, + html_info_x, + html_info, + ] + ) + parameters_copypaste.add_paste_fields("extras", extras_image, None) + + extras_image.change( + fn=modules.extras.clear_cache, + inputs=[], outputs=[] + ) + + with gr.Blocks(analytics_enabled=False) as pnginfo_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + image = gr.Image(elem_id="pnginfo_image", label="Source", source="upload", interactive=True, type="pil") + + with gr.Column(variant='panel'): + html = gr.HTML() + generation_info = gr.Textbox(visible=False, elem_id="pnginfo_generation_info") + html2 = gr.HTML() + with gr.Row(): + buttons = parameters_copypaste.create_buttons(["txt2img", "img2img", "inpaint", "extras"]) + parameters_copypaste.bind_buttons(buttons, image, generation_info) + + image.change( + fn=wrap_gradio_call(modules.extras.run_pnginfo), + inputs=[image], + outputs=[html, generation_info, html2], + ) + + with gr.Blocks(analytics_enabled=False) as modelmerger_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") + + with gr.Row(): + primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") + create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") + + secondary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_secondary_model_name", label="Secondary model (B)") + create_refresh_button(secondary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_B") + + tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") + create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") + + custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") + interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") + interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") + + with gr.Row(): + checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") + save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") + + modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') + + with gr.Column(variant='panel'): + submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) + + with gr.Blocks(analytics_enabled=False) as train_interface: + with gr.Row().style(equal_height=False): + gr.HTML(value="

See wiki for detailed explanation.

") + + with gr.Row().style(equal_height=False): + with gr.Tabs(elem_id="train_tabs"): + + with gr.Tab(label="Create embedding"): + new_embedding_name = gr.Textbox(label="Name", elem_id="train_new_embedding_name") + initialization_text = gr.Textbox(label="Initialization text", value="*", elem_id="train_initialization_text") + nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1, elem_id="train_nvpt") + overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding", elem_id="train_overwrite_old_embedding") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + create_embedding = gr.Button(value="Create embedding", variant='primary', elem_id="train_create_embedding") + + with gr.Tab(label="Create hypernetwork"): + new_hypernetwork_name = gr.Textbox(label="Name", elem_id="train_new_hypernetwork_name") + new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"], elem_id="train_new_hypernetwork_sizes") + new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'", elem_id="train_new_hypernetwork_layer_structure") + new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys, elem_id="train_new_hypernetwork_activation_func") + new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") + new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") + new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") + new_hypernetwork_dropout_structure = gr.Textbox("0, 0, 0", label="Enter hypernetwork Dropout structure (or empty). Recommended : 0~0.35 incrementing sequence: 0, 0.05, 0.15", placeholder="1st and last digit must be 0 and values should be between 0 and 1. ex:'0, 0.01, 0'") + overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary', elem_id="train_create_hypernetwork") + + with gr.Tab(label="Preprocess images"): + process_src = gr.Textbox(label='Source directory', elem_id="train_process_src") + process_dst = gr.Textbox(label='Destination directory', elem_id="train_process_dst") + process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_process_width") + process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_process_height") + preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"], elem_id="train_preprocess_txt_action") + + with gr.Row(): + process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") + process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") + process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") + process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") + process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") + + with gr.Row(visible=False) as process_split_extra_row: + process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_split_threshold") + process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05, elem_id="train_process_overlap_ratio") + + with gr.Row(visible=False) as process_focal_crop_row: + process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_face_weight") + process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") + process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") + process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + with gr.Row(): + interrupt_preprocessing = gr.Button("Interrupt", elem_id="train_interrupt_preprocessing") + run_preprocess = gr.Button(value="Preprocess", variant='primary', elem_id="train_run_preprocess") + + process_split.change( + fn=lambda show: gr_show(show), + inputs=[process_split], + outputs=[process_split_extra_row], + ) + + process_focal_crop.change( + fn=lambda show: gr_show(show), + inputs=[process_focal_crop], + outputs=[process_focal_crop_row], + ) + + def get_textual_inversion_template_names(): + return sorted([x for x in textual_inversion.textual_inversion_templates]) + + with gr.Tab(label="Train"): + gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") + with FormRow(): + train_embedding_name = gr.Dropdown(label='Embedding', elem_id="train_embedding", choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())) + create_refresh_button(train_embedding_name, sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings, lambda: {"choices": sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())}, "refresh_train_embedding_name") + + train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) + create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") + + with FormRow(): + embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") + hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") + + with FormRow(): + clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) + clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) + + with FormRow(): + batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") + gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") + + dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") + log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") + + with FormRow(): + template_file = gr.Dropdown(label='Prompt template', value="style_filewords.txt", elem_id="train_template_file", choices=get_textual_inversion_template_names()) + create_refresh_button(template_file, textual_inversion.list_textual_inversion_templates, lambda: {"choices": get_textual_inversion_template_names()}, "refrsh_train_template_file") + + training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") + training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") + varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") + steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") + + with FormRow(): + create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") + save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") + + save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") + preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") + + shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") + tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") + + latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") + + with gr.Row(): + train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") + interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") + train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") + + params = script_callbacks.UiTrainTabParams(txt2img_preview_params) + + script_callbacks.ui_train_tabs_callback(params) + + with gr.Column(): + progressbar = gr.HTML(elem_id="ti_progressbar") + ti_output = gr.Text(elem_id="ti_output", value="", show_label=False) + + ti_gallery = gr.Gallery(label='Output', show_label=False, elem_id='ti_gallery').style(grid=4) + ti_preview = gr.Image(elem_id='ti_preview', visible=False) + ti_progress = gr.HTML(elem_id="ti_progress", value="") + ti_outcome = gr.HTML(elem_id="ti_error", value="") + setup_progressbar(progressbar, ti_preview, 'ti', textinfo=ti_progress) + + create_embedding.click( + fn=modules.textual_inversion.ui.create_embedding, + inputs=[ + new_embedding_name, + initialization_text, + nvpt, + overwrite_old_embedding, + ], + outputs=[ + train_embedding_name, + ti_output, + ti_outcome, + ] + ) + + create_hypernetwork.click( + fn=modules.hypernetworks.ui.create_hypernetwork, + inputs=[ + new_hypernetwork_name, + new_hypernetwork_sizes, + overwrite_old_hypernetwork, + new_hypernetwork_layer_structure, + new_hypernetwork_activation_func, + new_hypernetwork_initialization_option, + new_hypernetwork_add_layer_norm, + new_hypernetwork_use_dropout, + new_hypernetwork_dropout_structure + ], + outputs=[ + train_hypernetwork_name, + ti_output, + ti_outcome, + ] + ) + + run_preprocess.click( + fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.preprocess, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + process_src, + process_dst, + process_width, + process_height, + preprocess_txt_action, + process_flip, + process_split, + process_caption, + process_caption_deepbooru, + process_split_threshold, + process_overlap_ratio, + process_focal_crop, + process_focal_crop_face_weight, + process_focal_crop_entropy_weight, + process_focal_crop_edges_weight, + process_focal_crop_debug, + ], + outputs=[ + ti_output, + ti_outcome, + ], + ) + + train_embedding.click( + fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.train_embedding, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + train_embedding_name, + embedding_learn_rate, + batch_size, + gradient_step, + dataset_directory, + log_directory, + training_width, + training_height, + varsize, + steps, + clip_grad_mode, + clip_grad_value, + shuffle_tags, + tag_drop_out, + latent_sampling_method, + create_image_every, + save_embedding_every, + template_file, + save_image_with_stored_embedding, + preview_from_txt2img, + *txt2img_preview_params, + ], + outputs=[ + ti_output, + ti_outcome, + ] + ) + + train_hypernetwork.click( + fn=wrap_gradio_gpu_call(modules.hypernetworks.ui.train_hypernetwork, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + train_hypernetwork_name, + hypernetwork_learn_rate, + batch_size, + gradient_step, + dataset_directory, + log_directory, + training_width, + training_height, + varsize, + steps, + clip_grad_mode, + clip_grad_value, + shuffle_tags, + tag_drop_out, + latent_sampling_method, + create_image_every, + save_embedding_every, + template_file, + preview_from_txt2img, + *txt2img_preview_params, + ], + outputs=[ + ti_output, + ti_outcome, + ] + ) + + interrupt_training.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + interrupt_preprocessing.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + def create_setting_component(key, is_quicksettings=False): + def fun(): + return opts.data[key] if key in opts.data else opts.data_labels[key].default + + info = opts.data_labels[key] + t = type(info.default) + + args = info.component_args() if callable(info.component_args) else info.component_args + + if info.component is not None: + comp = info.component + elif t == str: + comp = gr.Textbox + elif t == int: + comp = gr.Number + elif t == bool: + comp = gr.Checkbox + else: + raise Exception(f'bad options item type: {str(t)} for key {key}') + + elem_id = "setting_"+key + + if info.refresh is not None: + if is_quicksettings: + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) + else: + with FormRow(): + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) + else: + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + + return res + + components = [] + component_dict = {} + + script_callbacks.ui_settings_callback() + opts.reorder() + + def run_settings(*args): + changed = [] + + for key, value, comp in zip(opts.data_labels.keys(), args, components): + assert comp == dummy_component or opts.same_type(value, opts.data_labels[key].default), f"Bad value for setting {key}: {value}; expecting {type(opts.data_labels[key].default).__name__}" + + for key, value, comp in zip(opts.data_labels.keys(), args, components): + if comp == dummy_component: + continue + + if opts.set(key, value): + changed.append(key) + + try: + opts.save(shared.config_filename) + except RuntimeError: + return opts.dumpjson(), f'{len(changed)} settings changed without save: {", ".join(changed)}.' + return opts.dumpjson(), f'{len(changed)} settings changed{": " if len(changed) > 0 else ""}{", ".join(changed)}.' + + def run_settings_single(value, key): + if not opts.same_type(value, opts.data_labels[key].default): + return gr.update(visible=True), opts.dumpjson() + + if not opts.set(key, value): + return gr.update(value=getattr(opts, key)), opts.dumpjson() + + opts.save(shared.config_filename) + + return gr.update(value=value), opts.dumpjson() + + with gr.Blocks(analytics_enabled=False) as settings_interface: + with gr.Row(): + with gr.Column(scale=6): + settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") + with gr.Column(): + restart_gradio = gr.Button(value='Reload UI', variant='primary', elem_id="settings_restart_gradio") + + result = gr.HTML(elem_id="settings_result") + + quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] + quicksettings_names = {x: i for i, x in enumerate(quicksettings_names) if x != 'quicksettings'} + + quicksettings_list = [] + + previous_section = None + current_tab = None + with gr.Tabs(elem_id="settings"): + for i, (k, item) in enumerate(opts.data_labels.items()): + section_must_be_skipped = item.section[0] is None + + if previous_section != item.section and not section_must_be_skipped: + elem_id, text = item.section + + if current_tab is not None: + current_tab.__exit__() + + current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) + current_tab.__enter__() + + previous_section = item.section + + if k in quicksettings_names and not shared.cmd_opts.freeze_settings: + quicksettings_list.append((i, k, item)) + components.append(dummy_component) + elif section_must_be_skipped: + components.append(dummy_component) + else: + component = create_setting_component(k) + component_dict[k] = component + components.append(component) + + if current_tab is not None: + current_tab.__exit__() + + with gr.TabItem("Actions"): + request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") + download_localization = gr.Button(value='Download localization template', elem_id="download_localization") + reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") + + if os.path.exists("html/licenses.html"): + with open("html/licenses.html", encoding="utf8") as file: + with gr.TabItem("Licenses"): + gr.HTML(file.read(), elem_id="licenses") + + gr.Button(value="Show all pages", elem_id="settings_show_all_pages") + + request_notifications.click( + fn=lambda: None, + inputs=[], + outputs=[], + _js='function(){}' + ) + + download_localization.click( + fn=lambda: None, + inputs=[], + outputs=[], + _js='download_localization' + ) + + def reload_scripts(): + modules.scripts.reload_script_body_only() + reload_javascript() # need to refresh the html page + + reload_script_bodies.click( + fn=reload_scripts, + inputs=[], + outputs=[] + ) + + def request_restart(): + shared.state.interrupt() + shared.state.need_restart = True + + restart_gradio.click( + fn=request_restart, + _js='restart_reload', + inputs=[], + outputs=[], + ) + + interfaces = [ + (txt2img_interface, "txt2img", "txt2img"), + (img2img_interface, "img2img", "img2img"), + (extras_interface, "Extras", "extras"), + (pnginfo_interface, "PNG Info", "pnginfo"), + (modelmerger_interface, "Checkpoint Merger", "modelmerger"), + (train_interface, "Train", "ti"), + ] + + css = "" + + for cssfile in modules.scripts.list_files_with_name("style.css"): + if not os.path.isfile(cssfile): + continue + + with open(cssfile, "r", encoding="utf8") as file: + css += file.read() + "\n" + + if os.path.exists(os.path.join(script_path, "user.css")): + with open(os.path.join(script_path, "user.css"), "r", encoding="utf8") as file: + css += file.read() + "\n" + + if not cmd_opts.no_progressbar_hiding: + css += css_hide_progressbar + + interfaces += script_callbacks.ui_tabs_callback() + interfaces += [(settings_interface, "Settings", "settings")] + + extensions_interface = ui_extensions.create_ui() + interfaces += [(extensions_interface, "Extensions", "extensions")] + + with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: + with gr.Row(elem_id="quicksettings"): + for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): + component = create_setting_component(k, is_quicksettings=True) + component_dict[k] = component + + parameters_copypaste.integrate_settings_paste_fields(component_dict) + parameters_copypaste.run_bind() + + with gr.Tabs(elem_id="tabs") as tabs: + for interface, label, ifid in interfaces: + with gr.TabItem(label, id=ifid, elem_id='tab_' + ifid): + interface.render() + + if os.path.exists(os.path.join(script_path, "notification.mp3")): + audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) + + if os.path.exists("html/footer.html"): + with open("html/footer.html", encoding="utf8") as file: + footer = file.read() + footer = footer.format(versions=versions_html()) + gr.HTML(footer, elem_id="footer") + + text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) + settings_submit.click( + fn=wrap_gradio_call(run_settings, extra_outputs=[gr.update()]), + inputs=components, + outputs=[text_settings, result], + ) + + for i, k, item in quicksettings_list: + component = component_dict[k] + + component.change( + fn=lambda value, k=k: run_settings_single(value, key=k), + inputs=[component], + outputs=[component, text_settings], + ) + + component_keys = [k for k in opts.data_labels.keys() if k in component_dict] + + def get_settings_values(): + return [getattr(opts, key) for key in component_keys] + + demo.load( + fn=get_settings_values, + inputs=[], + outputs=[component_dict[k] for k in component_keys], + ) + + def modelmerger(*args): + try: + results = modules.extras.run_modelmerger(*args) + except Exception as e: + print("Error loading/saving model file:", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + modules.sd_models.list_models() # to remove the potentially missing models from the list + return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] + return results + + modelmerger_merge.click( + fn=modelmerger, + inputs=[ + primary_model_name, + secondary_model_name, + tertiary_model_name, + interp_method, + interp_amount, + save_as_half, + custom_name, + checkpoint_format, + ], + outputs=[ + submit_result, + primary_model_name, + secondary_model_name, + tertiary_model_name, + component_dict['sd_model_checkpoint'], + ] + ) + + ui_config_file = cmd_opts.ui_config_file + ui_settings = {} + settings_count = len(ui_settings) + error_loading = False + + try: + if os.path.exists(ui_config_file): + with open(ui_config_file, "r", encoding="utf8") as file: + ui_settings = json.load(file) + except Exception: + error_loading = True + print("Error loading settings:", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + + def loadsave(path, x): + def apply_field(obj, field, condition=None, init_field=None): + key = path + "/" + field + + if getattr(obj, 'custom_script_source', None) is not None: + key = 'customscript/' + obj.custom_script_source + '/' + key + + if getattr(obj, 'do_not_save_to_config', False): + return + + saved_value = ui_settings.get(key, None) + if saved_value is None: + ui_settings[key] = getattr(obj, field) + elif condition and not condition(saved_value): + print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') + else: + setattr(obj, field, saved_value) + if init_field is not None: + init_field(saved_value) + + if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number, gr.Dropdown] and x.visible: + apply_field(x, 'visible') + + if type(x) == gr.Slider: + apply_field(x, 'value') + apply_field(x, 'minimum') + apply_field(x, 'maximum') + apply_field(x, 'step') + + if type(x) == gr.Radio: + apply_field(x, 'value', lambda val: val in x.choices) + + if type(x) == gr.Checkbox: + apply_field(x, 'value') + + if type(x) == gr.Textbox: + apply_field(x, 'value') + + if type(x) == gr.Number: + apply_field(x, 'value') + + if type(x) == gr.Dropdown: + apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) + + visit(txt2img_interface, loadsave, "txt2img") + visit(img2img_interface, loadsave, "img2img") + visit(extras_interface, loadsave, "extras") + visit(modelmerger_interface, loadsave, "modelmerger") + visit(train_interface, loadsave, "train") + + if not error_loading and (not os.path.exists(ui_config_file) or settings_count != len(ui_settings)): + with open(ui_config_file, "w", encoding="utf8") as file: + json.dump(ui_settings, file, indent=4) + + return demo + + +def reload_javascript(): + with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: + javascript = f'' + + scripts_list = modules.scripts.list_scripts("javascript", ".js") + + for basedir, filename, path in scripts_list: + with open(path, "r", encoding="utf8") as jsfile: + javascript += f"\n" + + if cmd_opts.theme is not None: + javascript += f"\n\n" + + javascript += f"\n" + + def template_response(*args, **kwargs): + res = shared.GradioTemplateResponseOriginal(*args, **kwargs) + res.body = res.body.replace( + b'', f'{javascript}'.encode("utf8")) + res.init_headers() + return res + + gradio.routes.templates.TemplateResponse = template_response + + +if not hasattr(shared, 'GradioTemplateResponseOriginal'): + shared.GradioTemplateResponseOriginal = gradio.routes.templates.TemplateResponse + + +def versions_html(): + import torch + import launch + + python_version = ".".join([str(x) for x in sys.version_info[0:3]]) + commit = launch.commit_hash() + short_commit = commit[0:8] + + if shared.xformers_available: + import xformers + xformers_version = xformers.__version__ + else: + xformers_version = "N/A" + + return f""" +python: {python_version} + •  +torch: {torch.__version__} + •  +xformers: {xformers_version} + •  +gradio: {gr.__version__} + •  +commit: {short_commit} +""" diff --git a/modules/ui.py b/modules/ui.py deleted file mode 100644 index 9b9081b5..00000000 --- a/modules/ui.py +++ /dev/null @@ -1,1928 +0,0 @@ -import html -import json -import math -import mimetypes -import os -import platform -import random -import subprocess as sp -import sys -import tempfile -import time -import traceback -from functools import partial, reduce - -import gradio as gr -import gradio.routes -import gradio.utils -import numpy as np -from PIL import Image, PngImagePlugin -from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call - -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru -from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML -from modules.paths import script_path - -from modules.shared import opts, cmd_opts, restricted_opts - -import modules.codeformer_model -import modules.generation_parameters_copypaste as parameters_copypaste -import modules.gfpgan_model -import modules.hypernetworks.ui -import modules.scripts -import modules.shared as shared -import modules.styles -import modules.textual_inversion.ui -from modules import prompt_parser -from modules.images import save_image -from modules.sd_hijack import model_hijack -from modules.sd_samplers import samplers, samplers_for_img2img -from modules.textual_inversion import textual_inversion -import modules.hypernetworks.ui -from modules.generation_parameters_copypaste import image_from_url_text - -# this is a fix for Windows users. Without it, javascript files will be served with text/html content-type and the browser will not show any UI -mimetypes.init() -mimetypes.add_type('application/javascript', '.js') - -if not cmd_opts.share and not cmd_opts.listen: - # fix gradio phoning home - gradio.utils.version_check = lambda: None - gradio.utils.get_local_ip_address = lambda: '127.0.0.1' - -if cmd_opts.ngrok is not None: - import modules.ngrok as ngrok - print('ngrok authtoken detected, trying to connect...') - ngrok.connect( - cmd_opts.ngrok, - cmd_opts.port if cmd_opts.port is not None else 7860, - cmd_opts.ngrok_region - ) - - -def gr_show(visible=True): - return {"visible": visible, "__type__": "update"} - - -sample_img2img = "assets/stable-samples/img2img/sketch-mountains-input.jpg" -sample_img2img = sample_img2img if os.path.exists(sample_img2img) else None - -css_hide_progressbar = """ -.wrap .m-12 svg { display:none!important; } -.wrap .m-12::before { content:"Loading..." } -.wrap .z-20 svg { display:none!important; } -.wrap .z-20::before { content:"Loading..." } -.progress-bar { display:none!important; } -.meta-text { display:none!important; } -.meta-text-center { display:none!important; } -""" - -# Using constants for these since the variation selector isn't visible. -# Important that they exactly match script.js for tooltip to work. -random_symbol = '\U0001f3b2\ufe0f' # 🎲️ -reuse_symbol = '\u267b\ufe0f' # ♻️ -paste_symbol = '\u2199\ufe0f' # ↙ -folder_symbol = '\U0001f4c2' # 📂 -refresh_symbol = '\U0001f504' # 🔄 -save_style_symbol = '\U0001f4be' # 💾 -apply_style_symbol = '\U0001f4cb' # 📋 -clear_prompt_symbol = '\U0001F5D1' # 🗑️ - - -def plaintext_to_html(text): - text = "

" + "
\n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

" - return text - -def send_gradio_gallery_to_image(x): - if len(x) == 0: - return None - return image_from_url_text(x[0]) - -def save_files(js_data, images, do_make_zip, index): - import csv - filenames = [] - fullfns = [] - - #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it - class MyObject: - def __init__(self, d=None): - if d is not None: - for key, value in d.items(): - setattr(self, key, value) - - data = json.loads(js_data) - - p = MyObject(data) - path = opts.outdir_save - save_to_dirs = opts.use_save_to_dirs_for_ui - extension: str = opts.samples_format - start_index = 0 - - if index > -1 and opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only - - images = [images[index]] - start_index = index - - os.makedirs(opts.outdir_save, exist_ok=True) - - with open(os.path.join(opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: - at_start = file.tell() == 0 - writer = csv.writer(file) - if at_start: - writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) - - for image_index, filedata in enumerate(images, start_index): - image = image_from_url_text(filedata) - - is_grid = image_index < p.index_of_first_image - i = 0 if is_grid else (image_index - p.index_of_first_image) - - fullfn, txt_fullfn = save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) - - filename = os.path.relpath(fullfn, path) - filenames.append(filename) - fullfns.append(fullfn) - if txt_fullfn: - filenames.append(os.path.basename(txt_fullfn)) - fullfns.append(txt_fullfn) - - writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) - - # Make Zip - if do_make_zip: - zip_filepath = os.path.join(path, "images.zip") - - from zipfile import ZipFile - with ZipFile(zip_filepath, "w") as zip_file: - for i in range(len(fullfns)): - with open(fullfns[i], mode="rb") as f: - zip_file.writestr(filenames[i], f.read()) - fullfns.insert(0, zip_filepath) - - return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") - - -def calc_time_left(progress, threshold, label, force_display, show_eta): - if progress == 0: - return "" - else: - time_since_start = time.time() - shared.state.time_start - eta = (time_since_start/progress) - eta_relative = eta-time_since_start - if (eta_relative > threshold and show_eta) or force_display: - if eta_relative > 3600: - return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) - elif eta_relative > 60: - return label + time.strftime('%M:%S', time.gmtime(eta_relative)) - else: - return label + time.strftime('%Ss', time.gmtime(eta_relative)) - else: - return "" - - -def check_progress_call(id_part): - if shared.state.job_count == 0: - return "", gr_show(False), gr_show(False), gr_show(False) - - progress = 0 - - if shared.state.job_count > 0: - progress += shared.state.job_no / shared.state.job_count - if shared.state.sampling_steps > 0: - progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps - - # Show progress percentage and time left at the same moment, and base it also on steps done - show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 - - time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) - if time_left != "": - shared.state.time_left_force_display = True - - progress = min(progress, 1) - - progressbar = "" - if opts.show_progressbar: - progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" - - image = gr_show(False) - preview_visibility = gr_show(False) - - if opts.show_progress_every_n_steps != 0: - shared.state.set_current_image() - image = shared.state.current_image - - if image is None: - image = gr.update(value=None) - else: - preview_visibility = gr_show(True) - - if shared.state.textinfo is not None: - textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) - else: - textinfo_result = gr_show(False) - - return f"

{progressbar}

", preview_visibility, image, textinfo_result - - -def check_progress_call_initial(id_part): - shared.state.job_count = -1 - shared.state.current_latent = None - shared.state.current_image = None - shared.state.textinfo = None - shared.state.time_start = time.time() - shared.state.time_left_force_display = False - - return check_progress_call(id_part) - - -def visit(x, func, path=""): - if hasattr(x, 'children'): - for c in x.children: - visit(c, func, path) - elif x.label is not None: - func(path + "/" + str(x.label), x) - - -def add_style(name: str, prompt: str, negative_prompt: str): - if name is None: - return [gr_show() for x in range(4)] - - style = modules.styles.PromptStyle(name, prompt, negative_prompt) - shared.prompt_styles.styles[style.name] = style - # Save all loaded prompt styles: this allows us to update the storage format in the future more easily, because we - # reserialize all styles every time we save them - shared.prompt_styles.save_styles(shared.styles_filename) - - return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] - - -def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): - from modules import processing, devices - - if not enable: - return "" - - p = processing.StableDiffusionProcessingTxt2Img(width=width, height=height, enable_hr=True, hr_scale=hr_scale, hr_resize_x=hr_resize_x, hr_resize_y=hr_resize_y) - - with devices.autocast(): - p.init([""], [0], [0]) - - return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" - - -def apply_styles(prompt, prompt_neg, style1_name, style2_name): - prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) - prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, [style1_name, style2_name]) - - return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] - - -def interrogate(image): - prompt = shared.interrogator.interrogate(image.convert("RGB")) - - return gr_show(True) if prompt is None else prompt - - -def interrogate_deepbooru(image): - prompt = deepbooru.model.tag(image) - return gr_show(True) if prompt is None else prompt - - -def create_seed_inputs(target_interface): - with FormRow(elem_id=target_interface + '_seed_row'): - seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') - seed.style(container=False) - random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') - reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') - - with gr.Group(elem_id=target_interface + '_subseed_show_box'): - seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) - - # Components to show/hide based on the 'Extra' checkbox - seed_extras = [] - - with FormRow(visible=False, elem_id=target_interface + '_subseed_row') as seed_extra_row_1: - seed_extras.append(seed_extra_row_1) - subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') - subseed.style(container=False) - random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') - reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') - subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') - - with FormRow(visible=False) as seed_extra_row_2: - seed_extras.append(seed_extra_row_2) - seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') - seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') - - random_seed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[seed]) - random_subseed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[subseed]) - - def change_visibility(show): - return {comp: gr_show(show) for comp in seed_extras} - - seed_checkbox.change(change_visibility, show_progress=False, inputs=[seed_checkbox], outputs=seed_extras) - - return seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox - - - -def connect_clear_prompt(button): - """Given clear button, prompt, and token_counter objects, setup clear prompt button click event""" - button.click( - _js="clear_prompt", - fn=None, - inputs=[], - outputs=[], - ) - - -def connect_reuse_seed(seed: gr.Number, reuse_seed: gr.Button, generation_info: gr.Textbox, dummy_component, is_subseed): - """ Connects a 'reuse (sub)seed' button's click event so that it copies last used - (sub)seed value from generation info the to the seed field. If copying subseed and subseed strength - was 0, i.e. no variation seed was used, it copies the normal seed value instead.""" - def copy_seed(gen_info_string: str, index): - res = -1 - - try: - gen_info = json.loads(gen_info_string) - index -= gen_info.get('index_of_first_image', 0) - - if is_subseed and gen_info.get('subseed_strength', 0) > 0: - all_subseeds = gen_info.get('all_subseeds', [-1]) - res = all_subseeds[index if 0 <= index < len(all_subseeds) else 0] - else: - all_seeds = gen_info.get('all_seeds', [-1]) - res = all_seeds[index if 0 <= index < len(all_seeds) else 0] - - except json.decoder.JSONDecodeError as e: - if gen_info_string != '': - print("Error parsing JSON generation info:", file=sys.stderr) - print(gen_info_string, file=sys.stderr) - - return [res, gr_show(False)] - - reuse_seed.click( - fn=copy_seed, - _js="(x, y) => [x, selected_gallery_index()]", - show_progress=False, - inputs=[generation_info, dummy_component], - outputs=[seed, dummy_component] - ) - - -def update_token_counter(text, steps): - try: - _, prompt_flat_list, _ = prompt_parser.get_multicond_prompt_list([text]) - prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules(prompt_flat_list, steps) - - except Exception: - # a parsing error can happen here during typing, and we don't want to bother the user with - # messages related to it in console - prompt_schedules = [[[steps, text]]] - - flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) - prompts = [prompt_text for step, prompt_text in flat_prompts] - token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) - style_class = ' class="red"' if (token_count > max_length) else "" - return f"{token_count}/{max_length}" - - -def create_toprow(is_img2img): - id_part = "img2img" if is_img2img else "txt2img" - - with gr.Row(elem_id="toprow"): - with gr.Column(scale=6): - with gr.Row(): - with gr.Column(scale=80): - with gr.Row(): - prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, - placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) - - with gr.Row(): - with gr.Column(scale=80): - with gr.Row(): - negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, - placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) - - with gr.Column(scale=1, elem_id="roll_col"): - paste = gr.Button(value=paste_symbol, elem_id="paste") - save_style = gr.Button(value=save_style_symbol, elem_id="style_create") - prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") - clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") - token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") - token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") - - clear_prompt_button.click( - fn=lambda *x: x, - _js="confirm_clear_prompt", - inputs=[prompt, negative_prompt], - outputs=[prompt, negative_prompt], - ) - - button_interrogate = None - button_deepbooru = None - if is_img2img: - with gr.Column(scale=1, elem_id="interrogate_col"): - button_interrogate = gr.Button('Interrogate\nCLIP', elem_id="interrogate") - button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") - - with gr.Column(scale=1): - with gr.Row(): - skip = gr.Button('Skip', elem_id=f"{id_part}_skip") - interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") - submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') - - skip.click( - fn=lambda: shared.state.skip(), - inputs=[], - outputs=[], - ) - - interrupt.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - with gr.Row(): - with gr.Column(scale=1, elem_id="style_pos_col"): - prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - with gr.Column(scale=1, elem_id="style_neg_col"): - prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button - - -def setup_progressbar(progressbar, preview, id_part, textinfo=None): - if textinfo is None: - textinfo = gr.HTML(visible=False) - - check_progress = gr.Button('Check progress', elem_id=f"{id_part}_check_progress", visible=False) - check_progress.click( - fn=lambda: check_progress_call(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) - - check_progress_initial = gr.Button('Check progress (first)', elem_id=f"{id_part}_check_progress_initial", visible=False) - check_progress_initial.click( - fn=lambda: check_progress_call_initial(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) - - -def apply_setting(key, value): - if value is None: - return gr.update() - - if shared.cmd_opts.freeze_settings: - return gr.update() - - # dont allow model to be swapped when model hash exists in prompt - if key == "sd_model_checkpoint" and opts.disable_weights_auto_swap: - return gr.update() - - if key == "sd_model_checkpoint": - ckpt_info = sd_models.get_closet_checkpoint_match(value) - - if ckpt_info is not None: - value = ckpt_info.title - else: - return gr.update() - - comp_args = opts.data_labels[key].component_args - if comp_args and isinstance(comp_args, dict) and comp_args.get('visible') is False: - return - - valtype = type(opts.data_labels[key].default) - oldval = opts.data.get(key, None) - opts.data[key] = valtype(value) if valtype != type(None) else value - if oldval != value and opts.data_labels[key].onchange is not None: - opts.data_labels[key].onchange() - - opts.save(shared.config_filename) - return value - - -def update_generation_info(args): - generation_info, html_info, img_index = args - try: - generation_info = json.loads(generation_info) - if img_index < 0 or img_index >= len(generation_info["infotexts"]): - return html_info - return plaintext_to_html(generation_info["infotexts"][img_index]) - except Exception: - pass - # if the json parse or anything else fails, just return the old html_info - return html_info - - -def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): - def refresh(): - refresh_method() - args = refreshed_args() if callable(refreshed_args) else refreshed_args - - for k, v in args.items(): - setattr(refresh_component, k, v) - - return gr.update(**(args or {})) - - refresh_button = ToolButton(value=refresh_symbol, elem_id=elem_id) - refresh_button.click( - fn=refresh, - inputs=[], - outputs=[refresh_component] - ) - return refresh_button - - -def create_output_panel(tabname, outdir): - def open_folder(f): - if not os.path.exists(f): - print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') - return - elif not os.path.isdir(f): - print(f""" -WARNING -An open_folder request was made with an argument that is not a folder. -This could be an error or a malicious attempt to run code on your computer. -Requested path was: {f} -""", file=sys.stderr) - return - - if not shared.cmd_opts.hide_ui_dir_config: - path = os.path.normpath(f) - if platform.system() == "Windows": - os.startfile(path) - elif platform.system() == "Darwin": - sp.Popen(["open", path]) - elif "microsoft-standard-WSL2" in platform.uname().release: - sp.Popen(["wsl-open", path]) - else: - sp.Popen(["xdg-open", path]) - - with gr.Column(variant='panel'): - with gr.Group(): - result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) - - generation_info = None - with gr.Column(): - with gr.Row(elem_id=f"image_buttons_{tabname}"): - open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') - - if tabname != "extras": - save = gr.Button('Save', elem_id=f'save_{tabname}') - save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') - - buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) - - open_folder_button.click( - fn=lambda: open_folder(opts.outdir_samples or outdir), - inputs=[], - outputs=[], - ) - - if tabname != "extras": - with gr.Row(): - download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') - - with gr.Group(): - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') - if tabname == 'txt2img' or tabname == 'img2img': - generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") - generation_info_button.click( - fn=update_generation_info, - _js="(x, y) => [x, y, selected_gallery_index()]", - inputs=[generation_info, html_info], - outputs=[html_info], - preprocess=False - ) - - save.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - save_zip.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - else: - html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) - return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log - - -def create_sampler_and_steps_selection(choices, tabname): - if opts.samplers_in_dropdown: - with FormRow(elem_id=f"sampler_selection_{tabname}"): - sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) - else: - with FormGroup(elem_id=f"sampler_selection_{tabname}"): - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) - sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - - return steps, sampler_index - - -def ordered_ui_categories(): - user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} - - for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): - yield category - - -def create_ui(): - import modules.img2img - import modules.txt2img - - reload_javascript() - - parameters_copypaste.reset() - - modules.scripts.scripts_current = modules.scripts.scripts_txt2img - modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) - - with gr.Blocks(analytics_enabled=False) as txt2img_interface: - txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) - - dummy_component = gr.Label(visible=False) - txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Row(elem_id='txt2img_progress_row'): - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="txt2img_progressbar") - txt2img_preview = gr.Image(elem_id='txt2img_preview', visible=False) - setup_progressbar(progressbar, txt2img_preview, 'txt2img') - - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel', elem_id="txt2img_settings"): - for category in ordered_ui_categories(): - if category == "sampler": - steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") - - elif category == "dimensions": - with FormRow(): - with gr.Column(elem_id="txt2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") - - if opts.dimensions_and_batch_together: - with gr.Column(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - elif category == "cfg": - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") - - elif category == "seed": - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') - - elif category == "checkboxes": - with FormRow(elem_id="txt2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") - enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") - hr_final_resolution = FormHTML(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) - - elif category == "hires_fix": - with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: - with FormRow(elem_id="txt2img_hires_fix_row1"): - hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) - hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") - - with FormRow(elem_id="txt2img_hires_fix_row2"): - hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") - hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") - hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") - - elif category == "batch": - if not opts.dimensions_and_batch_together: - with FormRow(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - elif category == "scripts": - with FormGroup(elem_id="txt2img_script_container"): - custom_inputs = modules.scripts.scripts_txt2img.setup_ui() - - hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] - for input in hr_resolution_preview_inputs: - input.change( - fn=calc_resolution_hires, - inputs=hr_resolution_preview_inputs, - outputs=[hr_final_resolution], - show_progress=False, - ) - input.change( - None, - _js="onCalcResolutionHires", - inputs=hr_resolution_preview_inputs, - outputs=[], - show_progress=False, - ) - - txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) - parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) - - connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) - connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) - - txt2img_args = dict( - fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), - _js="submit", - inputs=[ - txt2img_prompt, - txt2img_negative_prompt, - txt2img_prompt_style, - txt2img_prompt_style2, - steps, - sampler_index, - restore_faces, - tiling, - batch_count, - batch_size, - cfg_scale, - seed, - subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, - height, - width, - enable_hr, - denoising_strength, - hr_scale, - hr_upscaler, - hr_second_pass_steps, - hr_resize_x, - hr_resize_y, - ] + custom_inputs, - - outputs=[ - txt2img_gallery, - generation_info, - html_info, - html_log, - ], - show_progress=False, - ) - - txt2img_prompt.submit(**txt2img_args) - submit.click(**txt2img_args) - - txt_prompt_img.change( - fn=modules.images.image_data, - inputs=[ - txt_prompt_img - ], - outputs=[ - txt2img_prompt, - txt_prompt_img - ] - ) - - enable_hr.change( - fn=lambda x: gr_show(x), - inputs=[enable_hr], - outputs=[hr_options], - show_progress = False, - ) - - txt2img_paste_fields = [ - (txt2img_prompt, "Prompt"), - (txt2img_negative_prompt, "Negative prompt"), - (steps, "Steps"), - (sampler_index, "Sampler"), - (restore_faces, "Face restoration"), - (cfg_scale, "CFG scale"), - (seed, "Seed"), - (width, "Size-1"), - (height, "Size-2"), - (batch_size, "Batch size"), - (subseed, "Variation seed"), - (subseed_strength, "Variation seed strength"), - (seed_resize_from_w, "Seed resize from-1"), - (seed_resize_from_h, "Seed resize from-2"), - (denoising_strength, "Denoising strength"), - (enable_hr, lambda d: "Denoising strength" in d), - (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), - (hr_scale, "Hires upscale"), - (hr_upscaler, "Hires upscaler"), - (hr_second_pass_steps, "Hires steps"), - (hr_resize_x, "Hires resize-1"), - (hr_resize_y, "Hires resize-2"), - *modules.scripts.scripts_txt2img.infotext_fields - ] - parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) - - txt2img_preview_params = [ - txt2img_prompt, - txt2img_negative_prompt, - steps, - sampler_index, - cfg_scale, - seed, - width, - height, - ] - - token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) - - modules.scripts.scripts_current = modules.scripts.scripts_img2img - modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) - - with gr.Blocks(analytics_enabled=False) as img2img_interface: - img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) - - with gr.Row(elem_id='img2img_progress_row'): - img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="img2img_progressbar") - img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) - setup_progressbar(progressbar, img2img_preview, 'img2img') - - with FormRow().style(equal_height=False): - with gr.Column(variant='panel', elem_id="img2img_settings"): - - with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: - with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): - init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) - - with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): - init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) - init_img_with_mask_orig = gr.State(None) - - use_color_sketch = cmd_opts.gradio_inpaint_tool == "color-sketch" - if use_color_sketch: - def update_orig(image, state): - if image is not None: - same_size = state is not None and state.size == image.size - has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) - edited = same_size and has_exact_match - return image if not edited or state is None else state - - init_img_with_mask.change(update_orig, [init_img_with_mask, init_img_with_mask_orig], init_img_with_mask_orig) - - init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") - init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") - - with FormRow(): - mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") - mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") - - with FormRow(): - mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") - inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") - - with FormRow(): - inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") - - with FormRow(): - with gr.Column(): - inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") - - with gr.Column(scale=4): - inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") - - with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): - hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' - gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") - img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") - img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") - - with FormRow(): - resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") - - for category in ordered_ui_categories(): - if category == "sampler": - steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") - - elif category == "dimensions": - with FormRow(): - with gr.Column(elem_id="img2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") - - if opts.dimensions_and_batch_together: - with gr.Column(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - elif category == "cfg": - with FormGroup(): - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") - - elif category == "seed": - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') - - elif category == "checkboxes": - with FormRow(elem_id="img2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") - - elif category == "batch": - if not opts.dimensions_and_batch_together: - with FormRow(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - elif category == "scripts": - with FormGroup(elem_id="img2img_script_container"): - custom_inputs = modules.scripts.scripts_img2img.setup_ui() - - img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) - parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) - - connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) - connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) - - img2img_prompt_img.change( - fn=modules.images.image_data, - inputs=[ - img2img_prompt_img - ], - outputs=[ - img2img_prompt, - img2img_prompt_img - ] - ) - - mask_mode.change( - lambda mode, img: { - init_img_with_mask: gr_show(mode == 0), - init_img_inpaint: gr_show(mode == 1), - init_mask_inpaint: gr_show(mode == 1), - }, - inputs=[mask_mode, init_img_with_mask], - outputs=[ - init_img_with_mask, - init_img_inpaint, - init_mask_inpaint, - ], - ) - - img2img_args = dict( - fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), - _js="submit_img2img", - inputs=[ - dummy_component, - img2img_prompt, - img2img_negative_prompt, - img2img_prompt_style, - img2img_prompt_style2, - init_img, - init_img_with_mask, - init_img_with_mask_orig, - init_img_inpaint, - init_mask_inpaint, - mask_mode, - steps, - sampler_index, - mask_blur, - mask_alpha, - inpainting_fill, - restore_faces, - tiling, - batch_count, - batch_size, - cfg_scale, - denoising_strength, - seed, - subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, - height, - width, - resize_mode, - inpaint_full_res, - inpaint_full_res_padding, - inpainting_mask_invert, - img2img_batch_input_dir, - img2img_batch_output_dir, - ] + custom_inputs, - outputs=[ - img2img_gallery, - generation_info, - html_info, - html_log, - ], - show_progress=False, - ) - - img2img_prompt.submit(**img2img_args) - submit.click(**img2img_args) - - img2img_interrogate.click( - fn=interrogate, - inputs=[init_img], - outputs=[img2img_prompt], - ) - - img2img_deepbooru.click( - fn=interrogate_deepbooru, - inputs=[init_img], - outputs=[img2img_prompt], - ) - - prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] - style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] - style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] - - for button, (prompt, negative_prompt) in zip([txt2img_save_style, img2img_save_style], prompts): - button.click( - fn=add_style, - _js="ask_for_style_name", - # Have to pass empty dummy component here, because the JavaScript and Python function have to accept - # the same number of parameters, but we only know the style-name after the JavaScript prompt - inputs=[dummy_component, prompt, negative_prompt], - outputs=[txt2img_prompt_style, img2img_prompt_style, txt2img_prompt_style2, img2img_prompt_style2], - ) - - for button, (prompt, negative_prompt), (style1, style2), js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): - button.click( - fn=apply_styles, - _js=js_func, - inputs=[prompt, negative_prompt, style1, style2], - outputs=[prompt, negative_prompt, style1, style2], - ) - - token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) - - img2img_paste_fields = [ - (img2img_prompt, "Prompt"), - (img2img_negative_prompt, "Negative prompt"), - (steps, "Steps"), - (sampler_index, "Sampler"), - (restore_faces, "Face restoration"), - (cfg_scale, "CFG scale"), - (seed, "Seed"), - (width, "Size-1"), - (height, "Size-2"), - (batch_size, "Batch size"), - (subseed, "Variation seed"), - (subseed_strength, "Variation seed strength"), - (seed_resize_from_w, "Seed resize from-1"), - (seed_resize_from_h, "Seed resize from-2"), - (denoising_strength, "Denoising strength"), - (mask_blur, "Mask blur"), - *modules.scripts.scripts_img2img.infotext_fields - ] - parameters_copypaste.add_paste_fields("img2img", init_img, img2img_paste_fields) - parameters_copypaste.add_paste_fields("inpaint", init_img_with_mask, img2img_paste_fields) - - modules.scripts.scripts_current = None - - with gr.Blocks(analytics_enabled=False) as extras_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - with gr.Tabs(elem_id="mode_extras"): - with gr.TabItem('Single Image', elem_id="extras_single_tab"): - extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") - - with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): - image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") - - with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): - extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") - extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") - show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") - - submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') - - with gr.Tabs(elem_id="extras_resize_mode"): - with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): - upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") - with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): - with gr.Group(): - with gr.Row(): - upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") - upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") - upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") - - with gr.Group(): - extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - - with gr.Group(): - extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") - - with gr.Group(): - gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") - - with gr.Group(): - codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") - codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") - - with gr.Group(): - upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") - - result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) - - submit.click( - fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), - _js="get_extras_tab_index", - inputs=[ - dummy_component, - dummy_component, - extras_image, - image_batch, - extras_batch_input_dir, - extras_batch_output_dir, - show_extras_results, - gfpgan_visibility, - codeformer_visibility, - codeformer_weight, - upscaling_resize, - upscaling_resize_w, - upscaling_resize_h, - upscaling_crop, - extras_upscaler_1, - extras_upscaler_2, - extras_upscaler_2_visibility, - upscale_before_face_fix, - ], - outputs=[ - result_images, - html_info_x, - html_info, - ] - ) - parameters_copypaste.add_paste_fields("extras", extras_image, None) - - extras_image.change( - fn=modules.extras.clear_cache, - inputs=[], outputs=[] - ) - - with gr.Blocks(analytics_enabled=False) as pnginfo_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - image = gr.Image(elem_id="pnginfo_image", label="Source", source="upload", interactive=True, type="pil") - - with gr.Column(variant='panel'): - html = gr.HTML() - generation_info = gr.Textbox(visible=False, elem_id="pnginfo_generation_info") - html2 = gr.HTML() - with gr.Row(): - buttons = parameters_copypaste.create_buttons(["txt2img", "img2img", "inpaint", "extras"]) - parameters_copypaste.bind_buttons(buttons, image, generation_info) - - image.change( - fn=wrap_gradio_call(modules.extras.run_pnginfo), - inputs=[image], - outputs=[html, generation_info, html2], - ) - - with gr.Blocks(analytics_enabled=False) as modelmerger_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") - - with gr.Row(): - primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") - create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") - - secondary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_secondary_model_name", label="Secondary model (B)") - create_refresh_button(secondary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_B") - - tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") - create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") - - custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") - interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") - interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") - - with gr.Row(): - checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") - save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") - - modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') - - with gr.Column(variant='panel'): - submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) - - with gr.Blocks(analytics_enabled=False) as train_interface: - with gr.Row().style(equal_height=False): - gr.HTML(value="

See wiki for detailed explanation.

") - - with gr.Row().style(equal_height=False): - with gr.Tabs(elem_id="train_tabs"): - - with gr.Tab(label="Create embedding"): - new_embedding_name = gr.Textbox(label="Name", elem_id="train_new_embedding_name") - initialization_text = gr.Textbox(label="Initialization text", value="*", elem_id="train_initialization_text") - nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1, elem_id="train_nvpt") - overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding", elem_id="train_overwrite_old_embedding") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - create_embedding = gr.Button(value="Create embedding", variant='primary', elem_id="train_create_embedding") - - with gr.Tab(label="Create hypernetwork"): - new_hypernetwork_name = gr.Textbox(label="Name", elem_id="train_new_hypernetwork_name") - new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"], elem_id="train_new_hypernetwork_sizes") - new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'", elem_id="train_new_hypernetwork_layer_structure") - new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys, elem_id="train_new_hypernetwork_activation_func") - new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") - new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") - new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") - new_hypernetwork_dropout_structure = gr.Textbox("0, 0, 0", label="Enter hypernetwork Dropout structure (or empty). Recommended : 0~0.35 incrementing sequence: 0, 0.05, 0.15", placeholder="1st and last digit must be 0 and values should be between 0 and 1. ex:'0, 0.01, 0'") - overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary', elem_id="train_create_hypernetwork") - - with gr.Tab(label="Preprocess images"): - process_src = gr.Textbox(label='Source directory', elem_id="train_process_src") - process_dst = gr.Textbox(label='Destination directory', elem_id="train_process_dst") - process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_process_width") - process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_process_height") - preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"], elem_id="train_preprocess_txt_action") - - with gr.Row(): - process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") - process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") - process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") - process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") - process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") - - with gr.Row(visible=False) as process_split_extra_row: - process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_split_threshold") - process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05, elem_id="train_process_overlap_ratio") - - with gr.Row(visible=False) as process_focal_crop_row: - process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_face_weight") - process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") - process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") - process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - with gr.Row(): - interrupt_preprocessing = gr.Button("Interrupt", elem_id="train_interrupt_preprocessing") - run_preprocess = gr.Button(value="Preprocess", variant='primary', elem_id="train_run_preprocess") - - process_split.change( - fn=lambda show: gr_show(show), - inputs=[process_split], - outputs=[process_split_extra_row], - ) - - process_focal_crop.change( - fn=lambda show: gr_show(show), - inputs=[process_focal_crop], - outputs=[process_focal_crop_row], - ) - - def get_textual_inversion_template_names(): - return sorted([x for x in textual_inversion.textual_inversion_templates]) - - with gr.Tab(label="Train"): - gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") - with FormRow(): - train_embedding_name = gr.Dropdown(label='Embedding', elem_id="train_embedding", choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())) - create_refresh_button(train_embedding_name, sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings, lambda: {"choices": sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())}, "refresh_train_embedding_name") - - train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) - create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") - - with FormRow(): - embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") - hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") - - with FormRow(): - clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) - clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) - - with FormRow(): - batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") - gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") - - dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") - log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") - - with FormRow(): - template_file = gr.Dropdown(label='Prompt template', value="style_filewords.txt", elem_id="train_template_file", choices=get_textual_inversion_template_names()) - create_refresh_button(template_file, textual_inversion.list_textual_inversion_templates, lambda: {"choices": get_textual_inversion_template_names()}, "refrsh_train_template_file") - - training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") - training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") - varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") - steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") - - with FormRow(): - create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") - save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") - - save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") - preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") - - shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") - tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") - - latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") - - with gr.Row(): - train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") - interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") - train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") - - params = script_callbacks.UiTrainTabParams(txt2img_preview_params) - - script_callbacks.ui_train_tabs_callback(params) - - with gr.Column(): - progressbar = gr.HTML(elem_id="ti_progressbar") - ti_output = gr.Text(elem_id="ti_output", value="", show_label=False) - - ti_gallery = gr.Gallery(label='Output', show_label=False, elem_id='ti_gallery').style(grid=4) - ti_preview = gr.Image(elem_id='ti_preview', visible=False) - ti_progress = gr.HTML(elem_id="ti_progress", value="") - ti_outcome = gr.HTML(elem_id="ti_error", value="") - setup_progressbar(progressbar, ti_preview, 'ti', textinfo=ti_progress) - - create_embedding.click( - fn=modules.textual_inversion.ui.create_embedding, - inputs=[ - new_embedding_name, - initialization_text, - nvpt, - overwrite_old_embedding, - ], - outputs=[ - train_embedding_name, - ti_output, - ti_outcome, - ] - ) - - create_hypernetwork.click( - fn=modules.hypernetworks.ui.create_hypernetwork, - inputs=[ - new_hypernetwork_name, - new_hypernetwork_sizes, - overwrite_old_hypernetwork, - new_hypernetwork_layer_structure, - new_hypernetwork_activation_func, - new_hypernetwork_initialization_option, - new_hypernetwork_add_layer_norm, - new_hypernetwork_use_dropout, - new_hypernetwork_dropout_structure - ], - outputs=[ - train_hypernetwork_name, - ti_output, - ti_outcome, - ] - ) - - run_preprocess.click( - fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.preprocess, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - process_src, - process_dst, - process_width, - process_height, - preprocess_txt_action, - process_flip, - process_split, - process_caption, - process_caption_deepbooru, - process_split_threshold, - process_overlap_ratio, - process_focal_crop, - process_focal_crop_face_weight, - process_focal_crop_entropy_weight, - process_focal_crop_edges_weight, - process_focal_crop_debug, - ], - outputs=[ - ti_output, - ti_outcome, - ], - ) - - train_embedding.click( - fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.train_embedding, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - train_embedding_name, - embedding_learn_rate, - batch_size, - gradient_step, - dataset_directory, - log_directory, - training_width, - training_height, - varsize, - steps, - clip_grad_mode, - clip_grad_value, - shuffle_tags, - tag_drop_out, - latent_sampling_method, - create_image_every, - save_embedding_every, - template_file, - save_image_with_stored_embedding, - preview_from_txt2img, - *txt2img_preview_params, - ], - outputs=[ - ti_output, - ti_outcome, - ] - ) - - train_hypernetwork.click( - fn=wrap_gradio_gpu_call(modules.hypernetworks.ui.train_hypernetwork, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - train_hypernetwork_name, - hypernetwork_learn_rate, - batch_size, - gradient_step, - dataset_directory, - log_directory, - training_width, - training_height, - varsize, - steps, - clip_grad_mode, - clip_grad_value, - shuffle_tags, - tag_drop_out, - latent_sampling_method, - create_image_every, - save_embedding_every, - template_file, - preview_from_txt2img, - *txt2img_preview_params, - ], - outputs=[ - ti_output, - ti_outcome, - ] - ) - - interrupt_training.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - interrupt_preprocessing.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - def create_setting_component(key, is_quicksettings=False): - def fun(): - return opts.data[key] if key in opts.data else opts.data_labels[key].default - - info = opts.data_labels[key] - t = type(info.default) - - args = info.component_args() if callable(info.component_args) else info.component_args - - if info.component is not None: - comp = info.component - elif t == str: - comp = gr.Textbox - elif t == int: - comp = gr.Number - elif t == bool: - comp = gr.Checkbox - else: - raise Exception(f'bad options item type: {str(t)} for key {key}') - - elem_id = "setting_"+key - - if info.refresh is not None: - if is_quicksettings: - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) - else: - with FormRow(): - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) - else: - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - - return res - - components = [] - component_dict = {} - - script_callbacks.ui_settings_callback() - opts.reorder() - - def run_settings(*args): - changed = [] - - for key, value, comp in zip(opts.data_labels.keys(), args, components): - assert comp == dummy_component or opts.same_type(value, opts.data_labels[key].default), f"Bad value for setting {key}: {value}; expecting {type(opts.data_labels[key].default).__name__}" - - for key, value, comp in zip(opts.data_labels.keys(), args, components): - if comp == dummy_component: - continue - - if opts.set(key, value): - changed.append(key) - - try: - opts.save(shared.config_filename) - except RuntimeError: - return opts.dumpjson(), f'{len(changed)} settings changed without save: {", ".join(changed)}.' - return opts.dumpjson(), f'{len(changed)} settings changed{": " if len(changed) > 0 else ""}{", ".join(changed)}.' - - def run_settings_single(value, key): - if not opts.same_type(value, opts.data_labels[key].default): - return gr.update(visible=True), opts.dumpjson() - - if not opts.set(key, value): - return gr.update(value=getattr(opts, key)), opts.dumpjson() - - opts.save(shared.config_filename) - - return gr.update(value=value), opts.dumpjson() - - with gr.Blocks(analytics_enabled=False) as settings_interface: - with gr.Row(): - with gr.Column(scale=6): - settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") - with gr.Column(): - restart_gradio = gr.Button(value='Reload UI', variant='primary', elem_id="settings_restart_gradio") - - result = gr.HTML(elem_id="settings_result") - - quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] - quicksettings_names = {x: i for i, x in enumerate(quicksettings_names) if x != 'quicksettings'} - - quicksettings_list = [] - - previous_section = None - current_tab = None - with gr.Tabs(elem_id="settings"): - for i, (k, item) in enumerate(opts.data_labels.items()): - section_must_be_skipped = item.section[0] is None - - if previous_section != item.section and not section_must_be_skipped: - elem_id, text = item.section - - if current_tab is not None: - current_tab.__exit__() - - current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) - current_tab.__enter__() - - previous_section = item.section - - if k in quicksettings_names and not shared.cmd_opts.freeze_settings: - quicksettings_list.append((i, k, item)) - components.append(dummy_component) - elif section_must_be_skipped: - components.append(dummy_component) - else: - component = create_setting_component(k) - component_dict[k] = component - components.append(component) - - if current_tab is not None: - current_tab.__exit__() - - with gr.TabItem("Actions"): - request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") - download_localization = gr.Button(value='Download localization template', elem_id="download_localization") - reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") - - if os.path.exists("html/licenses.html"): - with open("html/licenses.html", encoding="utf8") as file: - with gr.TabItem("Licenses"): - gr.HTML(file.read(), elem_id="licenses") - - gr.Button(value="Show all pages", elem_id="settings_show_all_pages") - - request_notifications.click( - fn=lambda: None, - inputs=[], - outputs=[], - _js='function(){}' - ) - - download_localization.click( - fn=lambda: None, - inputs=[], - outputs=[], - _js='download_localization' - ) - - def reload_scripts(): - modules.scripts.reload_script_body_only() - reload_javascript() # need to refresh the html page - - reload_script_bodies.click( - fn=reload_scripts, - inputs=[], - outputs=[] - ) - - def request_restart(): - shared.state.interrupt() - shared.state.need_restart = True - - restart_gradio.click( - fn=request_restart, - _js='restart_reload', - inputs=[], - outputs=[], - ) - - interfaces = [ - (txt2img_interface, "txt2img", "txt2img"), - (img2img_interface, "img2img", "img2img"), - (extras_interface, "Extras", "extras"), - (pnginfo_interface, "PNG Info", "pnginfo"), - (modelmerger_interface, "Checkpoint Merger", "modelmerger"), - (train_interface, "Train", "ti"), - ] - - css = "" - - for cssfile in modules.scripts.list_files_with_name("style.css"): - if not os.path.isfile(cssfile): - continue - - with open(cssfile, "r", encoding="utf8") as file: - css += file.read() + "\n" - - if os.path.exists(os.path.join(script_path, "user.css")): - with open(os.path.join(script_path, "user.css"), "r", encoding="utf8") as file: - css += file.read() + "\n" - - if not cmd_opts.no_progressbar_hiding: - css += css_hide_progressbar - - interfaces += script_callbacks.ui_tabs_callback() - interfaces += [(settings_interface, "Settings", "settings")] - - extensions_interface = ui_extensions.create_ui() - interfaces += [(extensions_interface, "Extensions", "extensions")] - - with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: - with gr.Row(elem_id="quicksettings"): - for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): - component = create_setting_component(k, is_quicksettings=True) - component_dict[k] = component - - parameters_copypaste.integrate_settings_paste_fields(component_dict) - parameters_copypaste.run_bind() - - with gr.Tabs(elem_id="tabs") as tabs: - for interface, label, ifid in interfaces: - with gr.TabItem(label, id=ifid, elem_id='tab_' + ifid): - interface.render() - - if os.path.exists(os.path.join(script_path, "notification.mp3")): - audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) - - if os.path.exists("html/footer.html"): - with open("html/footer.html", encoding="utf8") as file: - footer = file.read() - footer = footer.format(versions=versions_html()) - gr.HTML(footer, elem_id="footer") - - text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) - settings_submit.click( - fn=wrap_gradio_call(run_settings, extra_outputs=[gr.update()]), - inputs=components, - outputs=[text_settings, result], - ) - - for i, k, item in quicksettings_list: - component = component_dict[k] - - component.change( - fn=lambda value, k=k: run_settings_single(value, key=k), - inputs=[component], - outputs=[component, text_settings], - ) - - component_keys = [k for k in opts.data_labels.keys() if k in component_dict] - - def get_settings_values(): - return [getattr(opts, key) for key in component_keys] - - demo.load( - fn=get_settings_values, - inputs=[], - outputs=[component_dict[k] for k in component_keys], - ) - - def modelmerger(*args): - try: - results = modules.extras.run_modelmerger(*args) - except Exception as e: - print("Error loading/saving model file:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - modules.sd_models.list_models() # to remove the potentially missing models from the list - return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] - return results - - modelmerger_merge.click( - fn=modelmerger, - inputs=[ - primary_model_name, - secondary_model_name, - tertiary_model_name, - interp_method, - interp_amount, - save_as_half, - custom_name, - checkpoint_format, - ], - outputs=[ - submit_result, - primary_model_name, - secondary_model_name, - tertiary_model_name, - component_dict['sd_model_checkpoint'], - ] - ) - - ui_config_file = cmd_opts.ui_config_file - ui_settings = {} - settings_count = len(ui_settings) - error_loading = False - - try: - if os.path.exists(ui_config_file): - with open(ui_config_file, "r", encoding="utf8") as file: - ui_settings = json.load(file) - except Exception: - error_loading = True - print("Error loading settings:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - - def loadsave(path, x): - def apply_field(obj, field, condition=None, init_field=None): - key = path + "/" + field - - if getattr(obj, 'custom_script_source', None) is not None: - key = 'customscript/' + obj.custom_script_source + '/' + key - - if getattr(obj, 'do_not_save_to_config', False): - return - - saved_value = ui_settings.get(key, None) - if saved_value is None: - ui_settings[key] = getattr(obj, field) - elif condition and not condition(saved_value): - print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') - else: - setattr(obj, field, saved_value) - if init_field is not None: - init_field(saved_value) - - if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number, gr.Dropdown] and x.visible: - apply_field(x, 'visible') - - if type(x) == gr.Slider: - apply_field(x, 'value') - apply_field(x, 'minimum') - apply_field(x, 'maximum') - apply_field(x, 'step') - - if type(x) == gr.Radio: - apply_field(x, 'value', lambda val: val in x.choices) - - if type(x) == gr.Checkbox: - apply_field(x, 'value') - - if type(x) == gr.Textbox: - apply_field(x, 'value') - - if type(x) == gr.Number: - apply_field(x, 'value') - - if type(x) == gr.Dropdown: - apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) - - visit(txt2img_interface, loadsave, "txt2img") - visit(img2img_interface, loadsave, "img2img") - visit(extras_interface, loadsave, "extras") - visit(modelmerger_interface, loadsave, "modelmerger") - visit(train_interface, loadsave, "train") - - if not error_loading and (not os.path.exists(ui_config_file) or settings_count != len(ui_settings)): - with open(ui_config_file, "w", encoding="utf8") as file: - json.dump(ui_settings, file, indent=4) - - return demo - - -def reload_javascript(): - with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: - javascript = f'' - - scripts_list = modules.scripts.list_scripts("javascript", ".js") - - for basedir, filename, path in scripts_list: - with open(path, "r", encoding="utf8") as jsfile: - javascript += f"\n" - - if cmd_opts.theme is not None: - javascript += f"\n\n" - - javascript += f"\n" - - def template_response(*args, **kwargs): - res = shared.GradioTemplateResponseOriginal(*args, **kwargs) - res.body = res.body.replace( - b'', f'{javascript}'.encode("utf8")) - res.init_headers() - return res - - gradio.routes.templates.TemplateResponse = template_response - - -if not hasattr(shared, 'GradioTemplateResponseOriginal'): - shared.GradioTemplateResponseOriginal = gradio.routes.templates.TemplateResponse - - -def versions_html(): - import torch - import launch - - python_version = ".".join([str(x) for x in sys.version_info[0:3]]) - commit = launch.commit_hash() - short_commit = commit[0:8] - - if shared.xformers_available: - import xformers - xformers_version = xformers.__version__ - else: - xformers_version = "N/A" - - return f""" -python: {python_version} - •  -torch: {torch.__version__} - •  -xformers: {xformers_version} - •  -gradio: {gr.__version__} - •  -commit: {short_commit} -""" -- cgit v1.2.3 From 54dd5d6634ead25311a8bea0484675607601a21a Mon Sep 17 00:00:00 2001 From: Andrey <16777216c@gmail.com> Date: Tue, 10 Jan 2023 11:54:49 +0300 Subject: Split history ui.py to ui_progress.py --- modules/temp | 1928 --------------------------------------------------------- modules/ui.py | 1928 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 1928 insertions(+), 1928 deletions(-) delete mode 100644 modules/temp create mode 100644 modules/ui.py (limited to 'modules') diff --git a/modules/temp b/modules/temp deleted file mode 100644 index 9b9081b5..00000000 --- a/modules/temp +++ /dev/null @@ -1,1928 +0,0 @@ -import html -import json -import math -import mimetypes -import os -import platform -import random -import subprocess as sp -import sys -import tempfile -import time -import traceback -from functools import partial, reduce - -import gradio as gr -import gradio.routes -import gradio.utils -import numpy as np -from PIL import Image, PngImagePlugin -from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call - -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru -from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML -from modules.paths import script_path - -from modules.shared import opts, cmd_opts, restricted_opts - -import modules.codeformer_model -import modules.generation_parameters_copypaste as parameters_copypaste -import modules.gfpgan_model -import modules.hypernetworks.ui -import modules.scripts -import modules.shared as shared -import modules.styles -import modules.textual_inversion.ui -from modules import prompt_parser -from modules.images import save_image -from modules.sd_hijack import model_hijack -from modules.sd_samplers import samplers, samplers_for_img2img -from modules.textual_inversion import textual_inversion -import modules.hypernetworks.ui -from modules.generation_parameters_copypaste import image_from_url_text - -# this is a fix for Windows users. Without it, javascript files will be served with text/html content-type and the browser will not show any UI -mimetypes.init() -mimetypes.add_type('application/javascript', '.js') - -if not cmd_opts.share and not cmd_opts.listen: - # fix gradio phoning home - gradio.utils.version_check = lambda: None - gradio.utils.get_local_ip_address = lambda: '127.0.0.1' - -if cmd_opts.ngrok is not None: - import modules.ngrok as ngrok - print('ngrok authtoken detected, trying to connect...') - ngrok.connect( - cmd_opts.ngrok, - cmd_opts.port if cmd_opts.port is not None else 7860, - cmd_opts.ngrok_region - ) - - -def gr_show(visible=True): - return {"visible": visible, "__type__": "update"} - - -sample_img2img = "assets/stable-samples/img2img/sketch-mountains-input.jpg" -sample_img2img = sample_img2img if os.path.exists(sample_img2img) else None - -css_hide_progressbar = """ -.wrap .m-12 svg { display:none!important; } -.wrap .m-12::before { content:"Loading..." } -.wrap .z-20 svg { display:none!important; } -.wrap .z-20::before { content:"Loading..." } -.progress-bar { display:none!important; } -.meta-text { display:none!important; } -.meta-text-center { display:none!important; } -""" - -# Using constants for these since the variation selector isn't visible. -# Important that they exactly match script.js for tooltip to work. -random_symbol = '\U0001f3b2\ufe0f' # 🎲️ -reuse_symbol = '\u267b\ufe0f' # ♻️ -paste_symbol = '\u2199\ufe0f' # ↙ -folder_symbol = '\U0001f4c2' # 📂 -refresh_symbol = '\U0001f504' # 🔄 -save_style_symbol = '\U0001f4be' # 💾 -apply_style_symbol = '\U0001f4cb' # 📋 -clear_prompt_symbol = '\U0001F5D1' # 🗑️ - - -def plaintext_to_html(text): - text = "

" + "
\n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

" - return text - -def send_gradio_gallery_to_image(x): - if len(x) == 0: - return None - return image_from_url_text(x[0]) - -def save_files(js_data, images, do_make_zip, index): - import csv - filenames = [] - fullfns = [] - - #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it - class MyObject: - def __init__(self, d=None): - if d is not None: - for key, value in d.items(): - setattr(self, key, value) - - data = json.loads(js_data) - - p = MyObject(data) - path = opts.outdir_save - save_to_dirs = opts.use_save_to_dirs_for_ui - extension: str = opts.samples_format - start_index = 0 - - if index > -1 and opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only - - images = [images[index]] - start_index = index - - os.makedirs(opts.outdir_save, exist_ok=True) - - with open(os.path.join(opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: - at_start = file.tell() == 0 - writer = csv.writer(file) - if at_start: - writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) - - for image_index, filedata in enumerate(images, start_index): - image = image_from_url_text(filedata) - - is_grid = image_index < p.index_of_first_image - i = 0 if is_grid else (image_index - p.index_of_first_image) - - fullfn, txt_fullfn = save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) - - filename = os.path.relpath(fullfn, path) - filenames.append(filename) - fullfns.append(fullfn) - if txt_fullfn: - filenames.append(os.path.basename(txt_fullfn)) - fullfns.append(txt_fullfn) - - writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) - - # Make Zip - if do_make_zip: - zip_filepath = os.path.join(path, "images.zip") - - from zipfile import ZipFile - with ZipFile(zip_filepath, "w") as zip_file: - for i in range(len(fullfns)): - with open(fullfns[i], mode="rb") as f: - zip_file.writestr(filenames[i], f.read()) - fullfns.insert(0, zip_filepath) - - return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") - - -def calc_time_left(progress, threshold, label, force_display, show_eta): - if progress == 0: - return "" - else: - time_since_start = time.time() - shared.state.time_start - eta = (time_since_start/progress) - eta_relative = eta-time_since_start - if (eta_relative > threshold and show_eta) or force_display: - if eta_relative > 3600: - return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) - elif eta_relative > 60: - return label + time.strftime('%M:%S', time.gmtime(eta_relative)) - else: - return label + time.strftime('%Ss', time.gmtime(eta_relative)) - else: - return "" - - -def check_progress_call(id_part): - if shared.state.job_count == 0: - return "", gr_show(False), gr_show(False), gr_show(False) - - progress = 0 - - if shared.state.job_count > 0: - progress += shared.state.job_no / shared.state.job_count - if shared.state.sampling_steps > 0: - progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps - - # Show progress percentage and time left at the same moment, and base it also on steps done - show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 - - time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) - if time_left != "": - shared.state.time_left_force_display = True - - progress = min(progress, 1) - - progressbar = "" - if opts.show_progressbar: - progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" - - image = gr_show(False) - preview_visibility = gr_show(False) - - if opts.show_progress_every_n_steps != 0: - shared.state.set_current_image() - image = shared.state.current_image - - if image is None: - image = gr.update(value=None) - else: - preview_visibility = gr_show(True) - - if shared.state.textinfo is not None: - textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) - else: - textinfo_result = gr_show(False) - - return f"

{progressbar}

", preview_visibility, image, textinfo_result - - -def check_progress_call_initial(id_part): - shared.state.job_count = -1 - shared.state.current_latent = None - shared.state.current_image = None - shared.state.textinfo = None - shared.state.time_start = time.time() - shared.state.time_left_force_display = False - - return check_progress_call(id_part) - - -def visit(x, func, path=""): - if hasattr(x, 'children'): - for c in x.children: - visit(c, func, path) - elif x.label is not None: - func(path + "/" + str(x.label), x) - - -def add_style(name: str, prompt: str, negative_prompt: str): - if name is None: - return [gr_show() for x in range(4)] - - style = modules.styles.PromptStyle(name, prompt, negative_prompt) - shared.prompt_styles.styles[style.name] = style - # Save all loaded prompt styles: this allows us to update the storage format in the future more easily, because we - # reserialize all styles every time we save them - shared.prompt_styles.save_styles(shared.styles_filename) - - return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] - - -def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): - from modules import processing, devices - - if not enable: - return "" - - p = processing.StableDiffusionProcessingTxt2Img(width=width, height=height, enable_hr=True, hr_scale=hr_scale, hr_resize_x=hr_resize_x, hr_resize_y=hr_resize_y) - - with devices.autocast(): - p.init([""], [0], [0]) - - return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" - - -def apply_styles(prompt, prompt_neg, style1_name, style2_name): - prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) - prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, [style1_name, style2_name]) - - return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] - - -def interrogate(image): - prompt = shared.interrogator.interrogate(image.convert("RGB")) - - return gr_show(True) if prompt is None else prompt - - -def interrogate_deepbooru(image): - prompt = deepbooru.model.tag(image) - return gr_show(True) if prompt is None else prompt - - -def create_seed_inputs(target_interface): - with FormRow(elem_id=target_interface + '_seed_row'): - seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') - seed.style(container=False) - random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') - reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') - - with gr.Group(elem_id=target_interface + '_subseed_show_box'): - seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) - - # Components to show/hide based on the 'Extra' checkbox - seed_extras = [] - - with FormRow(visible=False, elem_id=target_interface + '_subseed_row') as seed_extra_row_1: - seed_extras.append(seed_extra_row_1) - subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') - subseed.style(container=False) - random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') - reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') - subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') - - with FormRow(visible=False) as seed_extra_row_2: - seed_extras.append(seed_extra_row_2) - seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') - seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') - - random_seed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[seed]) - random_subseed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[subseed]) - - def change_visibility(show): - return {comp: gr_show(show) for comp in seed_extras} - - seed_checkbox.change(change_visibility, show_progress=False, inputs=[seed_checkbox], outputs=seed_extras) - - return seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox - - - -def connect_clear_prompt(button): - """Given clear button, prompt, and token_counter objects, setup clear prompt button click event""" - button.click( - _js="clear_prompt", - fn=None, - inputs=[], - outputs=[], - ) - - -def connect_reuse_seed(seed: gr.Number, reuse_seed: gr.Button, generation_info: gr.Textbox, dummy_component, is_subseed): - """ Connects a 'reuse (sub)seed' button's click event so that it copies last used - (sub)seed value from generation info the to the seed field. If copying subseed and subseed strength - was 0, i.e. no variation seed was used, it copies the normal seed value instead.""" - def copy_seed(gen_info_string: str, index): - res = -1 - - try: - gen_info = json.loads(gen_info_string) - index -= gen_info.get('index_of_first_image', 0) - - if is_subseed and gen_info.get('subseed_strength', 0) > 0: - all_subseeds = gen_info.get('all_subseeds', [-1]) - res = all_subseeds[index if 0 <= index < len(all_subseeds) else 0] - else: - all_seeds = gen_info.get('all_seeds', [-1]) - res = all_seeds[index if 0 <= index < len(all_seeds) else 0] - - except json.decoder.JSONDecodeError as e: - if gen_info_string != '': - print("Error parsing JSON generation info:", file=sys.stderr) - print(gen_info_string, file=sys.stderr) - - return [res, gr_show(False)] - - reuse_seed.click( - fn=copy_seed, - _js="(x, y) => [x, selected_gallery_index()]", - show_progress=False, - inputs=[generation_info, dummy_component], - outputs=[seed, dummy_component] - ) - - -def update_token_counter(text, steps): - try: - _, prompt_flat_list, _ = prompt_parser.get_multicond_prompt_list([text]) - prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules(prompt_flat_list, steps) - - except Exception: - # a parsing error can happen here during typing, and we don't want to bother the user with - # messages related to it in console - prompt_schedules = [[[steps, text]]] - - flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) - prompts = [prompt_text for step, prompt_text in flat_prompts] - token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) - style_class = ' class="red"' if (token_count > max_length) else "" - return f"{token_count}/{max_length}" - - -def create_toprow(is_img2img): - id_part = "img2img" if is_img2img else "txt2img" - - with gr.Row(elem_id="toprow"): - with gr.Column(scale=6): - with gr.Row(): - with gr.Column(scale=80): - with gr.Row(): - prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, - placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) - - with gr.Row(): - with gr.Column(scale=80): - with gr.Row(): - negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, - placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) - - with gr.Column(scale=1, elem_id="roll_col"): - paste = gr.Button(value=paste_symbol, elem_id="paste") - save_style = gr.Button(value=save_style_symbol, elem_id="style_create") - prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") - clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") - token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") - token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") - - clear_prompt_button.click( - fn=lambda *x: x, - _js="confirm_clear_prompt", - inputs=[prompt, negative_prompt], - outputs=[prompt, negative_prompt], - ) - - button_interrogate = None - button_deepbooru = None - if is_img2img: - with gr.Column(scale=1, elem_id="interrogate_col"): - button_interrogate = gr.Button('Interrogate\nCLIP', elem_id="interrogate") - button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") - - with gr.Column(scale=1): - with gr.Row(): - skip = gr.Button('Skip', elem_id=f"{id_part}_skip") - interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") - submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') - - skip.click( - fn=lambda: shared.state.skip(), - inputs=[], - outputs=[], - ) - - interrupt.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - with gr.Row(): - with gr.Column(scale=1, elem_id="style_pos_col"): - prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - with gr.Column(scale=1, elem_id="style_neg_col"): - prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button - - -def setup_progressbar(progressbar, preview, id_part, textinfo=None): - if textinfo is None: - textinfo = gr.HTML(visible=False) - - check_progress = gr.Button('Check progress', elem_id=f"{id_part}_check_progress", visible=False) - check_progress.click( - fn=lambda: check_progress_call(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) - - check_progress_initial = gr.Button('Check progress (first)', elem_id=f"{id_part}_check_progress_initial", visible=False) - check_progress_initial.click( - fn=lambda: check_progress_call_initial(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) - - -def apply_setting(key, value): - if value is None: - return gr.update() - - if shared.cmd_opts.freeze_settings: - return gr.update() - - # dont allow model to be swapped when model hash exists in prompt - if key == "sd_model_checkpoint" and opts.disable_weights_auto_swap: - return gr.update() - - if key == "sd_model_checkpoint": - ckpt_info = sd_models.get_closet_checkpoint_match(value) - - if ckpt_info is not None: - value = ckpt_info.title - else: - return gr.update() - - comp_args = opts.data_labels[key].component_args - if comp_args and isinstance(comp_args, dict) and comp_args.get('visible') is False: - return - - valtype = type(opts.data_labels[key].default) - oldval = opts.data.get(key, None) - opts.data[key] = valtype(value) if valtype != type(None) else value - if oldval != value and opts.data_labels[key].onchange is not None: - opts.data_labels[key].onchange() - - opts.save(shared.config_filename) - return value - - -def update_generation_info(args): - generation_info, html_info, img_index = args - try: - generation_info = json.loads(generation_info) - if img_index < 0 or img_index >= len(generation_info["infotexts"]): - return html_info - return plaintext_to_html(generation_info["infotexts"][img_index]) - except Exception: - pass - # if the json parse or anything else fails, just return the old html_info - return html_info - - -def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): - def refresh(): - refresh_method() - args = refreshed_args() if callable(refreshed_args) else refreshed_args - - for k, v in args.items(): - setattr(refresh_component, k, v) - - return gr.update(**(args or {})) - - refresh_button = ToolButton(value=refresh_symbol, elem_id=elem_id) - refresh_button.click( - fn=refresh, - inputs=[], - outputs=[refresh_component] - ) - return refresh_button - - -def create_output_panel(tabname, outdir): - def open_folder(f): - if not os.path.exists(f): - print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') - return - elif not os.path.isdir(f): - print(f""" -WARNING -An open_folder request was made with an argument that is not a folder. -This could be an error or a malicious attempt to run code on your computer. -Requested path was: {f} -""", file=sys.stderr) - return - - if not shared.cmd_opts.hide_ui_dir_config: - path = os.path.normpath(f) - if platform.system() == "Windows": - os.startfile(path) - elif platform.system() == "Darwin": - sp.Popen(["open", path]) - elif "microsoft-standard-WSL2" in platform.uname().release: - sp.Popen(["wsl-open", path]) - else: - sp.Popen(["xdg-open", path]) - - with gr.Column(variant='panel'): - with gr.Group(): - result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) - - generation_info = None - with gr.Column(): - with gr.Row(elem_id=f"image_buttons_{tabname}"): - open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') - - if tabname != "extras": - save = gr.Button('Save', elem_id=f'save_{tabname}') - save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') - - buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) - - open_folder_button.click( - fn=lambda: open_folder(opts.outdir_samples or outdir), - inputs=[], - outputs=[], - ) - - if tabname != "extras": - with gr.Row(): - download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') - - with gr.Group(): - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') - if tabname == 'txt2img' or tabname == 'img2img': - generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") - generation_info_button.click( - fn=update_generation_info, - _js="(x, y) => [x, y, selected_gallery_index()]", - inputs=[generation_info, html_info], - outputs=[html_info], - preprocess=False - ) - - save.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - save_zip.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - else: - html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) - return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log - - -def create_sampler_and_steps_selection(choices, tabname): - if opts.samplers_in_dropdown: - with FormRow(elem_id=f"sampler_selection_{tabname}"): - sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) - else: - with FormGroup(elem_id=f"sampler_selection_{tabname}"): - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) - sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - - return steps, sampler_index - - -def ordered_ui_categories(): - user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} - - for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): - yield category - - -def create_ui(): - import modules.img2img - import modules.txt2img - - reload_javascript() - - parameters_copypaste.reset() - - modules.scripts.scripts_current = modules.scripts.scripts_txt2img - modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) - - with gr.Blocks(analytics_enabled=False) as txt2img_interface: - txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) - - dummy_component = gr.Label(visible=False) - txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Row(elem_id='txt2img_progress_row'): - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="txt2img_progressbar") - txt2img_preview = gr.Image(elem_id='txt2img_preview', visible=False) - setup_progressbar(progressbar, txt2img_preview, 'txt2img') - - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel', elem_id="txt2img_settings"): - for category in ordered_ui_categories(): - if category == "sampler": - steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") - - elif category == "dimensions": - with FormRow(): - with gr.Column(elem_id="txt2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") - - if opts.dimensions_and_batch_together: - with gr.Column(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - elif category == "cfg": - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") - - elif category == "seed": - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') - - elif category == "checkboxes": - with FormRow(elem_id="txt2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") - enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") - hr_final_resolution = FormHTML(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) - - elif category == "hires_fix": - with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: - with FormRow(elem_id="txt2img_hires_fix_row1"): - hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) - hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") - - with FormRow(elem_id="txt2img_hires_fix_row2"): - hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") - hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") - hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") - - elif category == "batch": - if not opts.dimensions_and_batch_together: - with FormRow(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - elif category == "scripts": - with FormGroup(elem_id="txt2img_script_container"): - custom_inputs = modules.scripts.scripts_txt2img.setup_ui() - - hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] - for input in hr_resolution_preview_inputs: - input.change( - fn=calc_resolution_hires, - inputs=hr_resolution_preview_inputs, - outputs=[hr_final_resolution], - show_progress=False, - ) - input.change( - None, - _js="onCalcResolutionHires", - inputs=hr_resolution_preview_inputs, - outputs=[], - show_progress=False, - ) - - txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) - parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) - - connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) - connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) - - txt2img_args = dict( - fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), - _js="submit", - inputs=[ - txt2img_prompt, - txt2img_negative_prompt, - txt2img_prompt_style, - txt2img_prompt_style2, - steps, - sampler_index, - restore_faces, - tiling, - batch_count, - batch_size, - cfg_scale, - seed, - subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, - height, - width, - enable_hr, - denoising_strength, - hr_scale, - hr_upscaler, - hr_second_pass_steps, - hr_resize_x, - hr_resize_y, - ] + custom_inputs, - - outputs=[ - txt2img_gallery, - generation_info, - html_info, - html_log, - ], - show_progress=False, - ) - - txt2img_prompt.submit(**txt2img_args) - submit.click(**txt2img_args) - - txt_prompt_img.change( - fn=modules.images.image_data, - inputs=[ - txt_prompt_img - ], - outputs=[ - txt2img_prompt, - txt_prompt_img - ] - ) - - enable_hr.change( - fn=lambda x: gr_show(x), - inputs=[enable_hr], - outputs=[hr_options], - show_progress = False, - ) - - txt2img_paste_fields = [ - (txt2img_prompt, "Prompt"), - (txt2img_negative_prompt, "Negative prompt"), - (steps, "Steps"), - (sampler_index, "Sampler"), - (restore_faces, "Face restoration"), - (cfg_scale, "CFG scale"), - (seed, "Seed"), - (width, "Size-1"), - (height, "Size-2"), - (batch_size, "Batch size"), - (subseed, "Variation seed"), - (subseed_strength, "Variation seed strength"), - (seed_resize_from_w, "Seed resize from-1"), - (seed_resize_from_h, "Seed resize from-2"), - (denoising_strength, "Denoising strength"), - (enable_hr, lambda d: "Denoising strength" in d), - (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), - (hr_scale, "Hires upscale"), - (hr_upscaler, "Hires upscaler"), - (hr_second_pass_steps, "Hires steps"), - (hr_resize_x, "Hires resize-1"), - (hr_resize_y, "Hires resize-2"), - *modules.scripts.scripts_txt2img.infotext_fields - ] - parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) - - txt2img_preview_params = [ - txt2img_prompt, - txt2img_negative_prompt, - steps, - sampler_index, - cfg_scale, - seed, - width, - height, - ] - - token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) - - modules.scripts.scripts_current = modules.scripts.scripts_img2img - modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) - - with gr.Blocks(analytics_enabled=False) as img2img_interface: - img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) - - with gr.Row(elem_id='img2img_progress_row'): - img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="img2img_progressbar") - img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) - setup_progressbar(progressbar, img2img_preview, 'img2img') - - with FormRow().style(equal_height=False): - with gr.Column(variant='panel', elem_id="img2img_settings"): - - with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: - with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): - init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) - - with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): - init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) - init_img_with_mask_orig = gr.State(None) - - use_color_sketch = cmd_opts.gradio_inpaint_tool == "color-sketch" - if use_color_sketch: - def update_orig(image, state): - if image is not None: - same_size = state is not None and state.size == image.size - has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) - edited = same_size and has_exact_match - return image if not edited or state is None else state - - init_img_with_mask.change(update_orig, [init_img_with_mask, init_img_with_mask_orig], init_img_with_mask_orig) - - init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") - init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") - - with FormRow(): - mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") - mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") - - with FormRow(): - mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") - inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") - - with FormRow(): - inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") - - with FormRow(): - with gr.Column(): - inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") - - with gr.Column(scale=4): - inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") - - with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): - hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' - gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") - img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") - img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") - - with FormRow(): - resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") - - for category in ordered_ui_categories(): - if category == "sampler": - steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") - - elif category == "dimensions": - with FormRow(): - with gr.Column(elem_id="img2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") - - if opts.dimensions_and_batch_together: - with gr.Column(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - elif category == "cfg": - with FormGroup(): - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") - - elif category == "seed": - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') - - elif category == "checkboxes": - with FormRow(elem_id="img2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") - - elif category == "batch": - if not opts.dimensions_and_batch_together: - with FormRow(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - elif category == "scripts": - with FormGroup(elem_id="img2img_script_container"): - custom_inputs = modules.scripts.scripts_img2img.setup_ui() - - img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) - parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) - - connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) - connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) - - img2img_prompt_img.change( - fn=modules.images.image_data, - inputs=[ - img2img_prompt_img - ], - outputs=[ - img2img_prompt, - img2img_prompt_img - ] - ) - - mask_mode.change( - lambda mode, img: { - init_img_with_mask: gr_show(mode == 0), - init_img_inpaint: gr_show(mode == 1), - init_mask_inpaint: gr_show(mode == 1), - }, - inputs=[mask_mode, init_img_with_mask], - outputs=[ - init_img_with_mask, - init_img_inpaint, - init_mask_inpaint, - ], - ) - - img2img_args = dict( - fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), - _js="submit_img2img", - inputs=[ - dummy_component, - img2img_prompt, - img2img_negative_prompt, - img2img_prompt_style, - img2img_prompt_style2, - init_img, - init_img_with_mask, - init_img_with_mask_orig, - init_img_inpaint, - init_mask_inpaint, - mask_mode, - steps, - sampler_index, - mask_blur, - mask_alpha, - inpainting_fill, - restore_faces, - tiling, - batch_count, - batch_size, - cfg_scale, - denoising_strength, - seed, - subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, - height, - width, - resize_mode, - inpaint_full_res, - inpaint_full_res_padding, - inpainting_mask_invert, - img2img_batch_input_dir, - img2img_batch_output_dir, - ] + custom_inputs, - outputs=[ - img2img_gallery, - generation_info, - html_info, - html_log, - ], - show_progress=False, - ) - - img2img_prompt.submit(**img2img_args) - submit.click(**img2img_args) - - img2img_interrogate.click( - fn=interrogate, - inputs=[init_img], - outputs=[img2img_prompt], - ) - - img2img_deepbooru.click( - fn=interrogate_deepbooru, - inputs=[init_img], - outputs=[img2img_prompt], - ) - - prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] - style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] - style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] - - for button, (prompt, negative_prompt) in zip([txt2img_save_style, img2img_save_style], prompts): - button.click( - fn=add_style, - _js="ask_for_style_name", - # Have to pass empty dummy component here, because the JavaScript and Python function have to accept - # the same number of parameters, but we only know the style-name after the JavaScript prompt - inputs=[dummy_component, prompt, negative_prompt], - outputs=[txt2img_prompt_style, img2img_prompt_style, txt2img_prompt_style2, img2img_prompt_style2], - ) - - for button, (prompt, negative_prompt), (style1, style2), js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): - button.click( - fn=apply_styles, - _js=js_func, - inputs=[prompt, negative_prompt, style1, style2], - outputs=[prompt, negative_prompt, style1, style2], - ) - - token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) - - img2img_paste_fields = [ - (img2img_prompt, "Prompt"), - (img2img_negative_prompt, "Negative prompt"), - (steps, "Steps"), - (sampler_index, "Sampler"), - (restore_faces, "Face restoration"), - (cfg_scale, "CFG scale"), - (seed, "Seed"), - (width, "Size-1"), - (height, "Size-2"), - (batch_size, "Batch size"), - (subseed, "Variation seed"), - (subseed_strength, "Variation seed strength"), - (seed_resize_from_w, "Seed resize from-1"), - (seed_resize_from_h, "Seed resize from-2"), - (denoising_strength, "Denoising strength"), - (mask_blur, "Mask blur"), - *modules.scripts.scripts_img2img.infotext_fields - ] - parameters_copypaste.add_paste_fields("img2img", init_img, img2img_paste_fields) - parameters_copypaste.add_paste_fields("inpaint", init_img_with_mask, img2img_paste_fields) - - modules.scripts.scripts_current = None - - with gr.Blocks(analytics_enabled=False) as extras_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - with gr.Tabs(elem_id="mode_extras"): - with gr.TabItem('Single Image', elem_id="extras_single_tab"): - extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") - - with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): - image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") - - with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): - extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") - extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") - show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") - - submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') - - with gr.Tabs(elem_id="extras_resize_mode"): - with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): - upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") - with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): - with gr.Group(): - with gr.Row(): - upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") - upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") - upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") - - with gr.Group(): - extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - - with gr.Group(): - extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") - - with gr.Group(): - gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") - - with gr.Group(): - codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") - codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") - - with gr.Group(): - upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") - - result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) - - submit.click( - fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), - _js="get_extras_tab_index", - inputs=[ - dummy_component, - dummy_component, - extras_image, - image_batch, - extras_batch_input_dir, - extras_batch_output_dir, - show_extras_results, - gfpgan_visibility, - codeformer_visibility, - codeformer_weight, - upscaling_resize, - upscaling_resize_w, - upscaling_resize_h, - upscaling_crop, - extras_upscaler_1, - extras_upscaler_2, - extras_upscaler_2_visibility, - upscale_before_face_fix, - ], - outputs=[ - result_images, - html_info_x, - html_info, - ] - ) - parameters_copypaste.add_paste_fields("extras", extras_image, None) - - extras_image.change( - fn=modules.extras.clear_cache, - inputs=[], outputs=[] - ) - - with gr.Blocks(analytics_enabled=False) as pnginfo_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - image = gr.Image(elem_id="pnginfo_image", label="Source", source="upload", interactive=True, type="pil") - - with gr.Column(variant='panel'): - html = gr.HTML() - generation_info = gr.Textbox(visible=False, elem_id="pnginfo_generation_info") - html2 = gr.HTML() - with gr.Row(): - buttons = parameters_copypaste.create_buttons(["txt2img", "img2img", "inpaint", "extras"]) - parameters_copypaste.bind_buttons(buttons, image, generation_info) - - image.change( - fn=wrap_gradio_call(modules.extras.run_pnginfo), - inputs=[image], - outputs=[html, generation_info, html2], - ) - - with gr.Blocks(analytics_enabled=False) as modelmerger_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") - - with gr.Row(): - primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") - create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") - - secondary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_secondary_model_name", label="Secondary model (B)") - create_refresh_button(secondary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_B") - - tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") - create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") - - custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") - interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") - interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") - - with gr.Row(): - checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") - save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") - - modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') - - with gr.Column(variant='panel'): - submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) - - with gr.Blocks(analytics_enabled=False) as train_interface: - with gr.Row().style(equal_height=False): - gr.HTML(value="

See wiki for detailed explanation.

") - - with gr.Row().style(equal_height=False): - with gr.Tabs(elem_id="train_tabs"): - - with gr.Tab(label="Create embedding"): - new_embedding_name = gr.Textbox(label="Name", elem_id="train_new_embedding_name") - initialization_text = gr.Textbox(label="Initialization text", value="*", elem_id="train_initialization_text") - nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1, elem_id="train_nvpt") - overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding", elem_id="train_overwrite_old_embedding") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - create_embedding = gr.Button(value="Create embedding", variant='primary', elem_id="train_create_embedding") - - with gr.Tab(label="Create hypernetwork"): - new_hypernetwork_name = gr.Textbox(label="Name", elem_id="train_new_hypernetwork_name") - new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"], elem_id="train_new_hypernetwork_sizes") - new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'", elem_id="train_new_hypernetwork_layer_structure") - new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys, elem_id="train_new_hypernetwork_activation_func") - new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") - new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") - new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") - new_hypernetwork_dropout_structure = gr.Textbox("0, 0, 0", label="Enter hypernetwork Dropout structure (or empty). Recommended : 0~0.35 incrementing sequence: 0, 0.05, 0.15", placeholder="1st and last digit must be 0 and values should be between 0 and 1. ex:'0, 0.01, 0'") - overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary', elem_id="train_create_hypernetwork") - - with gr.Tab(label="Preprocess images"): - process_src = gr.Textbox(label='Source directory', elem_id="train_process_src") - process_dst = gr.Textbox(label='Destination directory', elem_id="train_process_dst") - process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_process_width") - process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_process_height") - preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"], elem_id="train_preprocess_txt_action") - - with gr.Row(): - process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") - process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") - process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") - process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") - process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") - - with gr.Row(visible=False) as process_split_extra_row: - process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_split_threshold") - process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05, elem_id="train_process_overlap_ratio") - - with gr.Row(visible=False) as process_focal_crop_row: - process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_face_weight") - process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") - process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") - process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - with gr.Row(): - interrupt_preprocessing = gr.Button("Interrupt", elem_id="train_interrupt_preprocessing") - run_preprocess = gr.Button(value="Preprocess", variant='primary', elem_id="train_run_preprocess") - - process_split.change( - fn=lambda show: gr_show(show), - inputs=[process_split], - outputs=[process_split_extra_row], - ) - - process_focal_crop.change( - fn=lambda show: gr_show(show), - inputs=[process_focal_crop], - outputs=[process_focal_crop_row], - ) - - def get_textual_inversion_template_names(): - return sorted([x for x in textual_inversion.textual_inversion_templates]) - - with gr.Tab(label="Train"): - gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") - with FormRow(): - train_embedding_name = gr.Dropdown(label='Embedding', elem_id="train_embedding", choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())) - create_refresh_button(train_embedding_name, sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings, lambda: {"choices": sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())}, "refresh_train_embedding_name") - - train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) - create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") - - with FormRow(): - embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") - hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") - - with FormRow(): - clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) - clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) - - with FormRow(): - batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") - gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") - - dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") - log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") - - with FormRow(): - template_file = gr.Dropdown(label='Prompt template', value="style_filewords.txt", elem_id="train_template_file", choices=get_textual_inversion_template_names()) - create_refresh_button(template_file, textual_inversion.list_textual_inversion_templates, lambda: {"choices": get_textual_inversion_template_names()}, "refrsh_train_template_file") - - training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") - training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") - varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") - steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") - - with FormRow(): - create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") - save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") - - save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") - preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") - - shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") - tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") - - latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") - - with gr.Row(): - train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") - interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") - train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") - - params = script_callbacks.UiTrainTabParams(txt2img_preview_params) - - script_callbacks.ui_train_tabs_callback(params) - - with gr.Column(): - progressbar = gr.HTML(elem_id="ti_progressbar") - ti_output = gr.Text(elem_id="ti_output", value="", show_label=False) - - ti_gallery = gr.Gallery(label='Output', show_label=False, elem_id='ti_gallery').style(grid=4) - ti_preview = gr.Image(elem_id='ti_preview', visible=False) - ti_progress = gr.HTML(elem_id="ti_progress", value="") - ti_outcome = gr.HTML(elem_id="ti_error", value="") - setup_progressbar(progressbar, ti_preview, 'ti', textinfo=ti_progress) - - create_embedding.click( - fn=modules.textual_inversion.ui.create_embedding, - inputs=[ - new_embedding_name, - initialization_text, - nvpt, - overwrite_old_embedding, - ], - outputs=[ - train_embedding_name, - ti_output, - ti_outcome, - ] - ) - - create_hypernetwork.click( - fn=modules.hypernetworks.ui.create_hypernetwork, - inputs=[ - new_hypernetwork_name, - new_hypernetwork_sizes, - overwrite_old_hypernetwork, - new_hypernetwork_layer_structure, - new_hypernetwork_activation_func, - new_hypernetwork_initialization_option, - new_hypernetwork_add_layer_norm, - new_hypernetwork_use_dropout, - new_hypernetwork_dropout_structure - ], - outputs=[ - train_hypernetwork_name, - ti_output, - ti_outcome, - ] - ) - - run_preprocess.click( - fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.preprocess, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - process_src, - process_dst, - process_width, - process_height, - preprocess_txt_action, - process_flip, - process_split, - process_caption, - process_caption_deepbooru, - process_split_threshold, - process_overlap_ratio, - process_focal_crop, - process_focal_crop_face_weight, - process_focal_crop_entropy_weight, - process_focal_crop_edges_weight, - process_focal_crop_debug, - ], - outputs=[ - ti_output, - ti_outcome, - ], - ) - - train_embedding.click( - fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.train_embedding, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - train_embedding_name, - embedding_learn_rate, - batch_size, - gradient_step, - dataset_directory, - log_directory, - training_width, - training_height, - varsize, - steps, - clip_grad_mode, - clip_grad_value, - shuffle_tags, - tag_drop_out, - latent_sampling_method, - create_image_every, - save_embedding_every, - template_file, - save_image_with_stored_embedding, - preview_from_txt2img, - *txt2img_preview_params, - ], - outputs=[ - ti_output, - ti_outcome, - ] - ) - - train_hypernetwork.click( - fn=wrap_gradio_gpu_call(modules.hypernetworks.ui.train_hypernetwork, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - train_hypernetwork_name, - hypernetwork_learn_rate, - batch_size, - gradient_step, - dataset_directory, - log_directory, - training_width, - training_height, - varsize, - steps, - clip_grad_mode, - clip_grad_value, - shuffle_tags, - tag_drop_out, - latent_sampling_method, - create_image_every, - save_embedding_every, - template_file, - preview_from_txt2img, - *txt2img_preview_params, - ], - outputs=[ - ti_output, - ti_outcome, - ] - ) - - interrupt_training.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - interrupt_preprocessing.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - def create_setting_component(key, is_quicksettings=False): - def fun(): - return opts.data[key] if key in opts.data else opts.data_labels[key].default - - info = opts.data_labels[key] - t = type(info.default) - - args = info.component_args() if callable(info.component_args) else info.component_args - - if info.component is not None: - comp = info.component - elif t == str: - comp = gr.Textbox - elif t == int: - comp = gr.Number - elif t == bool: - comp = gr.Checkbox - else: - raise Exception(f'bad options item type: {str(t)} for key {key}') - - elem_id = "setting_"+key - - if info.refresh is not None: - if is_quicksettings: - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) - else: - with FormRow(): - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) - else: - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - - return res - - components = [] - component_dict = {} - - script_callbacks.ui_settings_callback() - opts.reorder() - - def run_settings(*args): - changed = [] - - for key, value, comp in zip(opts.data_labels.keys(), args, components): - assert comp == dummy_component or opts.same_type(value, opts.data_labels[key].default), f"Bad value for setting {key}: {value}; expecting {type(opts.data_labels[key].default).__name__}" - - for key, value, comp in zip(opts.data_labels.keys(), args, components): - if comp == dummy_component: - continue - - if opts.set(key, value): - changed.append(key) - - try: - opts.save(shared.config_filename) - except RuntimeError: - return opts.dumpjson(), f'{len(changed)} settings changed without save: {", ".join(changed)}.' - return opts.dumpjson(), f'{len(changed)} settings changed{": " if len(changed) > 0 else ""}{", ".join(changed)}.' - - def run_settings_single(value, key): - if not opts.same_type(value, opts.data_labels[key].default): - return gr.update(visible=True), opts.dumpjson() - - if not opts.set(key, value): - return gr.update(value=getattr(opts, key)), opts.dumpjson() - - opts.save(shared.config_filename) - - return gr.update(value=value), opts.dumpjson() - - with gr.Blocks(analytics_enabled=False) as settings_interface: - with gr.Row(): - with gr.Column(scale=6): - settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") - with gr.Column(): - restart_gradio = gr.Button(value='Reload UI', variant='primary', elem_id="settings_restart_gradio") - - result = gr.HTML(elem_id="settings_result") - - quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] - quicksettings_names = {x: i for i, x in enumerate(quicksettings_names) if x != 'quicksettings'} - - quicksettings_list = [] - - previous_section = None - current_tab = None - with gr.Tabs(elem_id="settings"): - for i, (k, item) in enumerate(opts.data_labels.items()): - section_must_be_skipped = item.section[0] is None - - if previous_section != item.section and not section_must_be_skipped: - elem_id, text = item.section - - if current_tab is not None: - current_tab.__exit__() - - current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) - current_tab.__enter__() - - previous_section = item.section - - if k in quicksettings_names and not shared.cmd_opts.freeze_settings: - quicksettings_list.append((i, k, item)) - components.append(dummy_component) - elif section_must_be_skipped: - components.append(dummy_component) - else: - component = create_setting_component(k) - component_dict[k] = component - components.append(component) - - if current_tab is not None: - current_tab.__exit__() - - with gr.TabItem("Actions"): - request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") - download_localization = gr.Button(value='Download localization template', elem_id="download_localization") - reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") - - if os.path.exists("html/licenses.html"): - with open("html/licenses.html", encoding="utf8") as file: - with gr.TabItem("Licenses"): - gr.HTML(file.read(), elem_id="licenses") - - gr.Button(value="Show all pages", elem_id="settings_show_all_pages") - - request_notifications.click( - fn=lambda: None, - inputs=[], - outputs=[], - _js='function(){}' - ) - - download_localization.click( - fn=lambda: None, - inputs=[], - outputs=[], - _js='download_localization' - ) - - def reload_scripts(): - modules.scripts.reload_script_body_only() - reload_javascript() # need to refresh the html page - - reload_script_bodies.click( - fn=reload_scripts, - inputs=[], - outputs=[] - ) - - def request_restart(): - shared.state.interrupt() - shared.state.need_restart = True - - restart_gradio.click( - fn=request_restart, - _js='restart_reload', - inputs=[], - outputs=[], - ) - - interfaces = [ - (txt2img_interface, "txt2img", "txt2img"), - (img2img_interface, "img2img", "img2img"), - (extras_interface, "Extras", "extras"), - (pnginfo_interface, "PNG Info", "pnginfo"), - (modelmerger_interface, "Checkpoint Merger", "modelmerger"), - (train_interface, "Train", "ti"), - ] - - css = "" - - for cssfile in modules.scripts.list_files_with_name("style.css"): - if not os.path.isfile(cssfile): - continue - - with open(cssfile, "r", encoding="utf8") as file: - css += file.read() + "\n" - - if os.path.exists(os.path.join(script_path, "user.css")): - with open(os.path.join(script_path, "user.css"), "r", encoding="utf8") as file: - css += file.read() + "\n" - - if not cmd_opts.no_progressbar_hiding: - css += css_hide_progressbar - - interfaces += script_callbacks.ui_tabs_callback() - interfaces += [(settings_interface, "Settings", "settings")] - - extensions_interface = ui_extensions.create_ui() - interfaces += [(extensions_interface, "Extensions", "extensions")] - - with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: - with gr.Row(elem_id="quicksettings"): - for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): - component = create_setting_component(k, is_quicksettings=True) - component_dict[k] = component - - parameters_copypaste.integrate_settings_paste_fields(component_dict) - parameters_copypaste.run_bind() - - with gr.Tabs(elem_id="tabs") as tabs: - for interface, label, ifid in interfaces: - with gr.TabItem(label, id=ifid, elem_id='tab_' + ifid): - interface.render() - - if os.path.exists(os.path.join(script_path, "notification.mp3")): - audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) - - if os.path.exists("html/footer.html"): - with open("html/footer.html", encoding="utf8") as file: - footer = file.read() - footer = footer.format(versions=versions_html()) - gr.HTML(footer, elem_id="footer") - - text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) - settings_submit.click( - fn=wrap_gradio_call(run_settings, extra_outputs=[gr.update()]), - inputs=components, - outputs=[text_settings, result], - ) - - for i, k, item in quicksettings_list: - component = component_dict[k] - - component.change( - fn=lambda value, k=k: run_settings_single(value, key=k), - inputs=[component], - outputs=[component, text_settings], - ) - - component_keys = [k for k in opts.data_labels.keys() if k in component_dict] - - def get_settings_values(): - return [getattr(opts, key) for key in component_keys] - - demo.load( - fn=get_settings_values, - inputs=[], - outputs=[component_dict[k] for k in component_keys], - ) - - def modelmerger(*args): - try: - results = modules.extras.run_modelmerger(*args) - except Exception as e: - print("Error loading/saving model file:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - modules.sd_models.list_models() # to remove the potentially missing models from the list - return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] - return results - - modelmerger_merge.click( - fn=modelmerger, - inputs=[ - primary_model_name, - secondary_model_name, - tertiary_model_name, - interp_method, - interp_amount, - save_as_half, - custom_name, - checkpoint_format, - ], - outputs=[ - submit_result, - primary_model_name, - secondary_model_name, - tertiary_model_name, - component_dict['sd_model_checkpoint'], - ] - ) - - ui_config_file = cmd_opts.ui_config_file - ui_settings = {} - settings_count = len(ui_settings) - error_loading = False - - try: - if os.path.exists(ui_config_file): - with open(ui_config_file, "r", encoding="utf8") as file: - ui_settings = json.load(file) - except Exception: - error_loading = True - print("Error loading settings:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - - def loadsave(path, x): - def apply_field(obj, field, condition=None, init_field=None): - key = path + "/" + field - - if getattr(obj, 'custom_script_source', None) is not None: - key = 'customscript/' + obj.custom_script_source + '/' + key - - if getattr(obj, 'do_not_save_to_config', False): - return - - saved_value = ui_settings.get(key, None) - if saved_value is None: - ui_settings[key] = getattr(obj, field) - elif condition and not condition(saved_value): - print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') - else: - setattr(obj, field, saved_value) - if init_field is not None: - init_field(saved_value) - - if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number, gr.Dropdown] and x.visible: - apply_field(x, 'visible') - - if type(x) == gr.Slider: - apply_field(x, 'value') - apply_field(x, 'minimum') - apply_field(x, 'maximum') - apply_field(x, 'step') - - if type(x) == gr.Radio: - apply_field(x, 'value', lambda val: val in x.choices) - - if type(x) == gr.Checkbox: - apply_field(x, 'value') - - if type(x) == gr.Textbox: - apply_field(x, 'value') - - if type(x) == gr.Number: - apply_field(x, 'value') - - if type(x) == gr.Dropdown: - apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) - - visit(txt2img_interface, loadsave, "txt2img") - visit(img2img_interface, loadsave, "img2img") - visit(extras_interface, loadsave, "extras") - visit(modelmerger_interface, loadsave, "modelmerger") - visit(train_interface, loadsave, "train") - - if not error_loading and (not os.path.exists(ui_config_file) or settings_count != len(ui_settings)): - with open(ui_config_file, "w", encoding="utf8") as file: - json.dump(ui_settings, file, indent=4) - - return demo - - -def reload_javascript(): - with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: - javascript = f'' - - scripts_list = modules.scripts.list_scripts("javascript", ".js") - - for basedir, filename, path in scripts_list: - with open(path, "r", encoding="utf8") as jsfile: - javascript += f"\n" - - if cmd_opts.theme is not None: - javascript += f"\n\n" - - javascript += f"\n" - - def template_response(*args, **kwargs): - res = shared.GradioTemplateResponseOriginal(*args, **kwargs) - res.body = res.body.replace( - b'', f'{javascript}'.encode("utf8")) - res.init_headers() - return res - - gradio.routes.templates.TemplateResponse = template_response - - -if not hasattr(shared, 'GradioTemplateResponseOriginal'): - shared.GradioTemplateResponseOriginal = gradio.routes.templates.TemplateResponse - - -def versions_html(): - import torch - import launch - - python_version = ".".join([str(x) for x in sys.version_info[0:3]]) - commit = launch.commit_hash() - short_commit = commit[0:8] - - if shared.xformers_available: - import xformers - xformers_version = xformers.__version__ - else: - xformers_version = "N/A" - - return f""" -python: {python_version} - •  -torch: {torch.__version__} - •  -xformers: {xformers_version} - •  -gradio: {gr.__version__} - •  -commit: {short_commit} -""" diff --git a/modules/ui.py b/modules/ui.py new file mode 100644 index 00000000..9b9081b5 --- /dev/null +++ b/modules/ui.py @@ -0,0 +1,1928 @@ +import html +import json +import math +import mimetypes +import os +import platform +import random +import subprocess as sp +import sys +import tempfile +import time +import traceback +from functools import partial, reduce + +import gradio as gr +import gradio.routes +import gradio.utils +import numpy as np +from PIL import Image, PngImagePlugin +from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call + +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru +from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML +from modules.paths import script_path + +from modules.shared import opts, cmd_opts, restricted_opts + +import modules.codeformer_model +import modules.generation_parameters_copypaste as parameters_copypaste +import modules.gfpgan_model +import modules.hypernetworks.ui +import modules.scripts +import modules.shared as shared +import modules.styles +import modules.textual_inversion.ui +from modules import prompt_parser +from modules.images import save_image +from modules.sd_hijack import model_hijack +from modules.sd_samplers import samplers, samplers_for_img2img +from modules.textual_inversion import textual_inversion +import modules.hypernetworks.ui +from modules.generation_parameters_copypaste import image_from_url_text + +# this is a fix for Windows users. Without it, javascript files will be served with text/html content-type and the browser will not show any UI +mimetypes.init() +mimetypes.add_type('application/javascript', '.js') + +if not cmd_opts.share and not cmd_opts.listen: + # fix gradio phoning home + gradio.utils.version_check = lambda: None + gradio.utils.get_local_ip_address = lambda: '127.0.0.1' + +if cmd_opts.ngrok is not None: + import modules.ngrok as ngrok + print('ngrok authtoken detected, trying to connect...') + ngrok.connect( + cmd_opts.ngrok, + cmd_opts.port if cmd_opts.port is not None else 7860, + cmd_opts.ngrok_region + ) + + +def gr_show(visible=True): + return {"visible": visible, "__type__": "update"} + + +sample_img2img = "assets/stable-samples/img2img/sketch-mountains-input.jpg" +sample_img2img = sample_img2img if os.path.exists(sample_img2img) else None + +css_hide_progressbar = """ +.wrap .m-12 svg { display:none!important; } +.wrap .m-12::before { content:"Loading..." } +.wrap .z-20 svg { display:none!important; } +.wrap .z-20::before { content:"Loading..." } +.progress-bar { display:none!important; } +.meta-text { display:none!important; } +.meta-text-center { display:none!important; } +""" + +# Using constants for these since the variation selector isn't visible. +# Important that they exactly match script.js for tooltip to work. +random_symbol = '\U0001f3b2\ufe0f' # 🎲️ +reuse_symbol = '\u267b\ufe0f' # ♻️ +paste_symbol = '\u2199\ufe0f' # ↙ +folder_symbol = '\U0001f4c2' # 📂 +refresh_symbol = '\U0001f504' # 🔄 +save_style_symbol = '\U0001f4be' # 💾 +apply_style_symbol = '\U0001f4cb' # 📋 +clear_prompt_symbol = '\U0001F5D1' # 🗑️ + + +def plaintext_to_html(text): + text = "

" + "
\n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

" + return text + +def send_gradio_gallery_to_image(x): + if len(x) == 0: + return None + return image_from_url_text(x[0]) + +def save_files(js_data, images, do_make_zip, index): + import csv + filenames = [] + fullfns = [] + + #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it + class MyObject: + def __init__(self, d=None): + if d is not None: + for key, value in d.items(): + setattr(self, key, value) + + data = json.loads(js_data) + + p = MyObject(data) + path = opts.outdir_save + save_to_dirs = opts.use_save_to_dirs_for_ui + extension: str = opts.samples_format + start_index = 0 + + if index > -1 and opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only + + images = [images[index]] + start_index = index + + os.makedirs(opts.outdir_save, exist_ok=True) + + with open(os.path.join(opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: + at_start = file.tell() == 0 + writer = csv.writer(file) + if at_start: + writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) + + for image_index, filedata in enumerate(images, start_index): + image = image_from_url_text(filedata) + + is_grid = image_index < p.index_of_first_image + i = 0 if is_grid else (image_index - p.index_of_first_image) + + fullfn, txt_fullfn = save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) + + filename = os.path.relpath(fullfn, path) + filenames.append(filename) + fullfns.append(fullfn) + if txt_fullfn: + filenames.append(os.path.basename(txt_fullfn)) + fullfns.append(txt_fullfn) + + writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) + + # Make Zip + if do_make_zip: + zip_filepath = os.path.join(path, "images.zip") + + from zipfile import ZipFile + with ZipFile(zip_filepath, "w") as zip_file: + for i in range(len(fullfns)): + with open(fullfns[i], mode="rb") as f: + zip_file.writestr(filenames[i], f.read()) + fullfns.insert(0, zip_filepath) + + return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") + + +def calc_time_left(progress, threshold, label, force_display, show_eta): + if progress == 0: + return "" + else: + time_since_start = time.time() - shared.state.time_start + eta = (time_since_start/progress) + eta_relative = eta-time_since_start + if (eta_relative > threshold and show_eta) or force_display: + if eta_relative > 3600: + return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) + elif eta_relative > 60: + return label + time.strftime('%M:%S', time.gmtime(eta_relative)) + else: + return label + time.strftime('%Ss', time.gmtime(eta_relative)) + else: + return "" + + +def check_progress_call(id_part): + if shared.state.job_count == 0: + return "", gr_show(False), gr_show(False), gr_show(False) + + progress = 0 + + if shared.state.job_count > 0: + progress += shared.state.job_no / shared.state.job_count + if shared.state.sampling_steps > 0: + progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps + + # Show progress percentage and time left at the same moment, and base it also on steps done + show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 + + time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) + if time_left != "": + shared.state.time_left_force_display = True + + progress = min(progress, 1) + + progressbar = "" + if opts.show_progressbar: + progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" + + image = gr_show(False) + preview_visibility = gr_show(False) + + if opts.show_progress_every_n_steps != 0: + shared.state.set_current_image() + image = shared.state.current_image + + if image is None: + image = gr.update(value=None) + else: + preview_visibility = gr_show(True) + + if shared.state.textinfo is not None: + textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) + else: + textinfo_result = gr_show(False) + + return f"

{progressbar}

", preview_visibility, image, textinfo_result + + +def check_progress_call_initial(id_part): + shared.state.job_count = -1 + shared.state.current_latent = None + shared.state.current_image = None + shared.state.textinfo = None + shared.state.time_start = time.time() + shared.state.time_left_force_display = False + + return check_progress_call(id_part) + + +def visit(x, func, path=""): + if hasattr(x, 'children'): + for c in x.children: + visit(c, func, path) + elif x.label is not None: + func(path + "/" + str(x.label), x) + + +def add_style(name: str, prompt: str, negative_prompt: str): + if name is None: + return [gr_show() for x in range(4)] + + style = modules.styles.PromptStyle(name, prompt, negative_prompt) + shared.prompt_styles.styles[style.name] = style + # Save all loaded prompt styles: this allows us to update the storage format in the future more easily, because we + # reserialize all styles every time we save them + shared.prompt_styles.save_styles(shared.styles_filename) + + return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] + + +def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): + from modules import processing, devices + + if not enable: + return "" + + p = processing.StableDiffusionProcessingTxt2Img(width=width, height=height, enable_hr=True, hr_scale=hr_scale, hr_resize_x=hr_resize_x, hr_resize_y=hr_resize_y) + + with devices.autocast(): + p.init([""], [0], [0]) + + return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" + + +def apply_styles(prompt, prompt_neg, style1_name, style2_name): + prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) + prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, [style1_name, style2_name]) + + return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] + + +def interrogate(image): + prompt = shared.interrogator.interrogate(image.convert("RGB")) + + return gr_show(True) if prompt is None else prompt + + +def interrogate_deepbooru(image): + prompt = deepbooru.model.tag(image) + return gr_show(True) if prompt is None else prompt + + +def create_seed_inputs(target_interface): + with FormRow(elem_id=target_interface + '_seed_row'): + seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') + seed.style(container=False) + random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') + reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') + + with gr.Group(elem_id=target_interface + '_subseed_show_box'): + seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) + + # Components to show/hide based on the 'Extra' checkbox + seed_extras = [] + + with FormRow(visible=False, elem_id=target_interface + '_subseed_row') as seed_extra_row_1: + seed_extras.append(seed_extra_row_1) + subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') + subseed.style(container=False) + random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') + reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') + subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') + + with FormRow(visible=False) as seed_extra_row_2: + seed_extras.append(seed_extra_row_2) + seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') + seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') + + random_seed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[seed]) + random_subseed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[subseed]) + + def change_visibility(show): + return {comp: gr_show(show) for comp in seed_extras} + + seed_checkbox.change(change_visibility, show_progress=False, inputs=[seed_checkbox], outputs=seed_extras) + + return seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox + + + +def connect_clear_prompt(button): + """Given clear button, prompt, and token_counter objects, setup clear prompt button click event""" + button.click( + _js="clear_prompt", + fn=None, + inputs=[], + outputs=[], + ) + + +def connect_reuse_seed(seed: gr.Number, reuse_seed: gr.Button, generation_info: gr.Textbox, dummy_component, is_subseed): + """ Connects a 'reuse (sub)seed' button's click event so that it copies last used + (sub)seed value from generation info the to the seed field. If copying subseed and subseed strength + was 0, i.e. no variation seed was used, it copies the normal seed value instead.""" + def copy_seed(gen_info_string: str, index): + res = -1 + + try: + gen_info = json.loads(gen_info_string) + index -= gen_info.get('index_of_first_image', 0) + + if is_subseed and gen_info.get('subseed_strength', 0) > 0: + all_subseeds = gen_info.get('all_subseeds', [-1]) + res = all_subseeds[index if 0 <= index < len(all_subseeds) else 0] + else: + all_seeds = gen_info.get('all_seeds', [-1]) + res = all_seeds[index if 0 <= index < len(all_seeds) else 0] + + except json.decoder.JSONDecodeError as e: + if gen_info_string != '': + print("Error parsing JSON generation info:", file=sys.stderr) + print(gen_info_string, file=sys.stderr) + + return [res, gr_show(False)] + + reuse_seed.click( + fn=copy_seed, + _js="(x, y) => [x, selected_gallery_index()]", + show_progress=False, + inputs=[generation_info, dummy_component], + outputs=[seed, dummy_component] + ) + + +def update_token_counter(text, steps): + try: + _, prompt_flat_list, _ = prompt_parser.get_multicond_prompt_list([text]) + prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules(prompt_flat_list, steps) + + except Exception: + # a parsing error can happen here during typing, and we don't want to bother the user with + # messages related to it in console + prompt_schedules = [[[steps, text]]] + + flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) + prompts = [prompt_text for step, prompt_text in flat_prompts] + token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) + style_class = ' class="red"' if (token_count > max_length) else "" + return f"{token_count}/{max_length}" + + +def create_toprow(is_img2img): + id_part = "img2img" if is_img2img else "txt2img" + + with gr.Row(elem_id="toprow"): + with gr.Column(scale=6): + with gr.Row(): + with gr.Column(scale=80): + with gr.Row(): + prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, + placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)" + ) + + with gr.Row(): + with gr.Column(scale=80): + with gr.Row(): + negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, + placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)" + ) + + with gr.Column(scale=1, elem_id="roll_col"): + paste = gr.Button(value=paste_symbol, elem_id="paste") + save_style = gr.Button(value=save_style_symbol, elem_id="style_create") + prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") + clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") + token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") + token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") + + clear_prompt_button.click( + fn=lambda *x: x, + _js="confirm_clear_prompt", + inputs=[prompt, negative_prompt], + outputs=[prompt, negative_prompt], + ) + + button_interrogate = None + button_deepbooru = None + if is_img2img: + with gr.Column(scale=1, elem_id="interrogate_col"): + button_interrogate = gr.Button('Interrogate\nCLIP', elem_id="interrogate") + button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") + + with gr.Column(scale=1): + with gr.Row(): + skip = gr.Button('Skip', elem_id=f"{id_part}_skip") + interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") + submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') + + skip.click( + fn=lambda: shared.state.skip(), + inputs=[], + outputs=[], + ) + + interrupt.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + with gr.Row(): + with gr.Column(scale=1, elem_id="style_pos_col"): + prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) + + with gr.Column(scale=1, elem_id="style_neg_col"): + prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) + + return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button + + +def setup_progressbar(progressbar, preview, id_part, textinfo=None): + if textinfo is None: + textinfo = gr.HTML(visible=False) + + check_progress = gr.Button('Check progress', elem_id=f"{id_part}_check_progress", visible=False) + check_progress.click( + fn=lambda: check_progress_call(id_part), + show_progress=False, + inputs=[], + outputs=[progressbar, preview, preview, textinfo], + ) + + check_progress_initial = gr.Button('Check progress (first)', elem_id=f"{id_part}_check_progress_initial", visible=False) + check_progress_initial.click( + fn=lambda: check_progress_call_initial(id_part), + show_progress=False, + inputs=[], + outputs=[progressbar, preview, preview, textinfo], + ) + + +def apply_setting(key, value): + if value is None: + return gr.update() + + if shared.cmd_opts.freeze_settings: + return gr.update() + + # dont allow model to be swapped when model hash exists in prompt + if key == "sd_model_checkpoint" and opts.disable_weights_auto_swap: + return gr.update() + + if key == "sd_model_checkpoint": + ckpt_info = sd_models.get_closet_checkpoint_match(value) + + if ckpt_info is not None: + value = ckpt_info.title + else: + return gr.update() + + comp_args = opts.data_labels[key].component_args + if comp_args and isinstance(comp_args, dict) and comp_args.get('visible') is False: + return + + valtype = type(opts.data_labels[key].default) + oldval = opts.data.get(key, None) + opts.data[key] = valtype(value) if valtype != type(None) else value + if oldval != value and opts.data_labels[key].onchange is not None: + opts.data_labels[key].onchange() + + opts.save(shared.config_filename) + return value + + +def update_generation_info(args): + generation_info, html_info, img_index = args + try: + generation_info = json.loads(generation_info) + if img_index < 0 or img_index >= len(generation_info["infotexts"]): + return html_info + return plaintext_to_html(generation_info["infotexts"][img_index]) + except Exception: + pass + # if the json parse or anything else fails, just return the old html_info + return html_info + + +def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): + def refresh(): + refresh_method() + args = refreshed_args() if callable(refreshed_args) else refreshed_args + + for k, v in args.items(): + setattr(refresh_component, k, v) + + return gr.update(**(args or {})) + + refresh_button = ToolButton(value=refresh_symbol, elem_id=elem_id) + refresh_button.click( + fn=refresh, + inputs=[], + outputs=[refresh_component] + ) + return refresh_button + + +def create_output_panel(tabname, outdir): + def open_folder(f): + if not os.path.exists(f): + print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') + return + elif not os.path.isdir(f): + print(f""" +WARNING +An open_folder request was made with an argument that is not a folder. +This could be an error or a malicious attempt to run code on your computer. +Requested path was: {f} +""", file=sys.stderr) + return + + if not shared.cmd_opts.hide_ui_dir_config: + path = os.path.normpath(f) + if platform.system() == "Windows": + os.startfile(path) + elif platform.system() == "Darwin": + sp.Popen(["open", path]) + elif "microsoft-standard-WSL2" in platform.uname().release: + sp.Popen(["wsl-open", path]) + else: + sp.Popen(["xdg-open", path]) + + with gr.Column(variant='panel'): + with gr.Group(): + result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) + + generation_info = None + with gr.Column(): + with gr.Row(elem_id=f"image_buttons_{tabname}"): + open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') + + if tabname != "extras": + save = gr.Button('Save', elem_id=f'save_{tabname}') + save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') + + buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) + + open_folder_button.click( + fn=lambda: open_folder(opts.outdir_samples or outdir), + inputs=[], + outputs=[], + ) + + if tabname != "extras": + with gr.Row(): + download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') + + with gr.Group(): + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') + + generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') + if tabname == 'txt2img' or tabname == 'img2img': + generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") + generation_info_button.click( + fn=update_generation_info, + _js="(x, y) => [x, y, selected_gallery_index()]", + inputs=[generation_info, html_info], + outputs=[html_info], + preprocess=False + ) + + save.click( + fn=wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ] + ) + + save_zip.click( + fn=wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ] + ) + + else: + html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') + + parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) + return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log + + +def create_sampler_and_steps_selection(choices, tabname): + if opts.samplers_in_dropdown: + with FormRow(elem_id=f"sampler_selection_{tabname}"): + sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) + else: + with FormGroup(elem_id=f"sampler_selection_{tabname}"): + steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) + sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") + + return steps, sampler_index + + +def ordered_ui_categories(): + user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} + + for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): + yield category + + +def create_ui(): + import modules.img2img + import modules.txt2img + + reload_javascript() + + parameters_copypaste.reset() + + modules.scripts.scripts_current = modules.scripts.scripts_txt2img + modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) + + with gr.Blocks(analytics_enabled=False) as txt2img_interface: + txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) + + dummy_component = gr.Label(visible=False) + txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) + + with gr.Row(elem_id='txt2img_progress_row'): + with gr.Column(scale=1): + pass + + with gr.Column(scale=1): + progressbar = gr.HTML(elem_id="txt2img_progressbar") + txt2img_preview = gr.Image(elem_id='txt2img_preview', visible=False) + setup_progressbar(progressbar, txt2img_preview, 'txt2img') + + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel', elem_id="txt2img_settings"): + for category in ordered_ui_categories(): + if category == "sampler": + steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") + + elif category == "dimensions": + with FormRow(): + with gr.Column(elem_id="txt2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") + + if opts.dimensions_and_batch_together: + with gr.Column(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") + + elif category == "cfg": + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") + + elif category == "seed": + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') + + elif category == "checkboxes": + with FormRow(elem_id="txt2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") + enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") + hr_final_resolution = FormHTML(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) + + elif category == "hires_fix": + with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: + with FormRow(elem_id="txt2img_hires_fix_row1"): + hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) + hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") + + with FormRow(elem_id="txt2img_hires_fix_row2"): + hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") + hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") + hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") + + elif category == "batch": + if not opts.dimensions_and_batch_together: + with FormRow(elem_id="txt2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") + + elif category == "scripts": + with FormGroup(elem_id="txt2img_script_container"): + custom_inputs = modules.scripts.scripts_txt2img.setup_ui() + + hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] + for input in hr_resolution_preview_inputs: + input.change( + fn=calc_resolution_hires, + inputs=hr_resolution_preview_inputs, + outputs=[hr_final_resolution], + show_progress=False, + ) + input.change( + None, + _js="onCalcResolutionHires", + inputs=hr_resolution_preview_inputs, + outputs=[], + show_progress=False, + ) + + txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) + parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) + + connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) + connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) + + txt2img_args = dict( + fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), + _js="submit", + inputs=[ + txt2img_prompt, + txt2img_negative_prompt, + txt2img_prompt_style, + txt2img_prompt_style2, + steps, + sampler_index, + restore_faces, + tiling, + batch_count, + batch_size, + cfg_scale, + seed, + subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, + height, + width, + enable_hr, + denoising_strength, + hr_scale, + hr_upscaler, + hr_second_pass_steps, + hr_resize_x, + hr_resize_y, + ] + custom_inputs, + + outputs=[ + txt2img_gallery, + generation_info, + html_info, + html_log, + ], + show_progress=False, + ) + + txt2img_prompt.submit(**txt2img_args) + submit.click(**txt2img_args) + + txt_prompt_img.change( + fn=modules.images.image_data, + inputs=[ + txt_prompt_img + ], + outputs=[ + txt2img_prompt, + txt_prompt_img + ] + ) + + enable_hr.change( + fn=lambda x: gr_show(x), + inputs=[enable_hr], + outputs=[hr_options], + show_progress = False, + ) + + txt2img_paste_fields = [ + (txt2img_prompt, "Prompt"), + (txt2img_negative_prompt, "Negative prompt"), + (steps, "Steps"), + (sampler_index, "Sampler"), + (restore_faces, "Face restoration"), + (cfg_scale, "CFG scale"), + (seed, "Seed"), + (width, "Size-1"), + (height, "Size-2"), + (batch_size, "Batch size"), + (subseed, "Variation seed"), + (subseed_strength, "Variation seed strength"), + (seed_resize_from_w, "Seed resize from-1"), + (seed_resize_from_h, "Seed resize from-2"), + (denoising_strength, "Denoising strength"), + (enable_hr, lambda d: "Denoising strength" in d), + (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), + (hr_scale, "Hires upscale"), + (hr_upscaler, "Hires upscaler"), + (hr_second_pass_steps, "Hires steps"), + (hr_resize_x, "Hires resize-1"), + (hr_resize_y, "Hires resize-2"), + *modules.scripts.scripts_txt2img.infotext_fields + ] + parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) + + txt2img_preview_params = [ + txt2img_prompt, + txt2img_negative_prompt, + steps, + sampler_index, + cfg_scale, + seed, + width, + height, + ] + + token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) + + modules.scripts.scripts_current = modules.scripts.scripts_img2img + modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) + + with gr.Blocks(analytics_enabled=False) as img2img_interface: + img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) + + with gr.Row(elem_id='img2img_progress_row'): + img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) + + with gr.Column(scale=1): + pass + + with gr.Column(scale=1): + progressbar = gr.HTML(elem_id="img2img_progressbar") + img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) + setup_progressbar(progressbar, img2img_preview, 'img2img') + + with FormRow().style(equal_height=False): + with gr.Column(variant='panel', elem_id="img2img_settings"): + + with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: + with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): + init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) + + with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): + init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) + init_img_with_mask_orig = gr.State(None) + + use_color_sketch = cmd_opts.gradio_inpaint_tool == "color-sketch" + if use_color_sketch: + def update_orig(image, state): + if image is not None: + same_size = state is not None and state.size == image.size + has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) + edited = same_size and has_exact_match + return image if not edited or state is None else state + + init_img_with_mask.change(update_orig, [init_img_with_mask, init_img_with_mask_orig], init_img_with_mask_orig) + + init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") + init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") + + with FormRow(): + mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") + mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") + + with FormRow(): + mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") + inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") + + with FormRow(): + inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") + + with FormRow(): + with gr.Column(): + inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") + + with gr.Column(scale=4): + inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") + + with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): + hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' + gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") + img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") + img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") + + with FormRow(): + resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") + + for category in ordered_ui_categories(): + if category == "sampler": + steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") + + elif category == "dimensions": + with FormRow(): + with gr.Column(elem_id="img2img_column_size", scale=4): + width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") + height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") + + if opts.dimensions_and_batch_together: + with gr.Column(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") + + elif category == "cfg": + with FormGroup(): + cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") + denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") + + elif category == "seed": + seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') + + elif category == "checkboxes": + with FormRow(elem_id="img2img_checkboxes"): + restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") + tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") + + elif category == "batch": + if not opts.dimensions_and_batch_together: + with FormRow(elem_id="img2img_column_batch"): + batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") + batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") + + elif category == "scripts": + with FormGroup(elem_id="img2img_script_container"): + custom_inputs = modules.scripts.scripts_img2img.setup_ui() + + img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) + parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) + + connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) + connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) + + img2img_prompt_img.change( + fn=modules.images.image_data, + inputs=[ + img2img_prompt_img + ], + outputs=[ + img2img_prompt, + img2img_prompt_img + ] + ) + + mask_mode.change( + lambda mode, img: { + init_img_with_mask: gr_show(mode == 0), + init_img_inpaint: gr_show(mode == 1), + init_mask_inpaint: gr_show(mode == 1), + }, + inputs=[mask_mode, init_img_with_mask], + outputs=[ + init_img_with_mask, + init_img_inpaint, + init_mask_inpaint, + ], + ) + + img2img_args = dict( + fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), + _js="submit_img2img", + inputs=[ + dummy_component, + img2img_prompt, + img2img_negative_prompt, + img2img_prompt_style, + img2img_prompt_style2, + init_img, + init_img_with_mask, + init_img_with_mask_orig, + init_img_inpaint, + init_mask_inpaint, + mask_mode, + steps, + sampler_index, + mask_blur, + mask_alpha, + inpainting_fill, + restore_faces, + tiling, + batch_count, + batch_size, + cfg_scale, + denoising_strength, + seed, + subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, + height, + width, + resize_mode, + inpaint_full_res, + inpaint_full_res_padding, + inpainting_mask_invert, + img2img_batch_input_dir, + img2img_batch_output_dir, + ] + custom_inputs, + outputs=[ + img2img_gallery, + generation_info, + html_info, + html_log, + ], + show_progress=False, + ) + + img2img_prompt.submit(**img2img_args) + submit.click(**img2img_args) + + img2img_interrogate.click( + fn=interrogate, + inputs=[init_img], + outputs=[img2img_prompt], + ) + + img2img_deepbooru.click( + fn=interrogate_deepbooru, + inputs=[init_img], + outputs=[img2img_prompt], + ) + + prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] + style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] + style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] + + for button, (prompt, negative_prompt) in zip([txt2img_save_style, img2img_save_style], prompts): + button.click( + fn=add_style, + _js="ask_for_style_name", + # Have to pass empty dummy component here, because the JavaScript and Python function have to accept + # the same number of parameters, but we only know the style-name after the JavaScript prompt + inputs=[dummy_component, prompt, negative_prompt], + outputs=[txt2img_prompt_style, img2img_prompt_style, txt2img_prompt_style2, img2img_prompt_style2], + ) + + for button, (prompt, negative_prompt), (style1, style2), js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): + button.click( + fn=apply_styles, + _js=js_func, + inputs=[prompt, negative_prompt, style1, style2], + outputs=[prompt, negative_prompt, style1, style2], + ) + + token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) + + img2img_paste_fields = [ + (img2img_prompt, "Prompt"), + (img2img_negative_prompt, "Negative prompt"), + (steps, "Steps"), + (sampler_index, "Sampler"), + (restore_faces, "Face restoration"), + (cfg_scale, "CFG scale"), + (seed, "Seed"), + (width, "Size-1"), + (height, "Size-2"), + (batch_size, "Batch size"), + (subseed, "Variation seed"), + (subseed_strength, "Variation seed strength"), + (seed_resize_from_w, "Seed resize from-1"), + (seed_resize_from_h, "Seed resize from-2"), + (denoising_strength, "Denoising strength"), + (mask_blur, "Mask blur"), + *modules.scripts.scripts_img2img.infotext_fields + ] + parameters_copypaste.add_paste_fields("img2img", init_img, img2img_paste_fields) + parameters_copypaste.add_paste_fields("inpaint", init_img_with_mask, img2img_paste_fields) + + modules.scripts.scripts_current = None + + with gr.Blocks(analytics_enabled=False) as extras_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + with gr.Tabs(elem_id="mode_extras"): + with gr.TabItem('Single Image', elem_id="extras_single_tab"): + extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") + + with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): + image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") + + with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): + extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") + extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") + show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") + + submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') + + with gr.Tabs(elem_id="extras_resize_mode"): + with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): + upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") + with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): + with gr.Group(): + with gr.Row(): + upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") + upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") + upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") + + with gr.Group(): + extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") + + with gr.Group(): + extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") + extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") + + with gr.Group(): + gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") + + with gr.Group(): + codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") + codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") + + with gr.Group(): + upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") + + result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) + + submit.click( + fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), + _js="get_extras_tab_index", + inputs=[ + dummy_component, + dummy_component, + extras_image, + image_batch, + extras_batch_input_dir, + extras_batch_output_dir, + show_extras_results, + gfpgan_visibility, + codeformer_visibility, + codeformer_weight, + upscaling_resize, + upscaling_resize_w, + upscaling_resize_h, + upscaling_crop, + extras_upscaler_1, + extras_upscaler_2, + extras_upscaler_2_visibility, + upscale_before_face_fix, + ], + outputs=[ + result_images, + html_info_x, + html_info, + ] + ) + parameters_copypaste.add_paste_fields("extras", extras_image, None) + + extras_image.change( + fn=modules.extras.clear_cache, + inputs=[], outputs=[] + ) + + with gr.Blocks(analytics_enabled=False) as pnginfo_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + image = gr.Image(elem_id="pnginfo_image", label="Source", source="upload", interactive=True, type="pil") + + with gr.Column(variant='panel'): + html = gr.HTML() + generation_info = gr.Textbox(visible=False, elem_id="pnginfo_generation_info") + html2 = gr.HTML() + with gr.Row(): + buttons = parameters_copypaste.create_buttons(["txt2img", "img2img", "inpaint", "extras"]) + parameters_copypaste.bind_buttons(buttons, image, generation_info) + + image.change( + fn=wrap_gradio_call(modules.extras.run_pnginfo), + inputs=[image], + outputs=[html, generation_info, html2], + ) + + with gr.Blocks(analytics_enabled=False) as modelmerger_interface: + with gr.Row().style(equal_height=False): + with gr.Column(variant='panel'): + gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") + + with gr.Row(): + primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") + create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") + + secondary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_secondary_model_name", label="Secondary model (B)") + create_refresh_button(secondary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_B") + + tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") + create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") + + custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") + interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") + interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") + + with gr.Row(): + checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") + save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") + + modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') + + with gr.Column(variant='panel'): + submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) + + with gr.Blocks(analytics_enabled=False) as train_interface: + with gr.Row().style(equal_height=False): + gr.HTML(value="

See wiki for detailed explanation.

") + + with gr.Row().style(equal_height=False): + with gr.Tabs(elem_id="train_tabs"): + + with gr.Tab(label="Create embedding"): + new_embedding_name = gr.Textbox(label="Name", elem_id="train_new_embedding_name") + initialization_text = gr.Textbox(label="Initialization text", value="*", elem_id="train_initialization_text") + nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1, elem_id="train_nvpt") + overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding", elem_id="train_overwrite_old_embedding") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + create_embedding = gr.Button(value="Create embedding", variant='primary', elem_id="train_create_embedding") + + with gr.Tab(label="Create hypernetwork"): + new_hypernetwork_name = gr.Textbox(label="Name", elem_id="train_new_hypernetwork_name") + new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"], elem_id="train_new_hypernetwork_sizes") + new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'", elem_id="train_new_hypernetwork_layer_structure") + new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys, elem_id="train_new_hypernetwork_activation_func") + new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") + new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") + new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") + new_hypernetwork_dropout_structure = gr.Textbox("0, 0, 0", label="Enter hypernetwork Dropout structure (or empty). Recommended : 0~0.35 incrementing sequence: 0, 0.05, 0.15", placeholder="1st and last digit must be 0 and values should be between 0 and 1. ex:'0, 0.01, 0'") + overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary', elem_id="train_create_hypernetwork") + + with gr.Tab(label="Preprocess images"): + process_src = gr.Textbox(label='Source directory', elem_id="train_process_src") + process_dst = gr.Textbox(label='Destination directory', elem_id="train_process_dst") + process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_process_width") + process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_process_height") + preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"], elem_id="train_preprocess_txt_action") + + with gr.Row(): + process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") + process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") + process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") + process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") + process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") + + with gr.Row(visible=False) as process_split_extra_row: + process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_split_threshold") + process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05, elem_id="train_process_overlap_ratio") + + with gr.Row(visible=False) as process_focal_crop_row: + process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_face_weight") + process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") + process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") + process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") + + with gr.Row(): + with gr.Column(scale=3): + gr.HTML(value="") + + with gr.Column(): + with gr.Row(): + interrupt_preprocessing = gr.Button("Interrupt", elem_id="train_interrupt_preprocessing") + run_preprocess = gr.Button(value="Preprocess", variant='primary', elem_id="train_run_preprocess") + + process_split.change( + fn=lambda show: gr_show(show), + inputs=[process_split], + outputs=[process_split_extra_row], + ) + + process_focal_crop.change( + fn=lambda show: gr_show(show), + inputs=[process_focal_crop], + outputs=[process_focal_crop_row], + ) + + def get_textual_inversion_template_names(): + return sorted([x for x in textual_inversion.textual_inversion_templates]) + + with gr.Tab(label="Train"): + gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") + with FormRow(): + train_embedding_name = gr.Dropdown(label='Embedding', elem_id="train_embedding", choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())) + create_refresh_button(train_embedding_name, sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings, lambda: {"choices": sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())}, "refresh_train_embedding_name") + + train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) + create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") + + with FormRow(): + embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") + hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") + + with FormRow(): + clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) + clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) + + with FormRow(): + batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") + gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") + + dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") + log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") + + with FormRow(): + template_file = gr.Dropdown(label='Prompt template', value="style_filewords.txt", elem_id="train_template_file", choices=get_textual_inversion_template_names()) + create_refresh_button(template_file, textual_inversion.list_textual_inversion_templates, lambda: {"choices": get_textual_inversion_template_names()}, "refrsh_train_template_file") + + training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") + training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") + varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") + steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") + + with FormRow(): + create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") + save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") + + save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") + preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") + + shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") + tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") + + latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") + + with gr.Row(): + train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") + interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") + train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") + + params = script_callbacks.UiTrainTabParams(txt2img_preview_params) + + script_callbacks.ui_train_tabs_callback(params) + + with gr.Column(): + progressbar = gr.HTML(elem_id="ti_progressbar") + ti_output = gr.Text(elem_id="ti_output", value="", show_label=False) + + ti_gallery = gr.Gallery(label='Output', show_label=False, elem_id='ti_gallery').style(grid=4) + ti_preview = gr.Image(elem_id='ti_preview', visible=False) + ti_progress = gr.HTML(elem_id="ti_progress", value="") + ti_outcome = gr.HTML(elem_id="ti_error", value="") + setup_progressbar(progressbar, ti_preview, 'ti', textinfo=ti_progress) + + create_embedding.click( + fn=modules.textual_inversion.ui.create_embedding, + inputs=[ + new_embedding_name, + initialization_text, + nvpt, + overwrite_old_embedding, + ], + outputs=[ + train_embedding_name, + ti_output, + ti_outcome, + ] + ) + + create_hypernetwork.click( + fn=modules.hypernetworks.ui.create_hypernetwork, + inputs=[ + new_hypernetwork_name, + new_hypernetwork_sizes, + overwrite_old_hypernetwork, + new_hypernetwork_layer_structure, + new_hypernetwork_activation_func, + new_hypernetwork_initialization_option, + new_hypernetwork_add_layer_norm, + new_hypernetwork_use_dropout, + new_hypernetwork_dropout_structure + ], + outputs=[ + train_hypernetwork_name, + ti_output, + ti_outcome, + ] + ) + + run_preprocess.click( + fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.preprocess, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + process_src, + process_dst, + process_width, + process_height, + preprocess_txt_action, + process_flip, + process_split, + process_caption, + process_caption_deepbooru, + process_split_threshold, + process_overlap_ratio, + process_focal_crop, + process_focal_crop_face_weight, + process_focal_crop_entropy_weight, + process_focal_crop_edges_weight, + process_focal_crop_debug, + ], + outputs=[ + ti_output, + ti_outcome, + ], + ) + + train_embedding.click( + fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.train_embedding, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + train_embedding_name, + embedding_learn_rate, + batch_size, + gradient_step, + dataset_directory, + log_directory, + training_width, + training_height, + varsize, + steps, + clip_grad_mode, + clip_grad_value, + shuffle_tags, + tag_drop_out, + latent_sampling_method, + create_image_every, + save_embedding_every, + template_file, + save_image_with_stored_embedding, + preview_from_txt2img, + *txt2img_preview_params, + ], + outputs=[ + ti_output, + ti_outcome, + ] + ) + + train_hypernetwork.click( + fn=wrap_gradio_gpu_call(modules.hypernetworks.ui.train_hypernetwork, extra_outputs=[gr.update()]), + _js="start_training_textual_inversion", + inputs=[ + train_hypernetwork_name, + hypernetwork_learn_rate, + batch_size, + gradient_step, + dataset_directory, + log_directory, + training_width, + training_height, + varsize, + steps, + clip_grad_mode, + clip_grad_value, + shuffle_tags, + tag_drop_out, + latent_sampling_method, + create_image_every, + save_embedding_every, + template_file, + preview_from_txt2img, + *txt2img_preview_params, + ], + outputs=[ + ti_output, + ti_outcome, + ] + ) + + interrupt_training.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + interrupt_preprocessing.click( + fn=lambda: shared.state.interrupt(), + inputs=[], + outputs=[], + ) + + def create_setting_component(key, is_quicksettings=False): + def fun(): + return opts.data[key] if key in opts.data else opts.data_labels[key].default + + info = opts.data_labels[key] + t = type(info.default) + + args = info.component_args() if callable(info.component_args) else info.component_args + + if info.component is not None: + comp = info.component + elif t == str: + comp = gr.Textbox + elif t == int: + comp = gr.Number + elif t == bool: + comp = gr.Checkbox + else: + raise Exception(f'bad options item type: {str(t)} for key {key}') + + elem_id = "setting_"+key + + if info.refresh is not None: + if is_quicksettings: + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) + else: + with FormRow(): + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) + else: + res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) + + return res + + components = [] + component_dict = {} + + script_callbacks.ui_settings_callback() + opts.reorder() + + def run_settings(*args): + changed = [] + + for key, value, comp in zip(opts.data_labels.keys(), args, components): + assert comp == dummy_component or opts.same_type(value, opts.data_labels[key].default), f"Bad value for setting {key}: {value}; expecting {type(opts.data_labels[key].default).__name__}" + + for key, value, comp in zip(opts.data_labels.keys(), args, components): + if comp == dummy_component: + continue + + if opts.set(key, value): + changed.append(key) + + try: + opts.save(shared.config_filename) + except RuntimeError: + return opts.dumpjson(), f'{len(changed)} settings changed without save: {", ".join(changed)}.' + return opts.dumpjson(), f'{len(changed)} settings changed{": " if len(changed) > 0 else ""}{", ".join(changed)}.' + + def run_settings_single(value, key): + if not opts.same_type(value, opts.data_labels[key].default): + return gr.update(visible=True), opts.dumpjson() + + if not opts.set(key, value): + return gr.update(value=getattr(opts, key)), opts.dumpjson() + + opts.save(shared.config_filename) + + return gr.update(value=value), opts.dumpjson() + + with gr.Blocks(analytics_enabled=False) as settings_interface: + with gr.Row(): + with gr.Column(scale=6): + settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") + with gr.Column(): + restart_gradio = gr.Button(value='Reload UI', variant='primary', elem_id="settings_restart_gradio") + + result = gr.HTML(elem_id="settings_result") + + quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] + quicksettings_names = {x: i for i, x in enumerate(quicksettings_names) if x != 'quicksettings'} + + quicksettings_list = [] + + previous_section = None + current_tab = None + with gr.Tabs(elem_id="settings"): + for i, (k, item) in enumerate(opts.data_labels.items()): + section_must_be_skipped = item.section[0] is None + + if previous_section != item.section and not section_must_be_skipped: + elem_id, text = item.section + + if current_tab is not None: + current_tab.__exit__() + + current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) + current_tab.__enter__() + + previous_section = item.section + + if k in quicksettings_names and not shared.cmd_opts.freeze_settings: + quicksettings_list.append((i, k, item)) + components.append(dummy_component) + elif section_must_be_skipped: + components.append(dummy_component) + else: + component = create_setting_component(k) + component_dict[k] = component + components.append(component) + + if current_tab is not None: + current_tab.__exit__() + + with gr.TabItem("Actions"): + request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") + download_localization = gr.Button(value='Download localization template', elem_id="download_localization") + reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") + + if os.path.exists("html/licenses.html"): + with open("html/licenses.html", encoding="utf8") as file: + with gr.TabItem("Licenses"): + gr.HTML(file.read(), elem_id="licenses") + + gr.Button(value="Show all pages", elem_id="settings_show_all_pages") + + request_notifications.click( + fn=lambda: None, + inputs=[], + outputs=[], + _js='function(){}' + ) + + download_localization.click( + fn=lambda: None, + inputs=[], + outputs=[], + _js='download_localization' + ) + + def reload_scripts(): + modules.scripts.reload_script_body_only() + reload_javascript() # need to refresh the html page + + reload_script_bodies.click( + fn=reload_scripts, + inputs=[], + outputs=[] + ) + + def request_restart(): + shared.state.interrupt() + shared.state.need_restart = True + + restart_gradio.click( + fn=request_restart, + _js='restart_reload', + inputs=[], + outputs=[], + ) + + interfaces = [ + (txt2img_interface, "txt2img", "txt2img"), + (img2img_interface, "img2img", "img2img"), + (extras_interface, "Extras", "extras"), + (pnginfo_interface, "PNG Info", "pnginfo"), + (modelmerger_interface, "Checkpoint Merger", "modelmerger"), + (train_interface, "Train", "ti"), + ] + + css = "" + + for cssfile in modules.scripts.list_files_with_name("style.css"): + if not os.path.isfile(cssfile): + continue + + with open(cssfile, "r", encoding="utf8") as file: + css += file.read() + "\n" + + if os.path.exists(os.path.join(script_path, "user.css")): + with open(os.path.join(script_path, "user.css"), "r", encoding="utf8") as file: + css += file.read() + "\n" + + if not cmd_opts.no_progressbar_hiding: + css += css_hide_progressbar + + interfaces += script_callbacks.ui_tabs_callback() + interfaces += [(settings_interface, "Settings", "settings")] + + extensions_interface = ui_extensions.create_ui() + interfaces += [(extensions_interface, "Extensions", "extensions")] + + with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: + with gr.Row(elem_id="quicksettings"): + for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): + component = create_setting_component(k, is_quicksettings=True) + component_dict[k] = component + + parameters_copypaste.integrate_settings_paste_fields(component_dict) + parameters_copypaste.run_bind() + + with gr.Tabs(elem_id="tabs") as tabs: + for interface, label, ifid in interfaces: + with gr.TabItem(label, id=ifid, elem_id='tab_' + ifid): + interface.render() + + if os.path.exists(os.path.join(script_path, "notification.mp3")): + audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) + + if os.path.exists("html/footer.html"): + with open("html/footer.html", encoding="utf8") as file: + footer = file.read() + footer = footer.format(versions=versions_html()) + gr.HTML(footer, elem_id="footer") + + text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) + settings_submit.click( + fn=wrap_gradio_call(run_settings, extra_outputs=[gr.update()]), + inputs=components, + outputs=[text_settings, result], + ) + + for i, k, item in quicksettings_list: + component = component_dict[k] + + component.change( + fn=lambda value, k=k: run_settings_single(value, key=k), + inputs=[component], + outputs=[component, text_settings], + ) + + component_keys = [k for k in opts.data_labels.keys() if k in component_dict] + + def get_settings_values(): + return [getattr(opts, key) for key in component_keys] + + demo.load( + fn=get_settings_values, + inputs=[], + outputs=[component_dict[k] for k in component_keys], + ) + + def modelmerger(*args): + try: + results = modules.extras.run_modelmerger(*args) + except Exception as e: + print("Error loading/saving model file:", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + modules.sd_models.list_models() # to remove the potentially missing models from the list + return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] + return results + + modelmerger_merge.click( + fn=modelmerger, + inputs=[ + primary_model_name, + secondary_model_name, + tertiary_model_name, + interp_method, + interp_amount, + save_as_half, + custom_name, + checkpoint_format, + ], + outputs=[ + submit_result, + primary_model_name, + secondary_model_name, + tertiary_model_name, + component_dict['sd_model_checkpoint'], + ] + ) + + ui_config_file = cmd_opts.ui_config_file + ui_settings = {} + settings_count = len(ui_settings) + error_loading = False + + try: + if os.path.exists(ui_config_file): + with open(ui_config_file, "r", encoding="utf8") as file: + ui_settings = json.load(file) + except Exception: + error_loading = True + print("Error loading settings:", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + + def loadsave(path, x): + def apply_field(obj, field, condition=None, init_field=None): + key = path + "/" + field + + if getattr(obj, 'custom_script_source', None) is not None: + key = 'customscript/' + obj.custom_script_source + '/' + key + + if getattr(obj, 'do_not_save_to_config', False): + return + + saved_value = ui_settings.get(key, None) + if saved_value is None: + ui_settings[key] = getattr(obj, field) + elif condition and not condition(saved_value): + print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') + else: + setattr(obj, field, saved_value) + if init_field is not None: + init_field(saved_value) + + if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number, gr.Dropdown] and x.visible: + apply_field(x, 'visible') + + if type(x) == gr.Slider: + apply_field(x, 'value') + apply_field(x, 'minimum') + apply_field(x, 'maximum') + apply_field(x, 'step') + + if type(x) == gr.Radio: + apply_field(x, 'value', lambda val: val in x.choices) + + if type(x) == gr.Checkbox: + apply_field(x, 'value') + + if type(x) == gr.Textbox: + apply_field(x, 'value') + + if type(x) == gr.Number: + apply_field(x, 'value') + + if type(x) == gr.Dropdown: + apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) + + visit(txt2img_interface, loadsave, "txt2img") + visit(img2img_interface, loadsave, "img2img") + visit(extras_interface, loadsave, "extras") + visit(modelmerger_interface, loadsave, "modelmerger") + visit(train_interface, loadsave, "train") + + if not error_loading and (not os.path.exists(ui_config_file) or settings_count != len(ui_settings)): + with open(ui_config_file, "w", encoding="utf8") as file: + json.dump(ui_settings, file, indent=4) + + return demo + + +def reload_javascript(): + with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: + javascript = f'' + + scripts_list = modules.scripts.list_scripts("javascript", ".js") + + for basedir, filename, path in scripts_list: + with open(path, "r", encoding="utf8") as jsfile: + javascript += f"\n" + + if cmd_opts.theme is not None: + javascript += f"\n\n" + + javascript += f"\n" + + def template_response(*args, **kwargs): + res = shared.GradioTemplateResponseOriginal(*args, **kwargs) + res.body = res.body.replace( + b'', f'{javascript}'.encode("utf8")) + res.init_headers() + return res + + gradio.routes.templates.TemplateResponse = template_response + + +if not hasattr(shared, 'GradioTemplateResponseOriginal'): + shared.GradioTemplateResponseOriginal = gradio.routes.templates.TemplateResponse + + +def versions_html(): + import torch + import launch + + python_version = ".".join([str(x) for x in sys.version_info[0:3]]) + commit = launch.commit_hash() + short_commit = commit[0:8] + + if shared.xformers_available: + import xformers + xformers_version = xformers.__version__ + else: + xformers_version = "N/A" + + return f""" +python: {python_version} + •  +torch: {torch.__version__} + •  +xformers: {xformers_version} + •  +gradio: {gr.__version__} + •  +commit: {short_commit} +""" -- cgit v1.2.3 From ef75c980536471c0729a2319440e3083cd57a4f0 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 10 Jan 2023 12:29:45 +0300 Subject: Split history ui.py to ui_progress.py --- modules/ui.py | 94 +-- modules/ui_progress.py | 1839 +----------------------------------------------- 2 files changed, 9 insertions(+), 1924 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 9b9081b5..3c458ce8 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -162,79 +162,6 @@ def save_files(js_data, images, do_make_zip, index): return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") -def calc_time_left(progress, threshold, label, force_display, show_eta): - if progress == 0: - return "" - else: - time_since_start = time.time() - shared.state.time_start - eta = (time_since_start/progress) - eta_relative = eta-time_since_start - if (eta_relative > threshold and show_eta) or force_display: - if eta_relative > 3600: - return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) - elif eta_relative > 60: - return label + time.strftime('%M:%S', time.gmtime(eta_relative)) - else: - return label + time.strftime('%Ss', time.gmtime(eta_relative)) - else: - return "" - - -def check_progress_call(id_part): - if shared.state.job_count == 0: - return "", gr_show(False), gr_show(False), gr_show(False) - - progress = 0 - - if shared.state.job_count > 0: - progress += shared.state.job_no / shared.state.job_count - if shared.state.sampling_steps > 0: - progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps - - # Show progress percentage and time left at the same moment, and base it also on steps done - show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 - - time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) - if time_left != "": - shared.state.time_left_force_display = True - - progress = min(progress, 1) - - progressbar = "" - if opts.show_progressbar: - progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" - - image = gr_show(False) - preview_visibility = gr_show(False) - - if opts.show_progress_every_n_steps != 0: - shared.state.set_current_image() - image = shared.state.current_image - - if image is None: - image = gr.update(value=None) - else: - preview_visibility = gr_show(True) - - if shared.state.textinfo is not None: - textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) - else: - textinfo_result = gr_show(False) - - return f"

{progressbar}

", preview_visibility, image, textinfo_result - - -def check_progress_call_initial(id_part): - shared.state.job_count = -1 - shared.state.current_latent = None - shared.state.current_image = None - shared.state.textinfo = None - shared.state.time_start = time.time() - shared.state.time_left_force_display = False - - return check_progress_call(id_part) - - def visit(x, func, path=""): if hasattr(x, 'children'): for c in x.children: @@ -456,25 +383,10 @@ def create_toprow(is_img2img): return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button -def setup_progressbar(progressbar, preview, id_part, textinfo=None): - if textinfo is None: - textinfo = gr.HTML(visible=False) +def setup_progressbar(*args, **kwargs): + import modules.ui_progress - check_progress = gr.Button('Check progress', elem_id=f"{id_part}_check_progress", visible=False) - check_progress.click( - fn=lambda: check_progress_call(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) - - check_progress_initial = gr.Button('Check progress (first)', elem_id=f"{id_part}_check_progress_initial", visible=False) - check_progress_initial.click( - fn=lambda: check_progress_call_initial(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) + modules.ui_progress.setup_progressbar(*args, **kwargs) def apply_setting(key, value): diff --git a/modules/ui_progress.py b/modules/ui_progress.py index 9b9081b5..592fda55 100644 --- a/modules/ui_progress.py +++ b/modules/ui_progress.py @@ -1,165 +1,10 @@ -import html -import json -import math -import mimetypes -import os -import platform -import random -import subprocess as sp -import sys -import tempfile import time -import traceback -from functools import partial, reduce import gradio as gr -import gradio.routes -import gradio.utils -import numpy as np -from PIL import Image, PngImagePlugin -from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru -from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML -from modules.paths import script_path +from modules.shared import opts -from modules.shared import opts, cmd_opts, restricted_opts - -import modules.codeformer_model -import modules.generation_parameters_copypaste as parameters_copypaste -import modules.gfpgan_model -import modules.hypernetworks.ui -import modules.scripts import modules.shared as shared -import modules.styles -import modules.textual_inversion.ui -from modules import prompt_parser -from modules.images import save_image -from modules.sd_hijack import model_hijack -from modules.sd_samplers import samplers, samplers_for_img2img -from modules.textual_inversion import textual_inversion -import modules.hypernetworks.ui -from modules.generation_parameters_copypaste import image_from_url_text - -# this is a fix for Windows users. Without it, javascript files will be served with text/html content-type and the browser will not show any UI -mimetypes.init() -mimetypes.add_type('application/javascript', '.js') - -if not cmd_opts.share and not cmd_opts.listen: - # fix gradio phoning home - gradio.utils.version_check = lambda: None - gradio.utils.get_local_ip_address = lambda: '127.0.0.1' - -if cmd_opts.ngrok is not None: - import modules.ngrok as ngrok - print('ngrok authtoken detected, trying to connect...') - ngrok.connect( - cmd_opts.ngrok, - cmd_opts.port if cmd_opts.port is not None else 7860, - cmd_opts.ngrok_region - ) - - -def gr_show(visible=True): - return {"visible": visible, "__type__": "update"} - - -sample_img2img = "assets/stable-samples/img2img/sketch-mountains-input.jpg" -sample_img2img = sample_img2img if os.path.exists(sample_img2img) else None - -css_hide_progressbar = """ -.wrap .m-12 svg { display:none!important; } -.wrap .m-12::before { content:"Loading..." } -.wrap .z-20 svg { display:none!important; } -.wrap .z-20::before { content:"Loading..." } -.progress-bar { display:none!important; } -.meta-text { display:none!important; } -.meta-text-center { display:none!important; } -""" - -# Using constants for these since the variation selector isn't visible. -# Important that they exactly match script.js for tooltip to work. -random_symbol = '\U0001f3b2\ufe0f' # 🎲️ -reuse_symbol = '\u267b\ufe0f' # ♻️ -paste_symbol = '\u2199\ufe0f' # ↙ -folder_symbol = '\U0001f4c2' # 📂 -refresh_symbol = '\U0001f504' # 🔄 -save_style_symbol = '\U0001f4be' # 💾 -apply_style_symbol = '\U0001f4cb' # 📋 -clear_prompt_symbol = '\U0001F5D1' # 🗑️ - - -def plaintext_to_html(text): - text = "

" + "
\n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

" - return text - -def send_gradio_gallery_to_image(x): - if len(x) == 0: - return None - return image_from_url_text(x[0]) - -def save_files(js_data, images, do_make_zip, index): - import csv - filenames = [] - fullfns = [] - - #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it - class MyObject: - def __init__(self, d=None): - if d is not None: - for key, value in d.items(): - setattr(self, key, value) - - data = json.loads(js_data) - - p = MyObject(data) - path = opts.outdir_save - save_to_dirs = opts.use_save_to_dirs_for_ui - extension: str = opts.samples_format - start_index = 0 - - if index > -1 and opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only - - images = [images[index]] - start_index = index - - os.makedirs(opts.outdir_save, exist_ok=True) - - with open(os.path.join(opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: - at_start = file.tell() == 0 - writer = csv.writer(file) - if at_start: - writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) - - for image_index, filedata in enumerate(images, start_index): - image = image_from_url_text(filedata) - - is_grid = image_index < p.index_of_first_image - i = 0 if is_grid else (image_index - p.index_of_first_image) - - fullfn, txt_fullfn = save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) - - filename = os.path.relpath(fullfn, path) - filenames.append(filename) - fullfns.append(fullfn) - if txt_fullfn: - filenames.append(os.path.basename(txt_fullfn)) - fullfns.append(txt_fullfn) - - writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) - - # Make Zip - if do_make_zip: - zip_filepath = os.path.join(path, "images.zip") - - from zipfile import ZipFile - with ZipFile(zip_filepath, "w") as zip_file: - for i in range(len(fullfns)): - with open(fullfns[i], mode="rb") as f: - zip_file.writestr(filenames[i], f.read()) - fullfns.insert(0, zip_filepath) - - return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") def calc_time_left(progress, threshold, label, force_display, show_eta): @@ -182,7 +27,7 @@ def calc_time_left(progress, threshold, label, force_display, show_eta): def check_progress_call(id_part): if shared.state.job_count == 0: - return "", gr_show(False), gr_show(False), gr_show(False) + return "", gr.update(visible=False), gr.update(visible=False), gr.update(visible=False) progress = 0 @@ -204,8 +49,8 @@ def check_progress_call(id_part): if opts.show_progressbar: progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" - image = gr_show(False) - preview_visibility = gr_show(False) + image = gr.update(visible=False) + preview_visibility = gr.update(visible=False) if opts.show_progress_every_n_steps != 0: shared.state.set_current_image() @@ -214,12 +59,12 @@ def check_progress_call(id_part): if image is None: image = gr.update(value=None) else: - preview_visibility = gr_show(True) + preview_visibility = gr.update(visible=True) if shared.state.textinfo is not None: textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) else: - textinfo_result = gr_show(False) + textinfo_result = gr.update(visible=False) return f"

{progressbar}

", preview_visibility, image, textinfo_result @@ -235,227 +80,6 @@ def check_progress_call_initial(id_part): return check_progress_call(id_part) -def visit(x, func, path=""): - if hasattr(x, 'children'): - for c in x.children: - visit(c, func, path) - elif x.label is not None: - func(path + "/" + str(x.label), x) - - -def add_style(name: str, prompt: str, negative_prompt: str): - if name is None: - return [gr_show() for x in range(4)] - - style = modules.styles.PromptStyle(name, prompt, negative_prompt) - shared.prompt_styles.styles[style.name] = style - # Save all loaded prompt styles: this allows us to update the storage format in the future more easily, because we - # reserialize all styles every time we save them - shared.prompt_styles.save_styles(shared.styles_filename) - - return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] - - -def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): - from modules import processing, devices - - if not enable: - return "" - - p = processing.StableDiffusionProcessingTxt2Img(width=width, height=height, enable_hr=True, hr_scale=hr_scale, hr_resize_x=hr_resize_x, hr_resize_y=hr_resize_y) - - with devices.autocast(): - p.init([""], [0], [0]) - - return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" - - -def apply_styles(prompt, prompt_neg, style1_name, style2_name): - prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) - prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, [style1_name, style2_name]) - - return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] - - -def interrogate(image): - prompt = shared.interrogator.interrogate(image.convert("RGB")) - - return gr_show(True) if prompt is None else prompt - - -def interrogate_deepbooru(image): - prompt = deepbooru.model.tag(image) - return gr_show(True) if prompt is None else prompt - - -def create_seed_inputs(target_interface): - with FormRow(elem_id=target_interface + '_seed_row'): - seed = (gr.Textbox if cmd_opts.use_textbox_seed else gr.Number)(label='Seed', value=-1, elem_id=target_interface + '_seed') - seed.style(container=False) - random_seed = gr.Button(random_symbol, elem_id=target_interface + '_random_seed') - reuse_seed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_seed') - - with gr.Group(elem_id=target_interface + '_subseed_show_box'): - seed_checkbox = gr.Checkbox(label='Extra', elem_id=target_interface + '_subseed_show', value=False) - - # Components to show/hide based on the 'Extra' checkbox - seed_extras = [] - - with FormRow(visible=False, elem_id=target_interface + '_subseed_row') as seed_extra_row_1: - seed_extras.append(seed_extra_row_1) - subseed = gr.Number(label='Variation seed', value=-1, elem_id=target_interface + '_subseed') - subseed.style(container=False) - random_subseed = gr.Button(random_symbol, elem_id=target_interface + '_random_subseed') - reuse_subseed = gr.Button(reuse_symbol, elem_id=target_interface + '_reuse_subseed') - subseed_strength = gr.Slider(label='Variation strength', value=0.0, minimum=0, maximum=1, step=0.01, elem_id=target_interface + '_subseed_strength') - - with FormRow(visible=False) as seed_extra_row_2: - seed_extras.append(seed_extra_row_2) - seed_resize_from_w = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from width", value=0, elem_id=target_interface + '_seed_resize_from_w') - seed_resize_from_h = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize seed from height", value=0, elem_id=target_interface + '_seed_resize_from_h') - - random_seed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[seed]) - random_subseed.click(fn=lambda: -1, show_progress=False, inputs=[], outputs=[subseed]) - - def change_visibility(show): - return {comp: gr_show(show) for comp in seed_extras} - - seed_checkbox.change(change_visibility, show_progress=False, inputs=[seed_checkbox], outputs=seed_extras) - - return seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox - - - -def connect_clear_prompt(button): - """Given clear button, prompt, and token_counter objects, setup clear prompt button click event""" - button.click( - _js="clear_prompt", - fn=None, - inputs=[], - outputs=[], - ) - - -def connect_reuse_seed(seed: gr.Number, reuse_seed: gr.Button, generation_info: gr.Textbox, dummy_component, is_subseed): - """ Connects a 'reuse (sub)seed' button's click event so that it copies last used - (sub)seed value from generation info the to the seed field. If copying subseed and subseed strength - was 0, i.e. no variation seed was used, it copies the normal seed value instead.""" - def copy_seed(gen_info_string: str, index): - res = -1 - - try: - gen_info = json.loads(gen_info_string) - index -= gen_info.get('index_of_first_image', 0) - - if is_subseed and gen_info.get('subseed_strength', 0) > 0: - all_subseeds = gen_info.get('all_subseeds', [-1]) - res = all_subseeds[index if 0 <= index < len(all_subseeds) else 0] - else: - all_seeds = gen_info.get('all_seeds', [-1]) - res = all_seeds[index if 0 <= index < len(all_seeds) else 0] - - except json.decoder.JSONDecodeError as e: - if gen_info_string != '': - print("Error parsing JSON generation info:", file=sys.stderr) - print(gen_info_string, file=sys.stderr) - - return [res, gr_show(False)] - - reuse_seed.click( - fn=copy_seed, - _js="(x, y) => [x, selected_gallery_index()]", - show_progress=False, - inputs=[generation_info, dummy_component], - outputs=[seed, dummy_component] - ) - - -def update_token_counter(text, steps): - try: - _, prompt_flat_list, _ = prompt_parser.get_multicond_prompt_list([text]) - prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules(prompt_flat_list, steps) - - except Exception: - # a parsing error can happen here during typing, and we don't want to bother the user with - # messages related to it in console - prompt_schedules = [[[steps, text]]] - - flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) - prompts = [prompt_text for step, prompt_text in flat_prompts] - token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) - style_class = ' class="red"' if (token_count > max_length) else "" - return f"{token_count}/{max_length}" - - -def create_toprow(is_img2img): - id_part = "img2img" if is_img2img else "txt2img" - - with gr.Row(elem_id="toprow"): - with gr.Column(scale=6): - with gr.Row(): - with gr.Column(scale=80): - with gr.Row(): - prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, - placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) - - with gr.Row(): - with gr.Column(scale=80): - with gr.Row(): - negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, - placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) - - with gr.Column(scale=1, elem_id="roll_col"): - paste = gr.Button(value=paste_symbol, elem_id="paste") - save_style = gr.Button(value=save_style_symbol, elem_id="style_create") - prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") - clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") - token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") - token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") - - clear_prompt_button.click( - fn=lambda *x: x, - _js="confirm_clear_prompt", - inputs=[prompt, negative_prompt], - outputs=[prompt, negative_prompt], - ) - - button_interrogate = None - button_deepbooru = None - if is_img2img: - with gr.Column(scale=1, elem_id="interrogate_col"): - button_interrogate = gr.Button('Interrogate\nCLIP', elem_id="interrogate") - button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") - - with gr.Column(scale=1): - with gr.Row(): - skip = gr.Button('Skip', elem_id=f"{id_part}_skip") - interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") - submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') - - skip.click( - fn=lambda: shared.state.skip(), - inputs=[], - outputs=[], - ) - - interrupt.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - with gr.Row(): - with gr.Column(scale=1, elem_id="style_pos_col"): - prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - with gr.Column(scale=1, elem_id="style_neg_col"): - prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button - - def setup_progressbar(progressbar, preview, id_part, textinfo=None): if textinfo is None: textinfo = gr.HTML(visible=False) @@ -475,1454 +99,3 @@ def setup_progressbar(progressbar, preview, id_part, textinfo=None): inputs=[], outputs=[progressbar, preview, preview, textinfo], ) - - -def apply_setting(key, value): - if value is None: - return gr.update() - - if shared.cmd_opts.freeze_settings: - return gr.update() - - # dont allow model to be swapped when model hash exists in prompt - if key == "sd_model_checkpoint" and opts.disable_weights_auto_swap: - return gr.update() - - if key == "sd_model_checkpoint": - ckpt_info = sd_models.get_closet_checkpoint_match(value) - - if ckpt_info is not None: - value = ckpt_info.title - else: - return gr.update() - - comp_args = opts.data_labels[key].component_args - if comp_args and isinstance(comp_args, dict) and comp_args.get('visible') is False: - return - - valtype = type(opts.data_labels[key].default) - oldval = opts.data.get(key, None) - opts.data[key] = valtype(value) if valtype != type(None) else value - if oldval != value and opts.data_labels[key].onchange is not None: - opts.data_labels[key].onchange() - - opts.save(shared.config_filename) - return value - - -def update_generation_info(args): - generation_info, html_info, img_index = args - try: - generation_info = json.loads(generation_info) - if img_index < 0 or img_index >= len(generation_info["infotexts"]): - return html_info - return plaintext_to_html(generation_info["infotexts"][img_index]) - except Exception: - pass - # if the json parse or anything else fails, just return the old html_info - return html_info - - -def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): - def refresh(): - refresh_method() - args = refreshed_args() if callable(refreshed_args) else refreshed_args - - for k, v in args.items(): - setattr(refresh_component, k, v) - - return gr.update(**(args or {})) - - refresh_button = ToolButton(value=refresh_symbol, elem_id=elem_id) - refresh_button.click( - fn=refresh, - inputs=[], - outputs=[refresh_component] - ) - return refresh_button - - -def create_output_panel(tabname, outdir): - def open_folder(f): - if not os.path.exists(f): - print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') - return - elif not os.path.isdir(f): - print(f""" -WARNING -An open_folder request was made with an argument that is not a folder. -This could be an error or a malicious attempt to run code on your computer. -Requested path was: {f} -""", file=sys.stderr) - return - - if not shared.cmd_opts.hide_ui_dir_config: - path = os.path.normpath(f) - if platform.system() == "Windows": - os.startfile(path) - elif platform.system() == "Darwin": - sp.Popen(["open", path]) - elif "microsoft-standard-WSL2" in platform.uname().release: - sp.Popen(["wsl-open", path]) - else: - sp.Popen(["xdg-open", path]) - - with gr.Column(variant='panel'): - with gr.Group(): - result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) - - generation_info = None - with gr.Column(): - with gr.Row(elem_id=f"image_buttons_{tabname}"): - open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') - - if tabname != "extras": - save = gr.Button('Save', elem_id=f'save_{tabname}') - save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') - - buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) - - open_folder_button.click( - fn=lambda: open_folder(opts.outdir_samples or outdir), - inputs=[], - outputs=[], - ) - - if tabname != "extras": - with gr.Row(): - download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') - - with gr.Group(): - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') - if tabname == 'txt2img' or tabname == 'img2img': - generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") - generation_info_button.click( - fn=update_generation_info, - _js="(x, y) => [x, y, selected_gallery_index()]", - inputs=[generation_info, html_info], - outputs=[html_info], - preprocess=False - ) - - save.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - save_zip.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - else: - html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) - return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log - - -def create_sampler_and_steps_selection(choices, tabname): - if opts.samplers_in_dropdown: - with FormRow(elem_id=f"sampler_selection_{tabname}"): - sampler_index = gr.Dropdown(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) - else: - with FormGroup(elem_id=f"sampler_selection_{tabname}"): - steps = gr.Slider(minimum=1, maximum=150, step=1, elem_id=f"{tabname}_steps", label="Sampling steps", value=20) - sampler_index = gr.Radio(label='Sampling method', elem_id=f"{tabname}_sampling", choices=[x.name for x in choices], value=choices[0].name, type="index") - - return steps, sampler_index - - -def ordered_ui_categories(): - user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} - - for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): - yield category - - -def create_ui(): - import modules.img2img - import modules.txt2img - - reload_javascript() - - parameters_copypaste.reset() - - modules.scripts.scripts_current = modules.scripts.scripts_txt2img - modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) - - with gr.Blocks(analytics_enabled=False) as txt2img_interface: - txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) - - dummy_component = gr.Label(visible=False) - txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Row(elem_id='txt2img_progress_row'): - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="txt2img_progressbar") - txt2img_preview = gr.Image(elem_id='txt2img_preview', visible=False) - setup_progressbar(progressbar, txt2img_preview, 'txt2img') - - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel', elem_id="txt2img_settings"): - for category in ordered_ui_categories(): - if category == "sampler": - steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") - - elif category == "dimensions": - with FormRow(): - with gr.Column(elem_id="txt2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="txt2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") - - if opts.dimensions_and_batch_together: - with gr.Column(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - elif category == "cfg": - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="txt2img_cfg_scale") - - elif category == "seed": - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') - - elif category == "checkboxes": - with FormRow(elem_id="txt2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") - enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") - hr_final_resolution = FormHTML(value="", elem_id="txtimg_hr_finalres", label="Upscaled resolution", interactive=False) - - elif category == "hires_fix": - with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: - with FormRow(elem_id="txt2img_hires_fix_row1"): - hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) - hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") - - with FormRow(elem_id="txt2img_hires_fix_row2"): - hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") - hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") - hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") - - elif category == "batch": - if not opts.dimensions_and_batch_together: - with FormRow(elem_id="txt2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") - - elif category == "scripts": - with FormGroup(elem_id="txt2img_script_container"): - custom_inputs = modules.scripts.scripts_txt2img.setup_ui() - - hr_resolution_preview_inputs = [enable_hr, width, height, hr_scale, hr_resize_x, hr_resize_y] - for input in hr_resolution_preview_inputs: - input.change( - fn=calc_resolution_hires, - inputs=hr_resolution_preview_inputs, - outputs=[hr_final_resolution], - show_progress=False, - ) - input.change( - None, - _js="onCalcResolutionHires", - inputs=hr_resolution_preview_inputs, - outputs=[], - show_progress=False, - ) - - txt2img_gallery, generation_info, html_info, html_log = create_output_panel("txt2img", opts.outdir_txt2img_samples) - parameters_copypaste.bind_buttons({"txt2img": txt2img_paste}, None, txt2img_prompt) - - connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) - connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) - - txt2img_args = dict( - fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), - _js="submit", - inputs=[ - txt2img_prompt, - txt2img_negative_prompt, - txt2img_prompt_style, - txt2img_prompt_style2, - steps, - sampler_index, - restore_faces, - tiling, - batch_count, - batch_size, - cfg_scale, - seed, - subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, - height, - width, - enable_hr, - denoising_strength, - hr_scale, - hr_upscaler, - hr_second_pass_steps, - hr_resize_x, - hr_resize_y, - ] + custom_inputs, - - outputs=[ - txt2img_gallery, - generation_info, - html_info, - html_log, - ], - show_progress=False, - ) - - txt2img_prompt.submit(**txt2img_args) - submit.click(**txt2img_args) - - txt_prompt_img.change( - fn=modules.images.image_data, - inputs=[ - txt_prompt_img - ], - outputs=[ - txt2img_prompt, - txt_prompt_img - ] - ) - - enable_hr.change( - fn=lambda x: gr_show(x), - inputs=[enable_hr], - outputs=[hr_options], - show_progress = False, - ) - - txt2img_paste_fields = [ - (txt2img_prompt, "Prompt"), - (txt2img_negative_prompt, "Negative prompt"), - (steps, "Steps"), - (sampler_index, "Sampler"), - (restore_faces, "Face restoration"), - (cfg_scale, "CFG scale"), - (seed, "Seed"), - (width, "Size-1"), - (height, "Size-2"), - (batch_size, "Batch size"), - (subseed, "Variation seed"), - (subseed_strength, "Variation seed strength"), - (seed_resize_from_w, "Seed resize from-1"), - (seed_resize_from_h, "Seed resize from-2"), - (denoising_strength, "Denoising strength"), - (enable_hr, lambda d: "Denoising strength" in d), - (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)), - (hr_scale, "Hires upscale"), - (hr_upscaler, "Hires upscaler"), - (hr_second_pass_steps, "Hires steps"), - (hr_resize_x, "Hires resize-1"), - (hr_resize_y, "Hires resize-2"), - *modules.scripts.scripts_txt2img.infotext_fields - ] - parameters_copypaste.add_paste_fields("txt2img", None, txt2img_paste_fields) - - txt2img_preview_params = [ - txt2img_prompt, - txt2img_negative_prompt, - steps, - sampler_index, - cfg_scale, - seed, - width, - height, - ] - - token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) - - modules.scripts.scripts_current = modules.scripts.scripts_img2img - modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) - - with gr.Blocks(analytics_enabled=False) as img2img_interface: - img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) - - with gr.Row(elem_id='img2img_progress_row'): - img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="img2img_progressbar") - img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) - setup_progressbar(progressbar, img2img_preview, 'img2img') - - with FormRow().style(equal_height=False): - with gr.Column(variant='panel', elem_id="img2img_settings"): - - with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: - with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): - init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) - - with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): - init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) - init_img_with_mask_orig = gr.State(None) - - use_color_sketch = cmd_opts.gradio_inpaint_tool == "color-sketch" - if use_color_sketch: - def update_orig(image, state): - if image is not None: - same_size = state is not None and state.size == image.size - has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) - edited = same_size and has_exact_match - return image if not edited or state is None else state - - init_img_with_mask.change(update_orig, [init_img_with_mask, init_img_with_mask_orig], init_img_with_mask_orig) - - init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") - init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") - - with FormRow(): - mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") - mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") - - with FormRow(): - mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") - inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") - - with FormRow(): - inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") - - with FormRow(): - with gr.Column(): - inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") - - with gr.Column(scale=4): - inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") - - with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): - hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' - gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") - img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") - img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") - - with FormRow(): - resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") - - for category in ordered_ui_categories(): - if category == "sampler": - steps, sampler_index = create_sampler_and_steps_selection(samplers_for_img2img, "img2img") - - elif category == "dimensions": - with FormRow(): - with gr.Column(elem_id="img2img_column_size", scale=4): - width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="img2img_width") - height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") - - if opts.dimensions_and_batch_together: - with gr.Column(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - elif category == "cfg": - with FormGroup(): - cfg_scale = gr.Slider(minimum=1.0, maximum=30.0, step=0.5, label='CFG Scale', value=7.0, elem_id="img2img_cfg_scale") - denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.75, elem_id="img2img_denoising_strength") - - elif category == "seed": - seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') - - elif category == "checkboxes": - with FormRow(elem_id="img2img_checkboxes"): - restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") - tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") - - elif category == "batch": - if not opts.dimensions_and_batch_together: - with FormRow(elem_id="img2img_column_batch"): - batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") - batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") - - elif category == "scripts": - with FormGroup(elem_id="img2img_script_container"): - custom_inputs = modules.scripts.scripts_img2img.setup_ui() - - img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) - parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) - - connect_reuse_seed(seed, reuse_seed, generation_info, dummy_component, is_subseed=False) - connect_reuse_seed(subseed, reuse_subseed, generation_info, dummy_component, is_subseed=True) - - img2img_prompt_img.change( - fn=modules.images.image_data, - inputs=[ - img2img_prompt_img - ], - outputs=[ - img2img_prompt, - img2img_prompt_img - ] - ) - - mask_mode.change( - lambda mode, img: { - init_img_with_mask: gr_show(mode == 0), - init_img_inpaint: gr_show(mode == 1), - init_mask_inpaint: gr_show(mode == 1), - }, - inputs=[mask_mode, init_img_with_mask], - outputs=[ - init_img_with_mask, - init_img_inpaint, - init_mask_inpaint, - ], - ) - - img2img_args = dict( - fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), - _js="submit_img2img", - inputs=[ - dummy_component, - img2img_prompt, - img2img_negative_prompt, - img2img_prompt_style, - img2img_prompt_style2, - init_img, - init_img_with_mask, - init_img_with_mask_orig, - init_img_inpaint, - init_mask_inpaint, - mask_mode, - steps, - sampler_index, - mask_blur, - mask_alpha, - inpainting_fill, - restore_faces, - tiling, - batch_count, - batch_size, - cfg_scale, - denoising_strength, - seed, - subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox, - height, - width, - resize_mode, - inpaint_full_res, - inpaint_full_res_padding, - inpainting_mask_invert, - img2img_batch_input_dir, - img2img_batch_output_dir, - ] + custom_inputs, - outputs=[ - img2img_gallery, - generation_info, - html_info, - html_log, - ], - show_progress=False, - ) - - img2img_prompt.submit(**img2img_args) - submit.click(**img2img_args) - - img2img_interrogate.click( - fn=interrogate, - inputs=[init_img], - outputs=[img2img_prompt], - ) - - img2img_deepbooru.click( - fn=interrogate_deepbooru, - inputs=[init_img], - outputs=[img2img_prompt], - ) - - prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] - style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] - style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] - - for button, (prompt, negative_prompt) in zip([txt2img_save_style, img2img_save_style], prompts): - button.click( - fn=add_style, - _js="ask_for_style_name", - # Have to pass empty dummy component here, because the JavaScript and Python function have to accept - # the same number of parameters, but we only know the style-name after the JavaScript prompt - inputs=[dummy_component, prompt, negative_prompt], - outputs=[txt2img_prompt_style, img2img_prompt_style, txt2img_prompt_style2, img2img_prompt_style2], - ) - - for button, (prompt, negative_prompt), (style1, style2), js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): - button.click( - fn=apply_styles, - _js=js_func, - inputs=[prompt, negative_prompt, style1, style2], - outputs=[prompt, negative_prompt, style1, style2], - ) - - token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) - - img2img_paste_fields = [ - (img2img_prompt, "Prompt"), - (img2img_negative_prompt, "Negative prompt"), - (steps, "Steps"), - (sampler_index, "Sampler"), - (restore_faces, "Face restoration"), - (cfg_scale, "CFG scale"), - (seed, "Seed"), - (width, "Size-1"), - (height, "Size-2"), - (batch_size, "Batch size"), - (subseed, "Variation seed"), - (subseed_strength, "Variation seed strength"), - (seed_resize_from_w, "Seed resize from-1"), - (seed_resize_from_h, "Seed resize from-2"), - (denoising_strength, "Denoising strength"), - (mask_blur, "Mask blur"), - *modules.scripts.scripts_img2img.infotext_fields - ] - parameters_copypaste.add_paste_fields("img2img", init_img, img2img_paste_fields) - parameters_copypaste.add_paste_fields("inpaint", init_img_with_mask, img2img_paste_fields) - - modules.scripts.scripts_current = None - - with gr.Blocks(analytics_enabled=False) as extras_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - with gr.Tabs(elem_id="mode_extras"): - with gr.TabItem('Single Image', elem_id="extras_single_tab"): - extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") - - with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): - image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") - - with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): - extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") - extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") - show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") - - submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') - - with gr.Tabs(elem_id="extras_resize_mode"): - with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): - upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") - with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): - with gr.Group(): - with gr.Row(): - upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") - upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") - upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") - - with gr.Group(): - extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - - with gr.Group(): - extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") - - with gr.Group(): - gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") - - with gr.Group(): - codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") - codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") - - with gr.Group(): - upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") - - result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) - - submit.click( - fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), - _js="get_extras_tab_index", - inputs=[ - dummy_component, - dummy_component, - extras_image, - image_batch, - extras_batch_input_dir, - extras_batch_output_dir, - show_extras_results, - gfpgan_visibility, - codeformer_visibility, - codeformer_weight, - upscaling_resize, - upscaling_resize_w, - upscaling_resize_h, - upscaling_crop, - extras_upscaler_1, - extras_upscaler_2, - extras_upscaler_2_visibility, - upscale_before_face_fix, - ], - outputs=[ - result_images, - html_info_x, - html_info, - ] - ) - parameters_copypaste.add_paste_fields("extras", extras_image, None) - - extras_image.change( - fn=modules.extras.clear_cache, - inputs=[], outputs=[] - ) - - with gr.Blocks(analytics_enabled=False) as pnginfo_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - image = gr.Image(elem_id="pnginfo_image", label="Source", source="upload", interactive=True, type="pil") - - with gr.Column(variant='panel'): - html = gr.HTML() - generation_info = gr.Textbox(visible=False, elem_id="pnginfo_generation_info") - html2 = gr.HTML() - with gr.Row(): - buttons = parameters_copypaste.create_buttons(["txt2img", "img2img", "inpaint", "extras"]) - parameters_copypaste.bind_buttons(buttons, image, generation_info) - - image.change( - fn=wrap_gradio_call(modules.extras.run_pnginfo), - inputs=[image], - outputs=[html, generation_info, html2], - ) - - with gr.Blocks(analytics_enabled=False) as modelmerger_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") - - with gr.Row(): - primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") - create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") - - secondary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_secondary_model_name", label="Secondary model (B)") - create_refresh_button(secondary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_B") - - tertiary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_tertiary_model_name", label="Tertiary model (C)") - create_refresh_button(tertiary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_C") - - custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") - interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") - interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") - - with gr.Row(): - checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") - save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") - - modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') - - with gr.Column(variant='panel'): - submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) - - with gr.Blocks(analytics_enabled=False) as train_interface: - with gr.Row().style(equal_height=False): - gr.HTML(value="

See wiki for detailed explanation.

") - - with gr.Row().style(equal_height=False): - with gr.Tabs(elem_id="train_tabs"): - - with gr.Tab(label="Create embedding"): - new_embedding_name = gr.Textbox(label="Name", elem_id="train_new_embedding_name") - initialization_text = gr.Textbox(label="Initialization text", value="*", elem_id="train_initialization_text") - nvpt = gr.Slider(label="Number of vectors per token", minimum=1, maximum=75, step=1, value=1, elem_id="train_nvpt") - overwrite_old_embedding = gr.Checkbox(value=False, label="Overwrite Old Embedding", elem_id="train_overwrite_old_embedding") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - create_embedding = gr.Button(value="Create embedding", variant='primary', elem_id="train_create_embedding") - - with gr.Tab(label="Create hypernetwork"): - new_hypernetwork_name = gr.Textbox(label="Name", elem_id="train_new_hypernetwork_name") - new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "1024", "320", "640", "1280"], elem_id="train_new_hypernetwork_sizes") - new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'", elem_id="train_new_hypernetwork_layer_structure") - new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys, elem_id="train_new_hypernetwork_activation_func") - new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"], elem_id="train_new_hypernetwork_initialization_option") - new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization", elem_id="train_new_hypernetwork_add_layer_norm") - new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout", elem_id="train_new_hypernetwork_use_dropout") - new_hypernetwork_dropout_structure = gr.Textbox("0, 0, 0", label="Enter hypernetwork Dropout structure (or empty). Recommended : 0~0.35 incrementing sequence: 0, 0.05, 0.15", placeholder="1st and last digit must be 0 and values should be between 0 and 1. ex:'0, 0.01, 0'") - overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork", elem_id="train_overwrite_old_hypernetwork") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - create_hypernetwork = gr.Button(value="Create hypernetwork", variant='primary', elem_id="train_create_hypernetwork") - - with gr.Tab(label="Preprocess images"): - process_src = gr.Textbox(label='Source directory', elem_id="train_process_src") - process_dst = gr.Textbox(label='Destination directory', elem_id="train_process_dst") - process_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_process_width") - process_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_process_height") - preprocess_txt_action = gr.Dropdown(label='Existing Caption txt Action', value="ignore", choices=["ignore", "copy", "prepend", "append"], elem_id="train_preprocess_txt_action") - - with gr.Row(): - process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") - process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") - process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") - process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") - process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") - - with gr.Row(visible=False) as process_split_extra_row: - process_split_threshold = gr.Slider(label='Split image threshold', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_split_threshold") - process_overlap_ratio = gr.Slider(label='Split image overlap ratio', value=0.2, minimum=0.0, maximum=0.9, step=0.05, elem_id="train_process_overlap_ratio") - - with gr.Row(visible=False) as process_focal_crop_row: - process_focal_crop_face_weight = gr.Slider(label='Focal point face weight', value=0.9, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_face_weight") - process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") - process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") - process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") - - with gr.Row(): - with gr.Column(scale=3): - gr.HTML(value="") - - with gr.Column(): - with gr.Row(): - interrupt_preprocessing = gr.Button("Interrupt", elem_id="train_interrupt_preprocessing") - run_preprocess = gr.Button(value="Preprocess", variant='primary', elem_id="train_run_preprocess") - - process_split.change( - fn=lambda show: gr_show(show), - inputs=[process_split], - outputs=[process_split_extra_row], - ) - - process_focal_crop.change( - fn=lambda show: gr_show(show), - inputs=[process_focal_crop], - outputs=[process_focal_crop_row], - ) - - def get_textual_inversion_template_names(): - return sorted([x for x in textual_inversion.textual_inversion_templates]) - - with gr.Tab(label="Train"): - gr.HTML(value="

Train an embedding or Hypernetwork; you must specify a directory with a set of 1:1 ratio images [wiki]

") - with FormRow(): - train_embedding_name = gr.Dropdown(label='Embedding', elem_id="train_embedding", choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())) - create_refresh_button(train_embedding_name, sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings, lambda: {"choices": sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys())}, "refresh_train_embedding_name") - - train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', elem_id="train_hypernetwork", choices=[x for x in shared.hypernetworks.keys()]) - create_refresh_button(train_hypernetwork_name, shared.reload_hypernetworks, lambda: {"choices": sorted([x for x in shared.hypernetworks.keys()])}, "refresh_train_hypernetwork_name") - - with FormRow(): - embedding_learn_rate = gr.Textbox(label='Embedding Learning rate', placeholder="Embedding Learning rate", value="0.005", elem_id="train_embedding_learn_rate") - hypernetwork_learn_rate = gr.Textbox(label='Hypernetwork Learning rate', placeholder="Hypernetwork Learning rate", value="0.00001", elem_id="train_hypernetwork_learn_rate") - - with FormRow(): - clip_grad_mode = gr.Dropdown(value="disabled", label="Gradient Clipping", choices=["disabled", "value", "norm"]) - clip_grad_value = gr.Textbox(placeholder="Gradient clip value", value="0.1", show_label=False) - - with FormRow(): - batch_size = gr.Number(label='Batch size', value=1, precision=0, elem_id="train_batch_size") - gradient_step = gr.Number(label='Gradient accumulation steps', value=1, precision=0, elem_id="train_gradient_step") - - dataset_directory = gr.Textbox(label='Dataset directory', placeholder="Path to directory with input images", elem_id="train_dataset_directory") - log_directory = gr.Textbox(label='Log directory', placeholder="Path to directory where to write outputs", value="textual_inversion", elem_id="train_log_directory") - - with FormRow(): - template_file = gr.Dropdown(label='Prompt template', value="style_filewords.txt", elem_id="train_template_file", choices=get_textual_inversion_template_names()) - create_refresh_button(template_file, textual_inversion.list_textual_inversion_templates, lambda: {"choices": get_textual_inversion_template_names()}, "refrsh_train_template_file") - - training_width = gr.Slider(minimum=64, maximum=2048, step=8, label="Width", value=512, elem_id="train_training_width") - training_height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="train_training_height") - varsize = gr.Checkbox(label="Do not resize images", value=False, elem_id="train_varsize") - steps = gr.Number(label='Max steps', value=100000, precision=0, elem_id="train_steps") - - with FormRow(): - create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_create_image_every") - save_embedding_every = gr.Number(label='Save a copy of embedding to log directory every N steps, 0 to disable', value=500, precision=0, elem_id="train_save_embedding_every") - - save_image_with_stored_embedding = gr.Checkbox(label='Save images with embedding in PNG chunks', value=True, elem_id="train_save_image_with_stored_embedding") - preview_from_txt2img = gr.Checkbox(label='Read parameters (prompt, etc...) from txt2img tab when making previews', value=False, elem_id="train_preview_from_txt2img") - - shuffle_tags = gr.Checkbox(label="Shuffle tags by ',' when creating prompts.", value=False, elem_id="train_shuffle_tags") - tag_drop_out = gr.Slider(minimum=0, maximum=1, step=0.1, label="Drop out tags when creating prompts.", value=0, elem_id="train_tag_drop_out") - - latent_sampling_method = gr.Radio(label='Choose latent sampling method', value="once", choices=['once', 'deterministic', 'random'], elem_id="train_latent_sampling_method") - - with gr.Row(): - train_embedding = gr.Button(value="Train Embedding", variant='primary', elem_id="train_train_embedding") - interrupt_training = gr.Button(value="Interrupt", elem_id="train_interrupt_training") - train_hypernetwork = gr.Button(value="Train Hypernetwork", variant='primary', elem_id="train_train_hypernetwork") - - params = script_callbacks.UiTrainTabParams(txt2img_preview_params) - - script_callbacks.ui_train_tabs_callback(params) - - with gr.Column(): - progressbar = gr.HTML(elem_id="ti_progressbar") - ti_output = gr.Text(elem_id="ti_output", value="", show_label=False) - - ti_gallery = gr.Gallery(label='Output', show_label=False, elem_id='ti_gallery').style(grid=4) - ti_preview = gr.Image(elem_id='ti_preview', visible=False) - ti_progress = gr.HTML(elem_id="ti_progress", value="") - ti_outcome = gr.HTML(elem_id="ti_error", value="") - setup_progressbar(progressbar, ti_preview, 'ti', textinfo=ti_progress) - - create_embedding.click( - fn=modules.textual_inversion.ui.create_embedding, - inputs=[ - new_embedding_name, - initialization_text, - nvpt, - overwrite_old_embedding, - ], - outputs=[ - train_embedding_name, - ti_output, - ti_outcome, - ] - ) - - create_hypernetwork.click( - fn=modules.hypernetworks.ui.create_hypernetwork, - inputs=[ - new_hypernetwork_name, - new_hypernetwork_sizes, - overwrite_old_hypernetwork, - new_hypernetwork_layer_structure, - new_hypernetwork_activation_func, - new_hypernetwork_initialization_option, - new_hypernetwork_add_layer_norm, - new_hypernetwork_use_dropout, - new_hypernetwork_dropout_structure - ], - outputs=[ - train_hypernetwork_name, - ti_output, - ti_outcome, - ] - ) - - run_preprocess.click( - fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.preprocess, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - process_src, - process_dst, - process_width, - process_height, - preprocess_txt_action, - process_flip, - process_split, - process_caption, - process_caption_deepbooru, - process_split_threshold, - process_overlap_ratio, - process_focal_crop, - process_focal_crop_face_weight, - process_focal_crop_entropy_weight, - process_focal_crop_edges_weight, - process_focal_crop_debug, - ], - outputs=[ - ti_output, - ti_outcome, - ], - ) - - train_embedding.click( - fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.train_embedding, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - train_embedding_name, - embedding_learn_rate, - batch_size, - gradient_step, - dataset_directory, - log_directory, - training_width, - training_height, - varsize, - steps, - clip_grad_mode, - clip_grad_value, - shuffle_tags, - tag_drop_out, - latent_sampling_method, - create_image_every, - save_embedding_every, - template_file, - save_image_with_stored_embedding, - preview_from_txt2img, - *txt2img_preview_params, - ], - outputs=[ - ti_output, - ti_outcome, - ] - ) - - train_hypernetwork.click( - fn=wrap_gradio_gpu_call(modules.hypernetworks.ui.train_hypernetwork, extra_outputs=[gr.update()]), - _js="start_training_textual_inversion", - inputs=[ - train_hypernetwork_name, - hypernetwork_learn_rate, - batch_size, - gradient_step, - dataset_directory, - log_directory, - training_width, - training_height, - varsize, - steps, - clip_grad_mode, - clip_grad_value, - shuffle_tags, - tag_drop_out, - latent_sampling_method, - create_image_every, - save_embedding_every, - template_file, - preview_from_txt2img, - *txt2img_preview_params, - ], - outputs=[ - ti_output, - ti_outcome, - ] - ) - - interrupt_training.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - interrupt_preprocessing.click( - fn=lambda: shared.state.interrupt(), - inputs=[], - outputs=[], - ) - - def create_setting_component(key, is_quicksettings=False): - def fun(): - return opts.data[key] if key in opts.data else opts.data_labels[key].default - - info = opts.data_labels[key] - t = type(info.default) - - args = info.component_args() if callable(info.component_args) else info.component_args - - if info.component is not None: - comp = info.component - elif t == str: - comp = gr.Textbox - elif t == int: - comp = gr.Number - elif t == bool: - comp = gr.Checkbox - else: - raise Exception(f'bad options item type: {str(t)} for key {key}') - - elem_id = "setting_"+key - - if info.refresh is not None: - if is_quicksettings: - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) - else: - with FormRow(): - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - create_refresh_button(res, info.refresh, info.component_args, "refresh_" + key) - else: - res = comp(label=info.label, value=fun(), elem_id=elem_id, **(args or {})) - - return res - - components = [] - component_dict = {} - - script_callbacks.ui_settings_callback() - opts.reorder() - - def run_settings(*args): - changed = [] - - for key, value, comp in zip(opts.data_labels.keys(), args, components): - assert comp == dummy_component or opts.same_type(value, opts.data_labels[key].default), f"Bad value for setting {key}: {value}; expecting {type(opts.data_labels[key].default).__name__}" - - for key, value, comp in zip(opts.data_labels.keys(), args, components): - if comp == dummy_component: - continue - - if opts.set(key, value): - changed.append(key) - - try: - opts.save(shared.config_filename) - except RuntimeError: - return opts.dumpjson(), f'{len(changed)} settings changed without save: {", ".join(changed)}.' - return opts.dumpjson(), f'{len(changed)} settings changed{": " if len(changed) > 0 else ""}{", ".join(changed)}.' - - def run_settings_single(value, key): - if not opts.same_type(value, opts.data_labels[key].default): - return gr.update(visible=True), opts.dumpjson() - - if not opts.set(key, value): - return gr.update(value=getattr(opts, key)), opts.dumpjson() - - opts.save(shared.config_filename) - - return gr.update(value=value), opts.dumpjson() - - with gr.Blocks(analytics_enabled=False) as settings_interface: - with gr.Row(): - with gr.Column(scale=6): - settings_submit = gr.Button(value="Apply settings", variant='primary', elem_id="settings_submit") - with gr.Column(): - restart_gradio = gr.Button(value='Reload UI', variant='primary', elem_id="settings_restart_gradio") - - result = gr.HTML(elem_id="settings_result") - - quicksettings_names = [x.strip() for x in opts.quicksettings.split(",")] - quicksettings_names = {x: i for i, x in enumerate(quicksettings_names) if x != 'quicksettings'} - - quicksettings_list = [] - - previous_section = None - current_tab = None - with gr.Tabs(elem_id="settings"): - for i, (k, item) in enumerate(opts.data_labels.items()): - section_must_be_skipped = item.section[0] is None - - if previous_section != item.section and not section_must_be_skipped: - elem_id, text = item.section - - if current_tab is not None: - current_tab.__exit__() - - current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) - current_tab.__enter__() - - previous_section = item.section - - if k in quicksettings_names and not shared.cmd_opts.freeze_settings: - quicksettings_list.append((i, k, item)) - components.append(dummy_component) - elif section_must_be_skipped: - components.append(dummy_component) - else: - component = create_setting_component(k) - component_dict[k] = component - components.append(component) - - if current_tab is not None: - current_tab.__exit__() - - with gr.TabItem("Actions"): - request_notifications = gr.Button(value='Request browser notifications', elem_id="request_notifications") - download_localization = gr.Button(value='Download localization template', elem_id="download_localization") - reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") - - if os.path.exists("html/licenses.html"): - with open("html/licenses.html", encoding="utf8") as file: - with gr.TabItem("Licenses"): - gr.HTML(file.read(), elem_id="licenses") - - gr.Button(value="Show all pages", elem_id="settings_show_all_pages") - - request_notifications.click( - fn=lambda: None, - inputs=[], - outputs=[], - _js='function(){}' - ) - - download_localization.click( - fn=lambda: None, - inputs=[], - outputs=[], - _js='download_localization' - ) - - def reload_scripts(): - modules.scripts.reload_script_body_only() - reload_javascript() # need to refresh the html page - - reload_script_bodies.click( - fn=reload_scripts, - inputs=[], - outputs=[] - ) - - def request_restart(): - shared.state.interrupt() - shared.state.need_restart = True - - restart_gradio.click( - fn=request_restart, - _js='restart_reload', - inputs=[], - outputs=[], - ) - - interfaces = [ - (txt2img_interface, "txt2img", "txt2img"), - (img2img_interface, "img2img", "img2img"), - (extras_interface, "Extras", "extras"), - (pnginfo_interface, "PNG Info", "pnginfo"), - (modelmerger_interface, "Checkpoint Merger", "modelmerger"), - (train_interface, "Train", "ti"), - ] - - css = "" - - for cssfile in modules.scripts.list_files_with_name("style.css"): - if not os.path.isfile(cssfile): - continue - - with open(cssfile, "r", encoding="utf8") as file: - css += file.read() + "\n" - - if os.path.exists(os.path.join(script_path, "user.css")): - with open(os.path.join(script_path, "user.css"), "r", encoding="utf8") as file: - css += file.read() + "\n" - - if not cmd_opts.no_progressbar_hiding: - css += css_hide_progressbar - - interfaces += script_callbacks.ui_tabs_callback() - interfaces += [(settings_interface, "Settings", "settings")] - - extensions_interface = ui_extensions.create_ui() - interfaces += [(extensions_interface, "Extensions", "extensions")] - - with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: - with gr.Row(elem_id="quicksettings"): - for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): - component = create_setting_component(k, is_quicksettings=True) - component_dict[k] = component - - parameters_copypaste.integrate_settings_paste_fields(component_dict) - parameters_copypaste.run_bind() - - with gr.Tabs(elem_id="tabs") as tabs: - for interface, label, ifid in interfaces: - with gr.TabItem(label, id=ifid, elem_id='tab_' + ifid): - interface.render() - - if os.path.exists(os.path.join(script_path, "notification.mp3")): - audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) - - if os.path.exists("html/footer.html"): - with open("html/footer.html", encoding="utf8") as file: - footer = file.read() - footer = footer.format(versions=versions_html()) - gr.HTML(footer, elem_id="footer") - - text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) - settings_submit.click( - fn=wrap_gradio_call(run_settings, extra_outputs=[gr.update()]), - inputs=components, - outputs=[text_settings, result], - ) - - for i, k, item in quicksettings_list: - component = component_dict[k] - - component.change( - fn=lambda value, k=k: run_settings_single(value, key=k), - inputs=[component], - outputs=[component, text_settings], - ) - - component_keys = [k for k in opts.data_labels.keys() if k in component_dict] - - def get_settings_values(): - return [getattr(opts, key) for key in component_keys] - - demo.load( - fn=get_settings_values, - inputs=[], - outputs=[component_dict[k] for k in component_keys], - ) - - def modelmerger(*args): - try: - results = modules.extras.run_modelmerger(*args) - except Exception as e: - print("Error loading/saving model file:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - modules.sd_models.list_models() # to remove the potentially missing models from the list - return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] - return results - - modelmerger_merge.click( - fn=modelmerger, - inputs=[ - primary_model_name, - secondary_model_name, - tertiary_model_name, - interp_method, - interp_amount, - save_as_half, - custom_name, - checkpoint_format, - ], - outputs=[ - submit_result, - primary_model_name, - secondary_model_name, - tertiary_model_name, - component_dict['sd_model_checkpoint'], - ] - ) - - ui_config_file = cmd_opts.ui_config_file - ui_settings = {} - settings_count = len(ui_settings) - error_loading = False - - try: - if os.path.exists(ui_config_file): - with open(ui_config_file, "r", encoding="utf8") as file: - ui_settings = json.load(file) - except Exception: - error_loading = True - print("Error loading settings:", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - - def loadsave(path, x): - def apply_field(obj, field, condition=None, init_field=None): - key = path + "/" + field - - if getattr(obj, 'custom_script_source', None) is not None: - key = 'customscript/' + obj.custom_script_source + '/' + key - - if getattr(obj, 'do_not_save_to_config', False): - return - - saved_value = ui_settings.get(key, None) - if saved_value is None: - ui_settings[key] = getattr(obj, field) - elif condition and not condition(saved_value): - print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') - else: - setattr(obj, field, saved_value) - if init_field is not None: - init_field(saved_value) - - if type(x) in [gr.Slider, gr.Radio, gr.Checkbox, gr.Textbox, gr.Number, gr.Dropdown] and x.visible: - apply_field(x, 'visible') - - if type(x) == gr.Slider: - apply_field(x, 'value') - apply_field(x, 'minimum') - apply_field(x, 'maximum') - apply_field(x, 'step') - - if type(x) == gr.Radio: - apply_field(x, 'value', lambda val: val in x.choices) - - if type(x) == gr.Checkbox: - apply_field(x, 'value') - - if type(x) == gr.Textbox: - apply_field(x, 'value') - - if type(x) == gr.Number: - apply_field(x, 'value') - - if type(x) == gr.Dropdown: - apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) - - visit(txt2img_interface, loadsave, "txt2img") - visit(img2img_interface, loadsave, "img2img") - visit(extras_interface, loadsave, "extras") - visit(modelmerger_interface, loadsave, "modelmerger") - visit(train_interface, loadsave, "train") - - if not error_loading and (not os.path.exists(ui_config_file) or settings_count != len(ui_settings)): - with open(ui_config_file, "w", encoding="utf8") as file: - json.dump(ui_settings, file, indent=4) - - return demo - - -def reload_javascript(): - with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: - javascript = f'' - - scripts_list = modules.scripts.list_scripts("javascript", ".js") - - for basedir, filename, path in scripts_list: - with open(path, "r", encoding="utf8") as jsfile: - javascript += f"\n" - - if cmd_opts.theme is not None: - javascript += f"\n\n" - - javascript += f"\n" - - def template_response(*args, **kwargs): - res = shared.GradioTemplateResponseOriginal(*args, **kwargs) - res.body = res.body.replace( - b'', f'{javascript}'.encode("utf8")) - res.init_headers() - return res - - gradio.routes.templates.TemplateResponse = template_response - - -if not hasattr(shared, 'GradioTemplateResponseOriginal'): - shared.GradioTemplateResponseOriginal = gradio.routes.templates.TemplateResponse - - -def versions_html(): - import torch - import launch - - python_version = ".".join([str(x) for x in sys.version_info[0:3]]) - commit = launch.commit_hash() - short_commit = commit[0:8] - - if shared.xformers_available: - import xformers - xformers_version = xformers.__version__ - else: - xformers_version = "N/A" - - return f""" -python: {python_version} - •  -torch: {torch.__version__} - •  -xformers: {xformers_version} - •  -gradio: {gr.__version__} - •  -commit: {short_commit} -""" -- cgit v1.2.3 From 0c3feb202c5714abd50d879c1db2cd9a71ce93e3 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 10 Jan 2023 14:08:29 +0300 Subject: disable torch weight initialization and CLIP downloading/reading checkpoint to speedup creating sd model from config --- modules/sd_disable_initialization.py | 44 ++++++++++++++++++++++++++++++++++++ modules/sd_models.py | 5 ++-- 2 files changed, 47 insertions(+), 2 deletions(-) create mode 100644 modules/sd_disable_initialization.py (limited to 'modules') diff --git a/modules/sd_disable_initialization.py b/modules/sd_disable_initialization.py new file mode 100644 index 00000000..c9a3b5e4 --- /dev/null +++ b/modules/sd_disable_initialization.py @@ -0,0 +1,44 @@ +import ldm.modules.encoders.modules +import open_clip +import torch + + +class DisableInitialization: + """ + When an object of this class enters a `with` block, it starts preventing torch's layer initialization + functions from working, and changes CLIP and OpenCLIP to not download model weights. When it leaves, + reverts everything to how it was. + + Use like this: + ``` + with DisableInitialization(): + do_things() + ``` + """ + + def __enter__(self): + def do_nothing(*args, **kwargs): + pass + + def create_model_and_transforms_without_pretrained(*args, pretrained=None, **kwargs): + return self.create_model_and_transforms(*args, pretrained=None, **kwargs) + + def CLIPTextModel_from_pretrained(pretrained_model_name_or_path, *model_args, **kwargs): + return self.CLIPTextModel_from_pretrained(None, *model_args, config=pretrained_model_name_or_path, state_dict={}, **kwargs) + + self.init_kaiming_uniform = torch.nn.init.kaiming_uniform_ + self.init_no_grad_normal = torch.nn.init._no_grad_normal_ + self.create_model_and_transforms = open_clip.create_model_and_transforms + self.CLIPTextModel_from_pretrained = ldm.modules.encoders.modules.CLIPTextModel.from_pretrained + + torch.nn.init.kaiming_uniform_ = do_nothing + torch.nn.init._no_grad_normal_ = do_nothing + open_clip.create_model_and_transforms = create_model_and_transforms_without_pretrained + ldm.modules.encoders.modules.CLIPTextModel.from_pretrained = CLIPTextModel_from_pretrained + + def __exit__(self, exc_type, exc_val, exc_tb): + torch.nn.init.kaiming_uniform_ = self.init_kaiming_uniform + torch.nn.init._no_grad_normal_ = self.init_no_grad_normal + open_clip.create_model_and_transforms = self.create_model_and_transforms + ldm.modules.encoders.modules.CLIPTextModel.from_pretrained = self.CLIPTextModel_from_pretrained + diff --git a/modules/sd_models.py b/modules/sd_models.py index 0a6d55ca..ee241032 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -13,7 +13,7 @@ import ldm.modules.midas as midas from ldm.util import instantiate_from_config -from modules import shared, modelloader, devices, script_callbacks, sd_vae +from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization from modules.paths import models_path from modules.sd_hijack_inpainting import do_inpainting_hijack, should_hijack_inpainting @@ -319,7 +319,8 @@ def load_model(checkpoint_info=None): if shared.cmd_opts.no_half: sd_config.model.params.unet_config.params.use_fp16 = False - sd_model = instantiate_from_config(sd_config.model) + with sd_disable_initialization.DisableInitialization(): + sd_model = instantiate_from_config(sd_config.model) load_model_weights(sd_model, checkpoint_info) -- cgit v1.2.3 From ce3f639ec8758ce2bc90483336361d2dc25acd3a Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 10 Jan 2023 16:51:04 +0300 Subject: add more stuff to ignore when creating model from config prevent .vae.safetensors files from being listed as stable diffusion models --- modules/modelloader.py | 4 +++- modules/sd_disable_initialization.py | 29 +++++++++++++++++++++++++---- modules/sd_models.py | 32 ++++++++++++++++++++++++++++---- 3 files changed, 56 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/modelloader.py b/modules/modelloader.py index 6a1a7ac8..e9aa514e 100644 --- a/modules/modelloader.py +++ b/modules/modelloader.py @@ -10,7 +10,7 @@ from modules.upscaler import Upscaler from modules.paths import script_path, models_path -def load_models(model_path: str, model_url: str = None, command_path: str = None, ext_filter=None, download_name=None) -> list: +def load_models(model_path: str, model_url: str = None, command_path: str = None, ext_filter=None, download_name=None, ext_blacklist=None) -> list: """ A one-and done loader to try finding the desired models in specified directories. @@ -45,6 +45,8 @@ def load_models(model_path: str, model_url: str = None, command_path: str = None full_path = file if os.path.isdir(full_path): continue + if ext_blacklist is not None and any([full_path.endswith(x) for x in ext_blacklist]): + continue if len(ext_filter) != 0: model_name, extension = os.path.splitext(file) if extension not in ext_filter: diff --git a/modules/sd_disable_initialization.py b/modules/sd_disable_initialization.py index c9a3b5e4..9942bd7e 100644 --- a/modules/sd_disable_initialization.py +++ b/modules/sd_disable_initialization.py @@ -1,15 +1,19 @@ import ldm.modules.encoders.modules import open_clip import torch +import transformers.utils.hub class DisableInitialization: """ - When an object of this class enters a `with` block, it starts preventing torch's layer initialization - functions from working, and changes CLIP and OpenCLIP to not download model weights. When it leaves, - reverts everything to how it was. + When an object of this class enters a `with` block, it starts: + - preventing torch's layer initialization functions from working + - changes CLIP and OpenCLIP to not download model weights + - changes CLIP to not make requests to check if there is a new version of a file you already have - Use like this: + When it leaves the block, it reverts everything to how it was before. + + Use it like this: ``` with DisableInitialization(): do_things() @@ -26,19 +30,36 @@ class DisableInitialization: def CLIPTextModel_from_pretrained(pretrained_model_name_or_path, *model_args, **kwargs): return self.CLIPTextModel_from_pretrained(None, *model_args, config=pretrained_model_name_or_path, state_dict={}, **kwargs) + def transformers_utils_hub_get_from_cache(url, *args, local_files_only=False, **kwargs): + + # this file is always 404, prevent making request + if url == 'https://huggingface.co/openai/clip-vit-large-patch14/resolve/main/added_tokens.json': + raise transformers.utils.hub.EntryNotFoundError + + try: + return self.transformers_utils_hub_get_from_cache(url, *args, local_files_only=True, **kwargs) + except Exception as e: + return self.transformers_utils_hub_get_from_cache(url, *args, local_files_only=False, **kwargs) + self.init_kaiming_uniform = torch.nn.init.kaiming_uniform_ self.init_no_grad_normal = torch.nn.init._no_grad_normal_ + self.init_no_grad_uniform_ = torch.nn.init._no_grad_uniform_ self.create_model_and_transforms = open_clip.create_model_and_transforms self.CLIPTextModel_from_pretrained = ldm.modules.encoders.modules.CLIPTextModel.from_pretrained + self.transformers_utils_hub_get_from_cache = transformers.utils.hub.get_from_cache torch.nn.init.kaiming_uniform_ = do_nothing torch.nn.init._no_grad_normal_ = do_nothing + torch.nn.init._no_grad_uniform_ = do_nothing open_clip.create_model_and_transforms = create_model_and_transforms_without_pretrained ldm.modules.encoders.modules.CLIPTextModel.from_pretrained = CLIPTextModel_from_pretrained + transformers.utils.hub.get_from_cache = transformers_utils_hub_get_from_cache def __exit__(self, exc_type, exc_val, exc_tb): torch.nn.init.kaiming_uniform_ = self.init_kaiming_uniform torch.nn.init._no_grad_normal_ = self.init_no_grad_normal + torch.nn.init._no_grad_uniform_ = self.init_no_grad_uniform_ open_clip.create_model_and_transforms = self.create_model_and_transforms ldm.modules.encoders.modules.CLIPTextModel.from_pretrained = self.CLIPTextModel_from_pretrained + transformers.utils.hub.get_from_cache = self.transformers_utils_hub_get_from_cache diff --git a/modules/sd_models.py b/modules/sd_models.py index ee241032..1bb9088b 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -2,6 +2,7 @@ import collections import os.path import sys import gc +import time from collections import namedtuple import torch import re @@ -61,7 +62,7 @@ def find_checkpoint_config(info): def list_models(): checkpoints_list.clear() - model_list = modelloader.load_models(model_path=model_path, command_path=shared.cmd_opts.ckpt_dir, ext_filter=[".ckpt", ".safetensors"]) + model_list = modelloader.load_models(model_path=model_path, command_path=shared.cmd_opts.ckpt_dir, ext_filter=[".ckpt", ".safetensors"], ext_blacklist=[".vae.safetensors"]) def modeltitle(path, shorthash): abspath = os.path.abspath(path) @@ -288,6 +289,17 @@ def enable_midas_autodownload(): midas.api.load_model = load_model_wrapper +class Timer: + def __init__(self): + self.start = time.time() + + def elapsed(self): + end = time.time() + res = end - self.start + self.start = end + return res + + def load_model(checkpoint_info=None): from modules import lowvram, sd_hijack checkpoint_info = checkpoint_info or select_checkpoint() @@ -319,11 +331,17 @@ def load_model(checkpoint_info=None): if shared.cmd_opts.no_half: sd_config.model.params.unet_config.params.use_fp16 = False + timer = Timer() + with sd_disable_initialization.DisableInitialization(): sd_model = instantiate_from_config(sd_config.model) + elapsed_create = timer.elapsed() + load_model_weights(sd_model, checkpoint_info) + elapsed_load_weights = timer.elapsed() + if shared.cmd_opts.lowvram or shared.cmd_opts.medvram: lowvram.setup_for_low_vram(sd_model, shared.cmd_opts.medvram) else: @@ -338,7 +356,9 @@ def load_model(checkpoint_info=None): script_callbacks.model_loaded_callback(sd_model) - print("Model loaded.") + elapsed_the_rest = timer.elapsed() + + print(f"Model loaded in {elapsed_create + elapsed_load_weights + elapsed_the_rest:.1f}s ({elapsed_create:.1f}s create model, {elapsed_load_weights:.1f}s load weights).") return sd_model @@ -349,7 +369,7 @@ def reload_model_weights(sd_model=None, info=None): if not sd_model: sd_model = shared.sd_model - if sd_model is None: # previous model load failed + if sd_model is None: # previous model load failed current_checkpoint_info = None else: current_checkpoint_info = sd_model.sd_checkpoint_info @@ -371,6 +391,8 @@ def reload_model_weights(sd_model=None, info=None): sd_hijack.model_hijack.undo_hijack(sd_model) + timer = Timer() + try: load_model_weights(sd_model, checkpoint_info) except Exception as e: @@ -384,6 +406,8 @@ def reload_model_weights(sd_model=None, info=None): if not shared.cmd_opts.lowvram and not shared.cmd_opts.medvram: sd_model.to(devices.device) - print("Weights loaded.") + elapsed = timer.elapsed() + + print(f"Weights loaded in {elapsed:.1f}s.") return sd_model -- cgit v1.2.3 From 0f8603a55988d22616b17140e6c4a7e9d0736af5 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 10 Jan 2023 17:46:59 +0300 Subject: add support for transformers==4.25.1 add fallback for when quick model creation fails --- modules/sd_disable_initialization.py | 42 ++++++++++++++++++++++++++++++------ modules/sd_models.py | 8 +++++-- 2 files changed, 42 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/sd_disable_initialization.py b/modules/sd_disable_initialization.py index 9942bd7e..088ac24b 100644 --- a/modules/sd_disable_initialization.py +++ b/modules/sd_disable_initialization.py @@ -30,30 +30,53 @@ class DisableInitialization: def CLIPTextModel_from_pretrained(pretrained_model_name_or_path, *model_args, **kwargs): return self.CLIPTextModel_from_pretrained(None, *model_args, config=pretrained_model_name_or_path, state_dict={}, **kwargs) - def transformers_utils_hub_get_from_cache(url, *args, local_files_only=False, **kwargs): + def transformers_modeling_utils_load_pretrained_model(*args, **kwargs): + args = args[0:3] + ('/', ) + args[4:] # resolved_archive_file; must set it to something to prevent what seems to be a bug + return self.transformers_modeling_utils_load_pretrained_model(*args, **kwargs) + + def transformers_utils_hub_get_file_from_cache(original, url, *args, **kwargs): # this file is always 404, prevent making request if url == 'https://huggingface.co/openai/clip-vit-large-patch14/resolve/main/added_tokens.json': raise transformers.utils.hub.EntryNotFoundError try: - return self.transformers_utils_hub_get_from_cache(url, *args, local_files_only=True, **kwargs) + return original(url, *args, local_files_only=True, **kwargs) except Exception as e: - return self.transformers_utils_hub_get_from_cache(url, *args, local_files_only=False, **kwargs) + return original(url, *args, local_files_only=False, **kwargs) + + def transformers_utils_hub_get_from_cache(url, *args, local_files_only=False, **kwargs): + return transformers_utils_hub_get_file_from_cache(self.transformers_utils_hub_get_from_cache, url, *args, **kwargs) + + def transformers_tokenization_utils_base_cached_file(url, *args, local_files_only=False, **kwargs): + return transformers_utils_hub_get_file_from_cache(self.transformers_tokenization_utils_base_cached_file, url, *args, **kwargs) + + def transformers_configuration_utils_cached_file(url, *args, local_files_only=False, **kwargs): + return transformers_utils_hub_get_file_from_cache(self.transformers_configuration_utils_cached_file, url, *args, **kwargs) self.init_kaiming_uniform = torch.nn.init.kaiming_uniform_ self.init_no_grad_normal = torch.nn.init._no_grad_normal_ self.init_no_grad_uniform_ = torch.nn.init._no_grad_uniform_ self.create_model_and_transforms = open_clip.create_model_and_transforms self.CLIPTextModel_from_pretrained = ldm.modules.encoders.modules.CLIPTextModel.from_pretrained - self.transformers_utils_hub_get_from_cache = transformers.utils.hub.get_from_cache + self.transformers_modeling_utils_load_pretrained_model = getattr(transformers.modeling_utils.PreTrainedModel, '_load_pretrained_model', None) + self.transformers_tokenization_utils_base_cached_file = getattr(transformers.tokenization_utils_base, 'cached_file', None) + self.transformers_configuration_utils_cached_file = getattr(transformers.configuration_utils, 'cached_file', None) + self.transformers_utils_hub_get_from_cache = getattr(transformers.utils.hub, 'get_from_cache', None) torch.nn.init.kaiming_uniform_ = do_nothing torch.nn.init._no_grad_normal_ = do_nothing torch.nn.init._no_grad_uniform_ = do_nothing open_clip.create_model_and_transforms = create_model_and_transforms_without_pretrained ldm.modules.encoders.modules.CLIPTextModel.from_pretrained = CLIPTextModel_from_pretrained - transformers.utils.hub.get_from_cache = transformers_utils_hub_get_from_cache + if self.transformers_modeling_utils_load_pretrained_model is not None: + transformers.modeling_utils.PreTrainedModel._load_pretrained_model = transformers_modeling_utils_load_pretrained_model + if self.transformers_tokenization_utils_base_cached_file is not None: + transformers.tokenization_utils_base.cached_file = transformers_tokenization_utils_base_cached_file + if self.transformers_configuration_utils_cached_file is not None: + transformers.configuration_utils.cached_file = transformers_configuration_utils_cached_file + if self.transformers_utils_hub_get_from_cache is not None: + transformers.utils.hub.get_from_cache = transformers_utils_hub_get_from_cache def __exit__(self, exc_type, exc_val, exc_tb): torch.nn.init.kaiming_uniform_ = self.init_kaiming_uniform @@ -61,5 +84,12 @@ class DisableInitialization: torch.nn.init._no_grad_uniform_ = self.init_no_grad_uniform_ open_clip.create_model_and_transforms = self.create_model_and_transforms ldm.modules.encoders.modules.CLIPTextModel.from_pretrained = self.CLIPTextModel_from_pretrained - transformers.utils.hub.get_from_cache = self.transformers_utils_hub_get_from_cache + if self.transformers_modeling_utils_load_pretrained_model is not None: + transformers.modeling_utils.PreTrainedModel._load_pretrained_model = self.transformers_modeling_utils_load_pretrained_model + if self.transformers_tokenization_utils_base_cached_file is not None: + transformers.utils.hub.cached_file = self.transformers_tokenization_utils_base_cached_file + if self.transformers_configuration_utils_cached_file is not None: + transformers.utils.hub.cached_file = self.transformers_configuration_utils_cached_file + if self.transformers_utils_hub_get_from_cache is not None: + transformers.utils.hub.get_from_cache = self.transformers_utils_hub_get_from_cache diff --git a/modules/sd_models.py b/modules/sd_models.py index 1bb9088b..b5bc12f0 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -14,7 +14,7 @@ import ldm.modules.midas as midas from ldm.util import instantiate_from_config -from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization +from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors from modules.paths import models_path from modules.sd_hijack_inpainting import do_inpainting_hijack, should_hijack_inpainting @@ -333,7 +333,11 @@ def load_model(checkpoint_info=None): timer = Timer() - with sd_disable_initialization.DisableInitialization(): + try: + with sd_disable_initialization.DisableInitialization(): + sd_model = instantiate_from_config(sd_config.model) + except Exception as e: + print('Failed to create model quickly; will retry using slow method.', file=sys.stderr) sd_model = instantiate_from_config(sd_config.model) elapsed_create = timer.elapsed() -- cgit v1.2.3 From 29fb5327640465fc83111e2170c5d8aa2b15266c Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 10 Jan 2023 23:47:02 +0300 Subject: change color selector in settings to be part of form --- modules/shared.py | 4 ++-- modules/ui_components.py | 6 ++++++ 2 files changed, 8 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index aa37c8ce..264264a6 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -14,7 +14,7 @@ import modules.interrogate import modules.memmon import modules.styles import modules.devices as devices -from modules import localization, sd_vae, extensions, script_loading, errors +from modules import localization, sd_vae, extensions, script_loading, errors, ui_components from modules.paths import models_path, script_path, sd_path @@ -387,7 +387,7 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "initial_noise_multiplier": OptionInfo(1.0, "Noise multiplier for img2img", gr.Slider, {"minimum": 0.5, "maximum": 1.5, "step": 0.01 }), "img2img_color_correction": OptionInfo(False, "Apply color correction to img2img results to match original colors."), "img2img_fix_steps": OptionInfo(False, "With img2img, do exactly the amount of steps the slider specifies (normally you'd do less with less denoising)."), - "img2img_background_color": OptionInfo("#ffffff", "With img2img, fill image's transparent parts with this color.", gr.ColorPicker, {}), + "img2img_background_color": OptionInfo("#ffffff", "With img2img, fill image's transparent parts with this color.", ui_components.FormColorPicker, {}), "enable_quantization": OptionInfo(False, "Enable quantization in K samplers for sharper and cleaner results. This may change existing seeds. Requires restart to apply."), "enable_emphasis": OptionInfo(True, "Emphasis: use (text) to make model pay more attention to text and [text] to make it pay less attention"), "enable_batch_seeds": OptionInfo(True, "Make K-diffusion samplers produce same images in a batch as when making a single image"), diff --git a/modules/ui_components.py b/modules/ui_components.py index cac001dc..97acff06 100644 --- a/modules/ui_components.py +++ b/modules/ui_components.py @@ -31,3 +31,9 @@ class FormHTML(gr.HTML, gr.components.FormComponent): def get_block_name(self): return "html" + +class FormColorPicker(gr.ColorPicker, gr.components.FormComponent): + """Same as gr.ColorPicker but fits inside gradio forms""" + + def get_block_name(self): + return "colorpicker" -- cgit v1.2.3 From 6be644fa04ce1542f3a01804310cbbc0a4a91620 Mon Sep 17 00:00:00 2001 From: dan Date: Wed, 11 Jan 2023 05:31:58 +0800 Subject: Enable batch_size>1 for mixed-sized training --- modules/textual_inversion/dataset.py | 36 ++++++++++++++++++++++++++++++++---- 1 file changed, 32 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py index fa48708e..b47414f3 100644 --- a/modules/textual_inversion/dataset.py +++ b/modules/textual_inversion/dataset.py @@ -3,8 +3,10 @@ import numpy as np import PIL import torch from PIL import Image -from torch.utils.data import Dataset, DataLoader +from torch.utils.data import Dataset, DataLoader, Sampler from torchvision import transforms +from collections import defaultdict +from random import shuffle, choices import random import tqdm @@ -45,12 +47,12 @@ class PersonalizedBase(Dataset): assert data_root, 'dataset directory not specified' assert os.path.isdir(data_root), "Dataset directory doesn't exist" assert os.listdir(data_root), "Dataset directory is empty" - assert batch_size == 1 or not varsize, 'variable img size must have batch size 1' self.image_paths = [os.path.join(data_root, file_path) for file_path in os.listdir(data_root)] self.shuffle_tags = shuffle_tags self.tag_drop_out = tag_drop_out + groups = defaultdict(list) print("Preparing dataset...") for path in tqdm.tqdm(self.image_paths): @@ -103,13 +105,14 @@ class PersonalizedBase(Dataset): if include_cond and not (self.tag_drop_out != 0 or self.shuffle_tags): with devices.autocast(): entry.cond = cond_model([entry.cond_text]).to(devices.cpu).squeeze(0) - + groups[image.size].append(len(self.dataset)) self.dataset.append(entry) del torchdata del latent_dist del latent_sample self.length = len(self.dataset) + self.groups = list(groups.values()) assert self.length > 0, "No images have been found in the dataset." self.batch_size = min(batch_size, self.length) self.gradient_step = min(gradient_step, self.length // self.batch_size) @@ -137,9 +140,34 @@ class PersonalizedBase(Dataset): entry.latent_sample = shared.sd_model.get_first_stage_encoding(entry.latent_dist).to(devices.cpu) return entry +class GroupedBatchSampler(Sampler): + def __init__(self, data_source: PersonalizedBase, batch_size: int): + n = len(data_source) + self.groups = data_source.groups + self.len = n_batch = n // batch_size + expected = [len(g) / n * n_batch * batch_size for g in data_source.groups] + self.base = [int(e) // batch_size for e in expected] + self.n_rand_batches = nrb = n_batch - sum(self.base) + self.probs = [e%batch_size/nrb/batch_size if nrb>0 else 0 for e in expected] + self.batch_size = batch_size + def __len__(self): + return self.len + def __iter__(self): + b = self.batch_size + for g in self.groups: + shuffle(g) + batches = [] + for g in self.groups: + batches.extend(g[i*b:(i+1)*b] for i in range(len(g) // b)) + for _ in range(self.n_rand_batches): + rand_group = choices(self.groups, self.probs)[0] + batches.append(choices(rand_group, k=b)) + shuffle(batches) + yield from batches + class PersonalizedDataLoader(DataLoader): def __init__(self, dataset, latent_sampling_method="once", batch_size=1, pin_memory=False): - super(PersonalizedDataLoader, self).__init__(dataset, shuffle=True, drop_last=True, batch_size=batch_size, pin_memory=pin_memory) + super(PersonalizedDataLoader, self).__init__(dataset, batch_sampler=GroupedBatchSampler(dataset, batch_size), pin_memory=pin_memory) if latent_sampling_method == "random": self.collate_fn = collate_wrapper_random else: -- cgit v1.2.3 From f9706acf431f77e0ce9e4270e5be7299922ee963 Mon Sep 17 00:00:00 2001 From: Lee Bousfield Date: Tue, 10 Jan 2023 18:40:34 -0700 Subject: Support loading textual inversion embeddings from safetensors files --- modules/textual_inversion/textual_inversion.py | 3 +++ 1 file changed, 3 insertions(+) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 5420903f..3866c154 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -9,6 +9,7 @@ import tqdm import html import datetime import csv +import safetensors.torch from PIL import Image, PngImagePlugin @@ -150,6 +151,8 @@ class EmbeddingDatabase: name = data.get('name', name) elif ext in ['.BIN', '.PT']: data = torch.load(path, map_location="cpu") + elif ext in ['.SAFETENSORS']: + data = safetensors.torch.load_file(path, device="cpu") else: return -- cgit v1.2.3 From 5830095b73515fc49b3fd567048470005191ec34 Mon Sep 17 00:00:00 2001 From: catboxanon <122327233+catboxanon@users.noreply.github.com> Date: Tue, 10 Jan 2023 21:43:24 -0500 Subject: Add old prompt parser compat option --- modules/shared.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 264264a6..b61bbd3f 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -400,6 +400,7 @@ options_templates.update(options_section(('compatibility', "Compatibility"), { "use_old_emphasis_implementation": OptionInfo(False, "Use old emphasis implementation. Can be useful to reproduce old seeds."), "use_old_karras_scheduler_sigmas": OptionInfo(False, "Use old karras scheduler sigmas (0.1 to 10)."), "use_old_hires_fix_width_height": OptionInfo(False, "For hires fix, use width/height sliders to set final resolution rather than first pass (disables Upscale by, Resize width/height to)."), + "use_old_prompt_parser_default_step_transformer": OptionInfo(False, "Use old prompt parser default step transformer. In particular, alternating words that contained emphasis were not parsed correctly. Useful to reproduce old seeds."), })) options_templates.update(options_section(('interrogate', "Interrogate Options"), { -- cgit v1.2.3 From 7e45fba55b24166501033a221e6268545fa47fbe Mon Sep 17 00:00:00 2001 From: catboxanon <122327233+catboxanon@users.noreply.github.com> Date: Tue, 10 Jan 2023 21:47:03 -0500 Subject: Fix prompt parser default step transformer w/ test --- modules/prompt_parser.py | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/prompt_parser.py b/modules/prompt_parser.py index f70872c4..b69f1425 100644 --- a/modules/prompt_parser.py +++ b/modules/prompt_parser.py @@ -3,6 +3,11 @@ from collections import namedtuple from typing import List import lark +try: + from modules.shared import opts +except: + pass + # a prompt like this: "fantasy landscape with a [mountain:lake:0.25] and [an oak:a christmas tree:0.75][ in foreground::0.6][ in background:0.25] [shoddy:masterful:0.5]" # will be represented with prompt_schedule like this (assuming steps=100): # [25, 'fantasy landscape with a mountain and an oak in foreground shoddy'] @@ -49,6 +54,8 @@ def get_learned_conditioning_prompt_schedules(prompts, steps): [[5, 'a c'], [10, 'a {b|d{ c']] >>> g("((a][:b:c [d:3]") [[3, '((a][:b:c '], [10, '((a][:b:c d']] + >>> g("[a|(b:1.1)]") + [[1, 'a'], [2, '(b:1.1)'], [3, 'a'], [4, '(b:1.1)'], [5, 'a'], [6, '(b:1.1)'], [7, 'a'], [8, '(b:1.1)'], [9, 'a'], [10, '(b:1.1)']] """ def collect_steps(steps, tree): @@ -84,7 +91,13 @@ def get_learned_conditioning_prompt_schedules(prompts, steps): yield args[0].value def __default__(self, data, children, meta): for child in children: - yield from child + try: + if opts.use_old_prompt_parser_default_step_transformer: + yield from child + else: + yield child + except: + yield child return AtStep().transform(tree) def get_schedule(prompt): -- cgit v1.2.3 From 37a230112198adcb3f24d59b399cff342a6d479e Mon Sep 17 00:00:00 2001 From: space-nuko <24979496+space-nuko@users.noreply.github.com> Date: Tue, 10 Jan 2023 20:30:09 -0800 Subject: Expose the compiled class module of scripts to extensions --- modules/scripts.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/scripts.py b/modules/scripts.py index 35164093..4ffc369b 100644 --- a/modules/scripts.py +++ b/modules/scripts.py @@ -152,7 +152,7 @@ def basedir(): scripts_data = [] ScriptFile = namedtuple("ScriptFile", ["basedir", "filename", "path"]) -ScriptClassData = namedtuple("ScriptClassData", ["script_class", "path", "basedir"]) +ScriptClassData = namedtuple("ScriptClassData", ["script_class", "path", "basedir", "module"]) def list_scripts(scriptdirname, extension): @@ -206,7 +206,7 @@ def load_scripts(): for key, script_class in module.__dict__.items(): if type(script_class) == type and issubclass(script_class, Script): - scripts_data.append(ScriptClassData(script_class, scriptfile.path, scriptfile.basedir)) + scripts_data.append(ScriptClassData(script_class, scriptfile.path, scriptfile.basedir, module)) except Exception: print(f"Error loading script: {scriptfile.filename}", file=sys.stderr) @@ -241,7 +241,7 @@ class ScriptRunner: self.alwayson_scripts.clear() self.selectable_scripts.clear() - for script_class, path, basedir in scripts_data: + for script_class, path, basedir, script_module in scripts_data: script = script_class() script.filename = path script.is_txt2img = not is_img2img -- cgit v1.2.3 From 954091697fce7a1b7997d5f3d73551f793f6bebc Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 11 Jan 2023 09:10:07 +0300 Subject: add an option to copy config from one of models in checkpoint merger --- modules/extras.py | 30 +++++++++++++++++++++++++++++- modules/ui.py | 9 ++++++--- 2 files changed, 35 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 7407bfe3..a03d558e 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -3,6 +3,7 @@ import math import os import sys import traceback +import shutil import numpy as np from PIL import Image @@ -248,7 +249,32 @@ def run_pnginfo(image): return '', geninfo, info -def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format): +def create_config(ckpt_result, config_source, a, b, c): + def config(x): + return sd_models.find_checkpoint_config(x) if x else None + + if config_source == 0: + cfg = config(a) or config(b) or config(c) + elif config_source == 1: + cfg = config(b) + elif config_source == 2: + cfg = config(c) + else: + cfg = None + + if cfg is None: + return + + filename, _ = os.path.splitext(ckpt_result) + checkpoint_filename = filename + ".yaml" + + print("Copying config:") + print(" from:", cfg) + print(" to:", checkpoint_filename) + shutil.copyfile(cfg, checkpoint_filename) + + +def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source): shared.state.begin() shared.state.job = 'model-merge' @@ -356,6 +382,8 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam sd_models.list_models() + create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) + print("Checkpoint saved.") shared.state.textinfo = "Checkpoint saved to " + output_modelname shared.state.end() diff --git a/modules/ui.py b/modules/ui.py index 3c458ce8..82f5dd7c 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1129,7 +1129,7 @@ def create_ui(): with gr.Column(variant='panel'): gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") - with gr.Row(): + with FormRow(): primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") @@ -1143,11 +1143,13 @@ def create_ui(): interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") - with gr.Row(): + with FormRow(): checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") - modelmerger_merge = gr.Button(elem_id="modelmerger_merge", label="Merge", variant='primary') + config_source = gr.Radio(choices=["A, B or C", "B", "C", "Don't"], value="A, B or C", label="Copy config from", type="index", elem_id="modelmerger_config_method") + + modelmerger_merge = gr.Button(elem_id="modelmerger_merge", value="Merge", variant='primary') with gr.Column(variant='panel'): submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) @@ -1703,6 +1705,7 @@ def create_ui(): save_as_half, custom_name, checkpoint_format, + config_source, ], outputs=[ submit_result, -- cgit v1.2.3 From 4fdacd31e48c6a7a35c1c25c559932585e8addde Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 11 Jan 2023 10:24:56 +0300 Subject: possible fix for fallback for fast model creation from config --- modules/sd_models.py | 3 +++ 1 file changed, 3 insertions(+) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index b5bc12f0..a0a8a909 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -337,6 +337,9 @@ def load_model(checkpoint_info=None): with sd_disable_initialization.DisableInitialization(): sd_model = instantiate_from_config(sd_config.model) except Exception as e: + pass + + if sd_model is None: print('Failed to create model quickly; will retry using slow method.', file=sys.stderr) sd_model = instantiate_from_config(sd_config.model) -- cgit v1.2.3 From 1a23dc32ac5e16fac10115cafd0b841abd06e59f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 11 Jan 2023 10:34:36 +0300 Subject: possible fix for fallback for fast model creation from config, attempt 2 --- modules/sd_models.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index a0a8a909..084ba7fa 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -333,6 +333,7 @@ def load_model(checkpoint_info=None): timer = Timer() + sd_model = None try: with sd_disable_initialization.DisableInitialization(): sd_model = instantiate_from_config(sd_config.model) -- cgit v1.2.3 From ab388d6f8bf51338de1950b3907c324b0ff6a872 Mon Sep 17 00:00:00 2001 From: catboxanon <122327233+catboxanon@users.noreply.github.com> Date: Wed, 11 Jan 2023 08:59:47 -0500 Subject: Remove compat option check for prompt parser --- modules/prompt_parser.py | 13 +------------ 1 file changed, 1 insertion(+), 12 deletions(-) (limited to 'modules') diff --git a/modules/prompt_parser.py b/modules/prompt_parser.py index b69f1425..870218db 100644 --- a/modules/prompt_parser.py +++ b/modules/prompt_parser.py @@ -3,11 +3,6 @@ from collections import namedtuple from typing import List import lark -try: - from modules.shared import opts -except: - pass - # a prompt like this: "fantasy landscape with a [mountain:lake:0.25] and [an oak:a christmas tree:0.75][ in foreground::0.6][ in background:0.25] [shoddy:masterful:0.5]" # will be represented with prompt_schedule like this (assuming steps=100): # [25, 'fantasy landscape with a mountain and an oak in foreground shoddy'] @@ -91,13 +86,7 @@ def get_learned_conditioning_prompt_schedules(prompts, steps): yield args[0].value def __default__(self, data, children, meta): for child in children: - try: - if opts.use_old_prompt_parser_default_step_transformer: - yield from child - else: - yield child - except: - yield child + yield child return AtStep().transform(tree) def get_schedule(prompt): -- cgit v1.2.3 From 0b38b72d31ead82c7d0998a29e50da90073831f7 Mon Sep 17 00:00:00 2001 From: catboxanon <122327233+catboxanon@users.noreply.github.com> Date: Wed, 11 Jan 2023 09:01:37 -0500 Subject: Remove compat option for prompt parser --- modules/shared.py | 1 - 1 file changed, 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index b61bbd3f..264264a6 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -400,7 +400,6 @@ options_templates.update(options_section(('compatibility', "Compatibility"), { "use_old_emphasis_implementation": OptionInfo(False, "Use old emphasis implementation. Can be useful to reproduce old seeds."), "use_old_karras_scheduler_sigmas": OptionInfo(False, "Use old karras scheduler sigmas (0.1 to 10)."), "use_old_hires_fix_width_height": OptionInfo(False, "For hires fix, use width/height sliders to set final resolution rather than first pass (disables Upscale by, Resize width/height to)."), - "use_old_prompt_parser_default_step_transformer": OptionInfo(False, "Use old prompt parser default step transformer. In particular, alternating words that contained emphasis were not parsed correctly. Useful to reproduce old seeds."), })) options_templates.update(options_section(('interrogate', "Interrogate Options"), { -- cgit v1.2.3 From 39ea251945d70efcf9b59d44eb0e71269d754aa4 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 11 Jan 2023 10:23:51 -0500 Subject: add textinfo to progress response --- modules/api/api.py | 4 ++-- modules/api/models.py | 1 + 2 files changed, 3 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 6c564ad8..5767ba90 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -286,7 +286,7 @@ class Api: # copy from check_progress_call of ui.py if shared.state.job_count == 0: - return ProgressResponse(progress=0, eta_relative=0, state=shared.state.dict()) + return ProgressResponse(progress=0, eta_relative=0, state=shared.state.dict(), textinfo=shared.state.textinfo) # avoid dividing zero progress = 0.01 @@ -308,7 +308,7 @@ class Api: if shared.state.current_image and not req.skip_current_image: current_image = encode_pil_to_base64(shared.state.current_image) - return ProgressResponse(progress=progress, eta_relative=eta_relative, state=shared.state.dict(), current_image=current_image) + return ProgressResponse(progress=progress, eta_relative=eta_relative, state=shared.state.dict(), current_image=current_image, textinfo=shared.state.textinfo) def interrogateapi(self, interrogatereq: InterrogateRequest): image_b64 = interrogatereq.image diff --git a/modules/api/models.py b/modules/api/models.py index 034b4aa0..c78095ca 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -168,6 +168,7 @@ class ProgressResponse(BaseModel): eta_relative: float = Field(title="ETA in secs") state: dict = Field(title="State", description="The current state snapshot") current_image: str = Field(default=None, title="Current image", description="The current image in base64 format. opts.show_progress_every_n_steps is required for this to work.") + textinfo: str = Field(default=None, title="Info text", description="Info text used by WebUI.") class InterrogateRequest(BaseModel): image: str = Field(default="", title="Image", description="Image to work on, must be a Base64 string containing the image's data.") -- cgit v1.2.3 From 3f43d8a966ba8462ba019a5ad573f94508cd45f8 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 11 Jan 2023 10:28:55 -0500 Subject: set descriptions --- modules/hypernetworks/hypernetwork.py | 4 +++- modules/textual_inversion/preprocess.py | 7 ++++++- modules/textual_inversion/textual_inversion.py | 4 +++- 3 files changed, 12 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 300d3975..194679e8 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -619,7 +619,9 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, epoch_num = hypernetwork.step // steps_per_epoch epoch_step = hypernetwork.step % steps_per_epoch - pbar.set_description(f"[Epoch {epoch_num}: {epoch_step+1}/{steps_per_epoch}]loss: {loss_step:.7f}") + description = f"Training hypernetwork [Epoch {epoch_num}: {epoch_step+1}/{steps_per_epoch}]loss: {loss_step:.7f}" + pbar.set_description(description) + shared.state.textinfo = description if hypernetwork_dir is not None and steps_done % save_hypernetwork_every == 0: # Before saving, change name to match current checkpoint. hypernetwork_name_every = f'{hypernetwork_name}-{steps_done}' diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py index feb876c6..3c1042ad 100644 --- a/modules/textual_inversion/preprocess.py +++ b/modules/textual_inversion/preprocess.py @@ -135,7 +135,8 @@ def preprocess_work(process_src, process_dst, process_width, process_height, pre params.process_caption_deepbooru = process_caption_deepbooru params.preprocess_txt_action = preprocess_txt_action - for index, imagefile in enumerate(tqdm.tqdm(files)): + pbar = tqdm.tqdm(files) + for index, imagefile in enumerate(pbar): params.subindex = 0 filename = os.path.join(src, imagefile) try: @@ -143,6 +144,10 @@ def preprocess_work(process_src, process_dst, process_width, process_height, pre except Exception: continue + description = f"Preprocessing [Image {index}/{len(files)}]" + pbar.set_description(description) + shared.state.textinfo = description + params.src = filename existing_caption = None diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 3866c154..b915b091 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -476,7 +476,9 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ epoch_num = embedding.step // steps_per_epoch epoch_step = embedding.step % steps_per_epoch - pbar.set_description(f"[Epoch {epoch_num}: {epoch_step+1}/{steps_per_epoch}]loss: {loss_step:.7f}") + description = f"Training textual inversion [Epoch {epoch_num}: {epoch_step+1}/{steps_per_epoch}]loss: {loss_step:.7f}" + pbar.set_description(description) + shared.state.textinfo = description if embedding_dir is not None and steps_done % save_embedding_every == 0: # Before saving, change name to match current checkpoint. embedding_name_every = f'{embedding_name}-{steps_done}' -- cgit v1.2.3 From 4bd490727e156ff53107d53416d6b89be86f2a62 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 11 Jan 2023 18:54:04 +0300 Subject: fix for an error caused by skipping initialization, for realsies this time: TypeError: expected str, bytes or os.PathLike object, not NoneType --- modules/sd_disable_initialization.py | 71 ++++++++++++++++-------------------- modules/sd_models.py | 1 + 2 files changed, 33 insertions(+), 39 deletions(-) (limited to 'modules') diff --git a/modules/sd_disable_initialization.py b/modules/sd_disable_initialization.py index 088ac24b..c72d8efc 100644 --- a/modules/sd_disable_initialization.py +++ b/modules/sd_disable_initialization.py @@ -20,6 +20,19 @@ class DisableInitialization: ``` """ + def __init__(self): + self.replaced = [] + + def replace(self, obj, field, func): + original = getattr(obj, field, None) + if original is None: + return None + + self.replaced.append((obj, field, original)) + setattr(obj, field, func) + + return original + def __enter__(self): def do_nothing(*args, **kwargs): pass @@ -37,11 +50,14 @@ class DisableInitialization: def transformers_utils_hub_get_file_from_cache(original, url, *args, **kwargs): # this file is always 404, prevent making request - if url == 'https://huggingface.co/openai/clip-vit-large-patch14/resolve/main/added_tokens.json': - raise transformers.utils.hub.EntryNotFoundError + if url == 'https://huggingface.co/openai/clip-vit-large-patch14/resolve/main/added_tokens.json' or url == 'openai/clip-vit-large-patch14' and args[0] == 'added_tokens.json': + return None try: - return original(url, *args, local_files_only=True, **kwargs) + res = original(url, *args, local_files_only=True, **kwargs) + if res is None: + res = original(url, *args, local_files_only=False, **kwargs) + return res except Exception as e: return original(url, *args, local_files_only=False, **kwargs) @@ -54,42 +70,19 @@ class DisableInitialization: def transformers_configuration_utils_cached_file(url, *args, local_files_only=False, **kwargs): return transformers_utils_hub_get_file_from_cache(self.transformers_configuration_utils_cached_file, url, *args, **kwargs) - self.init_kaiming_uniform = torch.nn.init.kaiming_uniform_ - self.init_no_grad_normal = torch.nn.init._no_grad_normal_ - self.init_no_grad_uniform_ = torch.nn.init._no_grad_uniform_ - self.create_model_and_transforms = open_clip.create_model_and_transforms - self.CLIPTextModel_from_pretrained = ldm.modules.encoders.modules.CLIPTextModel.from_pretrained - self.transformers_modeling_utils_load_pretrained_model = getattr(transformers.modeling_utils.PreTrainedModel, '_load_pretrained_model', None) - self.transformers_tokenization_utils_base_cached_file = getattr(transformers.tokenization_utils_base, 'cached_file', None) - self.transformers_configuration_utils_cached_file = getattr(transformers.configuration_utils, 'cached_file', None) - self.transformers_utils_hub_get_from_cache = getattr(transformers.utils.hub, 'get_from_cache', None) - - torch.nn.init.kaiming_uniform_ = do_nothing - torch.nn.init._no_grad_normal_ = do_nothing - torch.nn.init._no_grad_uniform_ = do_nothing - open_clip.create_model_and_transforms = create_model_and_transforms_without_pretrained - ldm.modules.encoders.modules.CLIPTextModel.from_pretrained = CLIPTextModel_from_pretrained - if self.transformers_modeling_utils_load_pretrained_model is not None: - transformers.modeling_utils.PreTrainedModel._load_pretrained_model = transformers_modeling_utils_load_pretrained_model - if self.transformers_tokenization_utils_base_cached_file is not None: - transformers.tokenization_utils_base.cached_file = transformers_tokenization_utils_base_cached_file - if self.transformers_configuration_utils_cached_file is not None: - transformers.configuration_utils.cached_file = transformers_configuration_utils_cached_file - if self.transformers_utils_hub_get_from_cache is not None: - transformers.utils.hub.get_from_cache = transformers_utils_hub_get_from_cache + self.replace(torch.nn.init, 'kaiming_uniform_', do_nothing) + self.replace(torch.nn.init, '_no_grad_normal_', do_nothing) + self.replace(torch.nn.init, '_no_grad_uniform_', do_nothing) + self.create_model_and_transforms = self.replace(open_clip, 'create_model_and_transforms', create_model_and_transforms_without_pretrained) + self.CLIPTextModel_from_pretrained = self.replace(ldm.modules.encoders.modules.CLIPTextModel, 'from_pretrained', CLIPTextModel_from_pretrained) + self.transformers_modeling_utils_load_pretrained_model = self.replace(transformers.modeling_utils.PreTrainedModel, '_load_pretrained_model', transformers_modeling_utils_load_pretrained_model) + self.transformers_tokenization_utils_base_cached_file = self.replace(transformers.tokenization_utils_base, 'cached_file', transformers_tokenization_utils_base_cached_file) + self.transformers_configuration_utils_cached_file = self.replace(transformers.configuration_utils, 'cached_file', transformers_configuration_utils_cached_file) + self.transformers_utils_hub_get_from_cache = self.replace(transformers.utils.hub, 'get_from_cache', transformers_utils_hub_get_from_cache) def __exit__(self, exc_type, exc_val, exc_tb): - torch.nn.init.kaiming_uniform_ = self.init_kaiming_uniform - torch.nn.init._no_grad_normal_ = self.init_no_grad_normal - torch.nn.init._no_grad_uniform_ = self.init_no_grad_uniform_ - open_clip.create_model_and_transforms = self.create_model_and_transforms - ldm.modules.encoders.modules.CLIPTextModel.from_pretrained = self.CLIPTextModel_from_pretrained - if self.transformers_modeling_utils_load_pretrained_model is not None: - transformers.modeling_utils.PreTrainedModel._load_pretrained_model = self.transformers_modeling_utils_load_pretrained_model - if self.transformers_tokenization_utils_base_cached_file is not None: - transformers.utils.hub.cached_file = self.transformers_tokenization_utils_base_cached_file - if self.transformers_configuration_utils_cached_file is not None: - transformers.utils.hub.cached_file = self.transformers_configuration_utils_cached_file - if self.transformers_utils_hub_get_from_cache is not None: - transformers.utils.hub.get_from_cache = self.transformers_utils_hub_get_from_cache + for obj, field, original in self.replaced: + setattr(obj, field, original) + + self.replaced.clear() diff --git a/modules/sd_models.py b/modules/sd_models.py index 084ba7fa..c466f273 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -334,6 +334,7 @@ def load_model(checkpoint_info=None): timer = Timer() sd_model = None + try: with sd_disable_initialization.DisableInitialization(): sd_model = instantiate_from_config(sd_config.model) -- cgit v1.2.3 From 0b8911d883118daa54f7735c5b753b5575d9f943 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 11 Jan 2023 20:33:24 +0300 Subject: img2img UI rework: obsolete --gradio-img2img-tool --gradio-inpaint-tool and always show all tools each in own tab --- modules/img2img.py | 58 ++++++++++++++---------------- modules/shared.py | 4 +-- modules/ui.py | 103 +++++++++++++++++++++++++++-------------------------- 3 files changed, 81 insertions(+), 84 deletions(-) (limited to 'modules') diff --git a/modules/img2img.py b/modules/img2img.py index ca58b5d8..f62783c6 100644 --- a/modules/img2img.py +++ b/modules/img2img.py @@ -59,38 +59,34 @@ def process_batch(p, input_dir, output_dir, args): processed_image.save(os.path.join(output_dir, filename)) -def img2img(mode: int, prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, init_img, init_img_with_mask, init_img_with_mask_orig, init_img_inpaint, init_mask_inpaint, mask_mode, steps: int, sampler_index: int, mask_blur: int, mask_alpha: float, inpainting_fill: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, denoising_strength: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, resize_mode: int, inpaint_full_res: bool, inpaint_full_res_padding: int, inpainting_mask_invert: int, img2img_batch_input_dir: str, img2img_batch_output_dir: str, *args): - is_inpaint = mode == 1 - is_batch = mode == 2 - - if is_inpaint: - # Drawn mask - if mask_mode == 0: - is_mask_sketch = isinstance(init_img_with_mask, dict) - is_mask_paint = not is_mask_sketch - if is_mask_sketch: - # Sketch: mask iff. not transparent - image, mask = init_img_with_mask["image"], init_img_with_mask["mask"] - alpha_mask = ImageOps.invert(image.split()[-1]).convert('L').point(lambda x: 255 if x > 0 else 0, mode='1') - mask = ImageChops.lighter(alpha_mask, mask.convert('L')).convert('L') - else: - # Color-sketch: mask iff. painted over - image = init_img_with_mask - orig = init_img_with_mask_orig or init_img_with_mask - pred = np.any(np.array(image) != np.array(orig), axis=-1) - mask = Image.fromarray(pred.astype(np.uint8) * 255, "L") - mask = ImageEnhance.Brightness(mask).enhance(1 - mask_alpha / 100) - blur = ImageFilter.GaussianBlur(mask_blur) - image = Image.composite(image.filter(blur), orig, mask.filter(blur)) - - image = image.convert("RGB") - # Uploaded mask - else: - image = init_img_inpaint - mask = init_mask_inpaint - # No mask +def img2img(mode: int, prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, init_img, sketch, init_img_with_mask, inpaint_color_sketch, inpaint_color_sketch_orig, init_img_inpaint, init_mask_inpaint, steps: int, sampler_index: int, mask_blur: int, mask_alpha: float, inpainting_fill: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, denoising_strength: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, resize_mode: int, inpaint_full_res: bool, inpaint_full_res_padding: int, inpainting_mask_invert: int, img2img_batch_input_dir: str, img2img_batch_output_dir: str, *args): + is_batch = mode == 5 + + if mode == 0: # img2img + image = init_img.convert("RGB") + mask = None + elif mode == 1: # img2img sketch + image = sketch.convert("RGB") + mask = None + elif mode == 2: # inpaint + image, mask = init_img_with_mask["image"], init_img_with_mask["mask"] + alpha_mask = ImageOps.invert(image.split()[-1]).convert('L').point(lambda x: 255 if x > 0 else 0, mode='1') + mask = ImageChops.lighter(alpha_mask, mask.convert('L')).convert('L') + image = image.convert("RGB") + elif mode == 3: # inpaint sketch + image = inpaint_color_sketch + orig = inpaint_color_sketch_orig or inpaint_color_sketch + pred = np.any(np.array(image) != np.array(orig), axis=-1) + mask = Image.fromarray(pred.astype(np.uint8) * 255, "L") + mask = ImageEnhance.Brightness(mask).enhance(1 - mask_alpha / 100) + blur = ImageFilter.GaussianBlur(mask_blur) + image = Image.composite(image.filter(blur), orig, mask.filter(blur)) + image = image.convert("RGB") + elif mode == 4: # inpaint upload mask + image = init_img_inpaint + mask = init_mask_inpaint else: - image = init_img + image = None mask = None # Use the EXIF orientation of photos taken by smartphones. diff --git a/modules/shared.py b/modules/shared.py index 264264a6..1c964237 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -74,8 +74,8 @@ parser.add_argument("--freeze-settings", action='store_true', help="disable edit parser.add_argument("--ui-settings-file", type=str, help="filename to use for ui settings", default=os.path.join(script_path, 'config.json')) parser.add_argument("--gradio-debug", action='store_true', help="launch gradio with --debug option") parser.add_argument("--gradio-auth", type=str, help='set gradio authentication like "username:password"; or comma-delimit multiple like "u1:p1,u2:p2,u3:p3"', default=None) -parser.add_argument("--gradio-img2img-tool", type=str, help='gradio image uploader tool: can be either editor for ctopping, or color-sketch for drawing', choices=["color-sketch", "editor"], default="editor") -parser.add_argument("--gradio-inpaint-tool", type=str, choices=["sketch", "color-sketch"], default="sketch", help="gradio inpainting editor: can be either sketch to only blur/noise the input, or color-sketch to paint over it") +parser.add_argument("--gradio-img2img-tool", type=str, help='does not do anything') +parser.add_argument("--gradio-inpaint-tool", type=str, help="does not do anything") parser.add_argument("--opt-channelslast", action='store_true', help="change memory type for stable diffusion to channels last") parser.add_argument("--styles-file", type=str, help="filename to use for styles", default=os.path.join(script_path, 'styles.csv')) parser.add_argument("--autolaunch", action='store_true', help="open the webui URL in the system's default browser upon launch", default=False) diff --git a/modules/ui.py b/modules/ui.py index 82f5dd7c..e86a624b 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -795,53 +795,67 @@ def create_ui(): with FormRow().style(equal_height=False): with gr.Column(variant='panel', elem_id="img2img_settings"): + with gr.Tabs(elem_id="mode_img2img"): + with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab") as tab_img2img: + init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool="editor", image_mode="RGBA").style(height=480) - with gr.Tabs(elem_id="mode_img2img") as tabs_img2img_mode: - with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab"): - init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_img2img_tool, image_mode="RGBA").style(height=480) + with gr.TabItem('Sketch', id='img2img_sketch', elem_id="img2img_img2img_sketch_tab") as tab_sketch: + sketch = gr.Image(label="Image for img2img", elem_id="img2img_sketch", show_label=False, source="upload", interactive=True, type="pil", tool="color-sketch", image_mode="RGBA").style(height=480) - with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab"): - init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool=cmd_opts.gradio_inpaint_tool, image_mode="RGBA").style(height=480) - init_img_with_mask_orig = gr.State(None) + with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab") as tab_inpaint: + init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool="sketch", image_mode="RGBA").style(height=480) - use_color_sketch = cmd_opts.gradio_inpaint_tool == "color-sketch" - if use_color_sketch: - def update_orig(image, state): - if image is not None: - same_size = state is not None and state.size == image.size - has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) - edited = same_size and has_exact_match - return image if not edited or state is None else state + with gr.TabItem('Inpaint sketch', id='inpaint_sketch', elem_id="img2img_inpaint_sketch_tab") as tab_inpaint_color: + inpaint_color_sketch = gr.Image(label="Color sketch inpainting", show_label=False, elem_id="inpaint_sketch", source="upload", interactive=True, type="pil", tool="color-sketch", image_mode="RGBA").style(height=480) + inpaint_color_sketch_orig = gr.State(None) - init_img_with_mask.change(update_orig, [init_img_with_mask, init_img_with_mask_orig], init_img_with_mask_orig) + def update_orig(image, state): + if image is not None: + same_size = state is not None and state.size == image.size + has_exact_match = np.any(np.all(np.array(image) == np.array(state), axis=-1)) + edited = same_size and has_exact_match + return image if not edited or state is None else state - init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_base") - init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", visible=False, elem_id="img_inpaint_mask") + inpaint_color_sketch.change(update_orig, [inpaint_color_sketch, inpaint_color_sketch_orig], inpaint_color_sketch_orig) - with FormRow(): - mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") - mask_alpha = gr.Slider(label="Mask transparency", interactive=use_color_sketch, visible=use_color_sketch, elem_id="img2img_mask_alpha") - - with FormRow(): - mask_mode = gr.Radio(label="Mask source", choices=["Draw mask", "Upload mask"], type="index", value="Draw mask", elem_id="mask_mode") - inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") - - with FormRow(): - inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") - - with FormRow(): - with gr.Column(): - inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") + with gr.TabItem('Inpaint upload', id='inpaint_upload', elem_id="img2img_inpaint_upload_tab") as tab_inpaint_upload: + init_img_inpaint = gr.Image(label="Image for img2img", show_label=False, source="upload", interactive=True, type="pil", elem_id="img_inpaint_base") + init_mask_inpaint = gr.Image(label="Mask", source="upload", interactive=True, type="pil", elem_id="img_inpaint_mask") - with gr.Column(scale=4): - inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") - - with gr.TabItem('Batch img2img', id='batch', elem_id="img2img_batch_tab"): + with gr.TabItem('Batch', id='batch', elem_id="img2img_batch_tab") as tab_batch: hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") + with FormGroup(elem_id="inpaint_controls", visible=False) as inpaint_controls: + with FormRow(): + mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") + mask_alpha = gr.Slider(label="Mask transparency", visible=False, elem_id="img2img_mask_alpha") + + with FormRow(): + inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") + + with FormRow(): + inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") + + with FormRow(): + with gr.Column(): + inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") + + with gr.Column(scale=4): + inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") + + def select_img2img_tab(tab): + return gr.update(visible=tab in [2, 3, 4]), gr.update(visible=tab == 3), + + for i, elem in enumerate([tab_img2img, tab_sketch, tab_inpaint, tab_inpaint_color, tab_inpaint_upload, tab_batch]): + elem.select( + fn=lambda tab=i: select_img2img_tab(tab), + inputs=[], + outputs=[inpaint_controls, mask_alpha], + ) + with FormRow(): resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") @@ -900,20 +914,6 @@ def create_ui(): ] ) - mask_mode.change( - lambda mode, img: { - init_img_with_mask: gr_show(mode == 0), - init_img_inpaint: gr_show(mode == 1), - init_mask_inpaint: gr_show(mode == 1), - }, - inputs=[mask_mode, init_img_with_mask], - outputs=[ - init_img_with_mask, - init_img_inpaint, - init_mask_inpaint, - ], - ) - img2img_args = dict( fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), _js="submit_img2img", @@ -924,11 +924,12 @@ def create_ui(): img2img_prompt_style, img2img_prompt_style2, init_img, + sketch, init_img_with_mask, - init_img_with_mask_orig, + inpaint_color_sketch, + inpaint_color_sketch_orig, init_img_inpaint, init_mask_inpaint, - mask_mode, steps, sampler_index, mask_blur, -- cgit v1.2.3 From d52a80f7f7da160c73afd067c8f1bf491391f994 Mon Sep 17 00:00:00 2001 From: Shondoit Date: Thu, 12 Jan 2023 09:22:29 +0100 Subject: Allow creation of zero vectors for TI --- modules/textual_inversion/textual_inversion.py | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index b915b091..853246a6 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -248,11 +248,14 @@ def create_embedding(name, num_vectors_per_token, overwrite_old, init_text='*'): with devices.autocast(): cond_model([""]) # will send cond model to GPU if lowvram/medvram is active - embedded = cond_model.encode_embedding_init_text(init_text, num_vectors_per_token) + #cond_model expects at least some text, so we provide '*' as backup. + embedded = cond_model.encode_embedding_init_text(init_text or '*', num_vectors_per_token) vec = torch.zeros((num_vectors_per_token, embedded.shape[1]), device=devices.device) - for i in range(num_vectors_per_token): - vec[i] = embedded[i * int(embedded.shape[0]) // num_vectors_per_token] + #Only copy if we provided an init_text, otherwise keep vectors as zeros + if init_text: + for i in range(num_vectors_per_token): + vec[i] = embedded[i * int(embedded.shape[0]) // num_vectors_per_token] # Remove illegal characters from name. name = "".join( x for x in name if (x.isalnum() or x in "._- ")) -- cgit v1.2.3 From 88416ab5ff787eec3b9962b43b5e544bb75fbad6 Mon Sep 17 00:00:00 2001 From: space-nuko <24979496+space-nuko@users.noreply.github.com> Date: Thu, 12 Jan 2023 13:46:59 -0800 Subject: Fix extension parameters not being saved to last used parameters --- modules/processing.py | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index f04a0e1e..ae04cab7 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -531,16 +531,16 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: def infotext(iteration=0, position_in_batch=0): return create_infotext(p, p.all_prompts, p.all_seeds, p.all_subseeds, comments, iteration, position_in_batch) - with open(os.path.join(shared.script_path, "params.txt"), "w", encoding="utf8") as file: - processed = Processed(p, [], p.seed, "") - file.write(processed.infotext(p, 0)) - if os.path.exists(cmd_opts.embeddings_dir) and not p.do_not_reload_embeddings: model_hijack.embedding_db.load_textual_inversion_embeddings() if p.scripts is not None: p.scripts.process(p) + with open(os.path.join(shared.script_path, "params.txt"), "w", encoding="utf8") as file: + processed = Processed(p, [], p.seed, "") + file.write(processed.infotext(p, 0)) + infotexts = [] output_images = [] -- cgit v1.2.3 From 6c88eaed4f5efca54a882eb1f8f30f01f350332a Mon Sep 17 00:00:00 2001 From: space-nuko <24979496+space-nuko@users.noreply.github.com> Date: Thu, 12 Jan 2023 13:50:09 -0800 Subject: Add script callback for fixing infotext parameters --- modules/generation_parameters_copypaste.py | 3 ++- modules/script_callbacks.py | 20 +++++++++++++++++++- 2 files changed, 21 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 620aa606..593d99ef 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -7,7 +7,7 @@ from pathlib import Path import gradio as gr from modules.shared import script_path -from modules import shared, ui_tempdir +from modules import shared, ui_tempdir, script_callbacks import tempfile from PIL import Image @@ -298,6 +298,7 @@ def connect_paste(button, paste_fields, input_comp, jsfunc=None): prompt = file.read() params = parse_generation_parameters(prompt) + script_callbacks.infotext_pasted_callback(prompt, params) res = [] for output, key in paste_fields: diff --git a/modules/script_callbacks.py b/modules/script_callbacks.py index 608c5300..a9e19236 100644 --- a/modules/script_callbacks.py +++ b/modules/script_callbacks.py @@ -2,7 +2,7 @@ import sys import traceback from collections import namedtuple import inspect -from typing import Optional +from typing import Optional, Dict, Any from fastapi import FastAPI from gradio import Blocks @@ -71,6 +71,7 @@ callback_map = dict( callbacks_before_component=[], callbacks_after_component=[], callbacks_image_grid=[], + callbacks_infotext_pasted=[], callbacks_script_unloaded=[], ) @@ -172,6 +173,14 @@ def image_grid_callback(params: ImageGridLoopParams): report_exception(c, 'image_grid') +def infotext_pasted_callback(infotext: str, params: Dict[str, Any]): + for c in callback_map['callbacks_infotext_pasted']: + try: + c.callback(infotext, params) + except Exception: + report_exception(c, 'infotext_pasted') + + def script_unloaded_callback(): for c in reversed(callback_map['callbacks_script_unloaded']): try: @@ -290,6 +299,15 @@ def on_image_grid(callback): add_callback(callback_map['callbacks_image_grid'], callback) +def on_infotext_pasted(callback): + """register a function to be called before applying an infotext. + The callback is called with two arguments: + - infotext: str - raw infotext. + - result: Dict[str, any] - parsed infotext parameters. + """ + add_callback(callback_map['callbacks_infotext_pasted'], callback) + + def on_script_unloaded(callback): """register a function to be called before the script is unloaded. Any hooks/hijacks/monkeying about that the script did should be reverted here""" -- cgit v1.2.3 From 0b262802b86a55c4f71faf377f2cb1aee2960b63 Mon Sep 17 00:00:00 2001 From: Josh R Date: Thu, 12 Jan 2023 17:31:05 -0800 Subject: add gradient settings to training settings log files --- modules/textual_inversion/logging.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/textual_inversion/logging.py b/modules/textual_inversion/logging.py index 8b1981d5..31e50b64 100644 --- a/modules/textual_inversion/logging.py +++ b/modules/textual_inversion/logging.py @@ -2,7 +2,7 @@ import datetime import json import os -saved_params_shared = {"model_name", "model_hash", "initial_step", "num_of_dataset_images", "learn_rate", "batch_size", "data_root", "log_directory", "training_width", "training_height", "steps", "create_image_every", "template_file"} +saved_params_shared = {"model_name", "model_hash", "initial_step", "num_of_dataset_images", "learn_rate", "batch_size", "clip_grad_mode", "clip_grad_value", "gradient_step", "data_root", "log_directory", "training_width", "training_height", "steps", "create_image_every", "template_file"} saved_params_ti = {"embedding_name", "num_vectors_per_token", "save_embedding_every", "save_image_with_stored_embedding"} saved_params_hypernet = {"hypernetwork_name", "layer_structure", "activation_func", "weight_init", "add_layer_norm", "use_dropout", "save_hypernetwork_every"} saved_params_all = saved_params_shared | saved_params_ti | saved_params_hypernet -- cgit v1.2.3 From a176d89487d92f5a5b152401e5c424b34ff43b96 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 13 Jan 2023 14:32:15 +0300 Subject: print bucket sizes for training without resizing images #6620 fix an error when generating a picture with embedding in it --- modules/textual_inversion/dataset.py | 16 ++++++++++++++++ modules/textual_inversion/image_embedding.py | 4 ++-- modules/textual_inversion/textual_inversion.py | 2 +- 3 files changed, 19 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py index b47414f3..d31963d4 100644 --- a/modules/textual_inversion/dataset.py +++ b/modules/textual_inversion/dataset.py @@ -118,6 +118,12 @@ class PersonalizedBase(Dataset): self.gradient_step = min(gradient_step, self.length // self.batch_size) self.latent_sampling_method = latent_sampling_method + if len(groups) > 1: + print("Buckets:") + for (w, h), ids in sorted(groups.items(), key=lambda x: x[0]): + print(f" {w}x{h}: {len(ids)}") + print() + def create_text(self, filename_text): text = random.choice(self.lines) tags = filename_text.split(',') @@ -140,8 +146,11 @@ class PersonalizedBase(Dataset): entry.latent_sample = shared.sd_model.get_first_stage_encoding(entry.latent_dist).to(devices.cpu) return entry + class GroupedBatchSampler(Sampler): def __init__(self, data_source: PersonalizedBase, batch_size: int): + super().__init__(data_source) + n = len(data_source) self.groups = data_source.groups self.len = n_batch = n // batch_size @@ -150,21 +159,28 @@ class GroupedBatchSampler(Sampler): self.n_rand_batches = nrb = n_batch - sum(self.base) self.probs = [e%batch_size/nrb/batch_size if nrb>0 else 0 for e in expected] self.batch_size = batch_size + def __len__(self): return self.len + def __iter__(self): b = self.batch_size + for g in self.groups: shuffle(g) + batches = [] for g in self.groups: batches.extend(g[i*b:(i+1)*b] for i in range(len(g) // b)) for _ in range(self.n_rand_batches): rand_group = choices(self.groups, self.probs)[0] batches.append(choices(rand_group, k=b)) + shuffle(batches) + yield from batches + class PersonalizedDataLoader(DataLoader): def __init__(self, dataset, latent_sampling_method="once", batch_size=1, pin_memory=False): super(PersonalizedDataLoader, self).__init__(dataset, batch_sampler=GroupedBatchSampler(dataset, batch_size), pin_memory=pin_memory) diff --git a/modules/textual_inversion/image_embedding.py b/modules/textual_inversion/image_embedding.py index ea653806..5593f88c 100644 --- a/modules/textual_inversion/image_embedding.py +++ b/modules/textual_inversion/image_embedding.py @@ -76,10 +76,10 @@ def insert_image_data_embed(image, data): next_size = data_np_low.shape[0] + (h-(data_np_low.shape[0] % h)) next_size = next_size + ((h*d)-(next_size % (h*d))) - data_np_low.resize(next_size) + data_np_low = np.resize(data_np_low, next_size) data_np_low = data_np_low.reshape((h, -1, d)) - data_np_high.resize(next_size) + data_np_high = np.resize(data_np_high, next_size) data_np_high = data_np_high.reshape((h, -1, d)) edge_style = list(data['string_to_param'].values())[0].cpu().detach().numpy().tolist()[0][:1024] diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 853246a6..e23906ca 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -479,7 +479,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ epoch_num = embedding.step // steps_per_epoch epoch_step = embedding.step % steps_per_epoch - description = f"Training textual inversion [Epoch {epoch_num}: {epoch_step+1}/{steps_per_epoch}]loss: {loss_step:.7f}" + description = f"Training textual inversion [Epoch {epoch_num}: {epoch_step+1}/{steps_per_epoch}] loss: {loss_step:.7f}" pbar.set_description(description) shared.state.textinfo = description if embedding_dir is not None and steps_done % save_embedding_every == 0: -- cgit v1.2.3 From 82725f0ac439f7e3b67858d55900e95330bbd326 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 13 Jan 2023 15:04:37 +0300 Subject: fix a bug caused by merge --- modules/textual_inversion/textual_inversion.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 85210b0e..6939efcc 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -11,6 +11,7 @@ import datetime import csv import safetensors.torch +import numpy as np from PIL import Image, PngImagePlugin from torch.utils.tensorboard import SummaryWriter -- cgit v1.2.3 From a95f1353089bdeaccd7c266b40cdd79efedfe632 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 09:56:59 +0300 Subject: change hash to sha256 --- modules/api/api.py | 2 +- modules/api/models.py | 3 +- modules/hashes.py | 72 +++++++++++++++ modules/hypernetworks/hypernetwork.py | 4 +- modules/sd_models.py | 116 ++++++++++++++++--------- modules/shared.py | 2 +- modules/textual_inversion/textual_inversion.py | 6 +- 7 files changed, 155 insertions(+), 50 deletions(-) create mode 100644 modules/hashes.py (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 5767ba90..9814bbc2 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -371,7 +371,7 @@ class Api: return upscalers def get_sd_models(self): - return [{"title":x.title, "model_name":x.model_name, "hash":x.hash, "filename": x.filename, "config": find_checkpoint_config(x)} for x in checkpoints_list.values()] + return [{"title": x.title, "model_name": x.model_name, "hash": x.shorthash, "sha256": x.sha256, "filename": x.filename, "config": find_checkpoint_config(x)} for x in checkpoints_list.values()] def get_hypernetworks(self): return [{"name": name, "path": shared.hypernetworks[name]} for name in shared.hypernetworks] diff --git a/modules/api/models.py b/modules/api/models.py index c78095ca..1eb1fcf1 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -224,7 +224,8 @@ class UpscalerItem(BaseModel): class SDModelItem(BaseModel): title: str = Field(title="Title") model_name: str = Field(title="Model Name") - hash: str = Field(title="Hash") + hash: Optional[str] = Field(title="Short hash") + sha256: Optional[str] = Field(title="sha256 hash") filename: str = Field(title="Filename") config: str = Field(title="Config file") diff --git a/modules/hashes.py b/modules/hashes.py new file mode 100644 index 00000000..ebfbd90c --- /dev/null +++ b/modules/hashes.py @@ -0,0 +1,72 @@ +import hashlib +import json +import os.path + +import filelock + + +cache_filename = "cache.json" +cache_data = None + + +def dump_cache(): + with filelock.FileLock(cache_filename+".lock"): + with open(cache_filename, "w", encoding="utf8") as file: + json.dump(cache_data, file, indent=4) + + +def cache(subsection): + global cache_data + + if cache_data is None: + with filelock.FileLock(cache_filename+".lock"): + if not os.path.isfile(cache_filename): + cache_data = {} + else: + with open(cache_filename, "r", encoding="utf8") as file: + cache_data = json.load(file) + + s = cache_data.get(subsection, {}) + cache_data[subsection] = s + + return s + + +def calculate_sha256(filename): + hash_sha256 = hashlib.sha256() + + with open(filename, "rb") as f: + for chunk in iter(lambda: f.read(4096), b""): + hash_sha256.update(chunk) + + return hash_sha256.hexdigest() + + +def sha256(filename, title): + hashes = cache("hashes") + ondisk_mtime = os.path.getmtime(filename) + + if title in hashes: + cached_sha256 = hashes[title].get("sha256", None) + cached_mtime = hashes[title].get("mtime", 0) + + if ondisk_mtime <= cached_mtime and cached_sha256 is not None: + return cached_sha256 + + print(f"Calculating sha256 for {filename}: ", end='') + sha256_value = calculate_sha256(filename) + print(f"{sha256_value}") + + hashes[title] = { + "mtime": ondisk_mtime, + "sha256": sha256_value, + } + + dump_cache() + + return sha256_value + + + + + diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 83cbb4f0..9b5f2e79 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -509,7 +509,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, if shared.opts.save_training_settings_to_txt: saved_params = dict( - model_name=checkpoint.model_name, model_hash=checkpoint.hash, num_of_dataset_images=len(ds), + model_name=checkpoint.model_name, model_hash=checkpoint.shorthash, num_of_dataset_images=len(ds), **{field: getattr(hypernetwork, field) for field in ['layer_structure', 'activation_func', 'weight_init', 'add_layer_norm', 'use_dropout', ]} ) logging.save_settings_to_file(log_directory, {**saved_params, **locals()}) @@ -737,7 +737,7 @@ def save_hypernetwork(hypernetwork, checkpoint, hypernetwork_name, filename): old_sd_checkpoint = hypernetwork.sd_checkpoint if hasattr(hypernetwork, "sd_checkpoint") else None old_sd_checkpoint_name = hypernetwork.sd_checkpoint_name if hasattr(hypernetwork, "sd_checkpoint_name") else None try: - hypernetwork.sd_checkpoint = checkpoint.hash + hypernetwork.sd_checkpoint = checkpoint.shorthash hypernetwork.sd_checkpoint_name = checkpoint.model_name hypernetwork.name = hypernetwork_name hypernetwork.save(filename) diff --git a/modules/sd_models.py b/modules/sd_models.py index c466f273..7babb9ae 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -14,17 +14,56 @@ import ldm.modules.midas as midas from ldm.util import instantiate_from_config -from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors +from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors, hashes from modules.paths import models_path from modules.sd_hijack_inpainting import do_inpainting_hijack, should_hijack_inpainting model_dir = "Stable-diffusion" model_path = os.path.abspath(os.path.join(models_path, model_dir)) -CheckpointInfo = namedtuple("CheckpointInfo", ['filename', 'title', 'hash', 'model_name']) checkpoints_list = {} +checkpoint_alisases = {} checkpoints_loaded = collections.OrderedDict() + +class CheckpointInfo: + def __init__(self, filename): + self.filename = filename + abspath = os.path.abspath(filename) + + if shared.cmd_opts.ckpt_dir is not None and abspath.startswith(shared.cmd_opts.ckpt_dir): + name = abspath.replace(shared.cmd_opts.ckpt_dir, '') + elif abspath.startswith(model_path): + name = abspath.replace(model_path, '') + else: + name = os.path.basename(filename) + + if name.startswith("\\") or name.startswith("/"): + name = name[1:] + + self.title = name + self.model_name = os.path.splitext(name.replace("/", "_").replace("\\", "_"))[0] + self.hash = model_hash(filename) + self.ids = [self.hash, self.model_name, self.title, f'{name} [{self.hash}]'] + self.shorthash = None + self.sha256 = None + + def register(self): + checkpoints_list[self.title] = self + for id in self.ids: + checkpoint_alisases[id] = self + + def calculate_shorthash(self): + self.sha256 = hashes.sha256(self.filename, self.title) + self.shorthash = self.sha256[0:10] + + if self.shorthash not in self.ids: + self.ids += [self.shorthash, self.sha256] + self.register() + + return self.shorthash + + try: # this silences the annoying "Some weights of the model checkpoint were not used when initializing..." message at start. @@ -43,10 +82,14 @@ def setup_model(): enable_midas_autodownload() -def checkpoint_tiles(): - convert = lambda name: int(name) if name.isdigit() else name.lower() - alphanumeric_key = lambda key: [convert(c) for c in re.split('([0-9]+)', key)] - return sorted([x.title for x in checkpoints_list.values()], key = alphanumeric_key) +def checkpoint_tiles(): + def convert(name): + return int(name) if name.isdigit() else name.lower() + + def alphanumeric_key(key): + return [convert(c) for c in re.split('([0-9]+)', key)] + + return sorted([x.title for x in checkpoints_list.values()], key=alphanumeric_key) def find_checkpoint_config(info): @@ -62,48 +105,38 @@ def find_checkpoint_config(info): def list_models(): checkpoints_list.clear() + checkpoint_alisases.clear() model_list = modelloader.load_models(model_path=model_path, command_path=shared.cmd_opts.ckpt_dir, ext_filter=[".ckpt", ".safetensors"], ext_blacklist=[".vae.safetensors"]) - def modeltitle(path, shorthash): - abspath = os.path.abspath(path) - - if shared.cmd_opts.ckpt_dir is not None and abspath.startswith(shared.cmd_opts.ckpt_dir): - name = abspath.replace(shared.cmd_opts.ckpt_dir, '') - elif abspath.startswith(model_path): - name = abspath.replace(model_path, '') - else: - name = os.path.basename(path) - - if name.startswith("\\") or name.startswith("/"): - name = name[1:] - - shortname = os.path.splitext(name.replace("/", "_").replace("\\", "_"))[0] - - return f'{name} [{shorthash}]', shortname - cmd_ckpt = shared.cmd_opts.ckpt if os.path.exists(cmd_ckpt): - h = model_hash(cmd_ckpt) - title, short_model_name = modeltitle(cmd_ckpt, h) - checkpoints_list[title] = CheckpointInfo(cmd_ckpt, title, h, short_model_name) - shared.opts.data['sd_model_checkpoint'] = title + checkpoint_info = CheckpointInfo(cmd_ckpt) + checkpoint_info.register() + + shared.opts.data['sd_model_checkpoint'] = checkpoint_info.title elif cmd_ckpt is not None and cmd_ckpt != shared.default_sd_model_file: print(f"Checkpoint in --ckpt argument not found (Possible it was moved to {model_path}: {cmd_ckpt}", file=sys.stderr) + for filename in model_list: - h = model_hash(filename) - title, short_model_name = modeltitle(filename, h) + checkpoint_info = CheckpointInfo(filename) + checkpoint_info.register() + - checkpoints_list[title] = CheckpointInfo(filename, title, h, short_model_name) +def get_closet_checkpoint_match(search_string): + checkpoint_info = checkpoint_alisases.get(search_string, None) + if checkpoint_info is not None: + return + found = sorted([info for info in checkpoints_list.values() if search_string in info.title], key=lambda x: len(x.title)) + if found: + return found[0] -def get_closet_checkpoint_match(searchString): - applicable = sorted([info for info in checkpoints_list.values() if searchString in info.title], key = lambda x:len(x.title)) - if len(applicable) > 0: - return applicable[0] return None def model_hash(filename): + """old hash that only looks at a small part of the file and is prone to collisions""" + try: with open(filename, "rb") as file: import hashlib @@ -119,7 +152,7 @@ def model_hash(filename): def select_checkpoint(): model_checkpoint = shared.opts.sd_model_checkpoint - checkpoint_info = checkpoints_list.get(model_checkpoint, None) + checkpoint_info = checkpoint_alisases.get(model_checkpoint, None) if checkpoint_info is not None: return checkpoint_info @@ -189,9 +222,8 @@ def read_state_dict(checkpoint_file, print_global_state=False, map_location=None return sd -def load_model_weights(model, checkpoint_info, vae_file="auto"): - checkpoint_file = checkpoint_info.filename - sd_model_hash = checkpoint_info.hash +def load_model_weights(model, checkpoint_info: CheckpointInfo, vae_file="auto"): + sd_model_hash = checkpoint_info.calculate_shorthash() cache_enabled = shared.opts.sd_checkpoint_cache > 0 @@ -201,9 +233,9 @@ def load_model_weights(model, checkpoint_info, vae_file="auto"): model.load_state_dict(checkpoints_loaded[checkpoint_info]) else: # load from file - print(f"Loading weights [{sd_model_hash}] from {checkpoint_file}") + print(f"Loading weights [{sd_model_hash}] from {checkpoint_info.filename}") - sd = read_state_dict(checkpoint_file) + sd = read_state_dict(checkpoint_info.filename) model.load_state_dict(sd, strict=False) del sd @@ -235,14 +267,14 @@ def load_model_weights(model, checkpoint_info, vae_file="auto"): checkpoints_loaded.popitem(last=False) # LRU model.sd_model_hash = sd_model_hash - model.sd_model_checkpoint = checkpoint_file + model.sd_model_checkpoint = checkpoint_info.filename model.sd_checkpoint_info = checkpoint_info model.logvar = model.logvar.to(devices.device) # fix for training sd_vae.delete_base_vae() sd_vae.clear_loaded_vae() - vae_file = sd_vae.resolve_vae(checkpoint_file, vae_file=vae_file) + vae_file = sd_vae.resolve_vae(checkpoint_info.filename, vae_file=vae_file) sd_vae.load_vae(model, vae_file) diff --git a/modules/shared.py b/modules/shared.py index b90ded52..d74c069d 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -428,7 +428,7 @@ options_templates.update(options_section(('ui', "User interface"), { "return_grid": OptionInfo(True, "Show grid in results for web"), "do_not_show_images": OptionInfo(False, "Do not show any images in results for web"), "add_model_hash_to_info": OptionInfo(True, "Add model hash to generation information"), - "add_model_name_to_info": OptionInfo(False, "Add model name to generation information"), + "add_model_name_to_info": OptionInfo(True, "Add model name to generation information"), "disable_weights_auto_swap": OptionInfo(False, "When reading generation parameters from text into UI (from PNG info or pasted text), do not change the selected model/checkpoint."), "send_seed": OptionInfo(True, "Send seed when sending prompt or image to other interface"), "send_size": OptionInfo(True, "Send size when sending prompt or image to another interface"), diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 6939efcc..63935878 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -407,7 +407,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=embedding_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method, varsize=varsize) if shared.opts.save_training_settings_to_txt: - save_settings_to_file(log_directory, {**dict(model_name=checkpoint.model_name, model_hash=checkpoint.hash, num_of_dataset_images=len(ds), num_vectors_per_token=len(embedding.vec)), **locals()}) + save_settings_to_file(log_directory, {**dict(model_name=checkpoint.model_name, model_hash=checkpoint.shorthash, num_of_dataset_images=len(ds), num_vectors_per_token=len(embedding.vec)), **locals()}) latent_sampling_method = ds.latent_sampling_method @@ -584,7 +584,7 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ checkpoint = sd_models.select_checkpoint() footer_left = checkpoint.model_name - footer_mid = '[{}]'.format(checkpoint.hash) + footer_mid = '[{}]'.format(checkpoint.shorthash) footer_right = '{}v {}s'.format(vectorSize, steps_done) captioned_image = caption_image_overlay(image, title, footer_left, footer_mid, footer_right) @@ -626,7 +626,7 @@ def save_embedding(embedding, optimizer, checkpoint, embedding_name, filename, r old_sd_checkpoint_name = embedding.sd_checkpoint_name if hasattr(embedding, "sd_checkpoint_name") else None old_cached_checksum = embedding.cached_checksum if hasattr(embedding, "cached_checksum") else None try: - embedding.sd_checkpoint = checkpoint.hash + embedding.sd_checkpoint = checkpoint.shorthash embedding.sd_checkpoint_name = checkpoint.model_name if remove_cached_checksum: embedding.cached_checksum = None -- cgit v1.2.3 From f9ac3352cb66ce2bc0aa4325130fc7267fb35e4f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 10:25:21 +0300 Subject: change hypernets to use sha256 hashes --- modules/hypernetworks/hypernetwork.py | 40 ++++++++++++++++++++--------------- modules/processing.py | 2 +- modules/sd_models.py | 2 +- modules/shared.py | 1 + 4 files changed, 26 insertions(+), 19 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 9b5f2e79..3aebefa8 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -12,7 +12,7 @@ import torch import tqdm from einops import rearrange, repeat from ldm.util import default -from modules import devices, processing, sd_models, shared, sd_samplers +from modules import devices, processing, sd_models, shared, sd_samplers, hashes from modules.textual_inversion import textual_inversion, logging from modules.textual_inversion.learn_schedule import LearnRateScheduler from torch import einsum @@ -225,7 +225,7 @@ class Hypernetwork: torch.save(state_dict, filename) if shared.opts.save_optimizer_state and self.optimizer_state_dict: - optimizer_saved_dict['hash'] = sd_models.model_hash(filename) + optimizer_saved_dict['hash'] = self.shorthash() optimizer_saved_dict['optimizer_state_dict'] = self.optimizer_state_dict torch.save(optimizer_saved_dict, filename + '.optim') @@ -237,32 +237,33 @@ class Hypernetwork: state_dict = torch.load(filename, map_location='cpu') self.layer_structure = state_dict.get('layer_structure', [1, 2, 1]) - print(self.layer_structure) - optional_info = state_dict.get('optional_info', None) - if optional_info is not None: - print(f"INFO:\n {optional_info}\n") - self.optional_info = optional_info + self.optional_info = state_dict.get('optional_info', None) self.activation_func = state_dict.get('activation_func', None) - print(f"Activation function is {self.activation_func}") self.weight_init = state_dict.get('weight_initialization', 'Normal') - print(f"Weight initialization is {self.weight_init}") self.add_layer_norm = state_dict.get('is_layer_norm', False) - print(f"Layer norm is set to {self.add_layer_norm}") self.dropout_structure = state_dict.get('dropout_structure', None) self.use_dropout = True if self.dropout_structure is not None and any(self.dropout_structure) else state_dict.get('use_dropout', False) - print(f"Dropout usage is set to {self.use_dropout}" ) self.activate_output = state_dict.get('activate_output', True) - print(f"Activate last layer is set to {self.activate_output}") self.last_layer_dropout = state_dict.get('last_layer_dropout', False) # Dropout structure should have same length as layer structure, Every digits should be in [0,1), and last digit must be 0. if self.dropout_structure is None: - print("Using previous dropout structure") self.dropout_structure = parse_dropout_structure(self.layer_structure, self.use_dropout, self.last_layer_dropout) - print(f"Dropout structure is set to {self.dropout_structure}") - optimizer_saved_dict = torch.load(self.filename + '.optim', map_location = 'cpu') if os.path.exists(self.filename + '.optim') else {} + if shared.opts.print_hypernet_extra: + if self.optional_info is not None: + print(f" INFO:\n {self.optional_info}\n") - if sd_models.model_hash(filename) == optimizer_saved_dict.get('hash', None): + print(f" Layer structure: {self.layer_structure}") + print(f" Activation function: {self.activation_func}") + print(f" Weight initialization: {self.weight_init}") + print(f" Layer norm: {self.add_layer_norm}") + print(f" Dropout usage: {self.use_dropout}" ) + print(f" Activate last layer: {self.activate_output}") + print(f" Dropout structure: {self.dropout_structure}") + + optimizer_saved_dict = torch.load(self.filename + '.optim', map_location='cpu') if os.path.exists(self.filename + '.optim') else {} + + if self.shorthash() == optimizer_saved_dict.get('hash', None): self.optimizer_state_dict = optimizer_saved_dict.get('optimizer_state_dict', None) else: self.optimizer_state_dict = None @@ -289,6 +290,11 @@ class Hypernetwork: self.sd_checkpoint_name = state_dict.get('sd_checkpoint_name', None) self.eval() + def shorthash(self): + sha256 = hashes.sha256(self.filename, f'hypernet/{self.name}') + + return sha256[0:10] + def list_hypernetworks(path): res = {} @@ -296,7 +302,7 @@ def list_hypernetworks(path): name = os.path.splitext(os.path.basename(filename))[0] # Prevent a hypothetical "None.pt" from being listed. if name != "None": - res[name + f"({sd_models.model_hash(filename)})"] = filename + res[name] = filename return res diff --git a/modules/processing.py b/modules/processing.py index ae04cab7..849f6b19 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -437,7 +437,7 @@ def create_infotext(p, all_prompts, all_seeds, all_subseeds, comments=None, iter "Model hash": getattr(p, 'sd_model_hash', None if not opts.add_model_hash_to_info or not shared.sd_model.sd_model_hash else shared.sd_model.sd_model_hash), "Model": (None if not opts.add_model_name_to_info or not shared.sd_model.sd_checkpoint_info.model_name else shared.sd_model.sd_checkpoint_info.model_name.replace(',', '').replace(':', '')), "Hypernet": (None if shared.loaded_hypernetwork is None else shared.loaded_hypernetwork.name), - "Hypernet hash": (None if shared.loaded_hypernetwork is None else sd_models.model_hash(shared.loaded_hypernetwork.filename)), + "Hypernet hash": (None if shared.loaded_hypernetwork is None else shared.loaded_hypernetwork.shorthash()), "Hypernet strength": (None if shared.loaded_hypernetwork is None or shared.opts.sd_hypernetwork_strength >= 1 else shared.opts.sd_hypernetwork_strength), "Batch size": (None if p.batch_size < 2 else p.batch_size), "Batch pos": (None if p.batch_size < 2 else position_in_batch), diff --git a/modules/sd_models.py b/modules/sd_models.py index 7babb9ae..8f00191c 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -125,7 +125,7 @@ def list_models(): def get_closet_checkpoint_match(search_string): checkpoint_info = checkpoint_alisases.get(search_string, None) if checkpoint_info is not None: - return + return checkpoint_info found = sorted([info for info in checkpoints_list.values() if search_string in info.title], key=lambda x: len(x.title)) if found: diff --git a/modules/shared.py b/modules/shared.py index d74c069d..a6c61db3 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -361,6 +361,7 @@ options_templates.update(options_section(('system', "System"), { "memmon_poll_rate": OptionInfo(8, "VRAM usage polls per second during generation. Set to 0 to disable.", gr.Slider, {"minimum": 0, "maximum": 40, "step": 1}), "samples_log_stdout": OptionInfo(False, "Always print all generation info to standard output"), "multiple_tqdm": OptionInfo(True, "Add a second progress bar to the console that shows progress for an entire job."), + "print_hypernet_extra": OptionInfo(False, "Print extra hypernetwork information to console."), })) options_templates.update(options_section(('training', "Training"), { -- cgit v1.2.3 From febd2b722e80959b89a0e5966a159b4eb430c5a5 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 13:37:55 +0300 Subject: update key to use with checkpoints' sha256 in cache --- modules/sd_models.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index 8f00191c..1fe6d11b 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -54,7 +54,7 @@ class CheckpointInfo: checkpoint_alisases[id] = self def calculate_shorthash(self): - self.sha256 = hashes.sha256(self.filename, self.title) + self.sha256 = hashes.sha256(self.filename, "checkpoint/" + self.title) self.shorthash = self.sha256[0:10] if self.shorthash not in self.ids: -- cgit v1.2.3 From 6eb72fd13f34d94d5459290dd1a0bf0e9ddeda82 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 13:38:10 +0300 Subject: bump gradio to 3.16.1 --- modules/ui.py | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index e86a624b..202e84e5 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -605,7 +605,7 @@ def create_ui(): setup_progressbar(progressbar, txt2img_preview, 'txt2img') with gr.Row().style(equal_height=False): - with gr.Column(variant='panel', elem_id="txt2img_settings"): + with gr.Column(variant='compact', elem_id="txt2img_settings"): for category in ordered_ui_categories(): if category == "sampler": steps, sampler_index = create_sampler_and_steps_selection(samplers, "txt2img") @@ -794,7 +794,7 @@ def create_ui(): setup_progressbar(progressbar, img2img_preview, 'img2img') with FormRow().style(equal_height=False): - with gr.Column(variant='panel', elem_id="img2img_settings"): + with gr.Column(variant='compact', elem_id="img2img_settings"): with gr.Tabs(elem_id="mode_img2img"): with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab") as tab_img2img: init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool="editor", image_mode="RGBA").style(height=480) @@ -1026,7 +1026,7 @@ def create_ui(): with gr.Blocks(analytics_enabled=False) as extras_interface: with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): + with gr.Column(variant='compact'): with gr.Tabs(elem_id="mode_extras"): with gr.TabItem('Single Image', elem_id="extras_single_tab"): extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") @@ -1127,8 +1127,8 @@ def create_ui(): with gr.Blocks(analytics_enabled=False) as modelmerger_interface: with gr.Row().style(equal_height=False): - with gr.Column(variant='panel'): - gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") + with gr.Column(variant='compact'): + gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") with FormRow(): primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") @@ -1150,7 +1150,8 @@ def create_ui(): config_source = gr.Radio(choices=["A, B or C", "B", "C", "Don't"], value="A, B or C", label="Copy config from", type="index", elem_id="modelmerger_config_method") - modelmerger_merge = gr.Button(elem_id="modelmerger_merge", value="Merge", variant='primary') + with gr.Row(): + modelmerger_merge = gr.Button(elem_id="modelmerger_merge", value="Merge", variant='primary') with gr.Column(variant='panel'): submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) -- cgit v1.2.3 From 5f8685237ed6427c9a8e502124074c740ea7696a Mon Sep 17 00:00:00 2001 From: bbc_mc Date: Sat, 14 Jan 2023 20:00:00 +0900 Subject: Exclude clip index from merge --- modules/extras.py | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index a03d558e..22668fcd 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -326,8 +326,14 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam print("Merging...") + chckpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] + for key in tqdm.tqdm(theta_0.keys()): if 'model' in key and key in theta_1: + + if key in chckpoint_dict_skip_on_merge: + continue + a = theta_0[key] b = theta_1[key] @@ -352,6 +358,10 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam # I believe this part should be discarded, but I'll leave it for now until I am sure for key in theta_1.keys(): if 'model' in key and key not in theta_0: + + if key in chckpoint_dict_skip_on_merge: + continue + theta_0[key] = theta_1[key] if save_as_half: theta_0[key] = theta_0[key].half() -- cgit v1.2.3 From 865228a83736bea9ede33e98041f2a7d0ca5daaa Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 14:56:39 +0300 Subject: change style dropdowns to multiselect --- modules/img2img.py | 4 ++-- modules/styles.py | 12 +++++++++--- modules/txt2img.py | 4 ++-- modules/ui.py | 53 ++++++++++++++++++++++++++++++----------------------- 4 files changed, 43 insertions(+), 30 deletions(-) (limited to 'modules') diff --git a/modules/img2img.py b/modules/img2img.py index f62783c6..79382cc1 100644 --- a/modules/img2img.py +++ b/modules/img2img.py @@ -59,7 +59,7 @@ def process_batch(p, input_dir, output_dir, args): processed_image.save(os.path.join(output_dir, filename)) -def img2img(mode: int, prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, init_img, sketch, init_img_with_mask, inpaint_color_sketch, inpaint_color_sketch_orig, init_img_inpaint, init_mask_inpaint, steps: int, sampler_index: int, mask_blur: int, mask_alpha: float, inpainting_fill: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, denoising_strength: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, resize_mode: int, inpaint_full_res: bool, inpaint_full_res_padding: int, inpainting_mask_invert: int, img2img_batch_input_dir: str, img2img_batch_output_dir: str, *args): +def img2img(mode: int, prompt: str, negative_prompt: str, prompt_styles, init_img, sketch, init_img_with_mask, inpaint_color_sketch, inpaint_color_sketch_orig, init_img_inpaint, init_mask_inpaint, steps: int, sampler_index: int, mask_blur: int, mask_alpha: float, inpainting_fill: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, denoising_strength: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, resize_mode: int, inpaint_full_res: bool, inpaint_full_res_padding: int, inpainting_mask_invert: int, img2img_batch_input_dir: str, img2img_batch_output_dir: str, *args): is_batch = mode == 5 if mode == 0: # img2img @@ -101,7 +101,7 @@ def img2img(mode: int, prompt: str, negative_prompt: str, prompt_style: str, pro outpath_grids=opts.outdir_grids or opts.outdir_img2img_grids, prompt=prompt, negative_prompt=negative_prompt, - styles=[prompt_style, prompt_style2], + styles=prompt_styles, seed=seed, subseed=subseed, subseed_strength=subseed_strength, diff --git a/modules/styles.py b/modules/styles.py index ce6e71ca..990d5623 100644 --- a/modules/styles.py +++ b/modules/styles.py @@ -40,12 +40,18 @@ def apply_styles_to_prompt(prompt, styles): class StyleDatabase: def __init__(self, path: str): self.no_style = PromptStyle("None", "", "") - self.styles = {"None": self.no_style} + self.styles = {} + self.path = path - if not os.path.exists(path): + self.reload() + + def reload(self): + self.styles.clear() + + if not os.path.exists(self.path): return - with open(path, "r", encoding="utf-8-sig", newline='') as file: + with open(self.path, "r", encoding="utf-8-sig", newline='') as file: reader = csv.DictReader(file) for row in reader: # Support loading old CSV format with "name, text"-columns diff --git a/modules/txt2img.py b/modules/txt2img.py index 38b5f591..5a71793b 100644 --- a/modules/txt2img.py +++ b/modules/txt2img.py @@ -8,13 +8,13 @@ import modules.processing as processing from modules.ui import plaintext_to_html -def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, hr_second_pass_steps: int, hr_resize_x: int, hr_resize_y: int, *args): +def txt2img(prompt: str, negative_prompt: str, prompt_styles, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, hr_second_pass_steps: int, hr_resize_x: int, hr_resize_y: int, *args): p = StableDiffusionProcessingTxt2Img( sd_model=shared.sd_model, outpath_samples=opts.outdir_samples or opts.outdir_txt2img_samples, outpath_grids=opts.outdir_grids or opts.outdir_txt2img_grids, prompt=prompt, - styles=[prompt_style, prompt_style2], + styles=prompt_styles, negative_prompt=negative_prompt, seed=seed, subseed=subseed, diff --git a/modules/ui.py b/modules/ui.py index 202e84e5..db198a47 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -180,7 +180,7 @@ def add_style(name: str, prompt: str, negative_prompt: str): # reserialize all styles every time we save them shared.prompt_styles.save_styles(shared.styles_filename) - return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(4)] + return [gr.Dropdown.update(visible=True, choices=list(shared.prompt_styles.styles)) for _ in range(2)] def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resize_y): @@ -197,11 +197,11 @@ def calc_resolution_hires(enable, width, height, hr_scale, hr_resize_x, hr_resiz return f"resize: from {p.width}x{p.height} to {p.hr_resize_x or p.hr_upscale_to_x}x{p.hr_resize_y or p.hr_upscale_to_y}" -def apply_styles(prompt, prompt_neg, style1_name, style2_name): - prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, [style1_name, style2_name]) - prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, [style1_name, style2_name]) +def apply_styles(prompt, prompt_neg, styles): + prompt = shared.prompt_styles.apply_styles_to_prompt(prompt, styles) + prompt_neg = shared.prompt_styles.apply_negative_styles_to_prompt(prompt_neg, styles) - return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] + return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value=[])] def interrogate(image): @@ -374,13 +374,10 @@ def create_toprow(is_img2img): ) with gr.Row(): - with gr.Column(scale=1, elem_id="style_pos_col"): - prompt_style = gr.Dropdown(label="Style 1", elem_id=f"{id_part}_style_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) + prompt_styles = gr.Dropdown(label="Styles", elem_id=f"{id_part}_styles", choices=[k for k, v in shared.prompt_styles.styles.items()], value=[], multiselect=True) + create_refresh_button(prompt_styles, shared.prompt_styles.reload, lambda: {"choices": [k for k, v in shared.prompt_styles.styles.items()]}, f"refresh_{id_part}_styles") - with gr.Column(scale=1, elem_id="style_neg_col"): - prompt_style2 = gr.Dropdown(label="Style 2", elem_id=f"{id_part}_style2_index", choices=[k for k, v in shared.prompt_styles.styles.items()], value=next(iter(shared.prompt_styles.styles.keys()))) - - return prompt, prompt_style, negative_prompt, prompt_style2, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button + return prompt, prompt_styles, negative_prompt, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button def setup_progressbar(*args, **kwargs): @@ -590,7 +587,7 @@ def create_ui(): modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) with gr.Blocks(analytics_enabled=False) as txt2img_interface: - txt2img_prompt, txt2img_prompt_style, txt2img_negative_prompt, txt2img_prompt_style2, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) + txt2img_prompt, txt2img_prompt_styles, txt2img_negative_prompt, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) dummy_component = gr.Label(visible=False) txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) @@ -684,8 +681,7 @@ def create_ui(): inputs=[ txt2img_prompt, txt2img_negative_prompt, - txt2img_prompt_style, - txt2img_prompt_style2, + txt2img_prompt_styles, steps, sampler_index, restore_faces, @@ -780,7 +776,7 @@ def create_ui(): modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) with gr.Blocks(analytics_enabled=False) as img2img_interface: - img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) + img2img_prompt, img2img_prompt_styles, img2img_negative_prompt, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) with gr.Row(elem_id='img2img_progress_row'): img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) @@ -921,8 +917,7 @@ def create_ui(): dummy_component, img2img_prompt, img2img_negative_prompt, - img2img_prompt_style, - img2img_prompt_style2, + img2img_prompt_styles, init_img, sketch, init_img_with_mask, @@ -977,7 +972,7 @@ def create_ui(): ) prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] - style_dropdowns = [(txt2img_prompt_style, txt2img_prompt_style2), (img2img_prompt_style, img2img_prompt_style2)] + style_dropdowns = [txt2img_prompt_styles, img2img_prompt_styles] style_js_funcs = ["update_txt2img_tokens", "update_img2img_tokens"] for button, (prompt, negative_prompt) in zip([txt2img_save_style, img2img_save_style], prompts): @@ -987,15 +982,15 @@ def create_ui(): # Have to pass empty dummy component here, because the JavaScript and Python function have to accept # the same number of parameters, but we only know the style-name after the JavaScript prompt inputs=[dummy_component, prompt, negative_prompt], - outputs=[txt2img_prompt_style, img2img_prompt_style, txt2img_prompt_style2, img2img_prompt_style2], + outputs=[txt2img_prompt_styles, img2img_prompt_styles], ) - for button, (prompt, negative_prompt), (style1, style2), js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): + for button, (prompt, negative_prompt), styles, js_func in zip([txt2img_prompt_style_apply, img2img_prompt_style_apply], prompts, style_dropdowns, style_js_funcs): button.click( fn=apply_styles, _js=js_func, - inputs=[prompt, negative_prompt, style1, style2], - outputs=[prompt, negative_prompt, style1, style2], + inputs=[prompt, negative_prompt, styles], + outputs=[prompt, negative_prompt, styles], ) token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) @@ -1530,6 +1525,7 @@ def create_ui(): previous_section = None current_tab = None + current_row = None with gr.Tabs(elem_id="settings"): for i, (k, item) in enumerate(opts.data_labels.items()): section_must_be_skipped = item.section[0] is None @@ -1538,10 +1534,14 @@ def create_ui(): elem_id, text = item.section if current_tab is not None: + current_row.__exit__() current_tab.__exit__() + gr.Group() current_tab = gr.TabItem(elem_id="settings_{}".format(elem_id), label=text) current_tab.__enter__() + current_row = gr.Column(variant='compact') + current_row.__enter__() previous_section = item.section @@ -1556,6 +1556,7 @@ def create_ui(): components.append(component) if current_tab is not None: + current_row.__exit__() current_tab.__exit__() with gr.TabItem("Actions"): @@ -1774,7 +1775,13 @@ def create_ui(): apply_field(x, 'value') if type(x) == gr.Dropdown: - apply_field(x, 'value', lambda val: val in x.choices, getattr(x, 'init_field', None)) + def check_dropdown(val): + if x.multiselect: + return all([value in x.choices for value in val]) + else: + return val in x.choices + + apply_field(x, 'value', check_dropdown, getattr(x, 'init_field', None)) visit(txt2img_interface, loadsave, "txt2img") visit(img2img_interface, loadsave, "img2img") -- cgit v1.2.3 From 08c6f009a5ee92dd3218a942c08e8337c26352be Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 15:55:40 +0300 Subject: load hashes from cache for checkpoints that have them add checkpoint hash to footer --- modules/hashes.py | 26 +++++++++++++++++++------- modules/sd_models.py | 9 ++++++--- modules/shared.py | 1 + modules/ui.py | 2 ++ 4 files changed, 28 insertions(+), 10 deletions(-) (limited to 'modules') diff --git a/modules/hashes.py b/modules/hashes.py index ebfbd90c..14231771 100644 --- a/modules/hashes.py +++ b/modules/hashes.py @@ -42,23 +42,35 @@ def calculate_sha256(filename): return hash_sha256.hexdigest() -def sha256(filename, title): +def sha256_from_cache(filename, title): hashes = cache("hashes") ondisk_mtime = os.path.getmtime(filename) - if title in hashes: - cached_sha256 = hashes[title].get("sha256", None) - cached_mtime = hashes[title].get("mtime", 0) + if title not in hashes: + return None + + cached_sha256 = hashes[title].get("sha256", None) + cached_mtime = hashes[title].get("mtime", 0) + + if ondisk_mtime > cached_mtime or cached_sha256 is None: + return None + + return cached_sha256 + + +def sha256(filename, title): + hashes = cache("hashes") - if ondisk_mtime <= cached_mtime and cached_sha256 is not None: - return cached_sha256 + sha256_value = sha256_from_cache(filename, title) + if sha256_value is not None: + return sha256_value print(f"Calculating sha256 for {filename}: ", end='') sha256_value = calculate_sha256(filename) print(f"{sha256_value}") hashes[title] = { - "mtime": ondisk_mtime, + "mtime": os.path.getmtime(filename), "sha256": sha256_value, } diff --git a/modules/sd_models.py b/modules/sd_models.py index 1fe6d11b..e5a0bc63 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -44,9 +44,11 @@ class CheckpointInfo: self.title = name self.model_name = os.path.splitext(name.replace("/", "_").replace("\\", "_"))[0] self.hash = model_hash(filename) - self.ids = [self.hash, self.model_name, self.title, f'{name} [{self.hash}]'] - self.shorthash = None - self.sha256 = None + + self.sha256 = hashes.sha256_from_cache(self.filename, "checkpoint/" + self.title) + self.shorthash = self.sha256[0:10] if self.sha256 else None + + self.ids = [self.hash, self.model_name, self.title, f'{name} [{self.hash}]'] + ([self.shorthash, self.sha256] if self.shorthash else []) def register(self): checkpoints_list[self.title] = self @@ -269,6 +271,7 @@ def load_model_weights(model, checkpoint_info: CheckpointInfo, vae_file="auto"): model.sd_model_hash = sd_model_hash model.sd_model_checkpoint = checkpoint_info.filename model.sd_checkpoint_info = checkpoint_info + shared.opts.data["sd_checkpoint_hash"] = checkpoint_info.sha256 model.logvar = model.logvar.to(devices.device) # fix for training diff --git a/modules/shared.py b/modules/shared.py index a6c61db3..c9988d4d 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -458,6 +458,7 @@ options_templates.update(options_section(('sampler-params', "Sampler parameters" options_templates.update(options_section((None, "Hidden options"), { "disabled_extensions": OptionInfo([], "Disable those extensions"), + "sd_checkpoint_hash": OptionInfo("", "SHA256 hash of the current checkpoint"), })) options_templates.update() diff --git a/modules/ui.py b/modules/ui.py index e86a624b..2625ae32 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1841,4 +1841,6 @@ xformers: {xformers_version} gradio: {gr.__version__}  •  commit: {short_commit} + •  +checkpoint: N/A """ -- cgit v1.2.3 From f94a215abed85b34ae978853078812801d3e7738 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 16:29:23 +0300 Subject: add an option to choose what you want to see in live preview (Live preview subject) and moves live preview settings to its own tab --- modules/sd_samplers.py | 15 ++++++++++----- modules/shared.py | 13 +++++++++---- modules/ui_progress.py | 2 +- 3 files changed, 20 insertions(+), 10 deletions(-) (limited to 'modules') diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py index 01221b89..7616fded 100644 --- a/modules/sd_samplers.py +++ b/modules/sd_samplers.py @@ -138,7 +138,7 @@ def samples_to_image_grid(samples, approximation=None): def store_latent(decoded): state.current_latent = decoded - if opts.show_progress_every_n_steps > 0 and shared.state.sampling_step % opts.show_progress_every_n_steps == 0: + if opts.live_previews_enable and opts.show_progress_every_n_steps > 0 and shared.state.sampling_step % opts.show_progress_every_n_steps == 0: if not shared.parallel_processing_allowed: shared.state.current_image = sample_to_image(decoded) @@ -243,7 +243,7 @@ class VanillaStableDiffusionSampler: self.nmask = p.nmask if hasattr(p, 'nmask') else None def adjust_steps_if_invalid(self, p, num_steps): - if (self.config.name == 'DDIM' and p.ddim_discretize == 'uniform') or (self.config.name == 'PLMS'): + if (self.config.name == 'DDIM' and p.ddim_discretize == 'uniform') or (self.config.name == 'PLMS'): valid_step = 999 / (1000 // num_steps) if valid_step == floor(valid_step): return int(valid_step) + 1 @@ -266,8 +266,7 @@ class VanillaStableDiffusionSampler: if image_conditioning is not None: conditioning = {"c_concat": [image_conditioning], "c_crossattn": [conditioning]} unconditional_conditioning = {"c_concat": [image_conditioning], "c_crossattn": [unconditional_conditioning]} - - + samples = self.launch_sampling(t_enc + 1, lambda: self.sampler.decode(x1, conditioning, t_enc, unconditional_guidance_scale=p.cfg_scale, unconditional_conditioning=unconditional_conditioning)) return samples @@ -352,6 +351,11 @@ class CFGDenoiser(torch.nn.Module): x_out[-uncond.shape[0]:] = self.inner_model(x_in[-uncond.shape[0]:], sigma_in[-uncond.shape[0]:], cond={"c_crossattn": [uncond], "c_concat": [image_cond_in[-uncond.shape[0]:]]}) + if opts.live_preview_content == "Prompt": + store_latent(x_out[0:uncond.shape[0]]) + elif opts.live_preview_content == "Negative prompt": + store_latent(x_out[-uncond.shape[0]:]) + denoised = self.combine_denoised(x_out, conds_list, uncond, cond_scale) if self.mask is not None: @@ -423,7 +427,8 @@ class KDiffusionSampler: def callback_state(self, d): step = d['i'] latent = d["denoised"] - store_latent(latent) + if opts.live_preview_content == "Combined": + store_latent(latent) self.last_latent = latent if self.stop_at is not None and step > self.stop_at: diff --git a/modules/shared.py b/modules/shared.py index c9988d4d..e0ec3136 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -176,7 +176,7 @@ class State: self.interrupted = True def nextjob(self): - if opts.show_progress_every_n_steps == -1: + if opts.live_previews_enable and opts.show_progress_every_n_steps == -1: self.do_set_current_image() self.job_no += 1 @@ -224,7 +224,7 @@ class State: if not parallel_processing_allowed: return - if self.sampling_step - self.current_image_sampling_step >= opts.show_progress_every_n_steps and opts.show_progress_every_n_steps > 0: + if self.sampling_step - self.current_image_sampling_step >= opts.show_progress_every_n_steps and opts.live_previews_enable: self.do_set_current_image() def do_set_current_image(self): @@ -423,8 +423,6 @@ options_templates.update(options_section(('interrogate', "Interrogate Options"), options_templates.update(options_section(('ui', "User interface"), { "show_progressbar": OptionInfo(True, "Show progressbar"), - "show_progress_every_n_steps": OptionInfo(0, "Show image creation progress every N sampling steps. Set to 0 to disable. Set to -1 to show after completion of batch.", gr.Slider, {"minimum": -1, "maximum": 32, "step": 1}), - "show_progress_type": OptionInfo("Full", "Image creation progress preview mode", gr.Radio, {"choices": ["Full", "Approx NN", "Approx cheap"]}), "show_progress_grid": OptionInfo(True, "Show previews of all images generated in a batch as a grid"), "return_grid": OptionInfo(True, "Show grid in results for web"), "do_not_show_images": OptionInfo(False, "Do not show any images in results for web"), @@ -444,6 +442,13 @@ options_templates.update(options_section(('ui', "User interface"), { 'localization': OptionInfo("None", "Localization (requires restart)", gr.Dropdown, lambda: {"choices": ["None"] + list(localization.localizations.keys())}, refresh=lambda: localization.list_localizations(cmd_opts.localizations_dir)), })) +options_templates.update(options_section(('ui', "Live previews"), { + "live_previews_enable": OptionInfo(True, "Show live previews of the created image"), + "show_progress_every_n_steps": OptionInfo(10, "Show new live preview image every N sampling steps. Set to -1 to show after completion of batch.", gr.Slider, {"minimum": -1, "maximum": 32, "step": 1}), + "show_progress_type": OptionInfo("Approx NN", "Image creation progress preview mode", gr.Radio, {"choices": ["Full", "Approx NN", "Approx cheap"]}), + "live_preview_content": OptionInfo("Prompt", "Live preview subject", gr.Radio, {"choices": ["Combined", "Prompt", "Negative prompt"]}), +})) + options_templates.update(options_section(('sampler-params', "Sampler parameters"), { "hide_samplers": OptionInfo([], "Hide samplers in user interface (requires restart)", gr.CheckboxGroup, lambda: {"choices": [x.name for x in list_samplers()]}), "eta_ddim": OptionInfo(0.0, "eta (noise multiplier) for DDIM", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), diff --git a/modules/ui_progress.py b/modules/ui_progress.py index 592fda55..7cd312e4 100644 --- a/modules/ui_progress.py +++ b/modules/ui_progress.py @@ -52,7 +52,7 @@ def check_progress_call(id_part): image = gr.update(visible=False) preview_visibility = gr.update(visible=False) - if opts.show_progress_every_n_steps != 0: + if opts.live_previews_enable: shared.state.set_current_image() image = shared.state.current_image -- cgit v1.2.3 From fad850fc3d33e7cda2ce4b3a32ab7976c313db53 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Sat, 14 Jan 2023 11:18:05 -0500 Subject: add server_start to shared.state --- modules/shared.py | 2 ++ 1 file changed, 2 insertions(+) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index e0ec3136..ef93637c 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -168,6 +168,7 @@ class State: textinfo = None time_start = None need_restart = False + server_start = None def skip(self): self.skipped = True @@ -241,6 +242,7 @@ class State: state = State() +state.server_start = time.time() artist_db = modules.artists.ArtistsDatabase(os.path.join(script_path, 'artists.csv')) -- cgit v1.2.3 From a5bbcd215304e0c83ab2b9fe7f172f88536d7629 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 19:56:09 +0300 Subject: fix bug with "Ignore selected VAE for..." option completely disabling VAE election rework VAE resolving code to be more simple --- modules/sd_models.py | 6 +- modules/sd_vae.py | 194 ++++++++++++++++++++------------------------------- modules/shared.py | 4 +- 3 files changed, 82 insertions(+), 122 deletions(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index e5a0bc63..6a681cef 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -224,7 +224,7 @@ def read_state_dict(checkpoint_file, print_global_state=False, map_location=None return sd -def load_model_weights(model, checkpoint_info: CheckpointInfo, vae_file="auto"): +def load_model_weights(model, checkpoint_info: CheckpointInfo): sd_model_hash = checkpoint_info.calculate_shorthash() cache_enabled = shared.opts.sd_checkpoint_cache > 0 @@ -277,8 +277,8 @@ def load_model_weights(model, checkpoint_info: CheckpointInfo, vae_file="auto"): sd_vae.delete_base_vae() sd_vae.clear_loaded_vae() - vae_file = sd_vae.resolve_vae(checkpoint_info.filename, vae_file=vae_file) - sd_vae.load_vae(model, vae_file) + vae_file, vae_source = sd_vae.resolve_vae(checkpoint_info.filename) + sd_vae.load_vae(model, vae_file, vae_source) def enable_midas_autodownload(): diff --git a/modules/sd_vae.py b/modules/sd_vae.py index 0a49daa1..6ea92711 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -9,23 +9,9 @@ import glob from copy import deepcopy -model_dir = "Stable-diffusion" -model_path = os.path.abspath(os.path.join(models_path, model_dir)) -vae_dir = "VAE" -vae_path = os.path.abspath(os.path.join(models_path, vae_dir)) - - +vae_path = os.path.abspath(os.path.join(models_path, "VAE")) vae_ignore_keys = {"model_ema.decay", "model_ema.num_updates"} - - -default_vae_dict = {"auto": "auto", "None": None, None: None} -default_vae_list = ["auto", "None"] - - -default_vae_values = [default_vae_dict[x] for x in default_vae_list] -vae_dict = dict(default_vae_dict) -vae_list = list(default_vae_list) -first_load = True +vae_dict = {} base_vae = None @@ -64,100 +50,69 @@ def restore_base_vae(model): def get_filename(filepath): - return os.path.splitext(os.path.basename(filepath))[0] - - -def refresh_vae_list(vae_path=vae_path, model_path=model_path): - global vae_dict, vae_list - res = {} - candidates = [ - *glob.iglob(os.path.join(model_path, '**/*.vae.ckpt'), recursive=True), - *glob.iglob(os.path.join(model_path, '**/*.vae.pt'), recursive=True), - *glob.iglob(os.path.join(model_path, '**/*.vae.safetensors'), recursive=True), - *glob.iglob(os.path.join(vae_path, '**/*.ckpt'), recursive=True), - *glob.iglob(os.path.join(vae_path, '**/*.pt'), recursive=True), - *glob.iglob(os.path.join(vae_path, '**/*.safetensors'), recursive=True), + return os.path.basename(filepath) + + +def refresh_vae_list(): + vae_dict.clear() + + paths = [ + os.path.join(sd_models.model_path, '**/*.vae.ckpt'), + os.path.join(sd_models.model_path, '**/*.vae.pt'), + os.path.join(sd_models.model_path, '**/*.vae.safetensors'), + os.path.join(vae_path, '**/*.ckpt'), + os.path.join(vae_path, '**/*.pt'), + os.path.join(vae_path, '**/*.safetensors'), ] - if shared.cmd_opts.vae_path is not None and os.path.isfile(shared.cmd_opts.vae_path): - candidates.append(shared.cmd_opts.vae_path) + + if shared.cmd_opts.ckpt_dir is not None and os.path.isdir(shared.cmd_opts.ckpt_dir): + paths += [ + os.path.join(shared.cmd_opts.ckpt_dir, '**/*.vae.ckpt'), + os.path.join(shared.cmd_opts.ckpt_dir, '**/*.vae.pt'), + os.path.join(shared.cmd_opts.ckpt_dir, '**/*.vae.safetensors'), + ] + + candidates = [] + for path in paths: + candidates += glob.iglob(path, recursive=True) + for filepath in candidates: name = get_filename(filepath) - res[name] = filepath - vae_list.clear() - vae_list.extend(default_vae_list) - vae_list.extend(list(res.keys())) - vae_dict.clear() - vae_dict.update(res) - vae_dict.update(default_vae_dict) - return vae_list - - -def get_vae_from_settings(vae_file="auto"): - # else, we load from settings, if not set to be default - if vae_file == "auto" and shared.opts.sd_vae is not None: - # if saved VAE settings isn't recognized, fallback to auto - vae_file = vae_dict.get(shared.opts.sd_vae, "auto") - # if VAE selected but not found, fallback to auto - if vae_file not in default_vae_values and not os.path.isfile(vae_file): - vae_file = "auto" - print(f"Selected VAE doesn't exist: {vae_file}") - return vae_file - - -def resolve_vae(checkpoint_file=None, vae_file="auto"): - global first_load, vae_dict, vae_list - - # if vae_file argument is provided, it takes priority, but not saved - if vae_file and vae_file not in default_vae_list: - if not os.path.isfile(vae_file): - print(f"VAE provided as function argument doesn't exist: {vae_file}") - vae_file = "auto" - # for the first load, if vae-path is provided, it takes priority, saved, and failure is reported - if first_load and shared.cmd_opts.vae_path is not None: - if os.path.isfile(shared.cmd_opts.vae_path): - vae_file = shared.cmd_opts.vae_path - shared.opts.data['sd_vae'] = get_filename(vae_file) - else: - print(f"VAE provided as command line argument doesn't exist: {vae_file}") - # fallback to selector in settings, if vae selector not set to act as default fallback - if not shared.opts.sd_vae_as_default: - vae_file = get_vae_from_settings(vae_file) - # vae-path cmd arg takes priority for auto - if vae_file == "auto" and shared.cmd_opts.vae_path is not None: - if os.path.isfile(shared.cmd_opts.vae_path): - vae_file = shared.cmd_opts.vae_path - print(f"Using VAE provided as command line argument: {vae_file}") - # if still not found, try look for ".vae.pt" beside model - model_path = os.path.splitext(checkpoint_file)[0] - if vae_file == "auto": - vae_file_try = model_path + ".vae.pt" - if os.path.isfile(vae_file_try): - vae_file = vae_file_try - print(f"Using VAE found similar to selected model: {vae_file}") - # if still not found, try look for ".vae.ckpt" beside model - if vae_file == "auto": - vae_file_try = model_path + ".vae.ckpt" - if os.path.isfile(vae_file_try): - vae_file = vae_file_try - print(f"Using VAE found similar to selected model: {vae_file}") - # if still not found, try look for ".vae.safetensors" beside model - if vae_file == "auto": - vae_file_try = model_path + ".vae.safetensors" - if os.path.isfile(vae_file_try): - vae_file = vae_file_try - print(f"Using VAE found similar to selected model: {vae_file}") - # No more fallbacks for auto - if vae_file == "auto": - vae_file = None - # Last check, just because - if vae_file and not os.path.exists(vae_file): - vae_file = None - - return vae_file - - -def load_vae(model, vae_file=None): - global first_load, vae_dict, vae_list, loaded_vae_file + vae_dict[name] = filepath + + +def find_vae_near_checkpoint(checkpoint_file): + checkpoint_path = os.path.splitext(checkpoint_file)[0] + for vae_location in [checkpoint_path + ".vae.pt", checkpoint_path + ".vae.ckpt", checkpoint_path + ".vae.safetensors"]: + if os.path.isfile(vae_location): + return vae_location + + return None + + +def resolve_vae(checkpoint_file): + if shared.cmd_opts.vae_path is not None: + return shared.cmd_opts.vae_path, 'from commandline argument' + + vae_near_checkpoint = find_vae_near_checkpoint(checkpoint_file) + if vae_near_checkpoint is not None and (shared.opts.sd_vae_as_default or shared.opts.sd_vae == "auto"): + return vae_near_checkpoint, 'found near the checkpoint' + + if shared.opts.sd_vae == "None": + return None, None + + vae_from_options = vae_dict.get(shared.opts.sd_vae, None) + if vae_from_options is not None: + return vae_from_options, 'specified in settings' + + if shared.opts.sd_vae != "Automatic": + print(f"Couldn't find VAE named {shared.opts.sd_vae}; using None instead") + + return None, None + + +def load_vae(model, vae_file=None, vae_source="from unknown source"): + global vae_dict, loaded_vae_file # save_settings = False cache_enabled = shared.opts.sd_vae_checkpoint_cache > 0 @@ -165,12 +120,12 @@ def load_vae(model, vae_file=None): if vae_file: if cache_enabled and vae_file in checkpoints_loaded: # use vae checkpoint cache - print(f"Loading VAE weights [{get_filename(vae_file)}] from cache") + print(f"Loading VAE weights {vae_source}: cached {get_filename(vae_file)}") store_base_vae(model) _load_vae_dict(model, checkpoints_loaded[vae_file]) else: - assert os.path.isfile(vae_file), f"VAE file doesn't exist: {vae_file}" - print(f"Loading VAE weights from: {vae_file}") + assert os.path.isfile(vae_file), f"VAE {vae_source} doesn't exist: {vae_file}" + print(f"Loading VAE weights {vae_source}: {vae_file}") store_base_vae(model) vae_ckpt = sd_models.read_state_dict(vae_file, map_location=shared.weight_load_location) @@ -191,14 +146,12 @@ def load_vae(model, vae_file=None): vae_opt = get_filename(vae_file) if vae_opt not in vae_dict: vae_dict[vae_opt] = vae_file - vae_list.append(vae_opt) + elif loaded_vae_file: restore_base_vae(model) loaded_vae_file = vae_file - first_load = False - # don't call this from outside def _load_vae_dict(model, vae_dict_1): @@ -211,7 +164,10 @@ def clear_loaded_vae(): loaded_vae_file = None -def reload_vae_weights(sd_model=None, vae_file="auto"): +unspecified = object() + + +def reload_vae_weights(sd_model=None, vae_file=unspecified): from modules import lowvram, devices, sd_hijack if not sd_model: @@ -219,7 +175,11 @@ def reload_vae_weights(sd_model=None, vae_file="auto"): checkpoint_info = sd_model.sd_checkpoint_info checkpoint_file = checkpoint_info.filename - vae_file = resolve_vae(checkpoint_file, vae_file=vae_file) + + if vae_file == unspecified: + vae_file, vae_source = resolve_vae(checkpoint_file) + else: + vae_source = "from function argument" if loaded_vae_file == vae_file: return @@ -231,7 +191,7 @@ def reload_vae_weights(sd_model=None, vae_file="auto"): sd_hijack.model_hijack.undo_hijack(sd_model) - load_vae(sd_model, vae_file) + load_vae(sd_model, vae_file, vae_source) sd_hijack.model_hijack.hijack(sd_model) script_callbacks.model_loaded_callback(sd_model) @@ -239,5 +199,5 @@ def reload_vae_weights(sd_model=None, vae_file="auto"): if not shared.cmd_opts.lowvram and not shared.cmd_opts.medvram: sd_model.to(devices.device) - print("VAE Weights loaded.") + print("VAE weights loaded.") return sd_model diff --git a/modules/shared.py b/modules/shared.py index e0ec3136..9756adea 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -83,7 +83,7 @@ parser.add_argument("--theme", type=str, help="launches the UI with light or dar parser.add_argument("--use-textbox-seed", action='store_true', help="use textbox for seeds in UI (no up/down, but possible to input long seeds)", default=False) parser.add_argument("--disable-console-progressbars", action='store_true', help="do not output progressbars to console", default=False) parser.add_argument("--enable-console-prompts", action='store_true', help="print prompts to console when generating with txt2img and img2img", default=False) -parser.add_argument('--vae-path', type=str, help='Path to Variational Autoencoders model', default=None) +parser.add_argument('--vae-path', type=str, help='Checkpoint to use as VAE; setting this argument disables all settings related to VAE', default=None) parser.add_argument("--disable-safe-unpickle", action='store_true', help="disable checking pytorch models for malicious code", default=False) parser.add_argument("--api", action='store_true', help="use api=True to launch the API together with the webui (use --nowebui instead for only the API)") parser.add_argument("--api-auth", type=str, help='Set authentication for API like "username:password"; or comma-delimit multiple like "u1:p1,u2:p2,u3:p3"', default=None) @@ -383,7 +383,7 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "sd_model_checkpoint": OptionInfo(None, "Stable Diffusion checkpoint", gr.Dropdown, lambda: {"choices": list_checkpoint_tiles()}, refresh=refresh_checkpoints), "sd_checkpoint_cache": OptionInfo(0, "Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), "sd_vae_checkpoint_cache": OptionInfo(0, "VAE Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), - "sd_vae": OptionInfo("auto", "SD VAE", gr.Dropdown, lambda: {"choices": sd_vae.vae_list}, refresh=sd_vae.refresh_vae_list), + "sd_vae": OptionInfo("Automatic", "SD VAE", gr.Dropdown, lambda: {"choices": ["Automatic", "None"] + list(sd_vae.vae_dict)}, refresh=sd_vae.refresh_vae_list), "sd_vae_as_default": OptionInfo(False, "Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them"), "sd_hypernetwork": OptionInfo("None", "Hypernetwork", gr.Dropdown, lambda: {"choices": ["None"] + [x for x in hypernetworks.keys()]}, refresh=reload_hypernetworks), "sd_hypernetwork_strength": OptionInfo(1.0, "Hypernetwork strength", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.001}), -- cgit v1.2.3 From f8c512478568293155539f616dce26c5e4495055 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 20:00:12 +0300 Subject: typo? --- modules/sd_vae.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_vae.py b/modules/sd_vae.py index 6ea92711..add5cecf 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -95,7 +95,7 @@ def resolve_vae(checkpoint_file): return shared.cmd_opts.vae_path, 'from commandline argument' vae_near_checkpoint = find_vae_near_checkpoint(checkpoint_file) - if vae_near_checkpoint is not None and (shared.opts.sd_vae_as_default or shared.opts.sd_vae == "auto"): + if vae_near_checkpoint is not None and (shared.opts.sd_vae_as_default or shared.opts.sd_vae == "Automatic"): return vae_near_checkpoint, 'found near the checkpoint' if shared.opts.sd_vae == "None": -- cgit v1.2.3 From 86359535d6fb0899fa9e838d27f2006b929331d5 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 14 Jan 2023 22:43:01 +0300 Subject: add buttons to copy images between img2img tabs --- modules/ui.py | 42 +++++++++++++++++++++++++++++++++++++++++- 1 file changed, 41 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 2625ae32..2425c66f 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -795,19 +795,39 @@ def create_ui(): with FormRow().style(equal_height=False): with gr.Column(variant='panel', elem_id="img2img_settings"): + copy_image_buttons = [] + copy_image_destinations = {} + + def add_copy_image_controls(tab_name, elem): + with gr.Row(variant="compact", elem_id=f"img2img_copy_to_{tab_name}"): + gr.HTML("Copy image to: ", elem_id=f"img2img_label_copy_to_{tab_name}") + + for title, name in zip(['img2img', 'sketch', 'inpaint', 'inpaint sketch'], ['img2img', 'sketch', 'inpaint', 'inpaint_sketch']): + if name == tab_name: + gr.Button(title, interactive=False) + copy_image_destinations[name] = elem + continue + + button = gr.Button(title) + copy_image_buttons.append((button, name, elem)) + with gr.Tabs(elem_id="mode_img2img"): with gr.TabItem('img2img', id='img2img', elem_id="img2img_img2img_tab") as tab_img2img: init_img = gr.Image(label="Image for img2img", elem_id="img2img_image", show_label=False, source="upload", interactive=True, type="pil", tool="editor", image_mode="RGBA").style(height=480) + add_copy_image_controls('img2img', init_img) with gr.TabItem('Sketch', id='img2img_sketch', elem_id="img2img_img2img_sketch_tab") as tab_sketch: sketch = gr.Image(label="Image for img2img", elem_id="img2img_sketch", show_label=False, source="upload", interactive=True, type="pil", tool="color-sketch", image_mode="RGBA").style(height=480) + add_copy_image_controls('sketch', sketch) with gr.TabItem('Inpaint', id='inpaint', elem_id="img2img_inpaint_tab") as tab_inpaint: init_img_with_mask = gr.Image(label="Image for inpainting with mask", show_label=False, elem_id="img2maskimg", source="upload", interactive=True, type="pil", tool="sketch", image_mode="RGBA").style(height=480) + add_copy_image_controls('inpaint', init_img_with_mask) with gr.TabItem('Inpaint sketch', id='inpaint_sketch', elem_id="img2img_inpaint_sketch_tab") as tab_inpaint_color: inpaint_color_sketch = gr.Image(label="Color sketch inpainting", show_label=False, elem_id="inpaint_sketch", source="upload", interactive=True, type="pil", tool="color-sketch", image_mode="RGBA").style(height=480) inpaint_color_sketch_orig = gr.State(None) + add_copy_image_controls('inpaint_sketch', inpaint_color_sketch) def update_orig(image, state): if image is not None: @@ -824,10 +844,29 @@ def create_ui(): with gr.TabItem('Batch', id='batch', elem_id="img2img_batch_tab") as tab_batch: hidden = '
Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' - gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") + gr.HTML(f"

Process images in a directory on the same machine where the server is running.
Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

") img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") + def copy_image(img): + if isinstance(img, dict) and 'image' in img: + return img['image'] + + return img + + for button, name, elem in copy_image_buttons: + button.click( + fn=copy_image, + inputs=[elem], + outputs=[copy_image_destinations[name]], + ) + button.click( + fn=lambda: None, + _js="switch_to_"+name.replace(" ", "_"), + inputs=[], + outputs=[], + ) + with FormGroup(elem_id="inpaint_controls", visible=False) as inpaint_controls: with FormRow(): mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") @@ -856,6 +895,7 @@ def create_ui(): outputs=[inpaint_controls, mask_alpha], ) + with FormRow(): resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") -- cgit v1.2.3 From 9ef41df6f9043d58fbbeea1f06be8e5c8622248b Mon Sep 17 00:00:00 2001 From: Josh R Date: Sat, 14 Jan 2023 15:26:45 -0800 Subject: add inpaint masking controls to orderable section that the settings can order --- modules/shared.py | 1 + modules/ui.py | 58 +++++++++++++++++++++++++++---------------------------- 2 files changed, 30 insertions(+), 29 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 51df056c..7ce8003f 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -116,6 +116,7 @@ restricted_opts = { } ui_reorder_categories = [ + "masking", "sampler", "dimensions", "cfg", diff --git a/modules/ui.py b/modules/ui.py index 2425c66f..174930ab 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -867,35 +867,6 @@ def create_ui(): outputs=[], ) - with FormGroup(elem_id="inpaint_controls", visible=False) as inpaint_controls: - with FormRow(): - mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") - mask_alpha = gr.Slider(label="Mask transparency", visible=False, elem_id="img2img_mask_alpha") - - with FormRow(): - inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") - - with FormRow(): - inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") - - with FormRow(): - with gr.Column(): - inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") - - with gr.Column(scale=4): - inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") - - def select_img2img_tab(tab): - return gr.update(visible=tab in [2, 3, 4]), gr.update(visible=tab == 3), - - for i, elem in enumerate([tab_img2img, tab_sketch, tab_inpaint, tab_inpaint_color, tab_inpaint_upload, tab_batch]): - elem.select( - fn=lambda tab=i: select_img2img_tab(tab), - inputs=[], - outputs=[inpaint_controls, mask_alpha], - ) - - with FormRow(): resize_mode = gr.Radio(label="Resize mode", elem_id="resize_mode", choices=["Just resize", "Crop and resize", "Resize and fill", "Just resize (latent upscale)"], type="index", value="Just resize") @@ -937,6 +908,35 @@ def create_ui(): with FormGroup(elem_id="img2img_script_container"): custom_inputs = modules.scripts.scripts_img2img.setup_ui() + elif category == "masking": + with FormGroup(elem_id="inpaint_controls", visible=False) as inpaint_controls: + with FormRow(): + mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") + mask_alpha = gr.Slider(label="Mask transparency", visible=False, elem_id="img2img_mask_alpha") + + with FormRow(): + inpainting_mask_invert = gr.Radio(label='Mask mode', choices=['Inpaint masked', 'Inpaint not masked'], value='Inpaint masked', type="index", elem_id="img2img_mask_mode") + + with FormRow(): + inpainting_fill = gr.Radio(label='Masked content', choices=['fill', 'original', 'latent noise', 'latent nothing'], value='original', type="index", elem_id="img2img_inpainting_fill") + + with FormRow(): + with gr.Column(): + inpaint_full_res = gr.Radio(label="Inpaint area", choices=["Whole picture", "Only masked"], type="index", value="Whole picture", elem_id="img2img_inpaint_full_res") + + with gr.Column(scale=4): + inpaint_full_res_padding = gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32, elem_id="img2img_inpaint_full_res_padding") + + def select_img2img_tab(tab): + return gr.update(visible=tab in [2, 3, 4]), gr.update(visible=tab == 3), + + for i, elem in enumerate([tab_img2img, tab_sketch, tab_inpaint, tab_inpaint_color, tab_inpaint_upload, tab_batch]): + elem.select( + fn=lambda tab=i: select_img2img_tab(tab), + inputs=[], + outputs=[inpaint_controls, mask_alpha], + ) + img2img_gallery, generation_info, html_info, html_log = create_output_panel("img2img", opts.outdir_img2img_samples) parameters_copypaste.bind_buttons({"img2img": img2img_paste}, None, img2img_prompt) -- cgit v1.2.3 From eef1990a5e6c41ecb6943ff5529316ad5ededb2a Mon Sep 17 00:00:00 2001 From: brkirch Date: Sun, 15 Jan 2023 08:13:33 -0500 Subject: Fix Approx NN on devices other than CUDA --- modules/sd_vae_approx.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_vae_approx.py b/modules/sd_vae_approx.py index 0a58542d..0027343a 100644 --- a/modules/sd_vae_approx.py +++ b/modules/sd_vae_approx.py @@ -36,7 +36,7 @@ def model(): if sd_vae_approx_model is None: sd_vae_approx_model = VAEApprox() - sd_vae_approx_model.load_state_dict(torch.load(os.path.join(paths.models_path, "VAE-approx", "model.pt"))) + sd_vae_approx_model.load_state_dict(torch.load(os.path.join(paths.models_path, "VAE-approx", "model.pt"), map_location='cpu' if devices.device.type != 'cuda' else None)) sd_vae_approx_model.eval() sd_vae_approx_model.to(devices.device, devices.dtype) -- cgit v1.2.3 From f0312565e5b4d56a421af889a9a8eaea0ba92959 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Sun, 15 Jan 2023 09:42:34 -0500 Subject: increase block size --- modules/hashes.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/hashes.py b/modules/hashes.py index 14231771..b85a7580 100644 --- a/modules/hashes.py +++ b/modules/hashes.py @@ -34,9 +34,10 @@ def cache(subsection): def calculate_sha256(filename): hash_sha256 = hashlib.sha256() + blksize = 1024 * 1024 with open(filename, "rb") as f: - for chunk in iter(lambda: f.read(4096), b""): + for chunk in iter(lambda: f.read(blksize), b""): hash_sha256.update(chunk) return hash_sha256.hexdigest() -- cgit v1.2.3 From d8b90ac121cbf0c18b1dc9d56a5e1d14ca51e74e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 15 Jan 2023 18:50:56 +0300 Subject: big rework of progressbar/preview system to allow multiple users to prompts at the same time and do not get previews of each other --- modules/call_queue.py | 19 ++++- modules/hypernetworks/hypernetwork.py | 6 +- modules/img2img.py | 2 +- modules/progress.py | 96 +++++++++++++++++++++++ modules/sd_samplers.py | 2 +- modules/shared.py | 16 ++-- modules/textual_inversion/preprocess.py | 2 +- modules/textual_inversion/textual_inversion.py | 6 +- modules/txt2img.py | 2 +- modules/ui.py | 41 +++------- modules/ui_progress.py | 101 ------------------------- 11 files changed, 143 insertions(+), 150 deletions(-) create mode 100644 modules/progress.py delete mode 100644 modules/ui_progress.py (limited to 'modules') diff --git a/modules/call_queue.py b/modules/call_queue.py index 4cd49533..92097c15 100644 --- a/modules/call_queue.py +++ b/modules/call_queue.py @@ -4,7 +4,7 @@ import threading import traceback import time -from modules import shared +from modules import shared, progress queue_lock = threading.Lock() @@ -22,12 +22,23 @@ def wrap_queued_call(func): def wrap_gradio_gpu_call(func, extra_outputs=None): def f(*args, **kwargs): - shared.state.begin() + # if the first argument is a string that says "task(...)", it is treated as a job id + if len(args) > 0 and type(args[0]) == str and args[0][0:5] == "task(" and args[0][-1] == ")": + id_task = args[0] + progress.add_task_to_queue(id_task) + else: + id_task = None with queue_lock: - res = func(*args, **kwargs) + shared.state.begin() + progress.start_task(id_task) + + try: + res = func(*args, **kwargs) + finally: + progress.finish_task(id_task) - shared.state.end() + shared.state.end() return res diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 3aebefa8..ae6af516 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -453,7 +453,7 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, shared.reload_hypernetworks() -def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_filename, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): +def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_filename, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): # images allows training previews to have infotext. Importing it at the top causes a circular import problem. from modules import images @@ -629,7 +629,6 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, description = f"Training hypernetwork [Epoch {epoch_num}: {epoch_step+1}/{steps_per_epoch}]loss: {loss_step:.7f}" pbar.set_description(description) - shared.state.textinfo = description if hypernetwork_dir is not None and steps_done % save_hypernetwork_every == 0: # Before saving, change name to match current checkpoint. hypernetwork_name_every = f'{hypernetwork_name}-{steps_done}' @@ -701,7 +700,8 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, torch.cuda.set_rng_state_all(cuda_rng_state) hypernetwork.train() if image is not None: - shared.state.current_image = image + shared.state.assign_current_image(image) + last_saved_image, last_text_info = images.save_image(image, images_dir, "", p.seed, p.prompt, shared.opts.samples_format, processed.infotexts[0], p=p, forced_filename=forced_filename, save_to_dirs=False) last_saved_image += f", prompt: {preview_text}" diff --git a/modules/img2img.py b/modules/img2img.py index f62783c6..f4a03c57 100644 --- a/modules/img2img.py +++ b/modules/img2img.py @@ -59,7 +59,7 @@ def process_batch(p, input_dir, output_dir, args): processed_image.save(os.path.join(output_dir, filename)) -def img2img(mode: int, prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, init_img, sketch, init_img_with_mask, inpaint_color_sketch, inpaint_color_sketch_orig, init_img_inpaint, init_mask_inpaint, steps: int, sampler_index: int, mask_blur: int, mask_alpha: float, inpainting_fill: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, denoising_strength: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, resize_mode: int, inpaint_full_res: bool, inpaint_full_res_padding: int, inpainting_mask_invert: int, img2img_batch_input_dir: str, img2img_batch_output_dir: str, *args): +def img2img(id_task: str, mode: int, prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, init_img, sketch, init_img_with_mask, inpaint_color_sketch, inpaint_color_sketch_orig, init_img_inpaint, init_mask_inpaint, steps: int, sampler_index: int, mask_blur: int, mask_alpha: float, inpainting_fill: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, denoising_strength: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, resize_mode: int, inpaint_full_res: bool, inpaint_full_res_padding: int, inpainting_mask_invert: int, img2img_batch_input_dir: str, img2img_batch_output_dir: str, *args): is_batch = mode == 5 if mode == 0: # img2img diff --git a/modules/progress.py b/modules/progress.py new file mode 100644 index 00000000..3327b883 --- /dev/null +++ b/modules/progress.py @@ -0,0 +1,96 @@ +import base64 +import io +import time + +import gradio as gr +from pydantic import BaseModel, Field + +from modules.shared import opts + +import modules.shared as shared + + +current_task = None +pending_tasks = {} +finished_tasks = [] + + +def start_task(id_task): + global current_task + + current_task = id_task + pending_tasks.pop(id_task, None) + + +def finish_task(id_task): + global current_task + + if current_task == id_task: + current_task = None + + finished_tasks.append(id_task) + if len(finished_tasks) > 16: + finished_tasks.pop(0) + + +def add_task_to_queue(id_job): + pending_tasks[id_job] = time.time() + + +class ProgressRequest(BaseModel): + id_task: str = Field(default=None, title="Task ID", description="id of the task to get progress for") + id_live_preview: int = Field(default=-1, title="Live preview image ID", description="id of last received last preview image") + + +class ProgressResponse(BaseModel): + active: bool = Field(title="Whether the task is being worked on right now") + queued: bool = Field(title="Whether the task is in queue") + completed: bool = Field(title="Whether the task has already finished") + progress: float = Field(default=None, title="Progress", description="The progress with a range of 0 to 1") + eta: float = Field(default=None, title="ETA in secs") + live_preview: str = Field(default=None, title="Live preview image", description="Current live preview; a data: uri") + id_live_preview: int = Field(default=None, title="Live preview image ID", description="Send this together with next request to prevent receiving same image") + textinfo: str = Field(default=None, title="Info text", description="Info text used by WebUI.") + + +def setup_progress_api(app): + return app.add_api_route("/internal/progress", progressapi, methods=["POST"], response_model=ProgressResponse) + + +def progressapi(req: ProgressRequest): + active = req.id_task == current_task + queued = req.id_task in pending_tasks + completed = req.id_task in finished_tasks + + if not active: + return ProgressResponse(active=active, queued=queued, completed=completed, id_live_preview=-1, textinfo="In queue..." if queued else "Waiting...") + + progress = 0 + + if shared.state.job_count > 0: + progress += shared.state.job_no / shared.state.job_count + if shared.state.sampling_steps > 0: + progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps + + progress = min(progress, 1) + + elapsed_since_start = time.time() - shared.state.time_start + predicted_duration = elapsed_since_start / progress if progress > 0 else None + eta = predicted_duration - elapsed_since_start if predicted_duration is not None else None + + id_live_preview = req.id_live_preview + shared.state.set_current_image() + if opts.live_previews_enable and shared.state.id_live_preview != req.id_live_preview: + image = shared.state.current_image + if image is not None: + buffered = io.BytesIO() + image.save(buffered, format="png") + live_preview = 'data:image/png;base64,' + base64.b64encode(buffered.getvalue()).decode("ascii") + id_live_preview = shared.state.id_live_preview + else: + live_preview = None + else: + live_preview = None + + return ProgressResponse(active=active, queued=queued, completed=completed, progress=progress, eta=eta, live_preview=live_preview, id_live_preview=id_live_preview, textinfo=shared.state.textinfo) + diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py index 7616fded..76e0e0d5 100644 --- a/modules/sd_samplers.py +++ b/modules/sd_samplers.py @@ -140,7 +140,7 @@ def store_latent(decoded): if opts.live_previews_enable and opts.show_progress_every_n_steps > 0 and shared.state.sampling_step % opts.show_progress_every_n_steps == 0: if not shared.parallel_processing_allowed: - shared.state.current_image = sample_to_image(decoded) + shared.state.assign_current_image(sample_to_image(decoded)) class InterruptedException(BaseException): diff --git a/modules/shared.py b/modules/shared.py index 51df056c..de99aca9 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -152,6 +152,7 @@ def reload_hypernetworks(): hypernetwork.load_hypernetwork(opts.sd_hypernetwork) + class State: skipped = False interrupted = False @@ -165,6 +166,7 @@ class State: current_latent = None current_image = None current_image_sampling_step = 0 + id_live_preview = 0 textinfo = None time_start = None need_restart = False @@ -207,6 +209,7 @@ class State: self.current_latent = None self.current_image = None self.current_image_sampling_step = 0 + self.id_live_preview = 0 self.skipped = False self.interrupted = False self.textinfo = None @@ -220,8 +223,8 @@ class State: devices.torch_gc() - """sets self.current_image from self.current_latent if enough sampling steps have been made after the last call to this""" def set_current_image(self): + """sets self.current_image from self.current_latent if enough sampling steps have been made after the last call to this""" if not parallel_processing_allowed: return @@ -234,12 +237,16 @@ class State: import modules.sd_samplers if opts.show_progress_grid: - self.current_image = modules.sd_samplers.samples_to_image_grid(self.current_latent) + self.assign_current_image(modules.sd_samplers.samples_to_image_grid(self.current_latent)) else: - self.current_image = modules.sd_samplers.sample_to_image(self.current_latent) + self.assign_current_image(modules.sd_samplers.sample_to_image(self.current_latent)) self.current_image_sampling_step = self.sampling_step + def assign_current_image(self, image): + self.current_image = image + self.id_live_preview += 1 + state = State() state.server_start = time.time() @@ -424,8 +431,6 @@ options_templates.update(options_section(('interrogate', "Interrogate Options"), })) options_templates.update(options_section(('ui', "User interface"), { - "show_progressbar": OptionInfo(True, "Show progressbar"), - "show_progress_grid": OptionInfo(True, "Show previews of all images generated in a batch as a grid"), "return_grid": OptionInfo(True, "Show grid in results for web"), "do_not_show_images": OptionInfo(False, "Do not show any images in results for web"), "add_model_hash_to_info": OptionInfo(True, "Add model hash to generation information"), @@ -446,6 +451,7 @@ options_templates.update(options_section(('ui', "User interface"), { options_templates.update(options_section(('ui', "Live previews"), { "live_previews_enable": OptionInfo(True, "Show live previews of the created image"), + "show_progress_grid": OptionInfo(True, "Show previews of all images generated in a batch as a grid"), "show_progress_every_n_steps": OptionInfo(10, "Show new live preview image every N sampling steps. Set to -1 to show after completion of batch.", gr.Slider, {"minimum": -1, "maximum": 32, "step": 1}), "show_progress_type": OptionInfo("Approx NN", "Image creation progress preview mode", gr.Radio, {"choices": ["Full", "Approx NN", "Approx cheap"]}), "live_preview_content": OptionInfo("Prompt", "Live preview subject", gr.Radio, {"choices": ["Combined", "Prompt", "Negative prompt"]}), diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py index 3c1042ad..64abff4d 100644 --- a/modules/textual_inversion/preprocess.py +++ b/modules/textual_inversion/preprocess.py @@ -12,7 +12,7 @@ from modules.shared import opts, cmd_opts from modules.textual_inversion import autocrop -def preprocess(process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru=False, split_threshold=0.5, overlap_ratio=0.2, process_focal_crop=False, process_focal_crop_face_weight=0.9, process_focal_crop_entropy_weight=0.3, process_focal_crop_edges_weight=0.5, process_focal_crop_debug=False): +def preprocess(id_task, process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru=False, split_threshold=0.5, overlap_ratio=0.2, process_focal_crop=False, process_focal_crop_face_weight=0.9, process_focal_crop_entropy_weight=0.3, process_focal_crop_edges_weight=0.5, process_focal_crop_debug=False): try: if process_caption: shared.interrogator.load() diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 63935878..7e4a6d24 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -345,7 +345,7 @@ def validate_train_inputs(model_name, learn_rate, batch_size, gradient_step, dat assert log_directory, "Log directory is empty" -def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_embedding_every, template_filename, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): +def train_embedding(id_task, embedding_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, varsize, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_embedding_every, template_filename, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height): save_embedding_every = save_embedding_every or 0 create_image_every = create_image_every or 0 template_file = textual_inversion_templates.get(template_filename, None) @@ -510,7 +510,6 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ description = f"Training textual inversion [Epoch {epoch_num}: {epoch_step+1}/{steps_per_epoch}] loss: {loss_step:.7f}" pbar.set_description(description) - shared.state.textinfo = description if embedding_dir is not None and steps_done % save_embedding_every == 0: # Before saving, change name to match current checkpoint. embedding_name_every = f'{embedding_name}-{steps_done}' @@ -560,7 +559,8 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_ shared.sd_model.first_stage_model.to(devices.cpu) if image is not None: - shared.state.current_image = image + shared.state.assign_current_image(image) + last_saved_image, last_text_info = images.save_image(image, images_dir, "", p.seed, p.prompt, shared.opts.samples_format, processed.infotexts[0], p=p, forced_filename=forced_filename, save_to_dirs=False) last_saved_image += f", prompt: {preview_text}" diff --git a/modules/txt2img.py b/modules/txt2img.py index 38b5f591..ca5d4550 100644 --- a/modules/txt2img.py +++ b/modules/txt2img.py @@ -8,7 +8,7 @@ import modules.processing as processing from modules.ui import plaintext_to_html -def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, hr_second_pass_steps: int, hr_resize_x: int, hr_resize_y: int, *args): +def txt2img(id_task: str, prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, hr_second_pass_steps: int, hr_resize_x: int, hr_resize_y: int, *args): p = StableDiffusionProcessingTxt2Img( sd_model=shared.sd_model, outpath_samples=opts.outdir_samples or opts.outdir_txt2img_samples, diff --git a/modules/ui.py b/modules/ui.py index 2425c66f..ff33236b 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -356,7 +356,7 @@ def create_toprow(is_img2img): button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") with gr.Column(scale=1): - with gr.Row(): + with gr.Row(elem_id=f"{id_part}_generate_box"): skip = gr.Button('Skip', elem_id=f"{id_part}_skip") interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') @@ -384,9 +384,7 @@ def create_toprow(is_img2img): def setup_progressbar(*args, **kwargs): - import modules.ui_progress - - modules.ui_progress.setup_progressbar(*args, **kwargs) + pass def apply_setting(key, value): @@ -479,8 +477,8 @@ Requested path was: {f} else: sp.Popen(["xdg-open", path]) - with gr.Column(variant='panel'): - with gr.Group(): + with gr.Column(variant='panel', elem_id=f"{tabname}_results"): + with gr.Group(elem_id=f"{tabname}_gallery_container"): result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) generation_info = None @@ -595,15 +593,6 @@ def create_ui(): dummy_component = gr.Label(visible=False) txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) - with gr.Row(elem_id='txt2img_progress_row'): - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="txt2img_progressbar") - txt2img_preview = gr.Image(elem_id='txt2img_preview', visible=False) - setup_progressbar(progressbar, txt2img_preview, 'txt2img') - with gr.Row().style(equal_height=False): with gr.Column(variant='panel', elem_id="txt2img_settings"): for category in ordered_ui_categories(): @@ -682,6 +671,7 @@ def create_ui(): fn=wrap_gradio_gpu_call(modules.txt2img.txt2img, extra_outputs=[None, '', '']), _js="submit", inputs=[ + dummy_component, txt2img_prompt, txt2img_negative_prompt, txt2img_prompt_style, @@ -782,16 +772,7 @@ def create_ui(): with gr.Blocks(analytics_enabled=False) as img2img_interface: img2img_prompt, img2img_prompt_style, img2img_negative_prompt, img2img_prompt_style2, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) - with gr.Row(elem_id='img2img_progress_row'): - img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) - - with gr.Column(scale=1): - pass - - with gr.Column(scale=1): - progressbar = gr.HTML(elem_id="img2img_progressbar") - img2img_preview = gr.Image(elem_id='img2img_preview', visible=False) - setup_progressbar(progressbar, img2img_preview, 'img2img') + img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) with FormRow().style(equal_height=False): with gr.Column(variant='panel', elem_id="img2img_settings"): @@ -958,6 +939,7 @@ def create_ui(): fn=wrap_gradio_gpu_call(modules.img2img.img2img, extra_outputs=[None, '', '']), _js="submit_img2img", inputs=[ + dummy_component, dummy_component, img2img_prompt, img2img_negative_prompt, @@ -1335,15 +1317,11 @@ def create_ui(): script_callbacks.ui_train_tabs_callback(params) - with gr.Column(): - progressbar = gr.HTML(elem_id="ti_progressbar") + with gr.Column(elem_id='ti_gallery_container'): ti_output = gr.Text(elem_id="ti_output", value="", show_label=False) - ti_gallery = gr.Gallery(label='Output', show_label=False, elem_id='ti_gallery').style(grid=4) - ti_preview = gr.Image(elem_id='ti_preview', visible=False) ti_progress = gr.HTML(elem_id="ti_progress", value="") ti_outcome = gr.HTML(elem_id="ti_error", value="") - setup_progressbar(progressbar, ti_preview, 'ti', textinfo=ti_progress) create_embedding.click( fn=modules.textual_inversion.ui.create_embedding, @@ -1384,6 +1362,7 @@ def create_ui(): fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.preprocess, extra_outputs=[gr.update()]), _js="start_training_textual_inversion", inputs=[ + dummy_component, process_src, process_dst, process_width, @@ -1411,6 +1390,7 @@ def create_ui(): fn=wrap_gradio_gpu_call(modules.textual_inversion.ui.train_embedding, extra_outputs=[gr.update()]), _js="start_training_textual_inversion", inputs=[ + dummy_component, train_embedding_name, embedding_learn_rate, batch_size, @@ -1443,6 +1423,7 @@ def create_ui(): fn=wrap_gradio_gpu_call(modules.hypernetworks.ui.train_hypernetwork, extra_outputs=[gr.update()]), _js="start_training_textual_inversion", inputs=[ + dummy_component, train_hypernetwork_name, hypernetwork_learn_rate, batch_size, diff --git a/modules/ui_progress.py b/modules/ui_progress.py deleted file mode 100644 index 7cd312e4..00000000 --- a/modules/ui_progress.py +++ /dev/null @@ -1,101 +0,0 @@ -import time - -import gradio as gr - -from modules.shared import opts - -import modules.shared as shared - - -def calc_time_left(progress, threshold, label, force_display, show_eta): - if progress == 0: - return "" - else: - time_since_start = time.time() - shared.state.time_start - eta = (time_since_start/progress) - eta_relative = eta-time_since_start - if (eta_relative > threshold and show_eta) or force_display: - if eta_relative > 3600: - return label + time.strftime('%H:%M:%S', time.gmtime(eta_relative)) - elif eta_relative > 60: - return label + time.strftime('%M:%S', time.gmtime(eta_relative)) - else: - return label + time.strftime('%Ss', time.gmtime(eta_relative)) - else: - return "" - - -def check_progress_call(id_part): - if shared.state.job_count == 0: - return "", gr.update(visible=False), gr.update(visible=False), gr.update(visible=False) - - progress = 0 - - if shared.state.job_count > 0: - progress += shared.state.job_no / shared.state.job_count - if shared.state.sampling_steps > 0: - progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps - - # Show progress percentage and time left at the same moment, and base it also on steps done - show_eta = progress >= 0.01 or shared.state.sampling_step >= 10 - - time_left = calc_time_left(progress, 1, " ETA: ", shared.state.time_left_force_display, show_eta) - if time_left != "": - shared.state.time_left_force_display = True - - progress = min(progress, 1) - - progressbar = "" - if opts.show_progressbar: - progressbar = f"""
{" " * 2 + str(int(progress*100))+"%" + time_left if show_eta else ""}
""" - - image = gr.update(visible=False) - preview_visibility = gr.update(visible=False) - - if opts.live_previews_enable: - shared.state.set_current_image() - image = shared.state.current_image - - if image is None: - image = gr.update(value=None) - else: - preview_visibility = gr.update(visible=True) - - if shared.state.textinfo is not None: - textinfo_result = gr.HTML.update(value=shared.state.textinfo, visible=True) - else: - textinfo_result = gr.update(visible=False) - - return f"

{progressbar}

", preview_visibility, image, textinfo_result - - -def check_progress_call_initial(id_part): - shared.state.job_count = -1 - shared.state.current_latent = None - shared.state.current_image = None - shared.state.textinfo = None - shared.state.time_start = time.time() - shared.state.time_left_force_display = False - - return check_progress_call(id_part) - - -def setup_progressbar(progressbar, preview, id_part, textinfo=None): - if textinfo is None: - textinfo = gr.HTML(visible=False) - - check_progress = gr.Button('Check progress', elem_id=f"{id_part}_check_progress", visible=False) - check_progress.click( - fn=lambda: check_progress_call(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) - - check_progress_initial = gr.Button('Check progress (first)', elem_id=f"{id_part}_check_progress_initial", visible=False) - check_progress_initial.click( - fn=lambda: check_progress_call_initial(id_part), - show_progress=False, - inputs=[], - outputs=[progressbar, preview, preview, textinfo], - ) -- cgit v1.2.3 From 388708f7b13dfbc890135cad678bfbcebd7baf37 Mon Sep 17 00:00:00 2001 From: pangbo13 <373108669@qq.com> Date: Mon, 16 Jan 2023 00:56:24 +0800 Subject: fix when show_progress_every_n_steps == -1 --- modules/shared.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index de99aca9..f857ccde 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -228,7 +228,7 @@ class State: if not parallel_processing_allowed: return - if self.sampling_step - self.current_image_sampling_step >= opts.show_progress_every_n_steps and opts.live_previews_enable: + if self.sampling_step - self.current_image_sampling_step >= opts.show_progress_every_n_steps and opts.live_previews_enable and opts.show_progress_every_n_steps != -1: self.do_set_current_image() def do_set_current_image(self): -- cgit v1.2.3 From 16f410893eb96c7810cbbd812541ba35e0e92524 Mon Sep 17 00:00:00 2001 From: AngelBottomless <35677394+aria1th@users.noreply.github.com> Date: Mon, 16 Jan 2023 02:08:47 +0900 Subject: fix missing 'mean loss' for tensorboard integration --- modules/hypernetworks/hypernetwork.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index ae6af516..bbd1f673 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -644,7 +644,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi if shared.opts.training_enable_tensorboard: epoch_num = hypernetwork.step // len(ds) epoch_step = hypernetwork.step - (epoch_num * len(ds)) + 1 - + mean_loss = sum(sum(x) for x in loss_dict.values()) / sum(len(x) for x in loss_dict.values()) textual_inversion.tensorboard_add(tensorboard_writer, loss=mean_loss, global_step=hypernetwork.step, step=epoch_step, learn_rate=scheduler.learn_rate, epoch_num=epoch_num) textual_inversion.write_loss(log_directory, "hypernetwork_loss.csv", hypernetwork.step, steps_per_epoch, { -- cgit v1.2.3 From a534bdfc801e0c83e378dfaa2d04cf865d7109f9 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 15 Jan 2023 20:27:39 +0300 Subject: add setting for progressbar update period --- modules/shared.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index de99aca9..3483db1c 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -455,6 +455,7 @@ options_templates.update(options_section(('ui', "Live previews"), { "show_progress_every_n_steps": OptionInfo(10, "Show new live preview image every N sampling steps. Set to -1 to show after completion of batch.", gr.Slider, {"minimum": -1, "maximum": 32, "step": 1}), "show_progress_type": OptionInfo("Approx NN", "Image creation progress preview mode", gr.Radio, {"choices": ["Full", "Approx NN", "Approx cheap"]}), "live_preview_content": OptionInfo("Prompt", "Live preview subject", gr.Radio, {"choices": ["Combined", "Prompt", "Negative prompt"]}), + "live_preview_refresh_period": OptionInfo(1000, "Progressbar/preview update period, in milliseconds") })) options_templates.update(options_section(('sampler-params', "Sampler parameters"), { -- cgit v1.2.3 From b6ce041cdf722b400df9b5eac306d0cb049923d7 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 15 Jan 2023 20:29:48 +0300 Subject: put interrupt and skip buttons back where they were --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index ff33236b..7a357f9a 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -357,8 +357,8 @@ def create_toprow(is_img2img): with gr.Column(scale=1): with gr.Row(elem_id=f"{id_part}_generate_box"): - skip = gr.Button('Skip', elem_id=f"{id_part}_skip") interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") + skip = gr.Button('Skip', elem_id=f"{id_part}_skip") submit = gr.Button('Generate', elem_id=f"{id_part}_generate", variant='primary') skip.click( -- cgit v1.2.3 From 110d1a2d598bcfacffe3d524df1a3422b4cbd8ec Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Sun, 15 Jan 2023 12:41:00 -0500 Subject: add fields to settings file --- modules/textual_inversion/logging.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/textual_inversion/logging.py b/modules/textual_inversion/logging.py index 31e50b64..734a4b6f 100644 --- a/modules/textual_inversion/logging.py +++ b/modules/textual_inversion/logging.py @@ -2,7 +2,7 @@ import datetime import json import os -saved_params_shared = {"model_name", "model_hash", "initial_step", "num_of_dataset_images", "learn_rate", "batch_size", "clip_grad_mode", "clip_grad_value", "gradient_step", "data_root", "log_directory", "training_width", "training_height", "steps", "create_image_every", "template_file"} +saved_params_shared = {"model_name", "model_hash", "initial_step", "num_of_dataset_images", "learn_rate", "batch_size", "clip_grad_mode", "clip_grad_value", "gradient_step", "data_root", "log_directory", "training_width", "training_height", "steps", "create_image_every", "template_file", "gradient_step", "latent_sampling_method"} saved_params_ti = {"embedding_name", "num_vectors_per_token", "save_embedding_every", "save_image_with_stored_embedding"} saved_params_hypernet = {"hypernetwork_name", "layer_structure", "activation_func", "weight_init", "add_layer_norm", "use_dropout", "save_hypernetwork_every"} saved_params_all = saved_params_shared | saved_params_ti | saved_params_hypernet -- cgit v1.2.3 From 598f7fcd84f655dd204ad5e258dc1c41cc806cde Mon Sep 17 00:00:00 2001 From: aria1th <35677394+aria1th@users.noreply.github.com> Date: Mon, 16 Jan 2023 02:46:21 +0900 Subject: Fix loss_dict problem --- modules/hypernetworks/hypernetwork.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index bbd1f673..438e3e9f 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -561,6 +561,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi _loss_step = 0 #internal # size = len(ds.indexes) # loss_dict = defaultdict(lambda : deque(maxlen = 1024)) + loss_logging = [] # losses = torch.zeros((size,)) # previous_mean_losses = [0] # previous_mean_loss = 0 @@ -601,6 +602,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi else: c = stack_conds(batch.cond).to(devices.device, non_blocking=pin_memory) loss = shared.sd_model(x, c)[0] / gradient_step + loss_logging.append(loss.item()) del x del c @@ -644,7 +646,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi if shared.opts.training_enable_tensorboard: epoch_num = hypernetwork.step // len(ds) epoch_step = hypernetwork.step - (epoch_num * len(ds)) + 1 - mean_loss = sum(sum(x) for x in loss_dict.values()) / sum(len(x) for x in loss_dict.values()) + mean_loss = sum(loss_logging) / len(loss_logging) textual_inversion.tensorboard_add(tensorboard_writer, loss=mean_loss, global_step=hypernetwork.step, step=epoch_step, learn_rate=scheduler.learn_rate, epoch_num=epoch_num) textual_inversion.write_loss(log_directory, "hypernetwork_loss.csv", hypernetwork.step, steps_per_epoch, { -- cgit v1.2.3 From 13445738d974edcca5ff2f4f8f3833c1f3433e5e Mon Sep 17 00:00:00 2001 From: aria1th <35677394+aria1th@users.noreply.github.com> Date: Mon, 16 Jan 2023 03:02:54 +0900 Subject: Fix tensorboard related functions --- modules/hypernetworks/hypernetwork.py | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 438e3e9f..c963fc40 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -561,7 +561,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi _loss_step = 0 #internal # size = len(ds.indexes) # loss_dict = defaultdict(lambda : deque(maxlen = 1024)) - loss_logging = [] + loss_logging = deque(maxlen=len(ds) * 3) # this should be configurable parameter, this is 3 * epoch(dataset size) # losses = torch.zeros((size,)) # previous_mean_losses = [0] # previous_mean_loss = 0 @@ -602,7 +602,6 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi else: c = stack_conds(batch.cond).to(devices.device, non_blocking=pin_memory) loss = shared.sd_model(x, c)[0] / gradient_step - loss_logging.append(loss.item()) del x del c @@ -612,7 +611,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi # go back until we reach gradient accumulation steps if (j + 1) % gradient_step != 0: continue - + loss_logging.append(_loss_step) if clip_grad: clip_grad(weights, clip_grad_sched.learn_rate) @@ -690,9 +689,6 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi processed = processing.process_images(p) image = processed.images[0] if len(processed.images) > 0 else None - - if shared.opts.training_enable_tensorboard and shared.opts.training_tensorboard_save_images: - textual_inversion.tensorboard_add_image(tensorboard_writer, f"Validation at epoch {epoch_num}", image, hypernetwork.step) if unload: shared.sd_model.cond_stage_model.to(devices.cpu) @@ -703,7 +699,10 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi hypernetwork.train() if image is not None: shared.state.assign_current_image(image) - + if shared.opts.training_enable_tensorboard and shared.opts.training_tensorboard_save_images: + textual_inversion.tensorboard_add_image(tensorboard_writer, + f"Validation at epoch {epoch_num}", image, + hypernetwork.step) last_saved_image, last_text_info = images.save_image(image, images_dir, "", p.seed, p.prompt, shared.opts.samples_format, processed.infotexts[0], p=p, forced_filename=forced_filename, save_to_dirs=False) last_saved_image += f", prompt: {preview_text}" -- cgit v1.2.3 From 8e2aeee4a127b295bfc880800e4a312e0f049b85 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 15 Jan 2023 22:29:53 +0300 Subject: add BREAK keyword to end current text chunk and start the next --- modules/prompt_parser.py | 7 ++++++- modules/sd_hijack_clip.py | 17 +++++++++++++---- 2 files changed, 19 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/prompt_parser.py b/modules/prompt_parser.py index 870218db..69665372 100644 --- a/modules/prompt_parser.py +++ b/modules/prompt_parser.py @@ -274,6 +274,7 @@ re_attention = re.compile(r""" : """, re.X) +re_break = re.compile(r"\s*\bBREAK\b\s*", re.S) def parse_prompt_attention(text): """ @@ -339,7 +340,11 @@ def parse_prompt_attention(text): elif text == ']' and len(square_brackets) > 0: multiply_range(square_brackets.pop(), square_bracket_multiplier) else: - res.append([text, 1.0]) + parts = re.split(re_break, text) + for i, part in enumerate(parts): + if i > 0: + res.append(["BREAK", -1]) + res.append([part, 1.0]) for pos in round_brackets: multiply_range(pos, round_bracket_multiplier) diff --git a/modules/sd_hijack_clip.py b/modules/sd_hijack_clip.py index 852afc66..9fa5c5c5 100644 --- a/modules/sd_hijack_clip.py +++ b/modules/sd_hijack_clip.py @@ -96,13 +96,18 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): token_count = 0 last_comma = -1 - def next_chunk(): - """puts current chunk into the list of results and produces the next one - empty""" + def next_chunk(is_last=False): + """puts current chunk into the list of results and produces the next one - empty; + if is_last is true, tokens tokens at the end won't add to token_count""" nonlocal token_count nonlocal last_comma nonlocal chunk - token_count += len(chunk.tokens) + if is_last: + token_count += len(chunk.tokens) + else: + token_count += self.chunk_length + to_add = self.chunk_length - len(chunk.tokens) if to_add > 0: chunk.tokens += [self.id_end] * to_add @@ -116,6 +121,10 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): chunk = PromptChunk() for tokens, (text, weight) in zip(tokenized, parsed): + if text == 'BREAK' and weight == -1: + next_chunk() + continue + position = 0 while position < len(tokens): token = tokens[position] @@ -159,7 +168,7 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module): position += embedding_length_in_tokens if len(chunk.tokens) > 0 or len(chunks) == 0: - next_chunk() + next_chunk(is_last=True) return chunks, token_count -- cgit v1.2.3 From 89314e79da21ac71ad3133ccf5ac3e85d4c24052 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 15 Jan 2023 23:23:16 +0300 Subject: fix an error that happens when you send an empty image from txt2img to img2img --- modules/generation_parameters_copypaste.py | 3 +++ 1 file changed, 3 insertions(+) (limited to 'modules') diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 593d99ef..a381ff59 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -37,6 +37,9 @@ def quote(text): def image_from_url_text(filedata): + if filedata is None: + return None + if type(filedata) == list and len(filedata) > 0 and type(filedata[0]) == dict and filedata[0].get("is_file", False): filedata = filedata[0] -- cgit v1.2.3 From 3db22e6ee45193559a2c3ba44ab672b067245f99 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 15 Jan 2023 23:32:38 +0300 Subject: rename masking to inpaint in UI make inpaint go to the right place for users who don't have it in config string --- modules/shared.py | 2 +- modules/ui.py | 6 +++--- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 3bdc375b..f06ae610 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -116,7 +116,7 @@ restricted_opts = { } ui_reorder_categories = [ - "masking", + "inpaint", "sampler", "dimensions", "cfg", diff --git a/modules/ui.py b/modules/ui.py index b3d4af3e..20b66165 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -570,9 +570,9 @@ def create_sampler_and_steps_selection(choices, tabname): def ordered_ui_categories(): - user_order = {x.strip(): i for i, x in enumerate(shared.opts.ui_reorder.split(","))} + user_order = {x.strip(): i * 2 + 1 for i, x in enumerate(shared.opts.ui_reorder.split(","))} - for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] + 1000)): + for i, category in sorted(enumerate(shared.ui_reorder_categories), key=lambda x: user_order.get(x[1], x[0] * 2 + 0)): yield category @@ -889,7 +889,7 @@ def create_ui(): with FormGroup(elem_id="img2img_script_container"): custom_inputs = modules.scripts.scripts_img2img.setup_ui() - elif category == "masking": + elif category == "inpaint": with FormGroup(elem_id="inpaint_controls", visible=False) as inpaint_controls: with FormRow(): mask_blur = gr.Slider(label='Mask blur', minimum=0, maximum=64, step=1, value=4, elem_id="img2img_mask_blur") -- cgit v1.2.3 From 3f887f7f61d69fa699a272166b79fdb787e9ce1d Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 16 Jan 2023 00:44:46 +0300 Subject: support old configs that say "auto" for ssd_vae change sd_vae_as_default to True by default as it's a more sensible setting --- modules/sd_vae.py | 6 ++++-- modules/shared.py | 2 +- 2 files changed, 5 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/sd_vae.py b/modules/sd_vae.py index add5cecf..e9c6bb40 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -94,8 +94,10 @@ def resolve_vae(checkpoint_file): if shared.cmd_opts.vae_path is not None: return shared.cmd_opts.vae_path, 'from commandline argument' + is_automatic = shared.opts.sd_vae in {"Automatic", "auto"} # "auto" for people with old config + vae_near_checkpoint = find_vae_near_checkpoint(checkpoint_file) - if vae_near_checkpoint is not None and (shared.opts.sd_vae_as_default or shared.opts.sd_vae == "Automatic"): + if vae_near_checkpoint is not None and (shared.opts.sd_vae_as_default or is_automatic): return vae_near_checkpoint, 'found near the checkpoint' if shared.opts.sd_vae == "None": @@ -105,7 +107,7 @@ def resolve_vae(checkpoint_file): if vae_from_options is not None: return vae_from_options, 'specified in settings' - if shared.opts.sd_vae != "Automatic": + if is_automatic: print(f"Couldn't find VAE named {shared.opts.sd_vae}; using None instead") return None, None diff --git a/modules/shared.py b/modules/shared.py index f06ae610..c5fc250e 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -394,7 +394,7 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "sd_checkpoint_cache": OptionInfo(0, "Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), "sd_vae_checkpoint_cache": OptionInfo(0, "VAE Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), "sd_vae": OptionInfo("Automatic", "SD VAE", gr.Dropdown, lambda: {"choices": ["Automatic", "None"] + list(sd_vae.vae_dict)}, refresh=sd_vae.refresh_vae_list), - "sd_vae_as_default": OptionInfo(False, "Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them"), + "sd_vae_as_default": OptionInfo(True, "Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them"), "sd_hypernetwork": OptionInfo("None", "Hypernetwork", gr.Dropdown, lambda: {"choices": ["None"] + [x for x in hypernetworks.keys()]}, refresh=reload_hypernetworks), "sd_hypernetwork_strength": OptionInfo(1.0, "Hypernetwork strength", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.001}), "inpainting_mask_weight": OptionInfo(1.0, "Inpainting conditioning mask strength", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), -- cgit v1.2.3 From ff6a5bcec1ce25aa8f08b157ea957d764be23d8d Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 16 Jan 2023 01:28:20 +0300 Subject: bugfix for previous commit --- modules/sd_vae.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_vae.py b/modules/sd_vae.py index e9c6bb40..b2af2ce7 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -107,7 +107,7 @@ def resolve_vae(checkpoint_file): if vae_from_options is not None: return vae_from_options, 'specified in settings' - if is_automatic: + if not is_automatic: print(f"Couldn't find VAE named {shared.opts.sd_vae}; using None instead") return None, None -- cgit v1.2.3 From 064983c0adb00cd9e88d2f06f66c9a1d5bc116c3 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 16 Jan 2023 12:56:30 +0300 Subject: return an option to hide progressbar --- modules/shared.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index c5fc250e..483c4c62 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -451,6 +451,7 @@ options_templates.update(options_section(('ui', "User interface"), { })) options_templates.update(options_section(('ui', "Live previews"), { + "show_progressbar": OptionInfo(True, "Show progressbar"), "live_previews_enable": OptionInfo(True, "Show live previews of the created image"), "show_progress_grid": OptionInfo(True, "Show previews of all images generated in a batch as a grid"), "show_progress_every_n_steps": OptionInfo(10, "Show new live preview image every N sampling steps. Set to -1 to show after completion of batch.", gr.Slider, {"minimum": -1, "maximum": 32, "step": 1}), -- cgit v1.2.3 From 9991967f40120b88a1dc925fdf7d747d5e016888 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 16 Jan 2023 22:59:46 +0300 Subject: Add a check and explanation for tensor with all NaNs. --- modules/devices.py | 28 ++++++++++++++++++++++++++++ modules/processing.py | 3 +++ modules/sd_samplers.py | 2 ++ 3 files changed, 33 insertions(+) (limited to 'modules') diff --git a/modules/devices.py b/modules/devices.py index caeb0276..6f034948 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -106,6 +106,33 @@ def autocast(disable=False): return torch.autocast("cuda") +class NansException(Exception): + pass + + +def test_for_nans(x, where): + from modules import shared + + if not torch.all(torch.isnan(x)).item(): + return + + if where == "unet": + message = "A tensor with all NaNs was produced in Unet." + + if not shared.cmd_opts.no_half: + message += " This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try using --no-half commandline argument to fix this." + + elif where == "vae": + message = "A tensor with all NaNs was produced in VAE." + + if not shared.cmd_opts.no_half and not shared.cmd_opts.no_half_vae: + message += " This could be because there's not enough precision to represent the picture. Try adding --no-half-vae commandline argument to fix this." + else: + message = "A tensor with all NaNs was produced." + + raise NansException(message) + + # MPS workaround for https://github.com/pytorch/pytorch/issues/79383 orig_tensor_to = torch.Tensor.to def tensor_to_fix(self, *args, **kwargs): @@ -156,3 +183,4 @@ if has_mps(): torch.Tensor.cumsum = lambda self, *args, **kwargs: ( cumsum_fix(self, orig_Tensor_cumsum, *args, **kwargs) ) orig_narrow = torch.narrow torch.narrow = lambda *args, **kwargs: ( orig_narrow(*args, **kwargs).clone() ) + diff --git a/modules/processing.py b/modules/processing.py index 849f6b19..ab7b3b7d 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -608,6 +608,9 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: samples_ddim = p.sample(conditioning=c, unconditional_conditioning=uc, seeds=seeds, subseeds=subseeds, subseed_strength=p.subseed_strength, prompts=prompts) x_samples_ddim = [decode_first_stage(p.sd_model, samples_ddim[i:i+1].to(dtype=devices.dtype_vae))[0].cpu() for i in range(samples_ddim.size(0))] + for x in x_samples_ddim: + devices.test_for_nans(x, "vae") + x_samples_ddim = torch.stack(x_samples_ddim).float() x_samples_ddim = torch.clamp((x_samples_ddim + 1.0) / 2.0, min=0.0, max=1.0) diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py index 76e0e0d5..6261d1f7 100644 --- a/modules/sd_samplers.py +++ b/modules/sd_samplers.py @@ -351,6 +351,8 @@ class CFGDenoiser(torch.nn.Module): x_out[-uncond.shape[0]:] = self.inner_model(x_in[-uncond.shape[0]:], sigma_in[-uncond.shape[0]:], cond={"c_crossattn": [uncond], "c_concat": [image_cond_in[-uncond.shape[0]:]]}) + devices.test_for_nans(x_out, "unet") + if opts.live_preview_content == "Prompt": store_latent(x_out[0:uncond.shape[0]]) elif opts.live_preview_content == "Negative prompt": -- cgit v1.2.3 From e0e80050091ea7f58ae17c69f31d1b5de5e0ae20 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 16 Jan 2023 23:09:08 +0300 Subject: make StableDiffusionProcessing class not hold a reference to shared.sd_model object --- modules/processing.py | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index ab7b3b7d..9c3673de 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -94,7 +94,7 @@ def txt2img_image_conditioning(sd_model, x, width, height): return image_conditioning -class StableDiffusionProcessing(): +class StableDiffusionProcessing: """ The first set of paramaters: sd_models -> do_not_reload_embeddings represent the minimum required to create a StableDiffusionProcessing """ @@ -102,7 +102,6 @@ class StableDiffusionProcessing(): if sampler_index is not None: print("sampler_index argument for StableDiffusionProcessing does not do anything; use sampler_name", file=sys.stderr) - self.sd_model = sd_model self.outpath_samples: str = outpath_samples self.outpath_grids: str = outpath_grids self.prompt: str = prompt @@ -156,6 +155,10 @@ class StableDiffusionProcessing(): self.all_subseeds = None self.iteration = 0 + @property + def sd_model(self): + return shared.sd_model + def txt2img_image_conditioning(self, x, width=None, height=None): self.is_using_inpainting_conditioning = self.sd_model.model.conditioning_key in {'hybrid', 'concat'} @@ -236,7 +239,6 @@ class StableDiffusionProcessing(): raise NotImplementedError() def close(self): - self.sd_model = None self.sampler = None @@ -471,7 +473,6 @@ def process_images(p: StableDiffusionProcessing) -> Processed: if k == 'sd_model_checkpoint': sd_models.reload_model_weights() # make onchange call for changing SD model - p.sd_model = shared.sd_model if k == 'sd_vae': sd_vae.reload_vae_weights() # make onchange call for changing VAE -- cgit v1.2.3 From eb2223340cfdd58efaa157662c279fbf6c90c7d9 Mon Sep 17 00:00:00 2001 From: fuggy <45698918+nonetrix@users.noreply.github.com> Date: Mon, 16 Jan 2023 21:50:30 -0600 Subject: Fix typo --- modules/errors.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/errors.py b/modules/errors.py index a668c014..a10e8708 100644 --- a/modules/errors.py +++ b/modules/errors.py @@ -19,7 +19,7 @@ def display(e: Exception, task): message = str(e) if "copying a param with shape torch.Size([640, 1024]) from checkpoint, the shape in current model is torch.Size([640, 768])" in message: print_error_explanation(""" -The most likely cause of this is you are trying to load Stable Diffusion 2.0 model without specifying its connfig file. +The most likely cause of this is you are trying to load Stable Diffusion 2.0 model without specifying its config file. See https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Features#stable-diffusion-20 for how to solve this. """) -- cgit v1.2.3 From c361b89026442f3412162657f330d500b803e052 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 17 Jan 2023 11:04:56 +0300 Subject: disable the new NaN check for the CI --- modules/devices.py | 3 +++ modules/shared.py | 1 + 2 files changed, 4 insertions(+) (limited to 'modules') diff --git a/modules/devices.py b/modules/devices.py index 6f034948..206184fb 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -113,6 +113,9 @@ class NansException(Exception): def test_for_nans(x, where): from modules import shared + if shared.cmd_opts.disable_nan_check: + return + if not torch.all(torch.isnan(x)).item(): return diff --git a/modules/shared.py b/modules/shared.py index 483c4c62..a708f23c 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -64,6 +64,7 @@ parser.add_argument("--sub-quad-chunk-threshold", type=int, help="the percentage parser.add_argument("--opt-split-attention-invokeai", action='store_true', help="force-enables InvokeAI's cross-attention layer optimization. By default, it's on when cuda is unavailable.") parser.add_argument("--opt-split-attention-v1", action='store_true', help="enable older version of split attention optimization that does not consume all the VRAM it can find") parser.add_argument("--disable-opt-split-attention", action='store_true', help="force-disables cross-attention layer optimization") +parser.add_argument("--disable-nan-check", action='store_true', help="do not check if produced images/latent spaces have nans; useful for running without a checkpoint in CI") parser.add_argument("--use-cpu", nargs='+', help="use CPU as torch device for specified modules", default=[], type=str.lower) parser.add_argument("--listen", action='store_true', help="launch gradio with 0.0.0.0 as server name, allowing to respond to network requests") parser.add_argument("--port", type=int, help="launch gradio with given server port, you need root/admin rights for ports < 1024, defaults to 7860 if available", default=None) -- cgit v1.2.3 From 4688bfff55dd6607e6608524fb219f97dc6fe8bb Mon Sep 17 00:00:00 2001 From: dan Date: Tue, 17 Jan 2023 17:16:43 +0800 Subject: Add auto-sized cropping UI --- modules/textual_inversion/preprocess.py | 38 ++++++++++++++++++++++++++++++--- modules/ui.py | 28 +++++++++++++++++++++++- 2 files changed, 62 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py index 64abff4d..86c1cd33 100644 --- a/modules/textual_inversion/preprocess.py +++ b/modules/textual_inversion/preprocess.py @@ -12,7 +12,7 @@ from modules.shared import opts, cmd_opts from modules.textual_inversion import autocrop -def preprocess(id_task, process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru=False, split_threshold=0.5, overlap_ratio=0.2, process_focal_crop=False, process_focal_crop_face_weight=0.9, process_focal_crop_entropy_weight=0.3, process_focal_crop_edges_weight=0.5, process_focal_crop_debug=False): +def preprocess(id_task, process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru=False, split_threshold=0.5, overlap_ratio=0.2, process_focal_crop=False, process_focal_crop_face_weight=0.9, process_focal_crop_entropy_weight=0.3, process_focal_crop_edges_weight=0.5, process_focal_crop_debug=False, process_multicrop=None, process_multicrop_mindim=None, process_multicrop_maxdim=None, process_multicrop_minarea=None, process_multicrop_maxarea=None, process_multicrop_objective=None, process_multicrop_threshold=None): try: if process_caption: shared.interrogator.load() @@ -20,7 +20,7 @@ def preprocess(id_task, process_src, process_dst, process_width, process_height, if process_caption_deepbooru: deepbooru.model.start() - preprocess_work(process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru, split_threshold, overlap_ratio, process_focal_crop, process_focal_crop_face_weight, process_focal_crop_entropy_weight, process_focal_crop_edges_weight, process_focal_crop_debug) + preprocess_work(process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru, split_threshold, overlap_ratio, process_focal_crop, process_focal_crop_face_weight, process_focal_crop_entropy_weight, process_focal_crop_edges_weight, process_focal_crop_debug, process_multicrop, process_multicrop_mindim, process_multicrop_maxdim, process_multicrop_minarea, process_multicrop_maxarea, process_multicrop_objective, process_multicrop_threshold) finally: @@ -109,8 +109,32 @@ def split_pic(image, inverse_xy, width, height, overlap_ratio): splitted = image.crop((0, y, to_w, y + to_h)) yield splitted +# not using torchvision.transforms.CenterCrop because it doesn't allow float regions +def center_crop(image: Image, w: int, h: int): + iw, ih = image.size + if ih / h < iw / w: + sw = w * ih / h + box = (iw - sw) / 2, 0, iw - (iw - sw) / 2, ih + else: + sh = h * iw / w + box = 0, (ih - sh) / 2, iw, ih - (ih - sh) / 2 + return image.resize((w, h), Image.Resampling.LANCZOS, box) + -def preprocess_work(process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru=False, split_threshold=0.5, overlap_ratio=0.2, process_focal_crop=False, process_focal_crop_face_weight=0.9, process_focal_crop_entropy_weight=0.3, process_focal_crop_edges_weight=0.5, process_focal_crop_debug=False): +def multicrop_pic(image: Image, mindim, maxdim, minarea, maxarea, objective, threshold): + iw, ih = image.size + err = lambda w, h: 1-(lambda x: x if x < 1 else 1/x)(iw/ih/(w/h)) + try: + w, h = max(((w, h) for w in range(mindim, maxdim+1, 64) for h in range(mindim, maxdim+1, 64) + if minarea <= w * h <= maxarea and err(w, h) <= threshold), + key= lambda wh: ((objective=='Maximize area')*wh[0]*wh[1], -err(*wh)) + ) + except ValueError: + return + return center_crop(image, w, h) + + +def preprocess_work(process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru=False, split_threshold=0.5, overlap_ratio=0.2, process_focal_crop=False, process_focal_crop_face_weight=0.9, process_focal_crop_entropy_weight=0.3, process_focal_crop_edges_weight=0.5, process_focal_crop_debug=False, process_multicrop=None, process_multicrop_mindim=None, process_multicrop_maxdim=None, process_multicrop_minarea=None, process_multicrop_maxarea=None, process_multicrop_objective=None, process_multicrop_threshold=None): width = process_width height = process_height src = os.path.abspath(process_src) @@ -194,6 +218,14 @@ def preprocess_work(process_src, process_dst, process_width, process_height, pre save_pic(focal, index, params, existing_caption=existing_caption) process_default_resize = False + if process_multicrop: + cropped = multicrop_pic(img, process_multicrop_mindim, process_multicrop_maxdim, process_multicrop_minarea, process_multicrop_maxarea, process_multicrop_objective, process_multicrop_threshold) + if cropped is not None: + save_pic(cropped, index, params, existing_caption=existing_caption) + else: + print(f"skipped {img.width}x{img.height} image {filename} (can't find suitable size within error threshold)") + process_default_resize = False + if process_default_resize: img = images.resize_image(1, img, width, height) save_pic(img, index, params, existing_caption=existing_caption) diff --git a/modules/ui.py b/modules/ui.py index 20b66165..bbce9acd 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1226,6 +1226,7 @@ def create_ui(): process_flip = gr.Checkbox(label='Create flipped copies', elem_id="train_process_flip") process_split = gr.Checkbox(label='Split oversized images', elem_id="train_process_split") process_focal_crop = gr.Checkbox(label='Auto focal point crop', elem_id="train_process_focal_crop") + process_multicrop = gr.Checkbox(label='Auto-sized crop', elem_id="train_process_multicrop") process_caption = gr.Checkbox(label='Use BLIP for caption', elem_id="train_process_caption") process_caption_deepbooru = gr.Checkbox(label='Use deepbooru for caption', visible=True, elem_id="train_process_caption_deepbooru") @@ -1238,7 +1239,19 @@ def create_ui(): process_focal_crop_entropy_weight = gr.Slider(label='Focal point entropy weight', value=0.15, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_entropy_weight") process_focal_crop_edges_weight = gr.Slider(label='Focal point edges weight', value=0.5, minimum=0.0, maximum=1.0, step=0.05, elem_id="train_process_focal_crop_edges_weight") process_focal_crop_debug = gr.Checkbox(label='Create debug image', elem_id="train_process_focal_crop_debug") - + + with gr.Column(visible=False) as process_multicrop_col: + gr.Markdown('Each image is center-cropped with an automatically chosen width and height.') + with gr.Row(): + process_multicrop_mindim = gr.Slider(minimum=64, maximum=2048, step=8, label="Dimension lower bound", value=384, elem_id="train_process_multicrop_mindim") + process_multicrop_maxdim = gr.Slider(minimum=64, maximum=2048, step=8, label="Dimension upper bound", value=768, elem_id="train_process_multicrop_maxdim") + with gr.Row(): + process_multicrop_minarea = gr.Slider(minimum=64*64, maximum=2048*2048, step=1, label="Area lower bound", value=64*64, elem_id="train_process_multicrop_minarea") + process_multicrop_maxarea = gr.Slider(minimum=64*64, maximum=2048*2048, step=1, label="Area upper bound", value=640*640, elem_id="train_process_multicrop_maxarea") + with gr.Row(): + process_multicrop_objective = gr.Radio(["Maximize area", "Minimize error"], value="Maximize area", label="Resizing objective", elem_id="train_process_multicrop_objective") + process_multicrop_threshold = gr.Slider(minimum=0, maximum=1, step=0.01, label="Error threshold", value=0.1, elem_id="train_process_multicrop_threshold") + with gr.Row(): with gr.Column(scale=3): gr.HTML(value="") @@ -1260,6 +1273,12 @@ def create_ui(): outputs=[process_focal_crop_row], ) + process_multicrop.change( + fn=lambda show: gr_show(show), + inputs=[process_multicrop], + outputs=[process_multicrop_col], + ) + def get_textual_inversion_template_names(): return sorted([x for x in textual_inversion.textual_inversion_templates]) @@ -1379,6 +1398,13 @@ def create_ui(): process_focal_crop_entropy_weight, process_focal_crop_edges_weight, process_focal_crop_debug, + process_multicrop, + process_multicrop_mindim, + process_multicrop_maxdim, + process_multicrop_minarea, + process_multicrop_maxarea, + process_multicrop_objective, + process_multicrop_threshold, ], outputs=[ ti_output, -- cgit v1.2.3 From aede265f1d6d512ca9e51a305e98a96a215366c4 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 17 Jan 2023 13:57:55 +0300 Subject: Fix unable to find Real-ESRGAN model info error (AttributeError: 'NoneType' object has no attribute 'data_path') #6841 #5170 --- modules/realesrgan_model.py | 12 ++++-------- modules/upscaler.py | 1 + 2 files changed, 5 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/realesrgan_model.py b/modules/realesrgan_model.py index 3ac0b97a..47f70251 100644 --- a/modules/realesrgan_model.py +++ b/modules/realesrgan_model.py @@ -38,13 +38,13 @@ class UpscalerRealESRGAN(Upscaler): return img info = self.load_model(path) - if not os.path.exists(info.data_path): + if not os.path.exists(info.local_data_path): print("Unable to load RealESRGAN model: %s" % info.name) return img upsampler = RealESRGANer( scale=info.scale, - model_path=info.data_path, + model_path=info.local_data_path, model=info.model(), half=not cmd_opts.no_half, tile=opts.ESRGAN_tile, @@ -58,17 +58,13 @@ class UpscalerRealESRGAN(Upscaler): def load_model(self, path): try: - info = None - for scaler in self.scalers: - if scaler.data_path == path: - info = scaler + info = next(iter([scaler for scaler in self.scalers if scaler.data_path == path]), None) if info is None: print(f"Unable to find model info: {path}") return None - model_file = load_file_from_url(url=info.data_path, model_dir=self.model_path, progress=True) - info.data_path = model_file + info.local_data_path = load_file_from_url(url=info.data_path, model_dir=self.model_path, progress=True) return info except Exception as e: print(f"Error making Real-ESRGAN models list: {e}", file=sys.stderr) diff --git a/modules/upscaler.py b/modules/upscaler.py index 231680cb..a5bf5acb 100644 --- a/modules/upscaler.py +++ b/modules/upscaler.py @@ -95,6 +95,7 @@ class UpscalerData: def __init__(self, name: str, path: str, upscaler: Upscaler = None, scale: int = 4, model=None): self.name = name self.data_path = path + self.local_data_path = path self.scaler = upscaler self.scale = scale self.model = model -- cgit v1.2.3 From 6e08da2c315c346225aa834017f4e32cfc0de200 Mon Sep 17 00:00:00 2001 From: ddPn08 Date: Tue, 17 Jan 2023 23:50:41 +0900 Subject: Add `--vae-dir` argument --- modules/sd_vae.py | 7 +++++++ modules/shared.py | 1 + 2 files changed, 8 insertions(+) (limited to 'modules') diff --git a/modules/sd_vae.py b/modules/sd_vae.py index b2af2ce7..da1bf15c 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -72,6 +72,13 @@ def refresh_vae_list(): os.path.join(shared.cmd_opts.ckpt_dir, '**/*.vae.safetensors'), ] + if shared.cmd_opts.vae_dir is not None and os.path.isdir(shared.cmd_opts.vae_dir): + paths += [ + os.path.join(shared.cmd_opts.vae_dir, '**/*.ckpt'), + os.path.join(shared.cmd_opts.vae_dir, '**/*.pt'), + os.path.join(shared.cmd_opts.vae_dir, '**/*.safetensors'), + ] + candidates = [] for path in paths: candidates += glob.iglob(path, recursive=True) diff --git a/modules/shared.py b/modules/shared.py index a708f23c..a1345ad3 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -26,6 +26,7 @@ parser = argparse.ArgumentParser() parser.add_argument("--config", type=str, default=os.path.join(script_path, "configs/v1-inference.yaml"), help="path to config which constructs model",) parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; if specified, this checkpoint will be added to the list of checkpoints and loaded",) parser.add_argument("--ckpt-dir", type=str, default=None, help="Path to directory with stable diffusion checkpoints") +parser.add_argument("--vae-dir", type=str, default=None, help="Path to directory with stable VAE files") parser.add_argument("--gfpgan-dir", type=str, help="GFPGAN directory", default=('./src/gfpgan' if os.path.exists('./src/gfpgan') else './GFPGAN')) parser.add_argument("--gfpgan-model", type=str, help="GFPGAN model file name", default=None) parser.add_argument("--no-half", action='store_true', help="do not switch the model to 16-bit floats") -- cgit v1.2.3 From 5e15a0b422981c0b5484885d0b4d28af6913c76f Mon Sep 17 00:00:00 2001 From: EllangoK Date: Tue, 17 Jan 2023 11:42:44 -0500 Subject: Changed params.txt save to after manual init call --- modules/processing.py | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 9c3673de..4a1f033e 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -538,10 +538,6 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if p.scripts is not None: p.scripts.process(p) - with open(os.path.join(shared.script_path, "params.txt"), "w", encoding="utf8") as file: - processed = Processed(p, [], p.seed, "") - file.write(processed.infotext(p, 0)) - infotexts = [] output_images = [] @@ -572,6 +568,10 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: with devices.autocast(): p.init(p.all_prompts, p.all_seeds, p.all_subseeds) + with open(os.path.join(shared.script_path, "params.txt"), "w", encoding="utf8") as file: + processed = Processed(p, [], p.seed, "") + file.write(processed.infotext(p, 0)) + if state.job_count == -1: state.job_count = p.n_iter -- cgit v1.2.3 From 3a0d6b77295162146d0a8d04278804334da6f1b4 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 17 Jan 2023 23:54:23 +0300 Subject: make it so that PNG images with EXIF do not lose parameters in PNG info tab --- modules/images.py | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/images.py b/modules/images.py index c3a5fc8b..3b1c5f34 100644 --- a/modules/images.py +++ b/modules/images.py @@ -605,8 +605,9 @@ def read_info_from_image(image): except ValueError: exif_comment = exif_comment.decode('utf8', errors="ignore") - items['exif comment'] = exif_comment - geninfo = exif_comment + if exif_comment: + items['exif comment'] = exif_comment + geninfo = exif_comment for field in ['jfif', 'jfif_version', 'jfif_unit', 'jfif_density', 'dpi', 'exif', 'loop', 'background', 'timestamp', 'duration']: -- cgit v1.2.3 From d906f87043d809e6d4d8de3c9926e184169b330f Mon Sep 17 00:00:00 2001 From: ddPn08 Date: Wed, 18 Jan 2023 07:52:10 +0900 Subject: fix typo --- modules/shared.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index a1345ad3..a42279ec 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -26,7 +26,7 @@ parser = argparse.ArgumentParser() parser.add_argument("--config", type=str, default=os.path.join(script_path, "configs/v1-inference.yaml"), help="path to config which constructs model",) parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; if specified, this checkpoint will be added to the list of checkpoints and loaded",) parser.add_argument("--ckpt-dir", type=str, default=None, help="Path to directory with stable diffusion checkpoints") -parser.add_argument("--vae-dir", type=str, default=None, help="Path to directory with stable VAE files") +parser.add_argument("--vae-dir", type=str, default=None, help="Path to directory with VAE files") parser.add_argument("--gfpgan-dir", type=str, help="GFPGAN directory", default=('./src/gfpgan' if os.path.exists('./src/gfpgan') else './GFPGAN')) parser.add_argument("--gfpgan-model", type=str, help="GFPGAN model file name", default=None) parser.add_argument("--no-half", action='store_true', help="do not switch the model to 16-bit floats") -- cgit v1.2.3 From a255dac4f8c5ee11c15b634563d3df513f1834b4 Mon Sep 17 00:00:00 2001 From: brkirch Date: Thu, 12 Jan 2023 08:00:38 -0500 Subject: Fix cumsum for MPS in newer torch The prior fix assumed that testing int16 was enough to determine if a fix is needed, but a recent fix for cumsum has int16 working but not bool. --- modules/devices.py | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/devices.py b/modules/devices.py index caeb0276..ac3ae0c9 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -139,8 +139,10 @@ orig_Tensor_cumsum = torch.Tensor.cumsum def cumsum_fix(input, cumsum_func, *args, **kwargs): if input.device.type == 'mps': output_dtype = kwargs.get('dtype', input.dtype) - if any(output_dtype == broken_dtype for broken_dtype in [torch.bool, torch.int8, torch.int16, torch.int64]): + if output_dtype == torch.int64: return cumsum_func(input.cpu(), *args, **kwargs).to(input.device) + elif cumsum_needs_bool_fix and output_dtype == torch.bool or cumsum_needs_int_fix and (output_dtype == torch.int8 or output_dtype == torch.int16): + return cumsum_func(input.to(torch.int32), *args, **kwargs).to(torch.int64) return cumsum_func(input, *args, **kwargs) @@ -151,8 +153,9 @@ if has_mps(): torch.nn.functional.layer_norm = layer_norm_fix torch.Tensor.numpy = numpy_fix elif version.parse(torch.__version__) > version.parse("1.13.1"): - if not torch.Tensor([1,2]).to(torch.device("mps")).equal(torch.Tensor([1,1]).to(torch.device("mps")).cumsum(0, dtype=torch.int16)): - torch.cumsum = lambda input, *args, **kwargs: ( cumsum_fix(input, orig_cumsum, *args, **kwargs) ) - torch.Tensor.cumsum = lambda self, *args, **kwargs: ( cumsum_fix(self, orig_Tensor_cumsum, *args, **kwargs) ) + cumsum_needs_int_fix = not torch.Tensor([1,2]).to(torch.device("mps")).equal(torch.ShortTensor([1,1]).to(torch.device("mps")).cumsum(0)) + cumsum_needs_bool_fix = not torch.BoolTensor([True,True]).to(device=torch.device("mps"), dtype=torch.int64).equal(torch.BoolTensor([True,False]).to(torch.device("mps")).cumsum(0)) + torch.cumsum = lambda input, *args, **kwargs: ( cumsum_fix(input, orig_cumsum, *args, **kwargs) ) + torch.Tensor.cumsum = lambda self, *args, **kwargs: ( cumsum_fix(self, orig_Tensor_cumsum, *args, **kwargs) ) orig_narrow = torch.narrow torch.narrow = lambda *args, **kwargs: ( orig_narrow(*args, **kwargs).clone() ) -- cgit v1.2.3 From 6faae2323963f9b0e0086a85b9d0472a24fbaa73 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 18 Jan 2023 14:33:09 +0300 Subject: repair broken quicksettings when some form-requiring options are added to it --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index e1f98d23..6d70a795 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1659,7 +1659,7 @@ def create_ui(): interfaces += [(extensions_interface, "Extensions", "extensions")] with gr.Blocks(css=css, analytics_enabled=False, title="Stable Diffusion") as demo: - with gr.Row(elem_id="quicksettings"): + with gr.Row(elem_id="quicksettings", variant="compact"): for i, k, item in sorted(quicksettings_list, key=lambda x: quicksettings_names.get(x[1], x[0])): component = create_setting_component(k, is_quicksettings=True) component_dict[k] = component -- cgit v1.2.3 From 8683427bd9315d2fda0d2f9644c8b1f6a182da55 Mon Sep 17 00:00:00 2001 From: Vladimir Repin <32306715+mezotaken@users.noreply.github.com> Date: Wed, 18 Jan 2023 20:16:52 +0300 Subject: Process interrogation on all img2img subtabs --- modules/ui.py | 50 +++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 43 insertions(+), 7 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 20b66165..78c0c92a 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -204,9 +204,31 @@ def apply_styles(prompt, prompt_neg, style1_name, style2_name): return [gr.Textbox.update(value=prompt), gr.Textbox.update(value=prompt_neg), gr.Dropdown.update(value="None"), gr.Dropdown.update(value="None")] +def process_interrogate(interrogation_function, mode, ii_input_dir, ii_output_dir, *ii_singles): + if mode in {0, 1, 3, 4}: + return [interrogation_function(ii_singles[mode]), None] + elif mode == 2: + return [interrogation_function(ii_singles[mode]["image"]), None] + elif mode == 5: + assert not shared.cmd_opts.hide_ui_dir_config, "Launched with --hide-ui-dir-config, batch img2img disabled" + images = shared.listfiles(ii_input_dir) + print(f"Will process {len(images)} images.") + if ii_output_dir != "": + os.makedirs(ii_output_dir, exist_ok=True) + else: + ii_output_dir = ii_input_dir + + for image in images: + img = Image.open(image) + filename = os.path.basename(image) + left, _ = os.path.splitext(filename) + print(interrogation_function(img), file=open(os.path.join(ii_output_dir, left + ".txt"), 'a')) + + return [gr_show(True), None] + + def interrogate(image): prompt = shared.interrogator.interrogate(image.convert("RGB")) - return gr_show(True) if prompt is None else prompt @@ -983,19 +1005,33 @@ def create_ui(): show_progress=False, ) + interrogate_args = dict( + _js="get_img2img_tab_index", + inputs=[ + dummy_component, + img2img_batch_input_dir, + img2img_batch_output_dir, + init_img, + sketch, + init_img_with_mask, + inpaint_color_sketch, + init_img_inpaint, + ], + outputs=[img2img_prompt, dummy_component], + show_progress=False, + ) + img2img_prompt.submit(**img2img_args) submit.click(**img2img_args) img2img_interrogate.click( - fn=interrogate, - inputs=[init_img], - outputs=[img2img_prompt], + fn=lambda *args : process_interrogate(interrogate, *args), + **interrogate_args, ) img2img_deepbooru.click( - fn=interrogate_deepbooru, - inputs=[init_img], - outputs=[img2img_prompt], + fn=lambda *args : process_interrogate(interrogate_deepbooru, *args), + **interrogate_args, ) prompts = [(txt2img_prompt, txt2img_negative_prompt), (img2img_prompt, img2img_negative_prompt)] -- cgit v1.2.3 From 924e222004ab54273806c5f2ca7a0e7cfa76ad83 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 18 Jan 2023 23:04:24 +0300 Subject: add option to show/hide warnings removed hiding warnings from LDSR fixed/reworked few places that produced warnings --- modules/hypernetworks/hypernetwork.py | 7 ++++- modules/sd_hijack.py | 8 ------ modules/sd_hijack_checkpoint.py | 38 +++++++++++++++++++++++++- modules/shared.py | 1 + modules/textual_inversion/textual_inversion.py | 6 +++- modules/ui.py | 31 ++++++++++++--------- 6 files changed, 67 insertions(+), 24 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index c963fc40..74e78582 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -12,7 +12,7 @@ import torch import tqdm from einops import rearrange, repeat from ldm.util import default -from modules import devices, processing, sd_models, shared, sd_samplers, hashes +from modules import devices, processing, sd_models, shared, sd_samplers, hashes, sd_hijack_checkpoint from modules.textual_inversion import textual_inversion, logging from modules.textual_inversion.learn_schedule import LearnRateScheduler from torch import einsum @@ -575,6 +575,8 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi pbar = tqdm.tqdm(total=steps - initial_step) try: + sd_hijack_checkpoint.add() + for i in range((steps-initial_step) * gradient_step): if scheduler.finished: break @@ -724,6 +726,9 @@ Last saved image: {html.escape(last_saved_image)}
pbar.close() hypernetwork.eval() #report_statistics(loss_dict) + sd_hijack_checkpoint.remove() + + filename = os.path.join(shared.cmd_opts.hypernetwork_dir, f'{hypernetwork_name}.pt') hypernetwork.optimizer_name = optimizer_name diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index 6b0d95af..870eba88 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -69,12 +69,6 @@ def undo_optimizations(): ldm.modules.diffusionmodules.model.AttnBlock.forward = diffusionmodules_model_AttnBlock_forward -def fix_checkpoint(): - ldm.modules.attention.BasicTransformerBlock.forward = sd_hijack_checkpoint.BasicTransformerBlock_forward - ldm.modules.diffusionmodules.openaimodel.ResBlock.forward = sd_hijack_checkpoint.ResBlock_forward - ldm.modules.diffusionmodules.openaimodel.AttentionBlock.forward = sd_hijack_checkpoint.AttentionBlock_forward - - class StableDiffusionModelHijack: fixes = None comments = [] @@ -106,8 +100,6 @@ class StableDiffusionModelHijack: self.optimization_method = apply_optimizations() self.clip = m.cond_stage_model - - fix_checkpoint() def flatten(el): flattened = [flatten(children) for children in el.children()] diff --git a/modules/sd_hijack_checkpoint.py b/modules/sd_hijack_checkpoint.py index 5712972f..2604d969 100644 --- a/modules/sd_hijack_checkpoint.py +++ b/modules/sd_hijack_checkpoint.py @@ -1,10 +1,46 @@ from torch.utils.checkpoint import checkpoint +import ldm.modules.attention +import ldm.modules.diffusionmodules.openaimodel + + def BasicTransformerBlock_forward(self, x, context=None): return checkpoint(self._forward, x, context) + def AttentionBlock_forward(self, x): return checkpoint(self._forward, x) + def ResBlock_forward(self, x, emb): - return checkpoint(self._forward, x, emb) \ No newline at end of file + return checkpoint(self._forward, x, emb) + + +stored = [] + + +def add(): + if len(stored) != 0: + return + + stored.extend([ + ldm.modules.attention.BasicTransformerBlock.forward, + ldm.modules.diffusionmodules.openaimodel.ResBlock.forward, + ldm.modules.diffusionmodules.openaimodel.AttentionBlock.forward + ]) + + ldm.modules.attention.BasicTransformerBlock.forward = BasicTransformerBlock_forward + ldm.modules.diffusionmodules.openaimodel.ResBlock.forward = ResBlock_forward + ldm.modules.diffusionmodules.openaimodel.AttentionBlock.forward = AttentionBlock_forward + + +def remove(): + if len(stored) == 0: + return + + ldm.modules.attention.BasicTransformerBlock.forward = stored[0] + ldm.modules.diffusionmodules.openaimodel.ResBlock.forward = stored[1] + ldm.modules.diffusionmodules.openaimodel.AttentionBlock.forward = stored[2] + + stored.clear() + diff --git a/modules/shared.py b/modules/shared.py index a708f23c..ddb97f99 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -369,6 +369,7 @@ options_templates.update(options_section(('face-restoration', "Face restoration" })) options_templates.update(options_section(('system', "System"), { + "show_warnings": OptionInfo(False, "Show warnings in console."), "memmon_poll_rate": OptionInfo(8, "VRAM usage polls per second during generation. Set to 0 to disable.", gr.Slider, {"minimum": 0, "maximum": 40, "step": 1}), "samples_log_stdout": OptionInfo(False, "Always print all generation info to standard output"), "multiple_tqdm": OptionInfo(True, "Add a second progress bar to the console that shows progress for an entire job."), diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 7e4a6d24..5a7be422 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -15,7 +15,7 @@ import numpy as np from PIL import Image, PngImagePlugin from torch.utils.tensorboard import SummaryWriter -from modules import shared, devices, sd_hijack, processing, sd_models, images, sd_samplers +from modules import shared, devices, sd_hijack, processing, sd_models, images, sd_samplers, sd_hijack_checkpoint import modules.textual_inversion.dataset from modules.textual_inversion.learn_schedule import LearnRateScheduler @@ -452,6 +452,8 @@ def train_embedding(id_task, embedding_name, learn_rate, batch_size, gradient_st pbar = tqdm.tqdm(total=steps - initial_step) try: + sd_hijack_checkpoint.add() + for i in range((steps-initial_step) * gradient_step): if scheduler.finished: break @@ -617,9 +619,11 @@ Last saved image: {html.escape(last_saved_image)}
pbar.close() shared.sd_model.first_stage_model.to(devices.device) shared.parallel_processing_allowed = old_parallel_processing_allowed + sd_hijack_checkpoint.remove() return embedding, filename + def save_embedding(embedding, optimizer, checkpoint, embedding_name, filename, remove_cached_checksum=True): old_embedding_name = embedding.name old_sd_checkpoint = embedding.sd_checkpoint if hasattr(embedding, "sd_checkpoint") else None diff --git a/modules/ui.py b/modules/ui.py index 6d70a795..25818fb0 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -11,6 +11,7 @@ import tempfile import time import traceback from functools import partial, reduce +import warnings import gradio as gr import gradio.routes @@ -41,6 +42,8 @@ from modules.textual_inversion import textual_inversion import modules.hypernetworks.ui from modules.generation_parameters_copypaste import image_from_url_text +warnings.filterwarnings("default" if opts.show_warnings else "ignore", category=UserWarning) + # this is a fix for Windows users. Without it, javascript files will be served with text/html content-type and the browser will not show any UI mimetypes.init() mimetypes.add_type('application/javascript', '.js') @@ -417,17 +420,16 @@ def apply_setting(key, value): return value -def update_generation_info(args): - generation_info, html_info, img_index = args +def update_generation_info(generation_info, html_info, img_index): try: generation_info = json.loads(generation_info) if img_index < 0 or img_index >= len(generation_info["infotexts"]): - return html_info - return plaintext_to_html(generation_info["infotexts"][img_index]) + return html_info, gr.update() + return plaintext_to_html(generation_info["infotexts"][img_index]), gr.update() except Exception: pass # if the json parse or anything else fails, just return the old html_info - return html_info + return html_info, gr.update() def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): @@ -508,10 +510,9 @@ Requested path was: {f} generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") generation_info_button.click( fn=update_generation_info, - _js="(x, y) => [x, y, selected_gallery_index()]", - inputs=[generation_info, html_info], - outputs=[html_info], - preprocess=False + _js="function(x, y, z){ console.log(x, y, z); return [x, y, selected_gallery_index()] }", + inputs=[generation_info, html_info, html_info], + outputs=[html_info, html_info], ) save.click( @@ -526,7 +527,8 @@ Requested path was: {f} outputs=[ download_files, html_log, - ] + ], + show_progress=False, ) save_zip.click( @@ -588,7 +590,7 @@ def create_ui(): txt2img_prompt, txt2img_prompt_styles, txt2img_negative_prompt, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) dummy_component = gr.Label(visible=False) - txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="bytes", visible=False) + txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="binary", visible=False) with gr.Row().style(equal_height=False): with gr.Column(variant='compact', elem_id="txt2img_settings"): @@ -768,7 +770,7 @@ def create_ui(): with gr.Blocks(analytics_enabled=False) as img2img_interface: img2img_prompt, img2img_prompt_styles, img2img_negative_prompt, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) - img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="bytes", visible=False) + img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="binary", visible=False) with FormRow().style(equal_height=False): with gr.Column(variant='compact', elem_id="img2img_settings"): @@ -1768,7 +1770,10 @@ def create_ui(): if saved_value is None: ui_settings[key] = getattr(obj, field) elif condition and not condition(saved_value): - print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') + pass + + # this warning is generally not useful; + # print(f'Warning: Bad ui setting value: {key}: {saved_value}; Default value "{getattr(obj, field)}" will be used instead.') else: setattr(obj, field, saved_value) if init_field is not None: -- cgit v1.2.3 From b186d44dcd0df9d127a663b297334a5bd8258b58 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 18 Jan 2023 23:20:23 +0300 Subject: use DDIM in hires fix is the sampler is PLMS --- modules/processing.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 9c3673de..8c18ac53 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -857,7 +857,8 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing): shared.state.nextjob() - self.sampler = sd_samplers.create_sampler(self.sampler_name, self.sd_model) + img2img_sampler_name = self.sampler_name if self.sampler_name != 'PLMS' else 'DDIM' # PLMS does not support img2img so we just silently switch ot DDIM + self.sampler = sd_samplers.create_sampler(img2img_sampler_name, self.sd_model) samples = samples[:, :, self.truncate_y//2:samples.shape[2]-(self.truncate_y+1)//2, self.truncate_x//2:samples.shape[3]-(self.truncate_x+1)//2] -- cgit v1.2.3 From bb0978ecfd3177d0bfd7cacd1ac8796d7eec2d79 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 00:44:51 +0300 Subject: fix hires fix ui weirdness caused by gradio update --- modules/ui.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 8b7f1dfb..09a3c92e 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -638,7 +638,7 @@ def create_ui(): seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('txt2img') elif category == "checkboxes": - with FormRow(elem_id="txt2img_checkboxes"): + with FormRow(elem_id="txt2img_checkboxes", variant="compact"): restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="txt2img_restore_faces") tiling = gr.Checkbox(label='Tiling', value=False, elem_id="txt2img_tiling") enable_hr = gr.Checkbox(label='Hires. fix', value=False, elem_id="txt2img_enable_hr") @@ -646,12 +646,12 @@ def create_ui(): elif category == "hires_fix": with FormGroup(visible=False, elem_id="txt2img_hires_fix") as hr_options: - with FormRow(elem_id="txt2img_hires_fix_row1"): + with FormRow(elem_id="txt2img_hires_fix_row1", variant="compact"): hr_upscaler = gr.Dropdown(label="Upscaler", elem_id="txt2img_hr_upscaler", choices=[*shared.latent_upscale_modes, *[x.name for x in shared.sd_upscalers]], value=shared.latent_upscale_default_mode) hr_second_pass_steps = gr.Slider(minimum=0, maximum=150, step=1, label='Hires steps', value=0, elem_id="txt2img_hires_steps") denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7, elem_id="txt2img_denoising_strength") - with FormRow(elem_id="txt2img_hires_fix_row2"): + with FormRow(elem_id="txt2img_hires_fix_row2", variant="compact"): hr_scale = gr.Slider(minimum=1.0, maximum=4.0, step=0.05, label="Upscale by", value=2.0, elem_id="txt2img_hr_scale") hr_resize_x = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize width to", value=0, elem_id="txt2img_hr_resize_x") hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y") -- cgit v1.2.3 From 99207bc816d027b522e1c49001748c63fd426b53 Mon Sep 17 00:00:00 2001 From: EllangoK Date: Wed, 18 Jan 2023 19:13:15 -0500 Subject: check model name values are set before merging --- modules/extras.py | 22 ++++++++++++++++++---- 1 file changed, 18 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 22668fcd..29eb1f07 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -287,10 +287,19 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam def add_difference(theta0, theta1_2_diff, alpha): return theta0 + (alpha * theta1_2_diff) + if not primary_model_name: + shared.state.textinfo = "Failed: Merging requires a primary model." + shared.state.end() + return ["Failed: Merging requires a primary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + primary_model_info = sd_models.checkpoints_list[primary_model_name] + + if not secondary_model_name: + shared.state.textinfo = "Failed: Merging requires a secondary model." + shared.state.end() + return ["Failed: Merging requires a secondary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + secondary_model_info = sd_models.checkpoints_list[secondary_model_name] - tertiary_model_info = sd_models.checkpoints_list.get(tertiary_model_name, None) - result_is_inpainting_model = False theta_funcs = { "Weighted sum": (None, weighted_sum), @@ -298,10 +307,15 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam } theta_func1, theta_func2 = theta_funcs[interp_method] - if theta_func1 and not tertiary_model_info: + tertiary_model_info = None + if theta_func1 and not tertiary_model_name: shared.state.textinfo = "Failed: Interpolation method requires a tertiary model." shared.state.end() - return ["Failed: Interpolation method requires a tertiary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + return [f"Failed: Interpolation method ({interp_method}) requires a tertiary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + else: + tertiary_model_info = sd_models.checkpoints_list.get(tertiary_model_name, None) + + result_is_inpainting_model = False shared.state.textinfo = f"Loading {secondary_model_info.filename}..." print(f"Loading {secondary_model_info.filename}...") -- cgit v1.2.3 From 26a6a78b16f88a6f88f4cca3f378db3b83fc94f8 Mon Sep 17 00:00:00 2001 From: EllangoK Date: Wed, 18 Jan 2023 21:21:52 -0500 Subject: only lookup tertiary model if theta_func1 is set --- modules/extras.py | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 29eb1f07..88eea22e 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -307,13 +307,12 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam } theta_func1, theta_func2 = theta_funcs[interp_method] - tertiary_model_info = None if theta_func1 and not tertiary_model_name: shared.state.textinfo = "Failed: Interpolation method requires a tertiary model." shared.state.end() return [f"Failed: Interpolation method ({interp_method}) requires a tertiary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] - else: - tertiary_model_info = sd_models.checkpoints_list.get(tertiary_model_name, None) + + tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None result_is_inpainting_model = False -- cgit v1.2.3 From 308b51012a5def38edb1c2e127e736c43aa6e1a3 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 08:41:37 +0300 Subject: fix an unlikely division by 0 error --- modules/progress.py | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/progress.py b/modules/progress.py index 3327b883..f9e005d3 100644 --- a/modules/progress.py +++ b/modules/progress.py @@ -67,10 +67,13 @@ def progressapi(req: ProgressRequest): progress = 0 - if shared.state.job_count > 0: - progress += shared.state.job_no / shared.state.job_count - if shared.state.sampling_steps > 0: - progress += 1 / shared.state.job_count * shared.state.sampling_step / shared.state.sampling_steps + job_count, job_no = shared.state.job_count, shared.state.job_no + sampling_steps, sampling_step = shared.state.sampling_steps, shared.state.sampling_step + + if job_count > 0: + progress += job_no / job_count + if sampling_steps > 0: + progress += 1 / job_count * sampling_step / sampling_steps progress = min(progress, 1) -- cgit v1.2.3 From 7cfc6450305125683799208fb7bc27c0b12586b3 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 08:53:50 +0300 Subject: eliminate repetition of code in #6910 --- modules/extras.py | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 88eea22e..367c15cc 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -278,6 +278,11 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam shared.state.begin() shared.state.job = 'model-merge' + def fail(message): + shared.state.textinfo = message + shared.state.end() + return [message, *[gr.update() for _ in range(4)]] + def weighted_sum(theta0, theta1, alpha): return ((1 - alpha) * theta0) + (alpha * theta1) @@ -288,16 +293,12 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam return theta0 + (alpha * theta1_2_diff) if not primary_model_name: - shared.state.textinfo = "Failed: Merging requires a primary model." - shared.state.end() - return ["Failed: Merging requires a primary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + return fail("Failed: Merging requires a primary model.") primary_model_info = sd_models.checkpoints_list[primary_model_name] if not secondary_model_name: - shared.state.textinfo = "Failed: Merging requires a secondary model." - shared.state.end() - return ["Failed: Merging requires a secondary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + return fail("Failed: Merging requires a secondary model.") secondary_model_info = sd_models.checkpoints_list[secondary_model_name] @@ -308,9 +309,7 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam theta_func1, theta_func2 = theta_funcs[interp_method] if theta_func1 and not tertiary_model_name: - shared.state.textinfo = "Failed: Interpolation method requires a tertiary model." - shared.state.end() - return [f"Failed: Interpolation method ({interp_method}) requires a tertiary model."] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None -- cgit v1.2.3 From c7e50425f63c07242068f8dcccce70a4ef28a17f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 09:25:37 +0300 Subject: add progress bar to modelmerger --- modules/extras.py | 18 +++++++++++++++--- modules/progress.py | 2 +- modules/ui.py | 13 ++++++++----- 3 files changed, 24 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 367c15cc..034f28e4 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -274,14 +274,15 @@ def create_config(ckpt_result, config_source, a, b, c): shutil.copyfile(cfg, checkpoint_filename) -def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source): +def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source): shared.state.begin() shared.state.job = 'model-merge' + shared.state.job_count = 1 def fail(message): shared.state.textinfo = message shared.state.end() - return [message, *[gr.update() for _ in range(4)]] + return [*[gr.update() for _ in range(4)], message] def weighted_sum(theta0, theta1, alpha): return ((1 - alpha) * theta0) + (alpha * theta1) @@ -320,9 +321,12 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') if theta_func1: + shared.state.job_count += 1 + print(f"Loading {tertiary_model_info.filename}...") theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') + shared.state.sampling_steps = len(theta_1.keys()) for key in tqdm.tqdm(theta_1.keys()): if 'model' in key: if key in theta_2: @@ -330,8 +334,12 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam theta_1[key] = theta_func1(theta_1[key], t2) else: theta_1[key] = torch.zeros_like(theta_1[key]) + + shared.state.sampling_step += 1 del theta_2 + shared.state.nextjob() + shared.state.textinfo = f"Loading {primary_model_info.filename}..." print(f"Loading {primary_model_info.filename}...") theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') @@ -340,6 +348,7 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam chckpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] + shared.state.sampling_steps = len(theta_0.keys()) for key in tqdm.tqdm(theta_0.keys()): if 'model' in key and key in theta_1: @@ -367,6 +376,8 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam if save_as_half: theta_0[key] = theta_0[key].half() + shared.state.sampling_step += 1 + # I believe this part should be discarded, but I'll leave it for now until I am sure for key in theta_1.keys(): if 'model' in key and key not in theta_0: @@ -393,6 +404,7 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam output_modelname = os.path.join(ckpt_dir, filename) + shared.state.nextjob() shared.state.textinfo = f"Saving to {output_modelname}..." print(f"Saving to {output_modelname}...") @@ -410,4 +422,4 @@ def run_modelmerger(primary_model_name, secondary_model_name, tertiary_model_nam shared.state.textinfo = "Checkpoint saved to " + output_modelname shared.state.end() - return ["Checkpoint saved to " + output_modelname] + [gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)] + return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] diff --git a/modules/progress.py b/modules/progress.py index f9e005d3..c69ecf3d 100644 --- a/modules/progress.py +++ b/modules/progress.py @@ -72,7 +72,7 @@ def progressapi(req: ProgressRequest): if job_count > 0: progress += job_no / job_count - if sampling_steps > 0: + if sampling_steps > 0 and job_count > 0: progress += 1 / job_count * sampling_step / sampling_steps progress = min(progress, 1) diff --git a/modules/ui.py b/modules/ui.py index 09a3c92e..aeee7853 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1208,8 +1208,9 @@ def create_ui(): with gr.Row(): modelmerger_merge = gr.Button(elem_id="modelmerger_merge", value="Merge", variant='primary') - with gr.Column(variant='panel'): - submit_result = gr.Textbox(elem_id="modelmerger_result", show_label=False) + with gr.Column(variant='compact', elem_id="modelmerger_results_container"): + with gr.Group(elem_id="modelmerger_results_panel"): + modelmerger_result = gr.HTML(elem_id="modelmerger_result", show_label=False) with gr.Blocks(analytics_enabled=False) as train_interface: with gr.Row().style(equal_height=False): @@ -1753,12 +1754,14 @@ def create_ui(): print("Error loading/saving model file:", file=sys.stderr) print(traceback.format_exc(), file=sys.stderr) modules.sd_models.list_models() # to remove the potentially missing models from the list - return [f"Error merging checkpoints: {e}"] + [gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)] + return [*[gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)], f"Error merging checkpoints: {e}"] return results modelmerger_merge.click( - fn=modelmerger, + fn=wrap_gradio_gpu_call(modelmerger, extra_outputs=lambda: [gr.update() for _ in range(4)]), + _js='modelmerger', inputs=[ + dummy_component, primary_model_name, secondary_model_name, tertiary_model_name, @@ -1770,11 +1773,11 @@ def create_ui(): config_source, ], outputs=[ - submit_result, primary_model_name, secondary_model_name, tertiary_model_name, component_dict['sd_model_checkpoint'], + modelmerger_result, ] ) -- cgit v1.2.3 From 0f5dbfffd0b7202a48e404d8e74b5cc9a3e5b135 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 10:39:51 +0300 Subject: allow baking in VAE in checkpoint merger tab do not save config if it's the default for checkpoint merger tab change file naming scheme for checkpoint merger tab allow just saving A without any merging for checkpoint merger tab some stylistic changes for UI in checkpoint merger tab --- modules/extras.py | 112 +++++++++++++++++++++++++++++++++--------------------- modules/sd_vae.py | 9 ++++- modules/shared.py | 3 +- modules/ui.py | 17 +++++++-- 4 files changed, 90 insertions(+), 51 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 034f28e4..fe701a0e 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -15,7 +15,7 @@ from typing import Callable, List, OrderedDict, Tuple from functools import partial from dataclasses import dataclass -from modules import processing, shared, images, devices, sd_models, sd_samplers +from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae from modules.shared import opts import modules.gfpgan_model from modules.ui import plaintext_to_html @@ -251,7 +251,8 @@ def run_pnginfo(image): def create_config(ckpt_result, config_source, a, b, c): def config(x): - return sd_models.find_checkpoint_config(x) if x else None + res = sd_models.find_checkpoint_config(x) if x else None + return res if res != shared.sd_default_config else None if config_source == 0: cfg = config(a) or config(b) or config(c) @@ -274,10 +275,12 @@ def create_config(ckpt_result, config_source, a, b, c): shutil.copyfile(cfg, checkpoint_filename) -def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source): +chckpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] + + +def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae): shared.state.begin() shared.state.job = 'model-merge' - shared.state.job_count = 1 def fail(message): shared.state.textinfo = message @@ -293,41 +296,68 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ def add_difference(theta0, theta1_2_diff, alpha): return theta0 + (alpha * theta1_2_diff) + def filename_weighed_sum(): + a = primary_model_info.model_name + b = secondary_model_info.model_name + Ma = round(1 - multiplier, 2) + Mb = round(multiplier, 2) + + return f"{Ma}({a}) + {Mb}({b})" + + def filename_add_differnece(): + a = primary_model_info.model_name + b = secondary_model_info.model_name + c = tertiary_model_info.model_name + M = round(multiplier, 2) + + return f"{a} + {M}({b} - {c})" + + def filename_nothing(): + return primary_model_info.model_name + + theta_funcs = { + "Weighted sum": (filename_weighed_sum, None, weighted_sum), + "Add difference": (filename_add_differnece, get_difference, add_difference), + "No interpolation": (filename_nothing, None, None), + } + filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] + shared.state.job_count = (1 if theta_func1 else 0) + (1 if theta_func2 else 0) + if not primary_model_name: return fail("Failed: Merging requires a primary model.") primary_model_info = sd_models.checkpoints_list[primary_model_name] - if not secondary_model_name: + if theta_func2 and not secondary_model_name: return fail("Failed: Merging requires a secondary model.") - - secondary_model_info = sd_models.checkpoints_list[secondary_model_name] - theta_funcs = { - "Weighted sum": (None, weighted_sum), - "Add difference": (get_difference, add_difference), - } - theta_func1, theta_func2 = theta_funcs[interp_method] + secondary_model_info = sd_models.checkpoints_list[secondary_model_name] if theta_func2 else None if theta_func1 and not tertiary_model_name: return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") - + tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None result_is_inpainting_model = False - shared.state.textinfo = f"Loading {secondary_model_info.filename}..." - print(f"Loading {secondary_model_info.filename}...") - theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') + if theta_func2: + shared.state.textinfo = f"Loading B" + print(f"Loading {secondary_model_info.filename}...") + theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') + else: + theta_1 = None if theta_func1: - shared.state.job_count += 1 - + shared.state.textinfo = f"Loading C" print(f"Loading {tertiary_model_info.filename}...") theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') + shared.state.textinfo = 'Merging B and C' shared.state.sampling_steps = len(theta_1.keys()) for key in tqdm.tqdm(theta_1.keys()): + if key in chckpoint_dict_skip_on_merge: + continue + if 'model' in key: if key in theta_2: t2 = theta_2.get(key, torch.zeros_like(theta_1[key])) @@ -345,12 +375,10 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') print("Merging...") - - chckpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] - + shared.state.textinfo = 'Merging A and B' shared.state.sampling_steps = len(theta_0.keys()) for key in tqdm.tqdm(theta_0.keys()): - if 'model' in key and key in theta_1: + if theta_1 and 'model' in key and key in theta_1: if key in chckpoint_dict_skip_on_merge: continue @@ -358,7 +386,6 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ a = theta_0[key] b = theta_1[key] - shared.state.textinfo = f'Merging layer {key}' # this enables merging an inpainting model (A) with another one (B); # where normal model would have 4 channels, for latenst space, inpainting model would # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 @@ -378,34 +405,31 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ shared.state.sampling_step += 1 - # I believe this part should be discarded, but I'll leave it for now until I am sure - for key in theta_1.keys(): - if 'model' in key and key not in theta_0: + del theta_1 + + bake_in_vae_filename = sd_vae.vae_dict.get(bake_in_vae, None) + if bake_in_vae_filename is not None: + print(f"Baking in VAE from {bake_in_vae_filename}") + shared.state.textinfo = 'Baking in VAE' + vae_dict = sd_vae.load_vae_dict(bake_in_vae_filename, map_location='cpu') - if key in chckpoint_dict_skip_on_merge: - continue + for key in vae_dict.keys(): + theta_0_key = 'first_stage_model.' + key + if theta_0_key in theta_0: + theta_0[theta_0_key] = vae_dict[key].half() if save_as_half else vae_dict[key] - theta_0[key] = theta_1[key] - if save_as_half: - theta_0[key] = theta_0[key].half() - del theta_1 + del vae_dict ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path - filename = \ - primary_model_info.model_name + '_' + str(round(1-multiplier, 2)) + '-' + \ - secondary_model_info.model_name + '_' + str(round(multiplier, 2)) + '-' + \ - interp_method.replace(" ", "_") + \ - '-merged.' + \ - ("inpainting." if result_is_inpainting_model else "") + \ - checkpoint_format - - filename = filename if custom_name == '' else (custom_name + '.' + checkpoint_format) + filename = filename_generator() if custom_name == '' else custom_name + filename += ".inpainting" if result_is_inpainting_model else "" + filename += "." + checkpoint_format output_modelname = os.path.join(ckpt_dir, filename) shared.state.nextjob() - shared.state.textinfo = f"Saving to {output_modelname}..." + shared.state.textinfo = "Saving" print(f"Saving to {output_modelname}...") _, extension = os.path.splitext(output_modelname) @@ -418,8 +442,8 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) - print("Checkpoint saved.") - shared.state.textinfo = "Checkpoint saved to " + output_modelname + print(f"Checkpoint saved to {output_modelname}.") + shared.state.textinfo = "Checkpoint saved" shared.state.end() return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] diff --git a/modules/sd_vae.py b/modules/sd_vae.py index da1bf15c..4ce238b8 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -120,6 +120,12 @@ def resolve_vae(checkpoint_file): return None, None +def load_vae_dict(filename, map_location): + vae_ckpt = sd_models.read_state_dict(filename, map_location=map_location) + vae_dict_1 = {k: v for k, v in vae_ckpt.items() if k[0:4] != "loss" and k not in vae_ignore_keys} + return vae_dict_1 + + def load_vae(model, vae_file=None, vae_source="from unknown source"): global vae_dict, loaded_vae_file # save_settings = False @@ -137,8 +143,7 @@ def load_vae(model, vae_file=None, vae_source="from unknown source"): print(f"Loading VAE weights {vae_source}: {vae_file}") store_base_vae(model) - vae_ckpt = sd_models.read_state_dict(vae_file, map_location=shared.weight_load_location) - vae_dict_1 = {k: v for k, v in vae_ckpt.items() if k[0:4] != "loss" and k not in vae_ignore_keys} + vae_dict_1 = load_vae_dict(vae_file, map_location=shared.weight_load_location) _load_vae_dict(model, vae_dict_1) if cache_enabled: diff --git a/modules/shared.py b/modules/shared.py index 77e5e91c..29b28bff 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -20,10 +20,11 @@ from modules.paths import models_path, script_path, sd_path demo = None +sd_default_config = os.path.join(script_path, "configs/v1-inference.yaml") sd_model_file = os.path.join(script_path, 'model.ckpt') default_sd_model_file = sd_model_file parser = argparse.ArgumentParser() -parser.add_argument("--config", type=str, default=os.path.join(script_path, "configs/v1-inference.yaml"), help="path to config which constructs model",) +parser.add_argument("--config", type=str, default=sd_default_config, help="path to config which constructs model",) parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; if specified, this checkpoint will be added to the list of checkpoints and loaded",) parser.add_argument("--ckpt-dir", type=str, default=None, help="Path to directory with stable diffusion checkpoints") parser.add_argument("--vae-dir", type=str, default=None, help="Path to directory with VAE files") diff --git a/modules/ui.py b/modules/ui.py index aeee7853..4e381a49 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -20,7 +20,7 @@ import numpy as np from PIL import Image, PngImagePlugin from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, sd_vae from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML from modules.paths import script_path @@ -1185,7 +1185,7 @@ def create_ui(): with gr.Column(variant='compact'): gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") - with FormRow(): + with FormRow(elem_id="modelmerger_models"): primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") create_refresh_button(primary_model_name, modules.sd_models.list_models, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, "refresh_checkpoint_A") @@ -1197,13 +1197,20 @@ def create_ui(): custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") - interp_method = gr.Radio(choices=["Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") + interp_method = gr.Radio(choices=["No interpolation", "Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") with FormRow(): checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") save_as_half = gr.Checkbox(value=False, label="Save as float16", elem_id="modelmerger_save_as_half") - config_source = gr.Radio(choices=["A, B or C", "B", "C", "Don't"], value="A, B or C", label="Copy config from", type="index", elem_id="modelmerger_config_method") + with FormRow(): + with gr.Column(): + config_source = gr.Radio(choices=["A, B or C", "B", "C", "Don't"], value="A, B or C", label="Copy config from", type="index", elem_id="modelmerger_config_method") + + with gr.Column(): + with FormRow(): + bake_in_vae = gr.Dropdown(choices=["None"] + list(sd_vae.vae_dict), value="None", label="Bake in VAE", elem_id="modelmerger_bake_in_vae") + create_refresh_button(bake_in_vae, sd_vae.refresh_vae_list, lambda: {"choices": ["None"] + list(sd_vae.vae_dict)}, "modelmerger_refresh_bake_in_vae") with gr.Row(): modelmerger_merge = gr.Button(elem_id="modelmerger_merge", value="Merge", variant='primary') @@ -1757,6 +1764,7 @@ def create_ui(): return [*[gr.Dropdown.update(choices=modules.sd_models.checkpoint_tiles()) for _ in range(4)], f"Error merging checkpoints: {e}"] return results + modelmerger_merge.click(fn=lambda: '', inputs=[], outputs=[modelmerger_result]) modelmerger_merge.click( fn=wrap_gradio_gpu_call(modelmerger, extra_outputs=lambda: [gr.update() for _ in range(4)]), _js='modelmerger', @@ -1771,6 +1779,7 @@ def create_ui(): custom_name, checkpoint_format, config_source, + bake_in_vae, ], outputs=[ primary_model_name, -- cgit v1.2.3 From 54674674b813894b908283531ddaab4ccfeac721 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 12:12:09 +0300 Subject: allow having at half precision when there is only one checkpoint in merger tab --- modules/extras.py | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index fe701a0e..d03f976e 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -278,6 +278,13 @@ def create_config(ckpt_result, config_source, a, b, c): chckpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] +def to_half(tensor, enable): + if enable and tensor.dtype == torch.float: + return tensor.half() + + return tensor + + def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae): shared.state.begin() shared.state.job = 'model-merge' @@ -400,8 +407,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ else: theta_0[key] = theta_func2(a, b, multiplier) - if save_as_half: - theta_0[key] = theta_0[key].half() + theta_0[key] = to_half(theta_0[key], save_as_half) shared.state.sampling_step += 1 @@ -416,10 +422,14 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ for key in vae_dict.keys(): theta_0_key = 'first_stage_model.' + key if theta_0_key in theta_0: - theta_0[theta_0_key] = vae_dict[key].half() if save_as_half else vae_dict[key] + theta_0[theta_0_key] = to_half(vae_dict[key], save_as_half) del vae_dict + if save_as_half and not theta_func2: + for key in theta_0.keys(): + theta_0[key] = to_half(theta_0[key], save_as_half) + ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path filename = filename_generator() if custom_name == '' else custom_name -- cgit v1.2.3 From 18a09c7e0032e2e655269e8e2b4f1ca6ed0cc7d3 Mon Sep 17 00:00:00 2001 From: dan Date: Thu, 19 Jan 2023 17:36:23 +0800 Subject: Simplification and bugfix --- modules/textual_inversion/preprocess.py | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py index 86c1cd33..454dcc36 100644 --- a/modules/textual_inversion/preprocess.py +++ b/modules/textual_inversion/preprocess.py @@ -124,13 +124,11 @@ def center_crop(image: Image, w: int, h: int): def multicrop_pic(image: Image, mindim, maxdim, minarea, maxarea, objective, threshold): iw, ih = image.size err = lambda w, h: 1-(lambda x: x if x < 1 else 1/x)(iw/ih/(w/h)) - try: - w, h = max(((w, h) for w in range(mindim, maxdim+1, 64) for h in range(mindim, maxdim+1, 64) - if minarea <= w * h <= maxarea and err(w, h) <= threshold), - key= lambda wh: ((objective=='Maximize area')*wh[0]*wh[1], -err(*wh)) - ) - except ValueError: - return + w, h = max(((w, h) for w in range(mindim, maxdim+1, 64) for h in range(mindim, maxdim+1, 64) + if minarea <= w * h <= maxarea and err(w, h) <= threshold), + key= lambda wh: (wh[0]*wh[1], -err(*wh))[::1 if objective=='Maximize area' else -1], + default=None + ) return center_crop(image, w, h) -- cgit v1.2.3 From 2985b317d719f0f0580d2ff93f3008ccabb9c251 Mon Sep 17 00:00:00 2001 From: dan Date: Thu, 19 Jan 2023 17:39:30 +0800 Subject: Fix of fix --- modules/textual_inversion/preprocess.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py index 454dcc36..c0ac11d3 100644 --- a/modules/textual_inversion/preprocess.py +++ b/modules/textual_inversion/preprocess.py @@ -124,12 +124,12 @@ def center_crop(image: Image, w: int, h: int): def multicrop_pic(image: Image, mindim, maxdim, minarea, maxarea, objective, threshold): iw, ih = image.size err = lambda w, h: 1-(lambda x: x if x < 1 else 1/x)(iw/ih/(w/h)) - w, h = max(((w, h) for w in range(mindim, maxdim+1, 64) for h in range(mindim, maxdim+1, 64) + wh = max(((w, h) for w in range(mindim, maxdim+1, 64) for h in range(mindim, maxdim+1, 64) if minarea <= w * h <= maxarea and err(w, h) <= threshold), key= lambda wh: (wh[0]*wh[1], -err(*wh))[::1 if objective=='Maximize area' else -1], default=None ) - return center_crop(image, w, h) + return wh and center_crop(image, *wh) def preprocess_work(process_src, process_dst, process_width, process_height, preprocess_txt_action, process_flip, process_split, process_caption, process_caption_deepbooru=False, split_threshold=0.5, overlap_ratio=0.2, process_focal_crop=False, process_focal_crop_face_weight=0.9, process_focal_crop_entropy_weight=0.3, process_focal_crop_edges_weight=0.5, process_focal_crop_debug=False, process_multicrop=None, process_multicrop_mindim=None, process_multicrop_maxdim=None, process_multicrop_minarea=None, process_multicrop_maxarea=None, process_multicrop_objective=None, process_multicrop_threshold=None): -- cgit v1.2.3 From b271e22f7ac1b2cabca8985b1e4437ab685a2c21 Mon Sep 17 00:00:00 2001 From: vt-idiot <81622808+vt-idiot@users.noreply.github.com> Date: Thu, 19 Jan 2023 06:12:19 -0500 Subject: Update shared.py `Witdth/Height` was driving me insane. -> `Width/Height` --- modules/shared.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 29b28bff..2f366454 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -448,7 +448,7 @@ options_templates.update(options_section(('ui', "User interface"), { "js_modal_lightbox_initially_zoomed": OptionInfo(True, "Show images zoomed in by default in full page image viewer"), "show_progress_in_title": OptionInfo(True, "Show generation progress in window title."), "samplers_in_dropdown": OptionInfo(True, "Use dropdown for sampler selection instead of radio group"), - "dimensions_and_batch_together": OptionInfo(True, "Show Witdth/Height and Batch sliders in same row"), + "dimensions_and_batch_together": OptionInfo(True, "Show Width/Height and Batch sliders in same row"), 'quicksettings': OptionInfo("sd_model_checkpoint", "Quicksettings list"), 'ui_reorder': OptionInfo(", ".join(ui_reorder_categories), "txt2img/img2img UI item order"), 'localization': OptionInfo("None", "Localization (requires restart)", gr.Dropdown, lambda: {"choices": ["None"] + list(localization.localizations.keys())}, refresh=lambda: localization.list_localizations(cmd_opts.localizations_dir)), -- cgit v1.2.3 From d1ea518dea3d7584be2927cc486d15ec3e18ddb0 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 18:07:37 +0300 Subject: remember the list of checkpoints after you press refresh button and reload the page --- modules/ui.py | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index af416d5f..0c5ba358 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1771,8 +1771,17 @@ def create_ui(): component_keys = [k for k in opts.data_labels.keys() if k in component_dict] + def get_value_for_setting(key): + value = getattr(opts, key) + + info = opts.data_labels[key] + args = info.component_args() if callable(info.component_args) else info.component_args or {} + args = {k: v for k, v in args.items() if k not in {'precision'}} + + return gr.update(value=value, **args) + def get_settings_values(): - return [getattr(opts, key) for key in component_keys] + return [get_value_for_setting(key) for key in component_keys] demo.load( fn=get_settings_values, -- cgit v1.2.3 From f2ae2529877072874ebaac0257fe4af48c5855a4 Mon Sep 17 00:00:00 2001 From: EllangoK Date: Thu, 19 Jan 2023 10:24:17 -0500 Subject: fixes minor typos around run_modelmerger --- modules/extras.py | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index d03f976e..1218f88f 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -275,7 +275,7 @@ def create_config(ckpt_result, config_source, a, b, c): shutil.copyfile(cfg, checkpoint_filename) -chckpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] +checkpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] def to_half(tensor, enable): @@ -303,7 +303,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ def add_difference(theta0, theta1_2_diff, alpha): return theta0 + (alpha * theta1_2_diff) - def filename_weighed_sum(): + def filename_weighted_sum(): a = primary_model_info.model_name b = secondary_model_info.model_name Ma = round(1 - multiplier, 2) @@ -311,7 +311,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ return f"{Ma}({a}) + {Mb}({b})" - def filename_add_differnece(): + def filename_add_difference(): a = primary_model_info.model_name b = secondary_model_info.model_name c = tertiary_model_info.model_name @@ -323,8 +323,8 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ return primary_model_info.model_name theta_funcs = { - "Weighted sum": (filename_weighed_sum, None, weighted_sum), - "Add difference": (filename_add_differnece, get_difference, add_difference), + "Weighted sum": (filename_weighted_sum, None, weighted_sum), + "Add difference": (filename_add_difference, get_difference, add_difference), "No interpolation": (filename_nothing, None, None), } filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] @@ -362,7 +362,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ shared.state.textinfo = 'Merging B and C' shared.state.sampling_steps = len(theta_1.keys()) for key in tqdm.tqdm(theta_1.keys()): - if key in chckpoint_dict_skip_on_merge: + if key in checkpoint_dict_skip_on_merge: continue if 'model' in key: @@ -387,7 +387,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ for key in tqdm.tqdm(theta_0.keys()): if theta_1 and 'model' in key and key in theta_1: - if key in chckpoint_dict_skip_on_merge: + if key in checkpoint_dict_skip_on_merge: continue a = theta_0[key] -- cgit v1.2.3 From c1928cdd6194928af0f53f70c51d59479b7025e2 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 18:58:08 +0300 Subject: bring back short hashes to sd checkpoint selection --- modules/sd_models.py | 15 +++++++++++---- modules/ui.py | 23 ++++++++++++----------- 2 files changed, 23 insertions(+), 15 deletions(-) (limited to 'modules') diff --git a/modules/sd_models.py b/modules/sd_models.py index 6a681cef..12083848 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -41,14 +41,16 @@ class CheckpointInfo: if name.startswith("\\") or name.startswith("/"): name = name[1:] - self.title = name + self.name = name self.model_name = os.path.splitext(name.replace("/", "_").replace("\\", "_"))[0] self.hash = model_hash(filename) - self.sha256 = hashes.sha256_from_cache(self.filename, "checkpoint/" + self.title) + self.sha256 = hashes.sha256_from_cache(self.filename, "checkpoint/" + name) self.shorthash = self.sha256[0:10] if self.sha256 else None - self.ids = [self.hash, self.model_name, self.title, f'{name} [{self.hash}]'] + ([self.shorthash, self.sha256] if self.shorthash else []) + self.title = name if self.shorthash is None else f'{name} [{self.shorthash}]' + + self.ids = [self.hash, self.model_name, self.title, name, f'{name} [{self.hash}]'] + ([self.shorthash, self.sha256, f'{self.name} [{self.shorthash}]'] if self.shorthash else []) def register(self): checkpoints_list[self.title] = self @@ -56,13 +58,15 @@ class CheckpointInfo: checkpoint_alisases[id] = self def calculate_shorthash(self): - self.sha256 = hashes.sha256(self.filename, "checkpoint/" + self.title) + self.sha256 = hashes.sha256(self.filename, "checkpoint/" + self.name) self.shorthash = self.sha256[0:10] if self.shorthash not in self.ids: self.ids += [self.shorthash, self.sha256] self.register() + self.title = f'{self.name} [{self.shorthash}]' + return self.shorthash @@ -225,7 +229,10 @@ def read_state_dict(checkpoint_file, print_global_state=False, map_location=None def load_model_weights(model, checkpoint_info: CheckpointInfo): + title = checkpoint_info.title sd_model_hash = checkpoint_info.calculate_shorthash() + if checkpoint_info.title != title: + shared.opts.data["sd_model_checkpoint"] = checkpoint_info.title cache_enabled = shared.opts.sd_checkpoint_cache > 0 diff --git a/modules/ui.py b/modules/ui.py index 0c5ba358..13d80ae2 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -439,7 +439,7 @@ def apply_setting(key, value): opts.data_labels[key].onchange() opts.save(shared.config_filename) - return value + return getattr(opts, key) def update_generation_info(generation_info, html_info, img_index): @@ -597,6 +597,16 @@ def ordered_ui_categories(): yield category +def get_value_for_setting(key): + value = getattr(opts, key) + + info = opts.data_labels[key] + args = info.component_args() if callable(info.component_args) else info.component_args or {} + args = {k: v for k, v in args.items() if k not in {'precision'}} + + return gr.update(value=value, **args) + + def create_ui(): import modules.img2img import modules.txt2img @@ -1600,7 +1610,7 @@ def create_ui(): opts.save(shared.config_filename) - return gr.update(value=value), opts.dumpjson() + return get_value_for_setting(key), opts.dumpjson() with gr.Blocks(analytics_enabled=False) as settings_interface: with gr.Row(): @@ -1771,15 +1781,6 @@ def create_ui(): component_keys = [k for k in opts.data_labels.keys() if k in component_dict] - def get_value_for_setting(key): - value = getattr(opts, key) - - info = opts.data_labels[key] - args = info.component_args() if callable(info.component_args) else info.component_args or {} - args = {k: v for k, v in args.items() if k not in {'precision'}} - - return gr.update(value=value, **args) - def get_settings_values(): return [get_value_for_setting(key) for key in component_keys] -- cgit v1.2.3 From 6073456c8348d15716b9bc5276d994fe8554e4ca Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 19 Jan 2023 20:39:03 +0300 Subject: write a comment for fix_checkpoint function --- modules/sd_hijack.py | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'modules') diff --git a/modules/sd_hijack.py b/modules/sd_hijack.py index 870eba88..f9652d21 100644 --- a/modules/sd_hijack.py +++ b/modules/sd_hijack.py @@ -69,6 +69,13 @@ def undo_optimizations(): ldm.modules.diffusionmodules.model.AttnBlock.forward = diffusionmodules_model_AttnBlock_forward +def fix_checkpoint(): + """checkpoints are now added and removed in embedding/hypernet code, since torch doesn't want + checkpoints to be added when not training (there's a warning)""" + + pass + + class StableDiffusionModelHijack: fixes = None comments = [] -- cgit v1.2.3 From 6c7a50d783c4e406d8597f9cf354bb8128026f6c Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 20 Jan 2023 08:36:30 +0300 Subject: remove some unnecessary logging to javascript console --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 13d80ae2..eb45a128 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -532,7 +532,7 @@ Requested path was: {f} generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") generation_info_button.click( fn=update_generation_info, - _js="function(x, y, z){ console.log(x, y, z); return [x, y, selected_gallery_index()] }", + _js="function(x, y, z){ return [x, y, selected_gallery_index()] }", inputs=[generation_info, html_info, html_info], outputs=[html_info, html_info], ) -- cgit v1.2.3 From 98466da4bc312c0fa9c8cea4c825afc64194cb58 Mon Sep 17 00:00:00 2001 From: EllangoK Date: Fri, 20 Jan 2023 00:48:15 -0500 Subject: adds descriptions for merging methods in ui --- modules/ui.py | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index eb45a128..ee434bde 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1190,10 +1190,19 @@ def create_ui(): outputs=[html, generation_info, html2], ) + def update_interp_description(value): + interp_description_css = "

{}

" + interp_descriptions = { + "No interpolation": interp_description_css.format("No interpolation will be used. Requires one model; A. Allows for format conversion and VAE baking."), + "Weighted sum": interp_description_css.format("A weighted sum will be used for interpolation. Requires two models; A and B. The result is calculated as A * (1 - M) + B * M"), + "Add difference": interp_description_css.format("The difference between the last two models will be added to the first. Requires three models; A, B and C. The result is calculated as A + (B - C) * M") + } + return interp_descriptions[value] + with gr.Blocks(analytics_enabled=False) as modelmerger_interface: with gr.Row().style(equal_height=False): with gr.Column(variant='compact'): - gr.HTML(value="

A merger of the two checkpoints will be generated in your checkpoint directory.

") + interp_description = gr.HTML(value=update_interp_description("Weighted sum"), elem_id="modelmerger_interp_description") with FormRow(elem_id="modelmerger_models"): primary_model_name = gr.Dropdown(modules.sd_models.checkpoint_tiles(), elem_id="modelmerger_primary_model_name", label="Primary model (A)") @@ -1208,6 +1217,7 @@ def create_ui(): custom_name = gr.Textbox(label="Custom Name (Optional)", elem_id="modelmerger_custom_name") interp_amount = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, label='Multiplier (M) - set to 0 to get model A', value=0.3, elem_id="modelmerger_interp_amount") interp_method = gr.Radio(choices=["No interpolation", "Weighted sum", "Add difference"], value="Weighted sum", label="Interpolation Method", elem_id="modelmerger_interp_method") + interp_method.change(fn=update_interp_description, inputs=[interp_method], outputs=[interp_description]) with FormRow(): checkpoint_format = gr.Radio(choices=["ckpt", "safetensors"], value="ckpt", label="Checkpoint format", elem_id="modelmerger_checkpoint_format") @@ -1903,6 +1913,9 @@ def create_ui(): with open(ui_config_file, "w", encoding="utf8") as file: json.dump(ui_settings, file, indent=4) + # Required as a workaround for change() event not triggering when loading values from ui-config.json + interp_description.value = update_interp_description(interp_method.value) + return demo -- cgit v1.2.3 From 20a59ab3b171f398abd09087108c1ed087dbea9b Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 20 Jan 2023 10:18:41 +0300 Subject: move token counter to the location of the prompt, add token counting for the negative prompt --- modules/ui.py | 25 ++++++++++++------------- 1 file changed, 12 insertions(+), 13 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index eb45a128..06c11848 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -335,28 +335,23 @@ def update_token_counter(text, steps): flat_prompts = reduce(lambda list1, list2: list1+list2, prompt_schedules) prompts = [prompt_text for step, prompt_text in flat_prompts] token_count, max_length = max([model_hijack.get_prompt_lengths(prompt) for prompt in prompts], key=lambda args: args[0]) - style_class = ' class="red"' if (token_count > max_length) else "" - return f"{token_count}/{max_length}" + return f"{token_count}/{max_length}" def create_toprow(is_img2img): id_part = "img2img" if is_img2img else "txt2img" - with gr.Row(elem_id="toprow"): - with gr.Column(scale=6): + with gr.Row(elem_id=f"{id_part}_toprow", variant="compact"): + with gr.Column(elem_id=f"{id_part}_prompt_container", scale=6): with gr.Row(): with gr.Column(scale=80): with gr.Row(): - prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, - placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) + prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)") with gr.Row(): with gr.Column(scale=80): with gr.Row(): - negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, - placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)" - ) + negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)") with gr.Column(scale=1, elem_id="roll_col"): paste = gr.Button(value=paste_symbol, elem_id="paste") @@ -365,6 +360,8 @@ def create_toprow(is_img2img): clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") + negative_token_counter = gr.HTML(value="", elem_id=f"{id_part}_negative_token_counter") + negative_token_button = gr.Button(visible=False, elem_id=f"{id_part}_negative_token_button") clear_prompt_button.click( fn=lambda *x: x, @@ -402,7 +399,7 @@ def create_toprow(is_img2img): prompt_styles = gr.Dropdown(label="Styles", elem_id=f"{id_part}_styles", choices=[k for k, v in shared.prompt_styles.styles.items()], value=[], multiselect=True) create_refresh_button(prompt_styles, shared.prompt_styles.reload, lambda: {"choices": [k for k, v in shared.prompt_styles.styles.items()]}, f"refresh_{id_part}_styles") - return prompt, prompt_styles, negative_prompt, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button + return prompt, prompt_styles, negative_prompt, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button, negative_token_counter, negative_token_button def setup_progressbar(*args, **kwargs): @@ -619,7 +616,7 @@ def create_ui(): modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) with gr.Blocks(analytics_enabled=False) as txt2img_interface: - txt2img_prompt, txt2img_prompt_styles, txt2img_negative_prompt, submit, _, _,txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button = create_toprow(is_img2img=False) + txt2img_prompt, txt2img_prompt_styles, txt2img_negative_prompt, submit, _, _, txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button, negative_token_counter, negative_token_button = create_toprow(is_img2img=False) dummy_component = gr.Label(visible=False) txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="binary", visible=False) @@ -795,12 +792,13 @@ def create_ui(): ] token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) + negative_token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_negative_prompt, steps], outputs=[negative_token_counter]) modules.scripts.scripts_current = modules.scripts.scripts_img2img modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) with gr.Blocks(analytics_enabled=False) as img2img_interface: - img2img_prompt, img2img_prompt_styles, img2img_negative_prompt, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste,token_counter, token_button = create_toprow(is_img2img=True) + img2img_prompt, img2img_prompt_styles, img2img_negative_prompt, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste, token_counter, token_button, negative_token_counter, negative_token_button = create_toprow(is_img2img=True) img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="binary", visible=False) @@ -1064,6 +1062,7 @@ def create_ui(): ) token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) + negative_token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_negative_prompt, steps], outputs=[negative_token_counter]) img2img_paste_fields = [ (img2img_prompt, "Prompt"), -- cgit v1.2.3 From 40ff6db5325fc34ad4fa35e80cb1e7768d9f7e75 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 08:36:07 +0300 Subject: extra networks UI rework of hypernets: rather than via settings, hypernets are added directly to prompt as --- modules/api/api.py | 7 +- modules/extra_networks.py | 147 ++++++++++++++++++++++++ modules/extra_networks_hypernet.py | 21 ++++ modules/generation_parameters_copypaste.py | 12 +- modules/hypernetworks/hypernetwork.py | 107 ++++++++++++------ modules/hypernetworks/ui.py | 5 +- modules/processing.py | 24 ++-- modules/sd_hijack_optimizations.py | 10 +- modules/shared.py | 21 +++- modules/textual_inversion/textual_inversion.py | 2 + modules/ui.py | 50 ++++++--- modules/ui_components.py | 10 ++ modules/ui_extra_networks.py | 149 +++++++++++++++++++++++++ modules/ui_extra_networks_hypernets.py | 34 ++++++ modules/ui_extra_networks_textual_inversion.py | 32 ++++++ 15 files changed, 544 insertions(+), 87 deletions(-) create mode 100644 modules/extra_networks.py create mode 100644 modules/extra_networks_hypernet.py create mode 100644 modules/ui_extra_networks.py create mode 100644 modules/ui_extra_networks_hypernets.py create mode 100644 modules/ui_extra_networks_textual_inversion.py (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 9814bbc2..2c371e6e 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -480,7 +480,7 @@ class Api: def train_hypernetwork(self, args: dict): try: shared.state.begin() - initial_hypernetwork = shared.loaded_hypernetwork + shared.loaded_hypernetworks = [] apply_optimizations = shared.opts.training_xattention_optimizations error = None filename = '' @@ -491,16 +491,15 @@ class Api: except Exception as e: error = e finally: - shared.loaded_hypernetwork = initial_hypernetwork shared.sd_model.cond_stage_model.to(devices.device) shared.sd_model.first_stage_model.to(devices.device) if not apply_optimizations: sd_hijack.apply_optimizations() shared.state.end() - return TrainResponse(info = "train embedding complete: filename: {filename} error: {error}".format(filename = filename, error = error)) + return TrainResponse(info="train embedding complete: filename: {filename} error: {error}".format(filename=filename, error=error)) except AssertionError as msg: shared.state.end() - return TrainResponse(info = "train embedding error: {error}".format(error = error)) + return TrainResponse(info="train embedding error: {error}".format(error=error)) def get_memory(self): try: diff --git a/modules/extra_networks.py b/modules/extra_networks.py new file mode 100644 index 00000000..1978673d --- /dev/null +++ b/modules/extra_networks.py @@ -0,0 +1,147 @@ +import re +from collections import defaultdict + +from modules import errors + +extra_network_registry = {} + + +def initialize(): + extra_network_registry.clear() + + +def register_extra_network(extra_network): + extra_network_registry[extra_network.name] = extra_network + + +class ExtraNetworkParams: + def __init__(self, items=None): + self.items = items or [] + + +class ExtraNetwork: + def __init__(self, name): + self.name = name + + def activate(self, p, params_list): + """ + Called by processing on every run. Whatever the extra network is meant to do should be activated here. + Passes arguments related to this extra network in params_list. + User passes arguments by specifying this in his prompt: + + + + Where name matches the name of this ExtraNetwork object, and arg1:arg2:arg3 are any natural number of text arguments + separated by colon. + + Even if the user does not mention this ExtraNetwork in his prompt, the call will stil be made, with empty params_list - + in this case, all effects of this extra networks should be disabled. + + Can be called multiple times before deactivate() - each new call should override the previous call completely. + + For example, if this ExtraNetwork's name is 'hypernet' and user's prompt is: + + > "1girl, " + + params_list will be: + + [ + ExtraNetworkParams(items=["agm", "1.1"]), + ExtraNetworkParams(items=["ray"]) + ] + + """ + raise NotImplementedError + + def deactivate(self, p): + """ + Called at the end of processing for housekeeping. No need to do anything here. + """ + + raise NotImplementedError + + +def activate(p, extra_network_data): + """call activate for extra networks in extra_network_data in specified order, then call + activate for all remaining registered networks with an empty argument list""" + + for extra_network_name, extra_network_args in extra_network_data.items(): + extra_network = extra_network_registry.get(extra_network_name, None) + if extra_network is None: + print(f"Skipping unknown extra network: {extra_network_name}") + continue + + try: + extra_network.activate(p, extra_network_args) + except Exception as e: + errors.display(e, f"activating extra network {extra_network_name} with arguments {extra_network_args}") + + for extra_network_name, extra_network in extra_network_registry.items(): + args = extra_network_data.get(extra_network_name, None) + if args is not None: + continue + + try: + extra_network.activate(p, []) + except Exception as e: + errors.display(e, f"activating extra network {extra_network_name}") + + +def deactivate(p, extra_network_data): + """call deactivate for extra networks in extra_network_data in specified order, then call + deactivate for all remaining registered networks""" + + for extra_network_name, extra_network_args in extra_network_data.items(): + extra_network = extra_network_registry.get(extra_network_name, None) + if extra_network is None: + continue + + try: + extra_network.deactivate(p) + except Exception as e: + errors.display(e, f"deactivating extra network {extra_network_name}") + + for extra_network_name, extra_network in extra_network_registry.items(): + args = extra_network_data.get(extra_network_name, None) + if args is not None: + continue + + try: + extra_network.deactivate(p) + except Exception as e: + errors.display(e, f"deactivating unmentioned extra network {extra_network_name}") + + +re_extra_net = re.compile(r"<(\w+):([^>]+)>") + + +def parse_prompt(prompt): + res = defaultdict(list) + + def found(m): + name = m.group(1) + args = m.group(2) + + res[name].append(ExtraNetworkParams(items=args.split(":"))) + + return "" + + prompt = re.sub(re_extra_net, found, prompt) + + return prompt, res + + +def parse_prompts(prompts): + res = [] + extra_data = None + + for prompt in prompts: + updated_prompt, parsed_extra_data = parse_prompt(prompt) + + if extra_data is None: + extra_data = parsed_extra_data + + res.append(updated_prompt) + + return res, extra_data + diff --git a/modules/extra_networks_hypernet.py b/modules/extra_networks_hypernet.py new file mode 100644 index 00000000..6a0c4ba8 --- /dev/null +++ b/modules/extra_networks_hypernet.py @@ -0,0 +1,21 @@ +from modules import extra_networks +from modules.hypernetworks import hypernetwork + + +class ExtraNetworkHypernet(extra_networks.ExtraNetwork): + def __init__(self): + super().__init__('hypernet') + + def activate(self, p, params_list): + names = [] + multipliers = [] + for params in params_list: + assert len(params.items) > 0 + + names.append(params.items[0]) + multipliers.append(float(params.items[1]) if len(params.items) > 1 else 1.0) + + hypernetwork.load_hypernetworks(names, multipliers) + + def deactivate(p, self): + pass diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index a381ff59..46e12dc6 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -79,8 +79,6 @@ def integrate_settings_paste_fields(component_dict): from modules import ui settings_map = { - 'sd_hypernetwork': 'Hypernet', - 'sd_hypernetwork_strength': 'Hypernet strength', 'CLIP_stop_at_last_layers': 'Clip skip', 'inpainting_mask_weight': 'Conditional mask weight', 'sd_model_checkpoint': 'Model hash', @@ -275,13 +273,9 @@ Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 965400086, Size: 512x512, Model if "Clip skip" not in res: res["Clip skip"] = "1" - if "Hypernet strength" not in res: - res["Hypernet strength"] = "1" - - if "Hypernet" in res: - hypernet_name = res["Hypernet"] - hypernet_hash = res.get("Hypernet hash", None) - res["Hypernet"] = find_hypernetwork_key(hypernet_name, hypernet_hash) + hypernet = res.get("Hypernet", None) + if hypernet is not None: + res["Prompt"] += f"""""" if "Hires resize-1" not in res: res["Hires resize-1"] = 0 diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 74e78582..80a47c79 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -25,7 +25,6 @@ from statistics import stdev, mean optimizer_dict = {optim_name : cls_obj for optim_name, cls_obj in inspect.getmembers(torch.optim, inspect.isclass) if optim_name != "Optimizer"} class HypernetworkModule(torch.nn.Module): - multiplier = 1.0 activation_dict = { "linear": torch.nn.Identity, "relu": torch.nn.ReLU, @@ -41,6 +40,8 @@ class HypernetworkModule(torch.nn.Module): add_layer_norm=False, activate_output=False, dropout_structure=None): super().__init__() + self.multiplier = 1.0 + assert layer_structure is not None, "layer_structure must not be None" assert layer_structure[0] == 1, "Multiplier Sequence should start with size 1!" assert layer_structure[-1] == 1, "Multiplier Sequence should end with size 1!" @@ -115,7 +116,7 @@ class HypernetworkModule(torch.nn.Module): state_dict[to] = x def forward(self, x): - return x + self.linear(x) * (HypernetworkModule.multiplier if not self.training else 1) + return x + self.linear(x) * (self.multiplier if not self.training else 1) def trainables(self): layer_structure = [] @@ -125,9 +126,6 @@ class HypernetworkModule(torch.nn.Module): return layer_structure -def apply_strength(value=None): - HypernetworkModule.multiplier = value if value is not None else shared.opts.sd_hypernetwork_strength - #param layer_structure : sequence used for length, use_dropout : controlling boolean, last_layer_dropout : for compatibility check. def parse_dropout_structure(layer_structure, use_dropout, last_layer_dropout): if layer_structure is None: @@ -192,6 +190,20 @@ class Hypernetwork: for param in layer.parameters(): param.requires_grad = mode + def to(self, device): + for k, layers in self.layers.items(): + for layer in layers: + layer.to(device) + + return self + + def set_multiplier(self, multiplier): + for k, layers in self.layers.items(): + for layer in layers: + layer.multiplier = multiplier + + return self + def eval(self): for k, layers in self.layers.items(): for layer in layers: @@ -269,11 +281,13 @@ class Hypernetwork: self.optimizer_state_dict = None if self.optimizer_state_dict: self.optimizer_name = optimizer_saved_dict.get('optimizer_name', 'AdamW') - print("Loaded existing optimizer from checkpoint") - print(f"Optimizer name is {self.optimizer_name}") + if shared.opts.print_hypernet_extra: + print("Loaded existing optimizer from checkpoint") + print(f"Optimizer name is {self.optimizer_name}") else: self.optimizer_name = "AdamW" - print("No saved optimizer exists in checkpoint") + if shared.opts.print_hypernet_extra: + print("No saved optimizer exists in checkpoint") for size, sd in state_dict.items(): if type(size) == int: @@ -306,23 +320,43 @@ def list_hypernetworks(path): return res -def load_hypernetwork(filename): - path = shared.hypernetworks.get(filename, None) - # Prevent any file named "None.pt" from being loaded. - if path is not None and filename != "None": - print(f"Loading hypernetwork {filename}") - try: - shared.loaded_hypernetwork = Hypernetwork() - shared.loaded_hypernetwork.load(path) +def load_hypernetwork(name): + path = shared.hypernetworks.get(name, None) - except Exception: - print(f"Error loading hypernetwork {path}", file=sys.stderr) - print(traceback.format_exc(), file=sys.stderr) - else: - if shared.loaded_hypernetwork is not None: - print("Unloading hypernetwork") + if path is None: + return None + + hypernetwork = Hypernetwork() + + try: + hypernetwork.load(path) + except Exception: + print(f"Error loading hypernetwork {path}", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + return None + + return hypernetwork + + +def load_hypernetworks(names, multipliers=None): + already_loaded = {} + + for hypernetwork in shared.loaded_hypernetworks: + if hypernetwork.name in names: + already_loaded[hypernetwork.name] = hypernetwork - shared.loaded_hypernetwork = None + shared.loaded_hypernetworks.clear() + + for i, name in enumerate(names): + hypernetwork = already_loaded.get(name, None) + if hypernetwork is None: + hypernetwork = load_hypernetwork(name) + + if hypernetwork is None: + continue + + hypernetwork.set_multiplier(multipliers[i] if multipliers else 1.0) + shared.loaded_hypernetworks.append(hypernetwork) def find_closest_hypernetwork_name(search: str): @@ -336,18 +370,27 @@ def find_closest_hypernetwork_name(search: str): return applicable[0] -def apply_hypernetwork(hypernetwork, context, layer=None): - hypernetwork_layers = (hypernetwork.layers if hypernetwork is not None else {}).get(context.shape[2], None) +def apply_single_hypernetwork(hypernetwork, context_k, context_v, layer=None): + hypernetwork_layers = (hypernetwork.layers if hypernetwork is not None else {}).get(context_k.shape[2], None) if hypernetwork_layers is None: - return context, context + return context_k, context_v if layer is not None: layer.hyper_k = hypernetwork_layers[0] layer.hyper_v = hypernetwork_layers[1] - context_k = hypernetwork_layers[0](context) - context_v = hypernetwork_layers[1](context) + context_k = hypernetwork_layers[0](context_k) + context_v = hypernetwork_layers[1](context_v) + return context_k, context_v + + +def apply_hypernetworks(hypernetworks, context, layer=None): + context_k = context + context_v = context + for hypernetwork in hypernetworks: + context_k, context_v = apply_single_hypernetwork(hypernetwork, context_k, context_v, layer) + return context_k, context_v @@ -357,7 +400,7 @@ def attention_CrossAttention_forward(self, x, context=None, mask=None): q = self.to_q(x) context = default(context, x) - context_k, context_v = apply_hypernetwork(shared.loaded_hypernetwork, context, self) + context_k, context_v = apply_hypernetworks(shared.loaded_hypernetworks, context, self) k = self.to_k(context_k) v = self.to_v(context_v) @@ -464,8 +507,9 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi template_file = template_file.path path = shared.hypernetworks.get(hypernetwork_name, None) - shared.loaded_hypernetwork = Hypernetwork() - shared.loaded_hypernetwork.load(path) + hypernetwork = Hypernetwork() + hypernetwork.load(path) + shared.loaded_hypernetworks = [hypernetwork] shared.state.job = "train-hypernetwork" shared.state.textinfo = "Initializing hypernetwork training..." @@ -489,7 +533,6 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi else: images_dir = None - hypernetwork = shared.loaded_hypernetwork checkpoint = sd_models.select_checkpoint() initial_step = hypernetwork.step or 0 diff --git a/modules/hypernetworks/ui.py b/modules/hypernetworks/ui.py index 81e3f519..76599f5a 100644 --- a/modules/hypernetworks/ui.py +++ b/modules/hypernetworks/ui.py @@ -9,6 +9,7 @@ from modules import devices, sd_hijack, shared not_available = ["hardswish", "multiheadattention"] keys = list(x for x in modules.hypernetworks.hypernetwork.HypernetworkModule.activation_dict.keys() if x not in not_available) + def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False, dropout_structure=None): filename = modules.hypernetworks.hypernetwork.create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure, activation_func, weight_init, add_layer_norm, use_dropout, dropout_structure) @@ -16,8 +17,7 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, def train_hypernetwork(*args): - - initial_hypernetwork = shared.loaded_hypernetwork + shared.loaded_hypernetworks = [] assert not shared.cmd_opts.lowvram, 'Training models with lowvram is not possible' @@ -34,7 +34,6 @@ Hypernetwork saved to {html.escape(filename)} except Exception: raise finally: - shared.loaded_hypernetwork = initial_hypernetwork shared.sd_model.cond_stage_model.to(devices.device) shared.sd_model.first_stage_model.to(devices.device) sd_hijack.apply_optimizations() diff --git a/modules/processing.py b/modules/processing.py index a3e9f709..b5deeacf 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -13,7 +13,7 @@ from skimage import exposure from typing import Any, Dict, List, Optional import modules.sd_hijack -from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_parameters_copypaste, script_callbacks +from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_parameters_copypaste, script_callbacks, extra_networks from modules.sd_hijack import model_hijack from modules.shared import opts, cmd_opts, state import modules.shared as shared @@ -438,9 +438,6 @@ def create_infotext(p, all_prompts, all_seeds, all_subseeds, comments=None, iter "Size": f"{p.width}x{p.height}", "Model hash": getattr(p, 'sd_model_hash', None if not opts.add_model_hash_to_info or not shared.sd_model.sd_model_hash else shared.sd_model.sd_model_hash), "Model": (None if not opts.add_model_name_to_info or not shared.sd_model.sd_checkpoint_info.model_name else shared.sd_model.sd_checkpoint_info.model_name.replace(',', '').replace(':', '')), - "Hypernet": (None if shared.loaded_hypernetwork is None else shared.loaded_hypernetwork.name), - "Hypernet hash": (None if shared.loaded_hypernetwork is None else shared.loaded_hypernetwork.shorthash()), - "Hypernet strength": (None if shared.loaded_hypernetwork is None or shared.opts.sd_hypernetwork_strength >= 1 else shared.opts.sd_hypernetwork_strength), "Batch size": (None if p.batch_size < 2 else p.batch_size), "Batch pos": (None if p.batch_size < 2 else position_in_batch), "Variation seed": (None if p.subseed_strength == 0 else all_subseeds[index]), @@ -468,14 +465,12 @@ def process_images(p: StableDiffusionProcessing) -> Processed: try: for k, v in p.override_settings.items(): setattr(opts, k, v) - if k == 'sd_hypernetwork': - shared.reload_hypernetworks() # make onchange call for changing hypernet if k == 'sd_model_checkpoint': - sd_models.reload_model_weights() # make onchange call for changing SD model + sd_models.reload_model_weights() if k == 'sd_vae': - sd_vae.reload_vae_weights() # make onchange call for changing VAE + sd_vae.reload_vae_weights() res = process_images_inner(p) @@ -484,9 +479,11 @@ def process_images(p: StableDiffusionProcessing) -> Processed: if p.override_settings_restore_afterwards: for k, v in stored_opts.items(): setattr(opts, k, v) - if k == 'sd_hypernetwork': shared.reload_hypernetworks() - if k == 'sd_model_checkpoint': sd_models.reload_model_weights() - if k == 'sd_vae': sd_vae.reload_vae_weights() + if k == 'sd_model_checkpoint': + sd_models.reload_model_weights() + + if k == 'sd_vae': + sd_vae.reload_vae_weights() return res @@ -564,10 +561,14 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: cache[0] = (required_prompts, steps) return cache[1] + p.all_prompts, extra_network_data = extra_networks.parse_prompts(p.all_prompts) + with torch.no_grad(), p.sd_model.ema_scope(): with devices.autocast(): p.init(p.all_prompts, p.all_seeds, p.all_subseeds) + extra_networks.activate(p, extra_network_data) + with open(os.path.join(shared.script_path, "params.txt"), "w", encoding="utf8") as file: processed = Processed(p, [], p.seed, "") file.write(processed.infotext(p, 0)) @@ -681,6 +682,7 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if opts.grid_save: images.save_image(grid, p.outpath_grids, "grid", p.all_seeds[0], p.all_prompts[0], opts.grid_format, info=infotext(), short_filename=not opts.grid_extended_filename, p=p, grid=True) + extra_networks.deactivate(p, extra_network_data) devices.torch_gc() res = Processed(p, output_images, p.all_seeds[0], infotext(), comments="".join(["\n\n" + x for x in comments]), subseed=p.all_subseeds[0], index_of_first_image=index_of_first_image, infotexts=infotexts) diff --git a/modules/sd_hijack_optimizations.py b/modules/sd_hijack_optimizations.py index cdc63ed7..4fa54329 100644 --- a/modules/sd_hijack_optimizations.py +++ b/modules/sd_hijack_optimizations.py @@ -44,7 +44,7 @@ def split_cross_attention_forward_v1(self, x, context=None, mask=None): q_in = self.to_q(x) context = default(context, x) - context_k, context_v = hypernetwork.apply_hypernetwork(shared.loaded_hypernetwork, context) + context_k, context_v = hypernetwork.apply_hypernetworks(shared.loaded_hypernetworks, context) k_in = self.to_k(context_k) v_in = self.to_v(context_v) del context, context_k, context_v, x @@ -78,7 +78,7 @@ def split_cross_attention_forward(self, x, context=None, mask=None): q_in = self.to_q(x) context = default(context, x) - context_k, context_v = hypernetwork.apply_hypernetwork(shared.loaded_hypernetwork, context) + context_k, context_v = hypernetwork.apply_hypernetworks(shared.loaded_hypernetworks, context) k_in = self.to_k(context_k) v_in = self.to_v(context_v) @@ -203,7 +203,7 @@ def split_cross_attention_forward_invokeAI(self, x, context=None, mask=None): q = self.to_q(x) context = default(context, x) - context_k, context_v = hypernetwork.apply_hypernetwork(shared.loaded_hypernetwork, context) + context_k, context_v = hypernetwork.apply_hypernetworks(shared.loaded_hypernetworks, context) k = self.to_k(context_k) * self.scale v = self.to_v(context_v) del context, context_k, context_v, x @@ -225,7 +225,7 @@ def sub_quad_attention_forward(self, x, context=None, mask=None): q = self.to_q(x) context = default(context, x) - context_k, context_v = hypernetwork.apply_hypernetwork(shared.loaded_hypernetwork, context) + context_k, context_v = hypernetwork.apply_hypernetworks(shared.loaded_hypernetworks, context) k = self.to_k(context_k) v = self.to_v(context_v) del context, context_k, context_v, x @@ -284,7 +284,7 @@ def xformers_attention_forward(self, x, context=None, mask=None): q_in = self.to_q(x) context = default(context, x) - context_k, context_v = hypernetwork.apply_hypernetwork(shared.loaded_hypernetwork, context) + context_k, context_v = hypernetwork.apply_hypernetworks(shared.loaded_hypernetworks, context) k_in = self.to_k(context_k) v_in = self.to_v(context_v) diff --git a/modules/shared.py b/modules/shared.py index 2f366454..c0e11f18 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -23,6 +23,7 @@ demo = None sd_default_config = os.path.join(script_path, "configs/v1-inference.yaml") sd_model_file = os.path.join(script_path, 'model.ckpt') default_sd_model_file = sd_model_file + parser = argparse.ArgumentParser() parser.add_argument("--config", type=str, default=sd_default_config, help="path to config which constructs model",) parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; if specified, this checkpoint will be added to the list of checkpoints and loaded",) @@ -145,7 +146,7 @@ config_filename = cmd_opts.ui_settings_file os.makedirs(cmd_opts.hypernetwork_dir, exist_ok=True) hypernetworks = {} -loaded_hypernetwork = None +loaded_hypernetworks = [] def reload_hypernetworks(): @@ -153,8 +154,6 @@ def reload_hypernetworks(): global hypernetworks hypernetworks = hypernetwork.list_hypernetworks(cmd_opts.hypernetwork_dir) - hypernetwork.load_hypernetwork(opts.sd_hypernetwork) - class State: @@ -399,8 +398,6 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "sd_vae_checkpoint_cache": OptionInfo(0, "VAE Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), "sd_vae": OptionInfo("Automatic", "SD VAE", gr.Dropdown, lambda: {"choices": ["Automatic", "None"] + list(sd_vae.vae_dict)}, refresh=sd_vae.refresh_vae_list), "sd_vae_as_default": OptionInfo(True, "Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them"), - "sd_hypernetwork": OptionInfo("None", "Hypernetwork", gr.Dropdown, lambda: {"choices": ["None"] + [x for x in hypernetworks.keys()]}, refresh=reload_hypernetworks), - "sd_hypernetwork_strength": OptionInfo(1.0, "Hypernetwork strength", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.001}), "inpainting_mask_weight": OptionInfo(1.0, "Inpainting conditioning mask strength", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), "initial_noise_multiplier": OptionInfo(1.0, "Noise multiplier for img2img", gr.Slider, {"minimum": 0.5, "maximum": 1.5, "step": 0.01 }), "img2img_color_correction": OptionInfo(False, "Apply color correction to img2img results to match original colors."), @@ -661,3 +658,17 @@ mem_mon.start() def listfiles(dirname): filenames = [os.path.join(dirname, x) for x in sorted(os.listdir(dirname)) if not x.startswith(".")] return [file for file in filenames if os.path.isfile(file)] + + +def html_path(filename): + return os.path.join(script_path, "html", filename) + + +def html(filename): + path = html_path(filename) + + if os.path.exists(path): + with open(path, encoding="utf8") as file: + return file.read() + + return "" diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 5a7be422..4e90f690 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -50,6 +50,7 @@ class Embedding: self.sd_checkpoint = None self.sd_checkpoint_name = None self.optimizer_state_dict = None + self.filename = None def save(self, filename): embedding_data = { @@ -182,6 +183,7 @@ class EmbeddingDatabase: embedding.sd_checkpoint_name = data.get('sd_checkpoint_name', None) embedding.vectors = vec.shape[0] embedding.shape = vec.shape[-1] + embedding.filename = path if self.expected_shape == -1 or self.expected_shape == embedding.shape: self.register_embedding(embedding, shared.sd_model) diff --git a/modules/ui.py b/modules/ui.py index 06c11848..d23b2b8e 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -20,7 +20,7 @@ import numpy as np from PIL import Image, PngImagePlugin from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, sd_vae +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, sd_vae, extra_networks from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML from modules.paths import script_path @@ -90,6 +90,7 @@ refresh_symbol = '\U0001f504' # 🔄 save_style_symbol = '\U0001f4be' # 💾 apply_style_symbol = '\U0001f4cb' # 📋 clear_prompt_symbol = '\U0001F5D1' # 🗑️ +extra_networks_symbol = '\U0001F3B4' # 🎴 def plaintext_to_html(text): @@ -324,6 +325,8 @@ def connect_reuse_seed(seed: gr.Number, reuse_seed: gr.Button, generation_info: def update_token_counter(text, steps): try: + text, _ = extra_networks.parse_prompt(text) + _, prompt_flat_list, _ = prompt_parser.get_multicond_prompt_list([text]) prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules(prompt_flat_list, steps) @@ -354,10 +357,10 @@ def create_toprow(is_img2img): negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)") with gr.Column(scale=1, elem_id="roll_col"): - paste = gr.Button(value=paste_symbol, elem_id="paste") - save_style = gr.Button(value=save_style_symbol, elem_id="style_create") - prompt_style_apply = gr.Button(value=apply_style_symbol, elem_id="style_apply") - clear_prompt_button = gr.Button(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") + paste = ToolButton(value=paste_symbol, elem_id="paste") + clear_prompt_button = ToolButton(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") + extra_networks_button = ToolButton(value=extra_networks_symbol, elem_id=f"{id_part}_extra_networks") + token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") negative_token_counter = gr.HTML(value="", elem_id=f"{id_part}_negative_token_counter") @@ -395,11 +398,14 @@ def create_toprow(is_img2img): outputs=[], ) - with gr.Row(): + with gr.Row(elem_id=f"{id_part}_styles_row"): prompt_styles = gr.Dropdown(label="Styles", elem_id=f"{id_part}_styles", choices=[k for k, v in shared.prompt_styles.styles.items()], value=[], multiselect=True) create_refresh_button(prompt_styles, shared.prompt_styles.reload, lambda: {"choices": [k for k, v in shared.prompt_styles.styles.items()]}, f"refresh_{id_part}_styles") - return prompt, prompt_styles, negative_prompt, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, token_counter, token_button, negative_token_counter, negative_token_button + prompt_style_apply = ToolButton(value=apply_style_symbol, elem_id="style_apply") + save_style = ToolButton(value=save_style_symbol, elem_id="style_create") + + return prompt, prompt_styles, negative_prompt, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, extra_networks_button, token_counter, token_button, negative_token_counter, negative_token_button def setup_progressbar(*args, **kwargs): @@ -616,11 +622,15 @@ def create_ui(): modules.scripts.scripts_txt2img.initialize_scripts(is_img2img=False) with gr.Blocks(analytics_enabled=False) as txt2img_interface: - txt2img_prompt, txt2img_prompt_styles, txt2img_negative_prompt, submit, _, _, txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, token_counter, token_button, negative_token_counter, negative_token_button = create_toprow(is_img2img=False) + txt2img_prompt, txt2img_prompt_styles, txt2img_negative_prompt, submit, _, _, txt2img_prompt_style_apply, txt2img_save_style, txt2img_paste, extra_networks_button, token_counter, token_button, negative_token_counter, negative_token_button = create_toprow(is_img2img=False) dummy_component = gr.Label(visible=False) txt_prompt_img = gr.File(label="", elem_id="txt2img_prompt_image", file_count="single", type="binary", visible=False) + with FormRow(variant='compact', elem_id="txt2img_extra_networks", visible=False) as extra_networks: + from modules import ui_extra_networks + extra_networks_ui = ui_extra_networks.create_ui(extra_networks, extra_networks_button, 'txt2img') + with gr.Row().style(equal_height=False): with gr.Column(variant='compact', elem_id="txt2img_settings"): for category in ordered_ui_categories(): @@ -794,14 +804,20 @@ def create_ui(): token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_prompt, steps], outputs=[token_counter]) negative_token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_negative_prompt, steps], outputs=[negative_token_counter]) + ui_extra_networks.setup_ui(extra_networks_ui, txt2img_gallery) + modules.scripts.scripts_current = modules.scripts.scripts_img2img modules.scripts.scripts_img2img.initialize_scripts(is_img2img=True) with gr.Blocks(analytics_enabled=False) as img2img_interface: - img2img_prompt, img2img_prompt_styles, img2img_negative_prompt, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste, token_counter, token_button, negative_token_counter, negative_token_button = create_toprow(is_img2img=True) + img2img_prompt, img2img_prompt_styles, img2img_negative_prompt, submit, img2img_interrogate, img2img_deepbooru, img2img_prompt_style_apply, img2img_save_style, img2img_paste, extra_networks_button, token_counter, token_button, negative_token_counter, negative_token_button = create_toprow(is_img2img=True) img2img_prompt_img = gr.File(label="", elem_id="img2img_prompt_image", file_count="single", type="binary", visible=False) + with FormRow(variant='compact', elem_id="img2img_extra_networks", visible=False) as extra_networks: + from modules import ui_extra_networks + extra_networks_ui_img2img = ui_extra_networks.create_ui(extra_networks, extra_networks_button, 'img2img') + with FormRow().style(equal_height=False): with gr.Column(variant='compact', elem_id="img2img_settings"): copy_image_buttons = [] @@ -1064,6 +1080,8 @@ def create_ui(): token_button.click(fn=update_token_counter, inputs=[img2img_prompt, steps], outputs=[token_counter]) negative_token_button.click(fn=wrap_queued_call(update_token_counter), inputs=[txt2img_negative_prompt, steps], outputs=[negative_token_counter]) + ui_extra_networks.setup_ui(extra_networks_ui_img2img, img2img_gallery) + img2img_paste_fields = [ (img2img_prompt, "Prompt"), (img2img_negative_prompt, "Negative prompt"), @@ -1666,10 +1684,8 @@ def create_ui(): download_localization = gr.Button(value='Download localization template', elem_id="download_localization") reload_script_bodies = gr.Button(value='Reload custom script bodies (No ui updates, No restart)', variant='secondary', elem_id="settings_reload_script_bodies") - if os.path.exists("html/licenses.html"): - with open("html/licenses.html", encoding="utf8") as file: - with gr.TabItem("Licenses"): - gr.HTML(file.read(), elem_id="licenses") + with gr.TabItem("Licenses"): + gr.HTML(shared.html("licenses.html"), elem_id="licenses") gr.Button(value="Show all pages", elem_id="settings_show_all_pages") @@ -1756,11 +1772,9 @@ def create_ui(): if os.path.exists(os.path.join(script_path, "notification.mp3")): audio_notification = gr.Audio(interactive=False, value=os.path.join(script_path, "notification.mp3"), elem_id="audio_notification", visible=False) - if os.path.exists("html/footer.html"): - with open("html/footer.html", encoding="utf8") as file: - footer = file.read() - footer = footer.format(versions=versions_html()) - gr.HTML(footer, elem_id="footer") + footer = shared.html("footer.html") + footer = footer.format(versions=versions_html()) + gr.HTML(footer, elem_id="footer") text_settings = gr.Textbox(elem_id="settings_json", value=lambda: opts.dumpjson(), visible=False) settings_submit.click( diff --git a/modules/ui_components.py b/modules/ui_components.py index 97acff06..46324425 100644 --- a/modules/ui_components.py +++ b/modules/ui_components.py @@ -11,6 +11,16 @@ class ToolButton(gr.Button, gr.components.FormComponent): return "button" +class ToolButtonTop(gr.Button, gr.components.FormComponent): + """Small button with single emoji as text, with extra margin at top, fits inside gradio forms""" + + def __init__(self, **kwargs): + super().__init__(variant="tool-top", **kwargs) + + def get_block_name(self): + return "button" + + class FormRow(gr.Row, gr.components.FormComponent): """Same as gr.Row but fits inside gradio forms""" diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py new file mode 100644 index 00000000..253e90f7 --- /dev/null +++ b/modules/ui_extra_networks.py @@ -0,0 +1,149 @@ +import os.path + +from modules import shared +import gradio as gr +import json + +from modules.generation_parameters_copypaste import image_from_url_text + +extra_pages = [] + + +def register_page(page): + """registers extra networks page for the UI; recommend doing it in on_app_started() callback for extensions""" + + extra_pages.append(page) + + +class ExtraNetworksPage: + def __init__(self, title): + self.title = title + self.card_page = shared.html("extra-networks-card.html") + self.allow_negative_prompt = False + + def refresh(self): + pass + + def create_html(self, tabname): + items_html = '' + + for item in self.list_items(): + items_html += self.create_html_for_item(item, tabname) + + if items_html == '': + dirs = "".join([f"
  • {x}
    • " for x in self.allowed_directories_for_previews()]) + items_html = shared.html("extra-networks-no-cards.html").format(dirs=dirs) + + res = "
    " + items_html + "
    " + + return res + + def list_items(self): + raise NotImplementedError() + + def allowed_directories_for_previews(self): + return [] + + def create_html_for_item(self, item, tabname): + preview = item.get("preview", None) + + args = { + "preview_html": "style='background-image: url(" + json.dumps(preview) + ")'" if preview else '', + "prompt": json.dumps(item["prompt"]), + "tabname": json.dumps(tabname), + "local_preview": json.dumps(item["local_preview"]), + "name": item["name"], + "allow_negative_prompt": "true" if self.allow_negative_prompt else "false", + } + + return self.card_page.format(**args) + + +def intialize(): + extra_pages.clear() + + +class ExtraNetworksUi: + def __init__(self): + self.pages = None + self.stored_extra_pages = None + + self.button_save_preview = None + self.preview_target_filename = None + + self.tabname = None + + +def create_ui(container, button, tabname): + ui = ExtraNetworksUi() + ui.pages = [] + ui.stored_extra_pages = extra_pages.copy() + ui.tabname = tabname + + with gr.Tabs(elem_id=tabname+"_extra_tabs") as tabs: + button_refresh = gr.Button('Refresh', elem_id=tabname+"_extra_refresh") + button_close = gr.Button('Close', elem_id=tabname+"_extra_close") + + for page in ui.stored_extra_pages: + with gr.Tab(page.title): + page_elem = gr.HTML(page.create_html(ui.tabname)) + ui.pages.append(page_elem) + + ui.button_save_preview = gr.Button('Save preview', elem_id=tabname+"_save_preview", visible=False) + ui.preview_target_filename = gr.Textbox('Preview save filename', elem_id=tabname+"_preview_filename", visible=False) + + button.click(fn=lambda: gr.update(visible=True), inputs=[], outputs=[container]) + button_close.click(fn=lambda: gr.update(visible=False), inputs=[], outputs=[container]) + + def refresh(): + res = [] + + for pg in ui.stored_extra_pages: + pg.refresh() + res.append(pg.create_html(ui.tabname)) + + return res + + button_refresh.click(fn=refresh, inputs=[], outputs=ui.pages) + + return ui + + +def path_is_parent(parent_path, child_path): + parent_path = os.path.abspath(parent_path) + child_path = os.path.abspath(child_path) + + return os.path.commonpath([parent_path]) == os.path.commonpath([parent_path, child_path]) + + +def setup_ui(ui, gallery): + def save_preview(index, images, filename): + if len(images) == 0: + print("There is no image in gallery to save as a preview.") + return [page.create_html(ui.tabname) for page in ui.stored_extra_pages] + + index = int(index) + index = 0 if index < 0 else index + index = len(images) - 1 if index >= len(images) else index + + img_info = images[index if index >= 0 else 0] + image = image_from_url_text(img_info) + + is_allowed = False + for extra_page in ui.stored_extra_pages: + if any([path_is_parent(x, filename) for x in extra_page.allowed_directories_for_previews()]): + is_allowed = True + break + + assert is_allowed, f'writing to {filename} is not allowed' + + image.save(filename) + + return [page.create_html(ui.tabname) for page in ui.stored_extra_pages] + + ui.button_save_preview.click( + fn=save_preview, + _js="function(x, y, z){console.log(x, y, z); return [selected_gallery_index(), y, z]}", + inputs=[ui.preview_target_filename, gallery, ui.preview_target_filename], + outputs=[*ui.pages] + ) diff --git a/modules/ui_extra_networks_hypernets.py b/modules/ui_extra_networks_hypernets.py new file mode 100644 index 00000000..312dbaf0 --- /dev/null +++ b/modules/ui_extra_networks_hypernets.py @@ -0,0 +1,34 @@ +import os + +from modules import shared, ui_extra_networks + + +class ExtraNetworksPageHypernetworks(ui_extra_networks.ExtraNetworksPage): + def __init__(self): + super().__init__('Hypernetworks') + + def refresh(self): + shared.reload_hypernetworks() + + def list_items(self): + for name, path in shared.hypernetworks.items(): + path, ext = os.path.splitext(path) + previews = [path + ".png", path + ".preview.png"] + + preview = None + for file in previews: + if os.path.isfile(file): + preview = "./file=" + file.replace('\\', '/') + "?mtime=" + str(os.path.getmtime(file)) + break + + yield { + "name": name, + "filename": path, + "preview": preview, + "prompt": f"", + "local_preview": path + ".png", + } + + def allowed_directories_for_previews(self): + return [shared.cmd_opts.hypernetwork_dir] + diff --git a/modules/ui_extra_networks_textual_inversion.py b/modules/ui_extra_networks_textual_inversion.py new file mode 100644 index 00000000..e4a6e3bf --- /dev/null +++ b/modules/ui_extra_networks_textual_inversion.py @@ -0,0 +1,32 @@ +import os + +from modules import ui_extra_networks, sd_hijack + + +class ExtraNetworksPageTextualInversion(ui_extra_networks.ExtraNetworksPage): + def __init__(self): + super().__init__('Textual Inversion') + self.allow_negative_prompt = True + + def refresh(self): + sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings(force_reload=True) + + def list_items(self): + for embedding in sd_hijack.model_hijack.embedding_db.word_embeddings.values(): + path, ext = os.path.splitext(embedding.filename) + preview_file = path + ".preview.png" + + preview = None + if os.path.isfile(preview_file): + preview = "./file=" + preview_file.replace('\\', '/') + "?mtime=" + str(os.path.getmtime(preview_file)) + + yield { + "name": embedding.name, + "filename": embedding.filename, + "preview": preview, + "prompt": embedding.name, + "local_preview": path + ".preview.png", + } + + def allowed_directories_for_previews(self): + return list(sd_hijack.model_hijack.embedding_db.embedding_dirs) -- cgit v1.2.3 From 6d805b669e86233432f56ee1892d062103abe501 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 09:14:27 +0300 Subject: make CLIP interrogator download original text files if the directory does not exist remove random artist built-in extension (to re-added as a normal extension on demand) remove artists.csv (but what does it mean????????????????????) make interrogate buttons show Loading... when you click them --- modules/api/api.py | 8 -------- modules/artists.py | 25 ----------------------- modules/interrogate.py | 55 +++++++++++++++++++++++++++++++++++++------------- modules/shared.py | 5 ----- modules/ui.py | 11 +++++----- 5 files changed, 46 insertions(+), 58 deletions(-) delete mode 100644 modules/artists.py (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 2c371e6e..f2e9e884 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -126,8 +126,6 @@ class Api: self.add_api_route("/sdapi/v1/face-restorers", self.get_face_restorers, methods=["GET"], response_model=List[FaceRestorerItem]) self.add_api_route("/sdapi/v1/realesrgan-models", self.get_realesrgan_models, methods=["GET"], response_model=List[RealesrganItem]) self.add_api_route("/sdapi/v1/prompt-styles", self.get_prompt_styles, methods=["GET"], response_model=List[PromptStyleItem]) - self.add_api_route("/sdapi/v1/artist-categories", self.get_artists_categories, methods=["GET"], response_model=List[str]) - self.add_api_route("/sdapi/v1/artists", self.get_artists, methods=["GET"], response_model=List[ArtistItem]) self.add_api_route("/sdapi/v1/embeddings", self.get_embeddings, methods=["GET"], response_model=EmbeddingsResponse) self.add_api_route("/sdapi/v1/refresh-checkpoints", self.refresh_checkpoints, methods=["POST"]) self.add_api_route("/sdapi/v1/create/embedding", self.create_embedding, methods=["POST"], response_model=CreateResponse) @@ -390,12 +388,6 @@ class Api: return styleList - def get_artists_categories(self): - return shared.artist_db.cats - - def get_artists(self): - return [{"name":x[0], "score":x[1], "category":x[2]} for x in shared.artist_db.artists] - def get_embeddings(self): db = sd_hijack.model_hijack.embedding_db diff --git a/modules/artists.py b/modules/artists.py deleted file mode 100644 index 3612758b..00000000 --- a/modules/artists.py +++ /dev/null @@ -1,25 +0,0 @@ -import os.path -import csv -from collections import namedtuple - -Artist = namedtuple("Artist", ['name', 'weight', 'category']) - - -class ArtistsDatabase: - def __init__(self, filename): - self.cats = set() - self.artists = [] - - if not os.path.exists(filename): - return - - with open(filename, "r", newline='', encoding="utf8") as file: - reader = csv.DictReader(file) - - for row in reader: - artist = Artist(row["artist"], float(row["score"]), row["category"]) - self.artists.append(artist) - self.cats.add(artist.category) - - def categories(self): - return sorted(self.cats) diff --git a/modules/interrogate.py b/modules/interrogate.py index 738d8ff7..19938cbb 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -5,12 +5,13 @@ from collections import namedtuple import re import torch +import torch.hub from torchvision import transforms from torchvision.transforms.functional import InterpolationMode import modules.shared as shared -from modules import devices, paths, lowvram, modelloader +from modules import devices, paths, lowvram, modelloader, errors blip_image_eval_size = 384 clip_model_name = 'ViT-L/14' @@ -20,27 +21,59 @@ Category = namedtuple("Category", ["name", "topn", "items"]) re_topn = re.compile(r"\.top(\d+)\.") +def download_default_clip_interrogate_categories(content_dir): + print("Downloading CLIP categories...") + + tmpdir = content_dir + "_tmp" + try: + os.makedirs(tmpdir) + + torch.hub.download_url_to_file("https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/artists.txt", os.path.join(tmpdir, "artists.txt")) + torch.hub.download_url_to_file("https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/flavors.txt", os.path.join(tmpdir, "flavors.top3.txt")) + torch.hub.download_url_to_file("https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/mediums.txt", os.path.join(tmpdir, "mediums.txt")) + torch.hub.download_url_to_file("https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/movements.txt", os.path.join(tmpdir, "movements.txt")) + + os.rename(tmpdir, content_dir) + + except Exception as e: + errors.display(e, "downloading default CLIP interrogate categories") + finally: + if os.path.exists(tmpdir): + os.remove(tmpdir) + + class InterrogateModels: blip_model = None clip_model = None clip_preprocess = None - categories = None dtype = None running_on_cpu = None def __init__(self, content_dir): - self.categories = [] + self.loaded_categories = None + self.content_dir = content_dir self.running_on_cpu = devices.device_interrogate == torch.device("cpu") - if os.path.exists(content_dir): - for filename in os.listdir(content_dir): + def categories(self): + if self.loaded_categories is not None: + return self.loaded_categories + + self.loaded_categories = [] + + if not os.path.exists(self.content_dir): + download_default_clip_interrogate_categories(self.content_dir) + + if os.path.exists(self.content_dir): + for filename in os.listdir(self.content_dir): m = re_topn.search(filename) topn = 1 if m is None else int(m.group(1)) - with open(os.path.join(content_dir, filename), "r", encoding="utf8") as file: + with open(os.path.join(self.content_dir, filename), "r", encoding="utf8") as file: lines = [x.strip() for x in file.readlines()] - self.categories.append(Category(name=filename, topn=topn, items=lines)) + self.loaded_categories.append(Category(name=filename, topn=topn, items=lines)) + + return self.loaded_categories def load_blip_model(self): import models.blip @@ -139,7 +172,6 @@ class InterrogateModels: shared.state.begin() shared.state.job = 'interrogate' try: - if shared.cmd_opts.lowvram or shared.cmd_opts.medvram: lowvram.send_everything_to_cpu() devices.torch_gc() @@ -159,12 +191,7 @@ class InterrogateModels: image_features /= image_features.norm(dim=-1, keepdim=True) - if shared.opts.interrogate_use_builtin_artists: - artist = self.rank(image_features, ["by " + artist.name for artist in shared.artist_db.artists])[0] - - res += ", " + artist[0] - - for name, topn, items in self.categories: + for name, topn, items in self.categories(): matches = self.rank(image_features, items, top_count=topn) for match, score in matches: if shared.opts.interrogate_return_ranks: diff --git a/modules/shared.py b/modules/shared.py index c0e11f18..72fb1934 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -9,7 +9,6 @@ from PIL import Image import gradio as gr import tqdm -import modules.artists import modules.interrogate import modules.memmon import modules.styles @@ -254,8 +253,6 @@ class State: state = State() state.server_start = time.time() -artist_db = modules.artists.ArtistsDatabase(os.path.join(script_path, 'artists.csv')) - styles_filename = cmd_opts.styles_file prompt_styles = modules.styles.StyleDatabase(styles_filename) @@ -408,7 +405,6 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "enable_batch_seeds": OptionInfo(True, "Make K-diffusion samplers produce same images in a batch as when making a single image"), "comma_padding_backtrack": OptionInfo(20, "Increase coherency by padding from the last comma within n tokens when using more than 75 tokens", gr.Slider, {"minimum": 0, "maximum": 74, "step": 1 }), 'CLIP_stop_at_last_layers': OptionInfo(1, "Clip skip", gr.Slider, {"minimum": 1, "maximum": 12, "step": 1}), - "random_artist_categories": OptionInfo([], "Allowed categories for random artists selection when using the Roll button", gr.CheckboxGroup, {"choices": artist_db.categories()}), })) options_templates.update(options_section(('compatibility', "Compatibility"), { @@ -419,7 +415,6 @@ options_templates.update(options_section(('compatibility', "Compatibility"), { options_templates.update(options_section(('interrogate', "Interrogate Options"), { "interrogate_keep_models_in_memory": OptionInfo(False, "Interrogate: keep models in VRAM"), - "interrogate_use_builtin_artists": OptionInfo(True, "Interrogate: use artists from artists.csv"), "interrogate_return_ranks": OptionInfo(False, "Interrogate: include ranks of model tags matches in results (Has no effect on caption-based interrogators)."), "interrogate_clip_num_beams": OptionInfo(1, "Interrogate: num_beams for BLIP", gr.Slider, {"minimum": 1, "maximum": 16, "step": 1}), "interrogate_clip_min_length": OptionInfo(24, "Interrogate: minimum description length (excluding artists, etc..)", gr.Slider, {"minimum": 1, "maximum": 128, "step": 1}), diff --git a/modules/ui.py b/modules/ui.py index d23b2b8e..164e0e93 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -228,17 +228,17 @@ def process_interrogate(interrogation_function, mode, ii_input_dir, ii_output_di left, _ = os.path.splitext(filename) print(interrogation_function(img), file=open(os.path.join(ii_output_dir, left + ".txt"), 'a')) - return [gr_show(True), None] + return [gr.update(), None] def interrogate(image): prompt = shared.interrogator.interrogate(image.convert("RGB")) - return gr_show(True) if prompt is None else prompt + return gr.update() if prompt is None else prompt def interrogate_deepbooru(image): prompt = deepbooru.model.tag(image) - return gr_show(True) if prompt is None else prompt + return gr.update() if prompt is None else prompt def create_seed_inputs(target_interface): @@ -1039,19 +1039,18 @@ def create_ui(): init_img_inpaint, ], outputs=[img2img_prompt, dummy_component], - show_progress=False, ) img2img_prompt.submit(**img2img_args) submit.click(**img2img_args) img2img_interrogate.click( - fn=lambda *args : process_interrogate(interrogate, *args), + fn=lambda *args: process_interrogate(interrogate, *args), **interrogate_args, ) img2img_deepbooru.click( - fn=lambda *args : process_interrogate(interrogate_deepbooru, *args), + fn=lambda *args: process_interrogate(interrogate_deepbooru, *args), **interrogate_args, ) -- cgit v1.2.3 From 184e23eb89c198b42f351a4d5ff862ee64917619 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 09:48:38 +0300 Subject: relocate tool buttons next to generate button prevent extra network tabs from putting images into wrong prompts prevent settings leaking into prompt --- modules/ui.py | 43 +++++++++++++++++++++---------------------- 1 file changed, 21 insertions(+), 22 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 164e0e93..fbc3efa0 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -349,30 +349,13 @@ def create_toprow(is_img2img): with gr.Row(): with gr.Column(scale=80): with gr.Row(): - prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=2, placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)") + prompt = gr.Textbox(label="Prompt", elem_id=f"{id_part}_prompt", show_label=False, lines=3, placeholder="Prompt (press Ctrl+Enter or Alt+Enter to generate)") with gr.Row(): with gr.Column(scale=80): with gr.Row(): negative_prompt = gr.Textbox(label="Negative prompt", elem_id=f"{id_part}_neg_prompt", show_label=False, lines=2, placeholder="Negative prompt (press Ctrl+Enter or Alt+Enter to generate)") - with gr.Column(scale=1, elem_id="roll_col"): - paste = ToolButton(value=paste_symbol, elem_id="paste") - clear_prompt_button = ToolButton(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") - extra_networks_button = ToolButton(value=extra_networks_symbol, elem_id=f"{id_part}_extra_networks") - - token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") - token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") - negative_token_counter = gr.HTML(value="", elem_id=f"{id_part}_negative_token_counter") - negative_token_button = gr.Button(visible=False, elem_id=f"{id_part}_negative_token_button") - - clear_prompt_button.click( - fn=lambda *x: x, - _js="confirm_clear_prompt", - inputs=[prompt, negative_prompt], - outputs=[prompt, negative_prompt], - ) - button_interrogate = None button_deepbooru = None if is_img2img: @@ -380,7 +363,7 @@ def create_toprow(is_img2img): button_interrogate = gr.Button('Interrogate\nCLIP', elem_id="interrogate") button_deepbooru = gr.Button('Interrogate\nDeepBooru', elem_id="deepbooru") - with gr.Column(scale=1): + with gr.Column(scale=1, elem_id=f"{id_part}_actions_column"): with gr.Row(elem_id=f"{id_part}_generate_box"): interrupt = gr.Button('Interrupt', elem_id=f"{id_part}_interrupt") skip = gr.Button('Skip', elem_id=f"{id_part}_skip") @@ -398,13 +381,29 @@ def create_toprow(is_img2img): outputs=[], ) + with gr.Row(elem_id=f"{id_part}_tools"): + paste = ToolButton(value=paste_symbol, elem_id="paste") + clear_prompt_button = ToolButton(value=clear_prompt_symbol, elem_id=f"{id_part}_clear_prompt") + extra_networks_button = ToolButton(value=extra_networks_symbol, elem_id=f"{id_part}_extra_networks") + prompt_style_apply = ToolButton(value=apply_style_symbol, elem_id=f"{id_part}_style_apply") + save_style = ToolButton(value=save_style_symbol, elem_id=f"{id_part}_style_create") + + token_counter = gr.HTML(value="", elem_id=f"{id_part}_token_counter") + token_button = gr.Button(visible=False, elem_id=f"{id_part}_token_button") + negative_token_counter = gr.HTML(value="", elem_id=f"{id_part}_negative_token_counter") + negative_token_button = gr.Button(visible=False, elem_id=f"{id_part}_negative_token_button") + + clear_prompt_button.click( + fn=lambda *x: x, + _js="confirm_clear_prompt", + inputs=[prompt, negative_prompt], + outputs=[prompt, negative_prompt], + ) + with gr.Row(elem_id=f"{id_part}_styles_row"): prompt_styles = gr.Dropdown(label="Styles", elem_id=f"{id_part}_styles", choices=[k for k, v in shared.prompt_styles.styles.items()], value=[], multiselect=True) create_refresh_button(prompt_styles, shared.prompt_styles.reload, lambda: {"choices": [k for k, v in shared.prompt_styles.styles.items()]}, f"refresh_{id_part}_styles") - prompt_style_apply = ToolButton(value=apply_style_symbol, elem_id="style_apply") - save_style = ToolButton(value=save_style_symbol, elem_id="style_create") - return prompt, prompt_styles, negative_prompt, submit, button_interrogate, button_deepbooru, prompt_style_apply, save_style, paste, extra_networks_button, token_counter, token_button, negative_token_counter, negative_token_button -- cgit v1.2.3 From 3262e825cc542ff634e6ba2e3a162eafdc6c1bba Mon Sep 17 00:00:00 2001 From: Takuma Mori Date: Sat, 21 Jan 2023 17:42:04 +0900 Subject: add --xformers-flash-attention option & impl --- modules/sd_hijack_optimizations.py | 26 ++++++++++++++++++++++++-- modules/shared.py | 1 + 2 files changed, 25 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/sd_hijack_optimizations.py b/modules/sd_hijack_optimizations.py index 4fa54329..9967359b 100644 --- a/modules/sd_hijack_optimizations.py +++ b/modules/sd_hijack_optimizations.py @@ -290,7 +290,19 @@ def xformers_attention_forward(self, x, context=None, mask=None): q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> b n h d', h=h), (q_in, k_in, v_in)) del q_in, k_in, v_in - out = xformers.ops.memory_efficient_attention(q, k, v, attn_bias=None) + + if shared.cmd_opts.xformers_flash_attention: + op = xformers.ops.MemoryEfficientAttentionFlashAttentionOp + fw, bw = op + if not fw.supports(xformers.ops.fmha.Inputs(query=q, key=k, value=v, attn_bias=None)): + # print('xformers_attention_forward', q.shape, k.shape, v.shape) + # Flash Attention is not availabe for the input arguments. + # Fallback to default xFormers' backend. + op = None + else: + op = None + + out = xformers.ops.memory_efficient_attention(q, k, v, attn_bias=None, op=op) out = rearrange(out, 'b n h d -> b n (h d)', h=h) return self.to_out(out) @@ -365,7 +377,17 @@ def xformers_attnblock_forward(self, x): q = q.contiguous() k = k.contiguous() v = v.contiguous() - out = xformers.ops.memory_efficient_attention(q, k, v) + if shared.cmd_opts.xformers_flash_attention: + op = xformers.ops.MemoryEfficientAttentionFlashAttentionOp + fw, bw = op + if not fw.supports(xformers.ops.fmha.Inputs(query=q, key=k, value=v)): + # print('xformers_attnblock_forward', q.shape, k.shape, v.shape) + # Flash Attention is not availabe for the input arguments. + # Fallback to default xFormers' backend. + op = None + else: + op = None + out = xformers.ops.memory_efficient_attention(q, k, v, op=op) out = rearrange(out, 'b (h w) c -> b c h w', h=h) out = self.proj_out(out) return x + out diff --git a/modules/shared.py b/modules/shared.py index 72fb1934..23328adf 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -57,6 +57,7 @@ parser.add_argument("--realesrgan-models-path", type=str, help="Path to director parser.add_argument("--clip-models-path", type=str, help="Path to directory with CLIP model file(s).", default=None) parser.add_argument("--xformers", action='store_true', help="enable xformers for cross attention layers") parser.add_argument("--force-enable-xformers", action='store_true', help="enable xformers for cross attention layers regardless of whether the checking code thinks you can run it; do not make bug reports if this fails to work") +parser.add_argument("--xformers-flash-attention", action='store_true', help="enable xformers with Flash Attention to improve reproducibility (supported for SD2.x or variant only)") parser.add_argument("--deepdanbooru", action='store_true', help="does not do anything") parser.add_argument("--opt-split-attention", action='store_true', help="force-enables Doggettx's cross-attention layer optimization. By default, it's on for torch cuda.") parser.add_argument("--opt-sub-quad-attention", action='store_true', help="enable memory efficient sub-quadratic cross-attention layer optimization") -- cgit v1.2.3 From 855b9e3d1c5a1bd8c2d815d38a38bc7c410be5a8 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 16:15:53 +0300 Subject: Lora support! update readme to reflect some recent changes --- modules/extra_networks_hypernet.py | 2 +- modules/script_callbacks.py | 15 +++++++++++++++ modules/ui_extra_networks.py | 2 +- 3 files changed, 17 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/extra_networks_hypernet.py b/modules/extra_networks_hypernet.py index 6a0c4ba8..ff279a1f 100644 --- a/modules/extra_networks_hypernet.py +++ b/modules/extra_networks_hypernet.py @@ -17,5 +17,5 @@ class ExtraNetworkHypernet(extra_networks.ExtraNetwork): hypernetwork.load_hypernetworks(names, multipliers) - def deactivate(p, self): + def deactivate(self, p): pass diff --git a/modules/script_callbacks.py b/modules/script_callbacks.py index a9e19236..4bb45ec7 100644 --- a/modules/script_callbacks.py +++ b/modules/script_callbacks.py @@ -73,6 +73,7 @@ callback_map = dict( callbacks_image_grid=[], callbacks_infotext_pasted=[], callbacks_script_unloaded=[], + callbacks_before_ui=[], ) @@ -189,6 +190,14 @@ def script_unloaded_callback(): report_exception(c, 'script_unloaded') +def before_ui_callback(): + for c in reversed(callback_map['callbacks_before_ui']): + try: + c.callback() + except Exception: + report_exception(c, 'before_ui') + + def add_callback(callbacks, fun): stack = [x for x in inspect.stack() if x.filename != __file__] filename = stack[0].filename if len(stack) > 0 else 'unknown file' @@ -313,3 +322,9 @@ def on_script_unloaded(callback): the script did should be reverted here""" add_callback(callback_map['callbacks_script_unloaded'], callback) + + +def on_before_ui(callback): + """register a function to be called before the UI is created.""" + + add_callback(callback_map['callbacks_before_ui'], callback) diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index 253e90f7..796e879c 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -10,7 +10,7 @@ extra_pages = [] def register_page(page): - """registers extra networks page for the UI; recommend doing it in on_app_started() callback for extensions""" + """registers extra networks page for the UI; recommend doing it in on_before_ui() callback for extensions""" extra_pages.append(page) -- cgit v1.2.3 From 92fb1096dbf6403e109a8eb7bc5d18ce487ae9b5 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 16:41:25 +0300 Subject: make it so that extra networks are not removed from infotext --- modules/processing.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index b5deeacf..241961ac 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -561,7 +561,7 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: cache[0] = (required_prompts, steps) return cache[1] - p.all_prompts, extra_network_data = extra_networks.parse_prompts(p.all_prompts) + _, extra_network_data = extra_networks.parse_prompts(p.all_prompts[0:1]) with torch.no_grad(), p.sd_model.ema_scope(): with devices.autocast(): @@ -593,6 +593,8 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if len(prompts) == 0: break + prompts, _ = extra_networks.parse_prompts(prompts) + if p.scripts is not None: p.scripts.process_batch(p, batch_number=n, prompts=prompts, seeds=seeds, subseeds=subseeds) -- cgit v1.2.3 From 424cefe11878c9c7d2663381441e7efe62532180 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 17:20:24 +0300 Subject: add search box to extra networks --- modules/ui_extra_networks.py | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index 796e879c..e2e060c8 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -18,6 +18,7 @@ def register_page(page): class ExtraNetworksPage: def __init__(self, title): self.title = title + self.name = title.lower() self.card_page = shared.html("extra-networks-card.html") self.allow_negative_prompt = False @@ -34,7 +35,11 @@ class ExtraNetworksPage: dirs = "".join([f"
  • {x}
    • " for x in self.allowed_directories_for_previews()]) items_html = shared.html("extra-networks-no-cards.html").format(dirs=dirs) - res = "
    " + items_html + "
    " + res = f""" +
    +{items_html} +
    +""" return res @@ -81,14 +86,15 @@ def create_ui(container, button, tabname): ui.tabname = tabname with gr.Tabs(elem_id=tabname+"_extra_tabs") as tabs: - button_refresh = gr.Button('Refresh', elem_id=tabname+"_extra_refresh") - button_close = gr.Button('Close', elem_id=tabname+"_extra_close") - for page in ui.stored_extra_pages: with gr.Tab(page.title): page_elem = gr.HTML(page.create_html(ui.tabname)) ui.pages.append(page_elem) + filter = gr.Textbox('', show_label=False, elem_id=tabname+"_extra_search", placeholder="Search...", visible=False) + button_refresh = gr.Button('Refresh', elem_id=tabname+"_extra_refresh") + button_close = gr.Button('Close', elem_id=tabname+"_extra_close") + ui.button_save_preview = gr.Button('Save preview', elem_id=tabname+"_save_preview", visible=False) ui.preview_target_filename = gr.Textbox('Preview save filename', elem_id=tabname+"_preview_filename", visible=False) -- cgit v1.2.3 From 63b824376c49013880ff44c260ea426e2899511e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 18:47:54 +0300 Subject: add --gradio-queue option to enable gradio queue --- modules/shared.py | 2 ++ 1 file changed, 2 insertions(+) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 72fb1934..52bbb807 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -100,6 +100,8 @@ parser.add_argument("--cors-allow-origins-regex", type=str, help="Allowed CORS o parser.add_argument("--tls-keyfile", type=str, help="Partially enables TLS, requires --tls-certfile to fully function", default=None) parser.add_argument("--tls-certfile", type=str, help="Partially enables TLS, requires --tls-keyfile to fully function", default=None) parser.add_argument("--server-name", type=str, help="Sets hostname of server", default=None) +parser.add_argument("--gradio-queue", action='store_true', help="Uses gradio queue; experimental option; breaks restart UI button") + script_loading.preload_extensions(extensions.extensions_dir, parser) script_loading.preload_extensions(extensions.extensions_builtin_dir, parser) -- cgit v1.2.3 From 3deea3413575db0ff71f20f4265f3bdc08e35453 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 19:36:08 +0300 Subject: extract extra network data from prompt earlier --- modules/processing.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 241961ac..6e6371a1 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -532,6 +532,8 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if os.path.exists(cmd_opts.embeddings_dir) and not p.do_not_reload_embeddings: model_hijack.embedding_db.load_textual_inversion_embeddings() + _, extra_network_data = extra_networks.parse_prompts(p.all_prompts[0:1]) + if p.scripts is not None: p.scripts.process(p) @@ -561,8 +563,6 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: cache[0] = (required_prompts, steps) return cache[1] - _, extra_network_data = extra_networks.parse_prompts(p.all_prompts[0:1]) - with torch.no_grad(), p.sd_model.ema_scope(): with devices.autocast(): p.init(p.all_prompts, p.all_seeds, p.all_subseeds) -- cgit v1.2.3 From f53527f7786575fe60da0223bd63ea3f0a06a754 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 20:07:14 +0300 Subject: make it run on gradio < 3.16.2 --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index fbc3efa0..b3105901 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1897,7 +1897,7 @@ def create_ui(): if type(x) == gr.Dropdown: def check_dropdown(val): - if x.multiselect: + if getattr(x, 'multiselect', False): return all([value in x.choices for value in val]) else: return val in x.choices -- cgit v1.2.3 From f726df8a2fd2620ba245de5702e2afe620f79b91 Mon Sep 17 00:00:00 2001 From: James Tolton Date: Sat, 21 Jan 2023 12:59:05 -0500 Subject: Compile and serve js from /statica instead of inline in html --- modules/ui.py | 35 ++++++++++++++++++++++++++++++----- 1 file changed, 30 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index fbc3efa0..d19eaf25 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -10,6 +10,7 @@ import sys import tempfile import time import traceback +from collections import OrderedDict from functools import partial, reduce import warnings @@ -1918,27 +1919,51 @@ def create_ui(): def reload_javascript(): + javascript_files = OrderedDict() with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: - javascript = f'' + contents = jsfile.read() + javascript_files["script.js"] = [contents] + # javascript = f'' scripts_list = modules.scripts.list_scripts("javascript", ".js") for basedir, filename, path in scripts_list: with open(path, "r", encoding="utf8") as jsfile: - javascript += f"\n" + contents = jsfile.read() + javascript_files[filename] = [contents] + # javascript += f"\n" if cmd_opts.theme is not None: - javascript += f"\n\n" + javascript_files["theme.js"] = [f"set_theme('{cmd_opts.theme}');"] + # javascript += f"\n\n" - javascript += f"\n" + # javascript += f"\n" + javascript_files["localization.js"] = [f"{localization.localization_js(shared.opts.localization)}"] + + compiled_name = "webui-compiled.js" + head = f""" + + """ def template_response(*args, **kwargs): res = shared.GradioTemplateResponseOriginal(*args, **kwargs) res.body = res.body.replace( - b'', f'{javascript}'.encode("utf8")) + b'', f'{head}'.encode("utf8")) res.init_headers() return res + for k in javascript_files: + javascript_files[k] = "\n".join(javascript_files[k]) + + # make static_path if not exists + statica_path = os.path.join(script_path, 'statica') + if not os.path.exists(statica_path): + os.mkdir(statica_path) + + javascript_out = "\n\n\n".join([f"// \n\n{v}" for k, v in javascript_files.items()]) + with open(os.path.join(script_path, "statica", compiled_name), "w", encoding="utf8") as jsfile: + jsfile.write(javascript_out) + gradio.routes.templates.TemplateResponse = template_response -- cgit v1.2.3 From 17af0fb95574068a1d5032ae96879dab145e173a Mon Sep 17 00:00:00 2001 From: James Tolton Date: Sat, 21 Jan 2023 13:27:05 -0500 Subject: remove commented out lines --- modules/ui.py | 4 ---- 1 file changed, 4 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index d19eaf25..ef85d43c 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1923,7 +1923,6 @@ def reload_javascript(): with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: contents = jsfile.read() javascript_files["script.js"] = [contents] - # javascript = f'' scripts_list = modules.scripts.list_scripts("javascript", ".js") @@ -1931,13 +1930,10 @@ def reload_javascript(): with open(path, "r", encoding="utf8") as jsfile: contents = jsfile.read() javascript_files[filename] = [contents] - # javascript += f"\n" if cmd_opts.theme is not None: javascript_files["theme.js"] = [f"set_theme('{cmd_opts.theme}');"] - # javascript += f"\n\n" - # javascript += f"\n" javascript_files["localization.js"] = [f"{localization.localization_js(shared.opts.localization)}"] compiled_name = "webui-compiled.js" -- cgit v1.2.3 From 50059ea661b63967b217e687819cf7a9081e4a0c Mon Sep 17 00:00:00 2001 From: James Tolton Date: Sat, 21 Jan 2023 14:07:48 -0500 Subject: server individually listed javascript files vs single compiled file --- modules/ui.py | 52 +++++++++++++++++++++++----------------------------- 1 file changed, 23 insertions(+), 29 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index ef85d43c..b372d29c 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1919,47 +1919,41 @@ def create_ui(): def reload_javascript(): - javascript_files = OrderedDict() - with open(os.path.join(script_path, "script.js"), "r", encoding="utf8") as jsfile: - contents = jsfile.read() - javascript_files["script.js"] = [contents] - scripts_list = modules.scripts.list_scripts("javascript", ".js") - + js_files = [] for basedir, filename, path in scripts_list: - with open(path, "r", encoding="utf8") as jsfile: - contents = jsfile.read() - javascript_files[filename] = [contents] + path = path[len(script_path) + 1:] + js_files.append(path) + inline = [f"{localization.localization_js(shared.opts.localization)};"] if cmd_opts.theme is not None: - javascript_files["theme.js"] = [f"set_theme('{cmd_opts.theme}');"] + inline.append(f"set_theme('{cmd_opts.theme}');", ) - javascript_files["localization.js"] = [f"{localization.localization_js(shared.opts.localization)}"] - - compiled_name = "webui-compiled.js" - head = f""" - - """ + t = int(time.time()) + head = [ + f""" + + """.strip() + ] + inline_code = "\n".join(inline) + head.append(f""" + + """.strip()) + for file in js_files: + head.append(f""" + + """.strip()) def template_response(*args, **kwargs): res = shared.GradioTemplateResponseOriginal(*args, **kwargs) + head_inject = "\n".join(head) res.body = res.body.replace( - b'', f'{head}'.encode("utf8")) + b'', f'{head_inject}'.encode("utf8")) res.init_headers() return res - for k in javascript_files: - javascript_files[k] = "\n".join(javascript_files[k]) - - # make static_path if not exists - statica_path = os.path.join(script_path, 'statica') - if not os.path.exists(statica_path): - os.mkdir(statica_path) - - javascript_out = "\n\n\n".join([f"// \n\n{v}" for k, v in javascript_files.items()]) - with open(os.path.join(script_path, "statica", compiled_name), "w", encoding="utf8") as jsfile: - jsfile.write(javascript_out) - gradio.routes.templates.TemplateResponse = template_response -- cgit v1.2.3 From 035459c9a22bebcf68ac454a1f178fefe8c82054 Mon Sep 17 00:00:00 2001 From: James Tolton Date: Sat, 21 Jan 2023 14:11:13 -0500 Subject: remove dead import --- modules/ui.py | 1 - 1 file changed, 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index b372d29c..5fde7fc5 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -10,7 +10,6 @@ import sys import tempfile import time import traceback -from collections import OrderedDict from functools import partial, reduce import warnings -- cgit v1.2.3 From e4e0918f58d382b5da400e680d743dcf0e66fd7f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 22:57:19 +0300 Subject: remove timestamp for js files, reformat code --- modules/ui.py | 34 ++++++++-------------------------- 1 file changed, 8 insertions(+), 26 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index b5581a06..ef7becc6 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1918,38 +1918,20 @@ def create_ui(): def reload_javascript(): - scripts_list = modules.scripts.list_scripts("javascript", ".js") - js_files = [] - for basedir, filename, path in scripts_list: - path = path[len(script_path) + 1:] - js_files.append(path) + head = f'\n' - inline = [f"{localization.localization_js(shared.opts.localization)};"] + inline = f"{localization.localization_js(shared.opts.localization)};" if cmd_opts.theme is not None: - inline.append(f"set_theme('{cmd_opts.theme}');", ) + inline += f"set_theme('{cmd_opts.theme}');" - t = int(time.time()) - head = [ - f""" - - """.strip() - ] - inline_code = "\n".join(inline) - head.append(f""" - - """.strip()) - for file in js_files: - head.append(f""" - - """.strip()) + head += f'\n' + + for script in modules.scripts.list_scripts("javascript", ".js"): + head += f'\n' def template_response(*args, **kwargs): res = shared.GradioTemplateResponseOriginal(*args, **kwargs) - head_inject = "\n".join(head) - res.body = res.body.replace( - b'', f'{head_inject}'.encode("utf8")) + res.body = res.body.replace(b'', f'{head}'.encode("utf8")) res.init_headers() return res -- cgit v1.2.3 From 4a8fe09652b304034708d967c47901312940e852 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 23:06:18 +0300 Subject: remove the double loading text --- modules/ui.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index ef7becc6..aa39a713 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -75,6 +75,7 @@ css_hide_progressbar = """ .wrap .m-12::before { content:"Loading..." } .wrap .z-20 svg { display:none!important; } .wrap .z-20::before { content:"Loading..." } +.wrap.cover-bg .z-20::before { content:"" } .progress-bar { display:none!important; } .meta-text { display:none!important; } .meta-text-center { display:none!important; } -- cgit v1.2.3 From 78f59a4e014d090bce7df3b218bfbcd7f11e0894 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 21 Jan 2023 23:40:13 +0300 Subject: enable compact view for train tab prevent previews from ruining hypernetwork training --- modules/hypernetworks/hypernetwork.py | 2 ++ modules/processing.py | 8 ++++++-- modules/ui.py | 2 +- 3 files changed, 9 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py index 80a47c79..503534e2 100644 --- a/modules/hypernetworks/hypernetwork.py +++ b/modules/hypernetworks/hypernetwork.py @@ -715,6 +715,8 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi do_not_save_samples=True, ) + p.disable_extra_networks = True + if preview_from_txt2img: p.prompt = preview_prompt p.negative_prompt = preview_negative_prompt diff --git a/modules/processing.py b/modules/processing.py index 6e6371a1..bc541e2f 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -140,6 +140,7 @@ class StableDiffusionProcessing: self.override_settings = {k: v for k, v in (override_settings or {}).items() if k not in shared.restricted_opts} self.override_settings_restore_afterwards = override_settings_restore_afterwards self.is_using_inpainting_conditioning = False + self.disable_extra_networks = False if not seed_enable_extras: self.subseed = -1 @@ -567,7 +568,8 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: with devices.autocast(): p.init(p.all_prompts, p.all_seeds, p.all_subseeds) - extra_networks.activate(p, extra_network_data) + if not p.disable_extra_networks: + extra_networks.activate(p, extra_network_data) with open(os.path.join(shared.script_path, "params.txt"), "w", encoding="utf8") as file: processed = Processed(p, [], p.seed, "") @@ -684,7 +686,9 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if opts.grid_save: images.save_image(grid, p.outpath_grids, "grid", p.all_seeds[0], p.all_prompts[0], opts.grid_format, info=infotext(), short_filename=not opts.grid_extended_filename, p=p, grid=True) - extra_networks.deactivate(p, extra_network_data) + if not p.disable_extra_networks: + extra_networks.deactivate(p, extra_network_data) + devices.torch_gc() res = Processed(p, output_images, p.all_seeds[0], infotext(), comments="".join(["\n\n" + x for x in comments]), subseed=p.all_subseeds[0], index_of_first_image=index_of_first_image, infotexts=infotexts) diff --git a/modules/ui.py b/modules/ui.py index daebbc9f..af6dfb21 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1259,7 +1259,7 @@ def create_ui(): with gr.Row().style(equal_height=False): gr.HTML(value="

    See wiki for detailed explanation.

    ") - with gr.Row().style(equal_height=False): + with gr.Row(variant="compact").style(equal_height=False): with gr.Tabs(elem_id="train_tabs"): with gr.Tab(label="Create embedding"): -- cgit v1.2.3 From fe7a623e6b7e04bab2cfc96e8fd6cf49b48daee1 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 22 Jan 2023 00:02:41 +0300 Subject: add a slider for default value of added extra networks --- modules/shared.py | 5 +++-- modules/ui_extra_networks.py | 2 +- modules/ui_extra_networks_hypernets.py | 3 ++- 3 files changed, 6 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 52bbb807..00a1d64c 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -398,7 +398,7 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "sd_vae": OptionInfo("Automatic", "SD VAE", gr.Dropdown, lambda: {"choices": ["Automatic", "None"] + list(sd_vae.vae_dict)}, refresh=sd_vae.refresh_vae_list), "sd_vae_as_default": OptionInfo(True, "Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them"), "inpainting_mask_weight": OptionInfo(1.0, "Inpainting conditioning mask strength", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), - "initial_noise_multiplier": OptionInfo(1.0, "Noise multiplier for img2img", gr.Slider, {"minimum": 0.5, "maximum": 1.5, "step": 0.01 }), + "initial_noise_multiplier": OptionInfo(1.0, "Noise multiplier for img2img", gr.Slider, {"minimum": 0.5, "maximum": 1.5, "step": 0.01}), "img2img_color_correction": OptionInfo(False, "Apply color correction to img2img results to match original colors."), "img2img_fix_steps": OptionInfo(False, "With img2img, do exactly the amount of steps the slider specifies (normally you'd do less with less denoising)."), "img2img_background_color": OptionInfo("#ffffff", "With img2img, fill image's transparent parts with this color.", ui_components.FormColorPicker, {}), @@ -406,7 +406,8 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "enable_emphasis": OptionInfo(True, "Emphasis: use (text) to make model pay more attention to text and [text] to make it pay less attention"), "enable_batch_seeds": OptionInfo(True, "Make K-diffusion samplers produce same images in a batch as when making a single image"), "comma_padding_backtrack": OptionInfo(20, "Increase coherency by padding from the last comma within n tokens when using more than 75 tokens", gr.Slider, {"minimum": 0, "maximum": 74, "step": 1 }), - 'CLIP_stop_at_last_layers': OptionInfo(1, "Clip skip", gr.Slider, {"minimum": 1, "maximum": 12, "step": 1}), + "CLIP_stop_at_last_layers": OptionInfo(1, "Clip skip", gr.Slider, {"minimum": 1, "maximum": 12, "step": 1}), + "extra_networks_default_multiplier": OptionInfo(1.0, "Multiplier for extra networks", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), })) options_templates.update(options_section(('compatibility', "Compatibility"), { diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index e2e060c8..4c88193f 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -54,7 +54,7 @@ class ExtraNetworksPage: args = { "preview_html": "style='background-image: url(" + json.dumps(preview) + ")'" if preview else '', - "prompt": json.dumps(item["prompt"]), + "prompt": item["prompt"], "tabname": json.dumps(tabname), "local_preview": json.dumps(item["local_preview"]), "name": item["name"], diff --git a/modules/ui_extra_networks_hypernets.py b/modules/ui_extra_networks_hypernets.py index 312dbaf0..65d000cf 100644 --- a/modules/ui_extra_networks_hypernets.py +++ b/modules/ui_extra_networks_hypernets.py @@ -1,3 +1,4 @@ +import json import os from modules import shared, ui_extra_networks @@ -25,7 +26,7 @@ class ExtraNetworksPageHypernetworks(ui_extra_networks.ExtraNetworksPage): "name": name, "filename": path, "preview": preview, - "prompt": f"", + "prompt": json.dumps(f""), "local_preview": path + ".png", } -- cgit v1.2.3 From f2eae6127d16a80d1516d3f6245b648eeca26330 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 22 Jan 2023 00:16:26 +0300 Subject: fix broken textual inversion extras tab --- modules/ui_extra_networks_textual_inversion.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui_extra_networks_textual_inversion.py b/modules/ui_extra_networks_textual_inversion.py index e4a6e3bf..dbd23d2d 100644 --- a/modules/ui_extra_networks_textual_inversion.py +++ b/modules/ui_extra_networks_textual_inversion.py @@ -1,3 +1,4 @@ +import json import os from modules import ui_extra_networks, sd_hijack @@ -24,7 +25,7 @@ class ExtraNetworksPageTextualInversion(ui_extra_networks.ExtraNetworksPage): "name": embedding.name, "filename": embedding.filename, "preview": preview, - "prompt": embedding.name, + "prompt": json.dumps(embedding.name), "local_preview": path + ".preview.png", } -- cgit v1.2.3 From bf457b30fbfedb4b6eb2a198cbaa9f2ba071d31f Mon Sep 17 00:00:00 2001 From: EllangoK Date: Sat, 21 Jan 2023 16:21:33 -0500 Subject: compact checkbox and fix copy image btn overflow also fixes type for #tab_extensions in style.css --- modules/ui.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index af6dfb21..12fc9e6a 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -919,7 +919,7 @@ def create_ui(): seed, reuse_seed, subseed, reuse_subseed, subseed_strength, seed_resize_from_h, seed_resize_from_w, seed_checkbox = create_seed_inputs('img2img') elif category == "checkboxes": - with FormRow(elem_id="img2img_checkboxes"): + with FormRow(elem_id="img2img_checkboxes", variant="compact"): restore_faces = gr.Checkbox(label='Restore faces', value=False, visible=len(shared.face_restorers) > 1, elem_id="img2img_restore_faces") tiling = gr.Checkbox(label='Tiling', value=False, elem_id="img2img_tiling") -- cgit v1.2.3 From 2621566153920eb70bfa439f3d7c126ee8d36ec8 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 22 Jan 2023 08:05:21 +0300 Subject: attention ctrl+up/down enhancements --- modules/shared.py | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index 00a1d64c..d68ac296 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -444,9 +444,11 @@ options_templates.update(options_section(('ui', "User interface"), { "show_progress_in_title": OptionInfo(True, "Show generation progress in window title."), "samplers_in_dropdown": OptionInfo(True, "Use dropdown for sampler selection instead of radio group"), "dimensions_and_batch_together": OptionInfo(True, "Show Width/Height and Batch sliders in same row"), - 'quicksettings': OptionInfo("sd_model_checkpoint", "Quicksettings list"), - 'ui_reorder': OptionInfo(", ".join(ui_reorder_categories), "txt2img/img2img UI item order"), - 'localization': OptionInfo("None", "Localization (requires restart)", gr.Dropdown, lambda: {"choices": ["None"] + list(localization.localizations.keys())}, refresh=lambda: localization.list_localizations(cmd_opts.localizations_dir)), + "keyedit_precision_attention": OptionInfo(0.1, "Ctrl+up/down precision when editing (attention:1.1)", gr.Slider, {"minimum": 0.01, "maximum": 0.2, "step": 0.001}), + "keyedit_precision_extra": OptionInfo(0.05, "Ctrl+up/down precision when editing ", gr.Slider, {"minimum": 0.01, "maximum": 0.2, "step": 0.001}), + "quicksettings": OptionInfo("sd_model_checkpoint", "Quicksettings list"), + "ui_reorder": OptionInfo(", ".join(ui_reorder_categories), "txt2img/img2img UI item order"), + "localization": OptionInfo("None", "Localization (requires restart)", gr.Dropdown, lambda: {"choices": ["None"] + list(localization.localizations.keys())}, refresh=lambda: localization.list_localizations(cmd_opts.localizations_dir)), })) options_templates.update(options_section(('ui', "Live previews"), { -- cgit v1.2.3 From 0792fae078ba362a5119f56d84e3f490a88690ae Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 22 Jan 2023 08:20:48 +0300 Subject: fix missing field for aesthetic embedding extension --- modules/sd_disable_initialization.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_disable_initialization.py b/modules/sd_disable_initialization.py index c72d8efc..e90aa9fe 100644 --- a/modules/sd_disable_initialization.py +++ b/modules/sd_disable_initialization.py @@ -41,7 +41,9 @@ class DisableInitialization: return self.create_model_and_transforms(*args, pretrained=None, **kwargs) def CLIPTextModel_from_pretrained(pretrained_model_name_or_path, *model_args, **kwargs): - return self.CLIPTextModel_from_pretrained(None, *model_args, config=pretrained_model_name_or_path, state_dict={}, **kwargs) + res = self.CLIPTextModel_from_pretrained(None, *model_args, config=pretrained_model_name_or_path, state_dict={}, **kwargs) + res.name_or_path = pretrained_model_name_or_path + return res def transformers_modeling_utils_load_pretrained_model(*args, **kwargs): args = args[0:3] + ('/', ) + args[4:] # resolved_archive_file; must set it to something to prevent what seems to be a bug -- cgit v1.2.3 From 112416d04171e4bee673f0adc9bd3aeba87ec71a Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 22 Jan 2023 10:17:12 +0300 Subject: add option to discard weights in checkpoint merger UI --- modules/extras.py | 9 ++++++++- modules/ui.py | 4 ++++ 2 files changed, 12 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 1218f88f..385430dc 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -1,6 +1,7 @@ from __future__ import annotations import math import os +import re import sys import traceback import shutil @@ -285,7 +286,7 @@ def to_half(tensor, enable): return tensor -def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae): +def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae, discard_weights): shared.state.begin() shared.state.job = 'model-merge' @@ -430,6 +431,12 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ for key in theta_0.keys(): theta_0[key] = to_half(theta_0[key], save_as_half) + if discard_weights: + regex = re.compile(discard_weights) + for key in list(theta_0): + if re.search(regex, key): + theta_0.pop(key, None) + ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path filename = filename_generator() if custom_name == '' else custom_name diff --git a/modules/ui.py b/modules/ui.py index af6dfb21..eb4b7e6b 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1248,6 +1248,9 @@ def create_ui(): bake_in_vae = gr.Dropdown(choices=["None"] + list(sd_vae.vae_dict), value="None", label="Bake in VAE", elem_id="modelmerger_bake_in_vae") create_refresh_button(bake_in_vae, sd_vae.refresh_vae_list, lambda: {"choices": ["None"] + list(sd_vae.vae_dict)}, "modelmerger_refresh_bake_in_vae") + with FormRow(): + discard_weights = gr.Textbox(value="", label="Discard weights with matching name", elem_id="modelmerger_discard_weights") + with gr.Row(): modelmerger_merge = gr.Button(elem_id="modelmerger_merge", value="Merge", variant='primary') @@ -1838,6 +1841,7 @@ def create_ui(): checkpoint_format, config_source, bake_in_vae, + discard_weights, ], outputs=[ primary_model_name, -- cgit v1.2.3 From 35419b274614984e2b511a6ad34f37e41481c809 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 22 Jan 2023 11:00:05 +0300 Subject: add an option to reorder tabs for extra networks --- modules/shared.py | 1 + modules/ui_extra_networks.py | 18 +++++++++++++++++- 2 files changed, 18 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index d68ac296..cd78e50a 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -448,6 +448,7 @@ options_templates.update(options_section(('ui', "User interface"), { "keyedit_precision_extra": OptionInfo(0.05, "Ctrl+up/down precision when editing ", gr.Slider, {"minimum": 0.01, "maximum": 0.2, "step": 0.001}), "quicksettings": OptionInfo("sd_model_checkpoint", "Quicksettings list"), "ui_reorder": OptionInfo(", ".join(ui_reorder_categories), "txt2img/img2img UI item order"), + "ui_extra_networks_tab_reorder": OptionInfo("", "Extra networks tab order"), "localization": OptionInfo("None", "Localization (requires restart)", gr.Dropdown, lambda: {"choices": ["None"] + list(localization.localizations.keys())}, refresh=lambda: localization.list_localizations(cmd_opts.localizations_dir)), })) diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index 4c88193f..285c8ffe 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -79,6 +79,22 @@ class ExtraNetworksUi: self.tabname = None +def pages_in_preferred_order(pages): + tab_order = [x.lower().strip() for x in shared.opts.ui_extra_networks_tab_reorder.split(",")] + + def tab_name_score(name): + name = name.lower() + for i, possible_match in enumerate(tab_order): + if possible_match in name: + return i + + return len(pages) + + tab_scores = {page.name: (tab_name_score(page.name), original_index) for original_index, page in enumerate(pages)} + + return sorted(pages, key=lambda x: tab_scores[x.name]) + + def create_ui(container, button, tabname): ui = ExtraNetworksUi() ui.pages = [] @@ -86,7 +102,7 @@ def create_ui(container, button, tabname): ui.tabname = tabname with gr.Tabs(elem_id=tabname+"_extra_tabs") as tabs: - for page in ui.stored_extra_pages: + for page in pages_in_preferred_order(ui.stored_extra_pages): with gr.Tab(page.title): page_elem = gr.HTML(page.create_html(ui.tabname)) ui.pages.append(page_elem) -- cgit v1.2.3 From c98cb0f8ecc904666f47684e238dd022039ca16f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 22 Jan 2023 11:04:02 +0300 Subject: amend previous commit to work in a proper fashion when saving previews --- modules/ui_extra_networks.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index 285c8ffe..af2b8071 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -98,11 +98,11 @@ def pages_in_preferred_order(pages): def create_ui(container, button, tabname): ui = ExtraNetworksUi() ui.pages = [] - ui.stored_extra_pages = extra_pages.copy() + ui.stored_extra_pages = pages_in_preferred_order(extra_pages.copy()) ui.tabname = tabname with gr.Tabs(elem_id=tabname+"_extra_tabs") as tabs: - for page in pages_in_preferred_order(ui.stored_extra_pages): + for page in ui.stored_extra_pages: with gr.Tab(page.title): page_elem = gr.HTML(page.create_html(ui.tabname)) ui.pages.append(page_elem) -- cgit v1.2.3 From 43ac9ff205910e8207dfd45a842577344d399a92 Mon Sep 17 00:00:00 2001 From: Andrey <16777216c@gmail.com> Date: Sun, 22 Jan 2023 15:26:40 +0300 Subject: Split history extras.py to postprocessing.py --- modules/extras.py | 466 ---------------------------------------------- modules/postprocessing.py | 466 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 466 insertions(+), 466 deletions(-) delete mode 100644 modules/extras.py create mode 100644 modules/postprocessing.py (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py deleted file mode 100644 index 385430dc..00000000 --- a/modules/extras.py +++ /dev/null @@ -1,466 +0,0 @@ -from __future__ import annotations -import math -import os -import re -import sys -import traceback -import shutil - -import numpy as np -from PIL import Image - -import torch -import tqdm - -from typing import Callable, List, OrderedDict, Tuple -from functools import partial -from dataclasses import dataclass - -from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae -from modules.shared import opts -import modules.gfpgan_model -from modules.ui import plaintext_to_html -import modules.codeformer_model -import gradio as gr -import safetensors.torch - -class LruCache(OrderedDict): - @dataclass(frozen=True) - class Key: - image_hash: int - info_hash: int - args_hash: int - - @dataclass - class Value: - image: Image.Image - info: str - - def __init__(self, max_size: int = 5, *args, **kwargs): - super().__init__(*args, **kwargs) - self._max_size = max_size - - def get(self, key: LruCache.Key) -> LruCache.Value: - ret = super().get(key) - if ret is not None: - self.move_to_end(key) # Move to end of eviction list - return ret - - def put(self, key: LruCache.Key, value: LruCache.Value) -> None: - self[key] = value - while len(self) > self._max_size: - self.popitem(last=False) - - -cached_images: LruCache = LruCache(max_size=5) - - -def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): - devices.torch_gc() - - shared.state.begin() - shared.state.job = 'extras' - - imageArr = [] - # Also keep track of original file names - imageNameArr = [] - outputs = [] - - if extras_mode == 1: - #convert file to pillow image - for img in image_folder: - image = Image.open(img) - imageArr.append(image) - imageNameArr.append(os.path.splitext(img.orig_name)[0]) - elif extras_mode == 2: - assert not shared.cmd_opts.hide_ui_dir_config, '--hide-ui-dir-config option must be disabled' - - if input_dir == '': - return outputs, "Please select an input directory.", '' - image_list = shared.listfiles(input_dir) - for img in image_list: - try: - image = Image.open(img) - except Exception: - continue - imageArr.append(image) - imageNameArr.append(img) - else: - imageArr.append(image) - imageNameArr.append(None) - - if extras_mode == 2 and output_dir != '': - outpath = output_dir - else: - outpath = opts.outdir_samples or opts.outdir_extras_samples - - # Extra operation definitions - - def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-gfpgan' - restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) - res = Image.fromarray(restored_img) - - if gfpgan_visibility < 1.0: - res = Image.blend(image, res, gfpgan_visibility) - - info += f"GFPGAN visibility:{round(gfpgan_visibility, 2)}\n" - return (res, info) - - def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-codeformer' - restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) - res = Image.fromarray(restored_img) - - if codeformer_visibility < 1.0: - res = Image.blend(image, res, codeformer_visibility) - - info += f"CodeFormer w: {round(codeformer_weight, 2)}, CodeFormer visibility:{round(codeformer_visibility, 2)}\n" - return (res, info) - - def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): - shared.state.job = 'extras-upscale' - upscaler = shared.sd_upscalers[scaler_index] - res = upscaler.scaler.upscale(image, resize, upscaler.data_path) - if mode == 1 and crop: - cropped = Image.new("RGB", (resize_w, resize_h)) - cropped.paste(res, box=(resize_w // 2 - res.width // 2, resize_h // 2 - res.height // 2)) - res = cropped - return res - - def run_prepare_crop(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - # Actual crop happens in run_upscalers_blend, this just sets upscaling_resize and adds info text - nonlocal upscaling_resize - if resize_mode == 1: - upscaling_resize = max(upscaling_resize_w/image.width, upscaling_resize_h/image.height) - crop_info = " (crop)" if upscaling_crop else "" - info += f"Resize to: {upscaling_resize_w:g}x{upscaling_resize_h:g}{crop_info}\n" - return (image, info) - - @dataclass - class UpscaleParams: - upscaler_idx: int - blend_alpha: float - - def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]: - blended_result: Image.Image = None - image_hash: str = hash(np.array(image.getdata()).tobytes()) - for upscaler in params: - upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode, - upscaling_resize_w, upscaling_resize_h, upscaling_crop) - cache_key = LruCache.Key(image_hash=image_hash, - info_hash=hash(info), - args_hash=hash(upscale_args)) - cached_entry = cached_images.get(cache_key) - if cached_entry is None: - res = upscale(image, *upscale_args) - info += f"Upscale: {round(upscaling_resize, 3)}, visibility: {upscaler.blend_alpha}, model:{shared.sd_upscalers[upscaler.upscaler_idx].name}\n" - cached_images.put(cache_key, LruCache.Value(image=res, info=info)) - else: - res, info = cached_entry.image, cached_entry.info - - if blended_result is None: - blended_result = res - else: - blended_result = Image.blend(blended_result, res, upscaler.blend_alpha) - return (blended_result, info) - - # Build a list of operations to run - facefix_ops: List[Callable] = [] - facefix_ops += [run_gfpgan] if gfpgan_visibility > 0 else [] - facefix_ops += [run_codeformer] if codeformer_visibility > 0 else [] - - upscale_ops: List[Callable] = [] - upscale_ops += [run_prepare_crop] if resize_mode == 1 else [] - - if upscaling_resize != 0: - step_params: List[UpscaleParams] = [] - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_1, blend_alpha=1.0)) - if extras_upscaler_2 != 0 and extras_upscaler_2_visibility > 0: - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_2, blend_alpha=extras_upscaler_2_visibility)) - - upscale_ops.append(partial(run_upscalers_blend, step_params)) - - extras_ops: List[Callable] = (upscale_ops + facefix_ops) if upscale_first else (facefix_ops + upscale_ops) - - for image, image_name in zip(imageArr, imageNameArr): - if image is None: - return outputs, "Please select an input image.", '' - - shared.state.textinfo = f'Processing image {image_name}' - - existing_pnginfo = image.info or {} - - image = image.convert("RGB") - info = "" - # Run each operation on each image - for op in extras_ops: - image, info = op(image, info) - - if opts.use_original_name_batch and image_name is not None: - basename = os.path.splitext(os.path.basename(image_name))[0] - else: - basename = '' - - if opts.enable_pnginfo: # append info before save - image.info = existing_pnginfo - image.info["extras"] = info - - if save_output: - # Add upscaler name as a suffix. - suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" - # Add second upscaler if applicable. - if suffix and extras_upscaler_2 and extras_upscaler_2_visibility: - suffix += f"-{shared.sd_upscalers[extras_upscaler_2].name}" - - images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, - no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) - - if extras_mode != 2 or show_extras_results : - outputs.append(image) - - devices.torch_gc() - - return outputs, plaintext_to_html(info), '' - -def clear_cache(): - cached_images.clear() - - -def run_pnginfo(image): - if image is None: - return '', '', '' - - geninfo, items = images.read_info_from_image(image) - items = {**{'parameters': geninfo}, **items} - - info = '' - for key, text in items.items(): - info += f""" -
    -

    {plaintext_to_html(str(key))}

    -

    {plaintext_to_html(str(text))}

    -
    -""".strip()+"\n" - - if len(info) == 0: - message = "Nothing found in the image." - info = f"

    {message}

    " - - return '', geninfo, info - - -def create_config(ckpt_result, config_source, a, b, c): - def config(x): - res = sd_models.find_checkpoint_config(x) if x else None - return res if res != shared.sd_default_config else None - - if config_source == 0: - cfg = config(a) or config(b) or config(c) - elif config_source == 1: - cfg = config(b) - elif config_source == 2: - cfg = config(c) - else: - cfg = None - - if cfg is None: - return - - filename, _ = os.path.splitext(ckpt_result) - checkpoint_filename = filename + ".yaml" - - print("Copying config:") - print(" from:", cfg) - print(" to:", checkpoint_filename) - shutil.copyfile(cfg, checkpoint_filename) - - -checkpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] - - -def to_half(tensor, enable): - if enable and tensor.dtype == torch.float: - return tensor.half() - - return tensor - - -def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae, discard_weights): - shared.state.begin() - shared.state.job = 'model-merge' - - def fail(message): - shared.state.textinfo = message - shared.state.end() - return [*[gr.update() for _ in range(4)], message] - - def weighted_sum(theta0, theta1, alpha): - return ((1 - alpha) * theta0) + (alpha * theta1) - - def get_difference(theta1, theta2): - return theta1 - theta2 - - def add_difference(theta0, theta1_2_diff, alpha): - return theta0 + (alpha * theta1_2_diff) - - def filename_weighted_sum(): - a = primary_model_info.model_name - b = secondary_model_info.model_name - Ma = round(1 - multiplier, 2) - Mb = round(multiplier, 2) - - return f"{Ma}({a}) + {Mb}({b})" - - def filename_add_difference(): - a = primary_model_info.model_name - b = secondary_model_info.model_name - c = tertiary_model_info.model_name - M = round(multiplier, 2) - - return f"{a} + {M}({b} - {c})" - - def filename_nothing(): - return primary_model_info.model_name - - theta_funcs = { - "Weighted sum": (filename_weighted_sum, None, weighted_sum), - "Add difference": (filename_add_difference, get_difference, add_difference), - "No interpolation": (filename_nothing, None, None), - } - filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] - shared.state.job_count = (1 if theta_func1 else 0) + (1 if theta_func2 else 0) - - if not primary_model_name: - return fail("Failed: Merging requires a primary model.") - - primary_model_info = sd_models.checkpoints_list[primary_model_name] - - if theta_func2 and not secondary_model_name: - return fail("Failed: Merging requires a secondary model.") - - secondary_model_info = sd_models.checkpoints_list[secondary_model_name] if theta_func2 else None - - if theta_func1 and not tertiary_model_name: - return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") - - tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None - - result_is_inpainting_model = False - - if theta_func2: - shared.state.textinfo = f"Loading B" - print(f"Loading {secondary_model_info.filename}...") - theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') - else: - theta_1 = None - - if theta_func1: - shared.state.textinfo = f"Loading C" - print(f"Loading {tertiary_model_info.filename}...") - theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') - - shared.state.textinfo = 'Merging B and C' - shared.state.sampling_steps = len(theta_1.keys()) - for key in tqdm.tqdm(theta_1.keys()): - if key in checkpoint_dict_skip_on_merge: - continue - - if 'model' in key: - if key in theta_2: - t2 = theta_2.get(key, torch.zeros_like(theta_1[key])) - theta_1[key] = theta_func1(theta_1[key], t2) - else: - theta_1[key] = torch.zeros_like(theta_1[key]) - - shared.state.sampling_step += 1 - del theta_2 - - shared.state.nextjob() - - shared.state.textinfo = f"Loading {primary_model_info.filename}..." - print(f"Loading {primary_model_info.filename}...") - theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') - - print("Merging...") - shared.state.textinfo = 'Merging A and B' - shared.state.sampling_steps = len(theta_0.keys()) - for key in tqdm.tqdm(theta_0.keys()): - if theta_1 and 'model' in key and key in theta_1: - - if key in checkpoint_dict_skip_on_merge: - continue - - a = theta_0[key] - b = theta_1[key] - - # this enables merging an inpainting model (A) with another one (B); - # where normal model would have 4 channels, for latenst space, inpainting model would - # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 - if a.shape != b.shape and a.shape[0:1] + a.shape[2:] == b.shape[0:1] + b.shape[2:]: - if a.shape[1] == 4 and b.shape[1] == 9: - raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") - - assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" - - theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) - result_is_inpainting_model = True - else: - theta_0[key] = theta_func2(a, b, multiplier) - - theta_0[key] = to_half(theta_0[key], save_as_half) - - shared.state.sampling_step += 1 - - del theta_1 - - bake_in_vae_filename = sd_vae.vae_dict.get(bake_in_vae, None) - if bake_in_vae_filename is not None: - print(f"Baking in VAE from {bake_in_vae_filename}") - shared.state.textinfo = 'Baking in VAE' - vae_dict = sd_vae.load_vae_dict(bake_in_vae_filename, map_location='cpu') - - for key in vae_dict.keys(): - theta_0_key = 'first_stage_model.' + key - if theta_0_key in theta_0: - theta_0[theta_0_key] = to_half(vae_dict[key], save_as_half) - - del vae_dict - - if save_as_half and not theta_func2: - for key in theta_0.keys(): - theta_0[key] = to_half(theta_0[key], save_as_half) - - if discard_weights: - regex = re.compile(discard_weights) - for key in list(theta_0): - if re.search(regex, key): - theta_0.pop(key, None) - - ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path - - filename = filename_generator() if custom_name == '' else custom_name - filename += ".inpainting" if result_is_inpainting_model else "" - filename += "." + checkpoint_format - - output_modelname = os.path.join(ckpt_dir, filename) - - shared.state.nextjob() - shared.state.textinfo = "Saving" - print(f"Saving to {output_modelname}...") - - _, extension = os.path.splitext(output_modelname) - if extension.lower() == ".safetensors": - safetensors.torch.save_file(theta_0, output_modelname, metadata={"format": "pt"}) - else: - torch.save(theta_0, output_modelname) - - sd_models.list_models() - - create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) - - print(f"Checkpoint saved to {output_modelname}.") - shared.state.textinfo = "Checkpoint saved" - shared.state.end() - - return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] diff --git a/modules/postprocessing.py b/modules/postprocessing.py new file mode 100644 index 00000000..385430dc --- /dev/null +++ b/modules/postprocessing.py @@ -0,0 +1,466 @@ +from __future__ import annotations +import math +import os +import re +import sys +import traceback +import shutil + +import numpy as np +from PIL import Image + +import torch +import tqdm + +from typing import Callable, List, OrderedDict, Tuple +from functools import partial +from dataclasses import dataclass + +from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae +from modules.shared import opts +import modules.gfpgan_model +from modules.ui import plaintext_to_html +import modules.codeformer_model +import gradio as gr +import safetensors.torch + +class LruCache(OrderedDict): + @dataclass(frozen=True) + class Key: + image_hash: int + info_hash: int + args_hash: int + + @dataclass + class Value: + image: Image.Image + info: str + + def __init__(self, max_size: int = 5, *args, **kwargs): + super().__init__(*args, **kwargs) + self._max_size = max_size + + def get(self, key: LruCache.Key) -> LruCache.Value: + ret = super().get(key) + if ret is not None: + self.move_to_end(key) # Move to end of eviction list + return ret + + def put(self, key: LruCache.Key, value: LruCache.Value) -> None: + self[key] = value + while len(self) > self._max_size: + self.popitem(last=False) + + +cached_images: LruCache = LruCache(max_size=5) + + +def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): + devices.torch_gc() + + shared.state.begin() + shared.state.job = 'extras' + + imageArr = [] + # Also keep track of original file names + imageNameArr = [] + outputs = [] + + if extras_mode == 1: + #convert file to pillow image + for img in image_folder: + image = Image.open(img) + imageArr.append(image) + imageNameArr.append(os.path.splitext(img.orig_name)[0]) + elif extras_mode == 2: + assert not shared.cmd_opts.hide_ui_dir_config, '--hide-ui-dir-config option must be disabled' + + if input_dir == '': + return outputs, "Please select an input directory.", '' + image_list = shared.listfiles(input_dir) + for img in image_list: + try: + image = Image.open(img) + except Exception: + continue + imageArr.append(image) + imageNameArr.append(img) + else: + imageArr.append(image) + imageNameArr.append(None) + + if extras_mode == 2 and output_dir != '': + outpath = output_dir + else: + outpath = opts.outdir_samples or opts.outdir_extras_samples + + # Extra operation definitions + + def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + shared.state.job = 'extras-gfpgan' + restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) + res = Image.fromarray(restored_img) + + if gfpgan_visibility < 1.0: + res = Image.blend(image, res, gfpgan_visibility) + + info += f"GFPGAN visibility:{round(gfpgan_visibility, 2)}\n" + return (res, info) + + def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + shared.state.job = 'extras-codeformer' + restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) + res = Image.fromarray(restored_img) + + if codeformer_visibility < 1.0: + res = Image.blend(image, res, codeformer_visibility) + + info += f"CodeFormer w: {round(codeformer_weight, 2)}, CodeFormer visibility:{round(codeformer_visibility, 2)}\n" + return (res, info) + + def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): + shared.state.job = 'extras-upscale' + upscaler = shared.sd_upscalers[scaler_index] + res = upscaler.scaler.upscale(image, resize, upscaler.data_path) + if mode == 1 and crop: + cropped = Image.new("RGB", (resize_w, resize_h)) + cropped.paste(res, box=(resize_w // 2 - res.width // 2, resize_h // 2 - res.height // 2)) + res = cropped + return res + + def run_prepare_crop(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + # Actual crop happens in run_upscalers_blend, this just sets upscaling_resize and adds info text + nonlocal upscaling_resize + if resize_mode == 1: + upscaling_resize = max(upscaling_resize_w/image.width, upscaling_resize_h/image.height) + crop_info = " (crop)" if upscaling_crop else "" + info += f"Resize to: {upscaling_resize_w:g}x{upscaling_resize_h:g}{crop_info}\n" + return (image, info) + + @dataclass + class UpscaleParams: + upscaler_idx: int + blend_alpha: float + + def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]: + blended_result: Image.Image = None + image_hash: str = hash(np.array(image.getdata()).tobytes()) + for upscaler in params: + upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode, + upscaling_resize_w, upscaling_resize_h, upscaling_crop) + cache_key = LruCache.Key(image_hash=image_hash, + info_hash=hash(info), + args_hash=hash(upscale_args)) + cached_entry = cached_images.get(cache_key) + if cached_entry is None: + res = upscale(image, *upscale_args) + info += f"Upscale: {round(upscaling_resize, 3)}, visibility: {upscaler.blend_alpha}, model:{shared.sd_upscalers[upscaler.upscaler_idx].name}\n" + cached_images.put(cache_key, LruCache.Value(image=res, info=info)) + else: + res, info = cached_entry.image, cached_entry.info + + if blended_result is None: + blended_result = res + else: + blended_result = Image.blend(blended_result, res, upscaler.blend_alpha) + return (blended_result, info) + + # Build a list of operations to run + facefix_ops: List[Callable] = [] + facefix_ops += [run_gfpgan] if gfpgan_visibility > 0 else [] + facefix_ops += [run_codeformer] if codeformer_visibility > 0 else [] + + upscale_ops: List[Callable] = [] + upscale_ops += [run_prepare_crop] if resize_mode == 1 else [] + + if upscaling_resize != 0: + step_params: List[UpscaleParams] = [] + step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_1, blend_alpha=1.0)) + if extras_upscaler_2 != 0 and extras_upscaler_2_visibility > 0: + step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_2, blend_alpha=extras_upscaler_2_visibility)) + + upscale_ops.append(partial(run_upscalers_blend, step_params)) + + extras_ops: List[Callable] = (upscale_ops + facefix_ops) if upscale_first else (facefix_ops + upscale_ops) + + for image, image_name in zip(imageArr, imageNameArr): + if image is None: + return outputs, "Please select an input image.", '' + + shared.state.textinfo = f'Processing image {image_name}' + + existing_pnginfo = image.info or {} + + image = image.convert("RGB") + info = "" + # Run each operation on each image + for op in extras_ops: + image, info = op(image, info) + + if opts.use_original_name_batch and image_name is not None: + basename = os.path.splitext(os.path.basename(image_name))[0] + else: + basename = '' + + if opts.enable_pnginfo: # append info before save + image.info = existing_pnginfo + image.info["extras"] = info + + if save_output: + # Add upscaler name as a suffix. + suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" + # Add second upscaler if applicable. + if suffix and extras_upscaler_2 and extras_upscaler_2_visibility: + suffix += f"-{shared.sd_upscalers[extras_upscaler_2].name}" + + images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, + no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) + + if extras_mode != 2 or show_extras_results : + outputs.append(image) + + devices.torch_gc() + + return outputs, plaintext_to_html(info), '' + +def clear_cache(): + cached_images.clear() + + +def run_pnginfo(image): + if image is None: + return '', '', '' + + geninfo, items = images.read_info_from_image(image) + items = {**{'parameters': geninfo}, **items} + + info = '' + for key, text in items.items(): + info += f""" +
    +

    {plaintext_to_html(str(key))}

    +

    {plaintext_to_html(str(text))}

    +
    +""".strip()+"\n" + + if len(info) == 0: + message = "Nothing found in the image." + info = f"

    {message}

    " + + return '', geninfo, info + + +def create_config(ckpt_result, config_source, a, b, c): + def config(x): + res = sd_models.find_checkpoint_config(x) if x else None + return res if res != shared.sd_default_config else None + + if config_source == 0: + cfg = config(a) or config(b) or config(c) + elif config_source == 1: + cfg = config(b) + elif config_source == 2: + cfg = config(c) + else: + cfg = None + + if cfg is None: + return + + filename, _ = os.path.splitext(ckpt_result) + checkpoint_filename = filename + ".yaml" + + print("Copying config:") + print(" from:", cfg) + print(" to:", checkpoint_filename) + shutil.copyfile(cfg, checkpoint_filename) + + +checkpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] + + +def to_half(tensor, enable): + if enable and tensor.dtype == torch.float: + return tensor.half() + + return tensor + + +def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae, discard_weights): + shared.state.begin() + shared.state.job = 'model-merge' + + def fail(message): + shared.state.textinfo = message + shared.state.end() + return [*[gr.update() for _ in range(4)], message] + + def weighted_sum(theta0, theta1, alpha): + return ((1 - alpha) * theta0) + (alpha * theta1) + + def get_difference(theta1, theta2): + return theta1 - theta2 + + def add_difference(theta0, theta1_2_diff, alpha): + return theta0 + (alpha * theta1_2_diff) + + def filename_weighted_sum(): + a = primary_model_info.model_name + b = secondary_model_info.model_name + Ma = round(1 - multiplier, 2) + Mb = round(multiplier, 2) + + return f"{Ma}({a}) + {Mb}({b})" + + def filename_add_difference(): + a = primary_model_info.model_name + b = secondary_model_info.model_name + c = tertiary_model_info.model_name + M = round(multiplier, 2) + + return f"{a} + {M}({b} - {c})" + + def filename_nothing(): + return primary_model_info.model_name + + theta_funcs = { + "Weighted sum": (filename_weighted_sum, None, weighted_sum), + "Add difference": (filename_add_difference, get_difference, add_difference), + "No interpolation": (filename_nothing, None, None), + } + filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] + shared.state.job_count = (1 if theta_func1 else 0) + (1 if theta_func2 else 0) + + if not primary_model_name: + return fail("Failed: Merging requires a primary model.") + + primary_model_info = sd_models.checkpoints_list[primary_model_name] + + if theta_func2 and not secondary_model_name: + return fail("Failed: Merging requires a secondary model.") + + secondary_model_info = sd_models.checkpoints_list[secondary_model_name] if theta_func2 else None + + if theta_func1 and not tertiary_model_name: + return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") + + tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None + + result_is_inpainting_model = False + + if theta_func2: + shared.state.textinfo = f"Loading B" + print(f"Loading {secondary_model_info.filename}...") + theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') + else: + theta_1 = None + + if theta_func1: + shared.state.textinfo = f"Loading C" + print(f"Loading {tertiary_model_info.filename}...") + theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') + + shared.state.textinfo = 'Merging B and C' + shared.state.sampling_steps = len(theta_1.keys()) + for key in tqdm.tqdm(theta_1.keys()): + if key in checkpoint_dict_skip_on_merge: + continue + + if 'model' in key: + if key in theta_2: + t2 = theta_2.get(key, torch.zeros_like(theta_1[key])) + theta_1[key] = theta_func1(theta_1[key], t2) + else: + theta_1[key] = torch.zeros_like(theta_1[key]) + + shared.state.sampling_step += 1 + del theta_2 + + shared.state.nextjob() + + shared.state.textinfo = f"Loading {primary_model_info.filename}..." + print(f"Loading {primary_model_info.filename}...") + theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') + + print("Merging...") + shared.state.textinfo = 'Merging A and B' + shared.state.sampling_steps = len(theta_0.keys()) + for key in tqdm.tqdm(theta_0.keys()): + if theta_1 and 'model' in key and key in theta_1: + + if key in checkpoint_dict_skip_on_merge: + continue + + a = theta_0[key] + b = theta_1[key] + + # this enables merging an inpainting model (A) with another one (B); + # where normal model would have 4 channels, for latenst space, inpainting model would + # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 + if a.shape != b.shape and a.shape[0:1] + a.shape[2:] == b.shape[0:1] + b.shape[2:]: + if a.shape[1] == 4 and b.shape[1] == 9: + raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") + + assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" + + theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) + result_is_inpainting_model = True + else: + theta_0[key] = theta_func2(a, b, multiplier) + + theta_0[key] = to_half(theta_0[key], save_as_half) + + shared.state.sampling_step += 1 + + del theta_1 + + bake_in_vae_filename = sd_vae.vae_dict.get(bake_in_vae, None) + if bake_in_vae_filename is not None: + print(f"Baking in VAE from {bake_in_vae_filename}") + shared.state.textinfo = 'Baking in VAE' + vae_dict = sd_vae.load_vae_dict(bake_in_vae_filename, map_location='cpu') + + for key in vae_dict.keys(): + theta_0_key = 'first_stage_model.' + key + if theta_0_key in theta_0: + theta_0[theta_0_key] = to_half(vae_dict[key], save_as_half) + + del vae_dict + + if save_as_half and not theta_func2: + for key in theta_0.keys(): + theta_0[key] = to_half(theta_0[key], save_as_half) + + if discard_weights: + regex = re.compile(discard_weights) + for key in list(theta_0): + if re.search(regex, key): + theta_0.pop(key, None) + + ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path + + filename = filename_generator() if custom_name == '' else custom_name + filename += ".inpainting" if result_is_inpainting_model else "" + filename += "." + checkpoint_format + + output_modelname = os.path.join(ckpt_dir, filename) + + shared.state.nextjob() + shared.state.textinfo = "Saving" + print(f"Saving to {output_modelname}...") + + _, extension = os.path.splitext(output_modelname) + if extension.lower() == ".safetensors": + safetensors.torch.save_file(theta_0, output_modelname, metadata={"format": "pt"}) + else: + torch.save(theta_0, output_modelname) + + sd_models.list_models() + + create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) + + print(f"Checkpoint saved to {output_modelname}.") + shared.state.textinfo = "Checkpoint saved" + shared.state.end() + + return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] -- cgit v1.2.3 From b238b14ee459486c4734cc2899b83f547813a467 Mon Sep 17 00:00:00 2001 From: Andrey <16777216c@gmail.com> Date: Sun, 22 Jan 2023 15:26:40 +0300 Subject: Split history extras.py to postprocessing.py --- modules/extras.py | 466 ------------------------------------------------------ modules/temp | 466 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 466 insertions(+), 466 deletions(-) delete mode 100644 modules/extras.py create mode 100644 modules/temp (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py deleted file mode 100644 index 385430dc..00000000 --- a/modules/extras.py +++ /dev/null @@ -1,466 +0,0 @@ -from __future__ import annotations -import math -import os -import re -import sys -import traceback -import shutil - -import numpy as np -from PIL import Image - -import torch -import tqdm - -from typing import Callable, List, OrderedDict, Tuple -from functools import partial -from dataclasses import dataclass - -from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae -from modules.shared import opts -import modules.gfpgan_model -from modules.ui import plaintext_to_html -import modules.codeformer_model -import gradio as gr -import safetensors.torch - -class LruCache(OrderedDict): - @dataclass(frozen=True) - class Key: - image_hash: int - info_hash: int - args_hash: int - - @dataclass - class Value: - image: Image.Image - info: str - - def __init__(self, max_size: int = 5, *args, **kwargs): - super().__init__(*args, **kwargs) - self._max_size = max_size - - def get(self, key: LruCache.Key) -> LruCache.Value: - ret = super().get(key) - if ret is not None: - self.move_to_end(key) # Move to end of eviction list - return ret - - def put(self, key: LruCache.Key, value: LruCache.Value) -> None: - self[key] = value - while len(self) > self._max_size: - self.popitem(last=False) - - -cached_images: LruCache = LruCache(max_size=5) - - -def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): - devices.torch_gc() - - shared.state.begin() - shared.state.job = 'extras' - - imageArr = [] - # Also keep track of original file names - imageNameArr = [] - outputs = [] - - if extras_mode == 1: - #convert file to pillow image - for img in image_folder: - image = Image.open(img) - imageArr.append(image) - imageNameArr.append(os.path.splitext(img.orig_name)[0]) - elif extras_mode == 2: - assert not shared.cmd_opts.hide_ui_dir_config, '--hide-ui-dir-config option must be disabled' - - if input_dir == '': - return outputs, "Please select an input directory.", '' - image_list = shared.listfiles(input_dir) - for img in image_list: - try: - image = Image.open(img) - except Exception: - continue - imageArr.append(image) - imageNameArr.append(img) - else: - imageArr.append(image) - imageNameArr.append(None) - - if extras_mode == 2 and output_dir != '': - outpath = output_dir - else: - outpath = opts.outdir_samples or opts.outdir_extras_samples - - # Extra operation definitions - - def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-gfpgan' - restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) - res = Image.fromarray(restored_img) - - if gfpgan_visibility < 1.0: - res = Image.blend(image, res, gfpgan_visibility) - - info += f"GFPGAN visibility:{round(gfpgan_visibility, 2)}\n" - return (res, info) - - def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-codeformer' - restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) - res = Image.fromarray(restored_img) - - if codeformer_visibility < 1.0: - res = Image.blend(image, res, codeformer_visibility) - - info += f"CodeFormer w: {round(codeformer_weight, 2)}, CodeFormer visibility:{round(codeformer_visibility, 2)}\n" - return (res, info) - - def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): - shared.state.job = 'extras-upscale' - upscaler = shared.sd_upscalers[scaler_index] - res = upscaler.scaler.upscale(image, resize, upscaler.data_path) - if mode == 1 and crop: - cropped = Image.new("RGB", (resize_w, resize_h)) - cropped.paste(res, box=(resize_w // 2 - res.width // 2, resize_h // 2 - res.height // 2)) - res = cropped - return res - - def run_prepare_crop(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - # Actual crop happens in run_upscalers_blend, this just sets upscaling_resize and adds info text - nonlocal upscaling_resize - if resize_mode == 1: - upscaling_resize = max(upscaling_resize_w/image.width, upscaling_resize_h/image.height) - crop_info = " (crop)" if upscaling_crop else "" - info += f"Resize to: {upscaling_resize_w:g}x{upscaling_resize_h:g}{crop_info}\n" - return (image, info) - - @dataclass - class UpscaleParams: - upscaler_idx: int - blend_alpha: float - - def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]: - blended_result: Image.Image = None - image_hash: str = hash(np.array(image.getdata()).tobytes()) - for upscaler in params: - upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode, - upscaling_resize_w, upscaling_resize_h, upscaling_crop) - cache_key = LruCache.Key(image_hash=image_hash, - info_hash=hash(info), - args_hash=hash(upscale_args)) - cached_entry = cached_images.get(cache_key) - if cached_entry is None: - res = upscale(image, *upscale_args) - info += f"Upscale: {round(upscaling_resize, 3)}, visibility: {upscaler.blend_alpha}, model:{shared.sd_upscalers[upscaler.upscaler_idx].name}\n" - cached_images.put(cache_key, LruCache.Value(image=res, info=info)) - else: - res, info = cached_entry.image, cached_entry.info - - if blended_result is None: - blended_result = res - else: - blended_result = Image.blend(blended_result, res, upscaler.blend_alpha) - return (blended_result, info) - - # Build a list of operations to run - facefix_ops: List[Callable] = [] - facefix_ops += [run_gfpgan] if gfpgan_visibility > 0 else [] - facefix_ops += [run_codeformer] if codeformer_visibility > 0 else [] - - upscale_ops: List[Callable] = [] - upscale_ops += [run_prepare_crop] if resize_mode == 1 else [] - - if upscaling_resize != 0: - step_params: List[UpscaleParams] = [] - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_1, blend_alpha=1.0)) - if extras_upscaler_2 != 0 and extras_upscaler_2_visibility > 0: - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_2, blend_alpha=extras_upscaler_2_visibility)) - - upscale_ops.append(partial(run_upscalers_blend, step_params)) - - extras_ops: List[Callable] = (upscale_ops + facefix_ops) if upscale_first else (facefix_ops + upscale_ops) - - for image, image_name in zip(imageArr, imageNameArr): - if image is None: - return outputs, "Please select an input image.", '' - - shared.state.textinfo = f'Processing image {image_name}' - - existing_pnginfo = image.info or {} - - image = image.convert("RGB") - info = "" - # Run each operation on each image - for op in extras_ops: - image, info = op(image, info) - - if opts.use_original_name_batch and image_name is not None: - basename = os.path.splitext(os.path.basename(image_name))[0] - else: - basename = '' - - if opts.enable_pnginfo: # append info before save - image.info = existing_pnginfo - image.info["extras"] = info - - if save_output: - # Add upscaler name as a suffix. - suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" - # Add second upscaler if applicable. - if suffix and extras_upscaler_2 and extras_upscaler_2_visibility: - suffix += f"-{shared.sd_upscalers[extras_upscaler_2].name}" - - images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, - no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) - - if extras_mode != 2 or show_extras_results : - outputs.append(image) - - devices.torch_gc() - - return outputs, plaintext_to_html(info), '' - -def clear_cache(): - cached_images.clear() - - -def run_pnginfo(image): - if image is None: - return '', '', '' - - geninfo, items = images.read_info_from_image(image) - items = {**{'parameters': geninfo}, **items} - - info = '' - for key, text in items.items(): - info += f""" -
    -

    {plaintext_to_html(str(key))}

    -

    {plaintext_to_html(str(text))}

    -
    -""".strip()+"\n" - - if len(info) == 0: - message = "Nothing found in the image." - info = f"

    {message}

    " - - return '', geninfo, info - - -def create_config(ckpt_result, config_source, a, b, c): - def config(x): - res = sd_models.find_checkpoint_config(x) if x else None - return res if res != shared.sd_default_config else None - - if config_source == 0: - cfg = config(a) or config(b) or config(c) - elif config_source == 1: - cfg = config(b) - elif config_source == 2: - cfg = config(c) - else: - cfg = None - - if cfg is None: - return - - filename, _ = os.path.splitext(ckpt_result) - checkpoint_filename = filename + ".yaml" - - print("Copying config:") - print(" from:", cfg) - print(" to:", checkpoint_filename) - shutil.copyfile(cfg, checkpoint_filename) - - -checkpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] - - -def to_half(tensor, enable): - if enable and tensor.dtype == torch.float: - return tensor.half() - - return tensor - - -def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae, discard_weights): - shared.state.begin() - shared.state.job = 'model-merge' - - def fail(message): - shared.state.textinfo = message - shared.state.end() - return [*[gr.update() for _ in range(4)], message] - - def weighted_sum(theta0, theta1, alpha): - return ((1 - alpha) * theta0) + (alpha * theta1) - - def get_difference(theta1, theta2): - return theta1 - theta2 - - def add_difference(theta0, theta1_2_diff, alpha): - return theta0 + (alpha * theta1_2_diff) - - def filename_weighted_sum(): - a = primary_model_info.model_name - b = secondary_model_info.model_name - Ma = round(1 - multiplier, 2) - Mb = round(multiplier, 2) - - return f"{Ma}({a}) + {Mb}({b})" - - def filename_add_difference(): - a = primary_model_info.model_name - b = secondary_model_info.model_name - c = tertiary_model_info.model_name - M = round(multiplier, 2) - - return f"{a} + {M}({b} - {c})" - - def filename_nothing(): - return primary_model_info.model_name - - theta_funcs = { - "Weighted sum": (filename_weighted_sum, None, weighted_sum), - "Add difference": (filename_add_difference, get_difference, add_difference), - "No interpolation": (filename_nothing, None, None), - } - filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] - shared.state.job_count = (1 if theta_func1 else 0) + (1 if theta_func2 else 0) - - if not primary_model_name: - return fail("Failed: Merging requires a primary model.") - - primary_model_info = sd_models.checkpoints_list[primary_model_name] - - if theta_func2 and not secondary_model_name: - return fail("Failed: Merging requires a secondary model.") - - secondary_model_info = sd_models.checkpoints_list[secondary_model_name] if theta_func2 else None - - if theta_func1 and not tertiary_model_name: - return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") - - tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None - - result_is_inpainting_model = False - - if theta_func2: - shared.state.textinfo = f"Loading B" - print(f"Loading {secondary_model_info.filename}...") - theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') - else: - theta_1 = None - - if theta_func1: - shared.state.textinfo = f"Loading C" - print(f"Loading {tertiary_model_info.filename}...") - theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') - - shared.state.textinfo = 'Merging B and C' - shared.state.sampling_steps = len(theta_1.keys()) - for key in tqdm.tqdm(theta_1.keys()): - if key in checkpoint_dict_skip_on_merge: - continue - - if 'model' in key: - if key in theta_2: - t2 = theta_2.get(key, torch.zeros_like(theta_1[key])) - theta_1[key] = theta_func1(theta_1[key], t2) - else: - theta_1[key] = torch.zeros_like(theta_1[key]) - - shared.state.sampling_step += 1 - del theta_2 - - shared.state.nextjob() - - shared.state.textinfo = f"Loading {primary_model_info.filename}..." - print(f"Loading {primary_model_info.filename}...") - theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') - - print("Merging...") - shared.state.textinfo = 'Merging A and B' - shared.state.sampling_steps = len(theta_0.keys()) - for key in tqdm.tqdm(theta_0.keys()): - if theta_1 and 'model' in key and key in theta_1: - - if key in checkpoint_dict_skip_on_merge: - continue - - a = theta_0[key] - b = theta_1[key] - - # this enables merging an inpainting model (A) with another one (B); - # where normal model would have 4 channels, for latenst space, inpainting model would - # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 - if a.shape != b.shape and a.shape[0:1] + a.shape[2:] == b.shape[0:1] + b.shape[2:]: - if a.shape[1] == 4 and b.shape[1] == 9: - raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") - - assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" - - theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) - result_is_inpainting_model = True - else: - theta_0[key] = theta_func2(a, b, multiplier) - - theta_0[key] = to_half(theta_0[key], save_as_half) - - shared.state.sampling_step += 1 - - del theta_1 - - bake_in_vae_filename = sd_vae.vae_dict.get(bake_in_vae, None) - if bake_in_vae_filename is not None: - print(f"Baking in VAE from {bake_in_vae_filename}") - shared.state.textinfo = 'Baking in VAE' - vae_dict = sd_vae.load_vae_dict(bake_in_vae_filename, map_location='cpu') - - for key in vae_dict.keys(): - theta_0_key = 'first_stage_model.' + key - if theta_0_key in theta_0: - theta_0[theta_0_key] = to_half(vae_dict[key], save_as_half) - - del vae_dict - - if save_as_half and not theta_func2: - for key in theta_0.keys(): - theta_0[key] = to_half(theta_0[key], save_as_half) - - if discard_weights: - regex = re.compile(discard_weights) - for key in list(theta_0): - if re.search(regex, key): - theta_0.pop(key, None) - - ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path - - filename = filename_generator() if custom_name == '' else custom_name - filename += ".inpainting" if result_is_inpainting_model else "" - filename += "." + checkpoint_format - - output_modelname = os.path.join(ckpt_dir, filename) - - shared.state.nextjob() - shared.state.textinfo = "Saving" - print(f"Saving to {output_modelname}...") - - _, extension = os.path.splitext(output_modelname) - if extension.lower() == ".safetensors": - safetensors.torch.save_file(theta_0, output_modelname, metadata={"format": "pt"}) - else: - torch.save(theta_0, output_modelname) - - sd_models.list_models() - - create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) - - print(f"Checkpoint saved to {output_modelname}.") - shared.state.textinfo = "Checkpoint saved" - shared.state.end() - - return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] diff --git a/modules/temp b/modules/temp new file mode 100644 index 00000000..385430dc --- /dev/null +++ b/modules/temp @@ -0,0 +1,466 @@ +from __future__ import annotations +import math +import os +import re +import sys +import traceback +import shutil + +import numpy as np +from PIL import Image + +import torch +import tqdm + +from typing import Callable, List, OrderedDict, Tuple +from functools import partial +from dataclasses import dataclass + +from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae +from modules.shared import opts +import modules.gfpgan_model +from modules.ui import plaintext_to_html +import modules.codeformer_model +import gradio as gr +import safetensors.torch + +class LruCache(OrderedDict): + @dataclass(frozen=True) + class Key: + image_hash: int + info_hash: int + args_hash: int + + @dataclass + class Value: + image: Image.Image + info: str + + def __init__(self, max_size: int = 5, *args, **kwargs): + super().__init__(*args, **kwargs) + self._max_size = max_size + + def get(self, key: LruCache.Key) -> LruCache.Value: + ret = super().get(key) + if ret is not None: + self.move_to_end(key) # Move to end of eviction list + return ret + + def put(self, key: LruCache.Key, value: LruCache.Value) -> None: + self[key] = value + while len(self) > self._max_size: + self.popitem(last=False) + + +cached_images: LruCache = LruCache(max_size=5) + + +def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): + devices.torch_gc() + + shared.state.begin() + shared.state.job = 'extras' + + imageArr = [] + # Also keep track of original file names + imageNameArr = [] + outputs = [] + + if extras_mode == 1: + #convert file to pillow image + for img in image_folder: + image = Image.open(img) + imageArr.append(image) + imageNameArr.append(os.path.splitext(img.orig_name)[0]) + elif extras_mode == 2: + assert not shared.cmd_opts.hide_ui_dir_config, '--hide-ui-dir-config option must be disabled' + + if input_dir == '': + return outputs, "Please select an input directory.", '' + image_list = shared.listfiles(input_dir) + for img in image_list: + try: + image = Image.open(img) + except Exception: + continue + imageArr.append(image) + imageNameArr.append(img) + else: + imageArr.append(image) + imageNameArr.append(None) + + if extras_mode == 2 and output_dir != '': + outpath = output_dir + else: + outpath = opts.outdir_samples or opts.outdir_extras_samples + + # Extra operation definitions + + def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + shared.state.job = 'extras-gfpgan' + restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) + res = Image.fromarray(restored_img) + + if gfpgan_visibility < 1.0: + res = Image.blend(image, res, gfpgan_visibility) + + info += f"GFPGAN visibility:{round(gfpgan_visibility, 2)}\n" + return (res, info) + + def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + shared.state.job = 'extras-codeformer' + restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) + res = Image.fromarray(restored_img) + + if codeformer_visibility < 1.0: + res = Image.blend(image, res, codeformer_visibility) + + info += f"CodeFormer w: {round(codeformer_weight, 2)}, CodeFormer visibility:{round(codeformer_visibility, 2)}\n" + return (res, info) + + def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): + shared.state.job = 'extras-upscale' + upscaler = shared.sd_upscalers[scaler_index] + res = upscaler.scaler.upscale(image, resize, upscaler.data_path) + if mode == 1 and crop: + cropped = Image.new("RGB", (resize_w, resize_h)) + cropped.paste(res, box=(resize_w // 2 - res.width // 2, resize_h // 2 - res.height // 2)) + res = cropped + return res + + def run_prepare_crop(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + # Actual crop happens in run_upscalers_blend, this just sets upscaling_resize and adds info text + nonlocal upscaling_resize + if resize_mode == 1: + upscaling_resize = max(upscaling_resize_w/image.width, upscaling_resize_h/image.height) + crop_info = " (crop)" if upscaling_crop else "" + info += f"Resize to: {upscaling_resize_w:g}x{upscaling_resize_h:g}{crop_info}\n" + return (image, info) + + @dataclass + class UpscaleParams: + upscaler_idx: int + blend_alpha: float + + def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]: + blended_result: Image.Image = None + image_hash: str = hash(np.array(image.getdata()).tobytes()) + for upscaler in params: + upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode, + upscaling_resize_w, upscaling_resize_h, upscaling_crop) + cache_key = LruCache.Key(image_hash=image_hash, + info_hash=hash(info), + args_hash=hash(upscale_args)) + cached_entry = cached_images.get(cache_key) + if cached_entry is None: + res = upscale(image, *upscale_args) + info += f"Upscale: {round(upscaling_resize, 3)}, visibility: {upscaler.blend_alpha}, model:{shared.sd_upscalers[upscaler.upscaler_idx].name}\n" + cached_images.put(cache_key, LruCache.Value(image=res, info=info)) + else: + res, info = cached_entry.image, cached_entry.info + + if blended_result is None: + blended_result = res + else: + blended_result = Image.blend(blended_result, res, upscaler.blend_alpha) + return (blended_result, info) + + # Build a list of operations to run + facefix_ops: List[Callable] = [] + facefix_ops += [run_gfpgan] if gfpgan_visibility > 0 else [] + facefix_ops += [run_codeformer] if codeformer_visibility > 0 else [] + + upscale_ops: List[Callable] = [] + upscale_ops += [run_prepare_crop] if resize_mode == 1 else [] + + if upscaling_resize != 0: + step_params: List[UpscaleParams] = [] + step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_1, blend_alpha=1.0)) + if extras_upscaler_2 != 0 and extras_upscaler_2_visibility > 0: + step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_2, blend_alpha=extras_upscaler_2_visibility)) + + upscale_ops.append(partial(run_upscalers_blend, step_params)) + + extras_ops: List[Callable] = (upscale_ops + facefix_ops) if upscale_first else (facefix_ops + upscale_ops) + + for image, image_name in zip(imageArr, imageNameArr): + if image is None: + return outputs, "Please select an input image.", '' + + shared.state.textinfo = f'Processing image {image_name}' + + existing_pnginfo = image.info or {} + + image = image.convert("RGB") + info = "" + # Run each operation on each image + for op in extras_ops: + image, info = op(image, info) + + if opts.use_original_name_batch and image_name is not None: + basename = os.path.splitext(os.path.basename(image_name))[0] + else: + basename = '' + + if opts.enable_pnginfo: # append info before save + image.info = existing_pnginfo + image.info["extras"] = info + + if save_output: + # Add upscaler name as a suffix. + suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" + # Add second upscaler if applicable. + if suffix and extras_upscaler_2 and extras_upscaler_2_visibility: + suffix += f"-{shared.sd_upscalers[extras_upscaler_2].name}" + + images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, + no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) + + if extras_mode != 2 or show_extras_results : + outputs.append(image) + + devices.torch_gc() + + return outputs, plaintext_to_html(info), '' + +def clear_cache(): + cached_images.clear() + + +def run_pnginfo(image): + if image is None: + return '', '', '' + + geninfo, items = images.read_info_from_image(image) + items = {**{'parameters': geninfo}, **items} + + info = '' + for key, text in items.items(): + info += f""" +
    +

    {plaintext_to_html(str(key))}

    +

    {plaintext_to_html(str(text))}

    +
    +""".strip()+"\n" + + if len(info) == 0: + message = "Nothing found in the image." + info = f"

    {message}

    " + + return '', geninfo, info + + +def create_config(ckpt_result, config_source, a, b, c): + def config(x): + res = sd_models.find_checkpoint_config(x) if x else None + return res if res != shared.sd_default_config else None + + if config_source == 0: + cfg = config(a) or config(b) or config(c) + elif config_source == 1: + cfg = config(b) + elif config_source == 2: + cfg = config(c) + else: + cfg = None + + if cfg is None: + return + + filename, _ = os.path.splitext(ckpt_result) + checkpoint_filename = filename + ".yaml" + + print("Copying config:") + print(" from:", cfg) + print(" to:", checkpoint_filename) + shutil.copyfile(cfg, checkpoint_filename) + + +checkpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] + + +def to_half(tensor, enable): + if enable and tensor.dtype == torch.float: + return tensor.half() + + return tensor + + +def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae, discard_weights): + shared.state.begin() + shared.state.job = 'model-merge' + + def fail(message): + shared.state.textinfo = message + shared.state.end() + return [*[gr.update() for _ in range(4)], message] + + def weighted_sum(theta0, theta1, alpha): + return ((1 - alpha) * theta0) + (alpha * theta1) + + def get_difference(theta1, theta2): + return theta1 - theta2 + + def add_difference(theta0, theta1_2_diff, alpha): + return theta0 + (alpha * theta1_2_diff) + + def filename_weighted_sum(): + a = primary_model_info.model_name + b = secondary_model_info.model_name + Ma = round(1 - multiplier, 2) + Mb = round(multiplier, 2) + + return f"{Ma}({a}) + {Mb}({b})" + + def filename_add_difference(): + a = primary_model_info.model_name + b = secondary_model_info.model_name + c = tertiary_model_info.model_name + M = round(multiplier, 2) + + return f"{a} + {M}({b} - {c})" + + def filename_nothing(): + return primary_model_info.model_name + + theta_funcs = { + "Weighted sum": (filename_weighted_sum, None, weighted_sum), + "Add difference": (filename_add_difference, get_difference, add_difference), + "No interpolation": (filename_nothing, None, None), + } + filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] + shared.state.job_count = (1 if theta_func1 else 0) + (1 if theta_func2 else 0) + + if not primary_model_name: + return fail("Failed: Merging requires a primary model.") + + primary_model_info = sd_models.checkpoints_list[primary_model_name] + + if theta_func2 and not secondary_model_name: + return fail("Failed: Merging requires a secondary model.") + + secondary_model_info = sd_models.checkpoints_list[secondary_model_name] if theta_func2 else None + + if theta_func1 and not tertiary_model_name: + return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") + + tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None + + result_is_inpainting_model = False + + if theta_func2: + shared.state.textinfo = f"Loading B" + print(f"Loading {secondary_model_info.filename}...") + theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') + else: + theta_1 = None + + if theta_func1: + shared.state.textinfo = f"Loading C" + print(f"Loading {tertiary_model_info.filename}...") + theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') + + shared.state.textinfo = 'Merging B and C' + shared.state.sampling_steps = len(theta_1.keys()) + for key in tqdm.tqdm(theta_1.keys()): + if key in checkpoint_dict_skip_on_merge: + continue + + if 'model' in key: + if key in theta_2: + t2 = theta_2.get(key, torch.zeros_like(theta_1[key])) + theta_1[key] = theta_func1(theta_1[key], t2) + else: + theta_1[key] = torch.zeros_like(theta_1[key]) + + shared.state.sampling_step += 1 + del theta_2 + + shared.state.nextjob() + + shared.state.textinfo = f"Loading {primary_model_info.filename}..." + print(f"Loading {primary_model_info.filename}...") + theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') + + print("Merging...") + shared.state.textinfo = 'Merging A and B' + shared.state.sampling_steps = len(theta_0.keys()) + for key in tqdm.tqdm(theta_0.keys()): + if theta_1 and 'model' in key and key in theta_1: + + if key in checkpoint_dict_skip_on_merge: + continue + + a = theta_0[key] + b = theta_1[key] + + # this enables merging an inpainting model (A) with another one (B); + # where normal model would have 4 channels, for latenst space, inpainting model would + # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 + if a.shape != b.shape and a.shape[0:1] + a.shape[2:] == b.shape[0:1] + b.shape[2:]: + if a.shape[1] == 4 and b.shape[1] == 9: + raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") + + assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" + + theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) + result_is_inpainting_model = True + else: + theta_0[key] = theta_func2(a, b, multiplier) + + theta_0[key] = to_half(theta_0[key], save_as_half) + + shared.state.sampling_step += 1 + + del theta_1 + + bake_in_vae_filename = sd_vae.vae_dict.get(bake_in_vae, None) + if bake_in_vae_filename is not None: + print(f"Baking in VAE from {bake_in_vae_filename}") + shared.state.textinfo = 'Baking in VAE' + vae_dict = sd_vae.load_vae_dict(bake_in_vae_filename, map_location='cpu') + + for key in vae_dict.keys(): + theta_0_key = 'first_stage_model.' + key + if theta_0_key in theta_0: + theta_0[theta_0_key] = to_half(vae_dict[key], save_as_half) + + del vae_dict + + if save_as_half and not theta_func2: + for key in theta_0.keys(): + theta_0[key] = to_half(theta_0[key], save_as_half) + + if discard_weights: + regex = re.compile(discard_weights) + for key in list(theta_0): + if re.search(regex, key): + theta_0.pop(key, None) + + ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path + + filename = filename_generator() if custom_name == '' else custom_name + filename += ".inpainting" if result_is_inpainting_model else "" + filename += "." + checkpoint_format + + output_modelname = os.path.join(ckpt_dir, filename) + + shared.state.nextjob() + shared.state.textinfo = "Saving" + print(f"Saving to {output_modelname}...") + + _, extension = os.path.splitext(output_modelname) + if extension.lower() == ".safetensors": + safetensors.torch.save_file(theta_0, output_modelname, metadata={"format": "pt"}) + else: + torch.save(theta_0, output_modelname) + + sd_models.list_models() + + create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) + + print(f"Checkpoint saved to {output_modelname}.") + shared.state.textinfo = "Checkpoint saved" + shared.state.end() + + return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] -- cgit v1.2.3 From c56b36712289020a98f0c77794b9045a251ecd55 Mon Sep 17 00:00:00 2001 From: Andrey <16777216c@gmail.com> Date: Sun, 22 Jan 2023 15:26:41 +0300 Subject: Split history extras.py to postprocessing.py --- modules/extras.py | 466 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ modules/temp | 466 ------------------------------------------------------ 2 files changed, 466 insertions(+), 466 deletions(-) create mode 100644 modules/extras.py delete mode 100644 modules/temp (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py new file mode 100644 index 00000000..385430dc --- /dev/null +++ b/modules/extras.py @@ -0,0 +1,466 @@ +from __future__ import annotations +import math +import os +import re +import sys +import traceback +import shutil + +import numpy as np +from PIL import Image + +import torch +import tqdm + +from typing import Callable, List, OrderedDict, Tuple +from functools import partial +from dataclasses import dataclass + +from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae +from modules.shared import opts +import modules.gfpgan_model +from modules.ui import plaintext_to_html +import modules.codeformer_model +import gradio as gr +import safetensors.torch + +class LruCache(OrderedDict): + @dataclass(frozen=True) + class Key: + image_hash: int + info_hash: int + args_hash: int + + @dataclass + class Value: + image: Image.Image + info: str + + def __init__(self, max_size: int = 5, *args, **kwargs): + super().__init__(*args, **kwargs) + self._max_size = max_size + + def get(self, key: LruCache.Key) -> LruCache.Value: + ret = super().get(key) + if ret is not None: + self.move_to_end(key) # Move to end of eviction list + return ret + + def put(self, key: LruCache.Key, value: LruCache.Value) -> None: + self[key] = value + while len(self) > self._max_size: + self.popitem(last=False) + + +cached_images: LruCache = LruCache(max_size=5) + + +def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): + devices.torch_gc() + + shared.state.begin() + shared.state.job = 'extras' + + imageArr = [] + # Also keep track of original file names + imageNameArr = [] + outputs = [] + + if extras_mode == 1: + #convert file to pillow image + for img in image_folder: + image = Image.open(img) + imageArr.append(image) + imageNameArr.append(os.path.splitext(img.orig_name)[0]) + elif extras_mode == 2: + assert not shared.cmd_opts.hide_ui_dir_config, '--hide-ui-dir-config option must be disabled' + + if input_dir == '': + return outputs, "Please select an input directory.", '' + image_list = shared.listfiles(input_dir) + for img in image_list: + try: + image = Image.open(img) + except Exception: + continue + imageArr.append(image) + imageNameArr.append(img) + else: + imageArr.append(image) + imageNameArr.append(None) + + if extras_mode == 2 and output_dir != '': + outpath = output_dir + else: + outpath = opts.outdir_samples or opts.outdir_extras_samples + + # Extra operation definitions + + def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + shared.state.job = 'extras-gfpgan' + restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) + res = Image.fromarray(restored_img) + + if gfpgan_visibility < 1.0: + res = Image.blend(image, res, gfpgan_visibility) + + info += f"GFPGAN visibility:{round(gfpgan_visibility, 2)}\n" + return (res, info) + + def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + shared.state.job = 'extras-codeformer' + restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) + res = Image.fromarray(restored_img) + + if codeformer_visibility < 1.0: + res = Image.blend(image, res, codeformer_visibility) + + info += f"CodeFormer w: {round(codeformer_weight, 2)}, CodeFormer visibility:{round(codeformer_visibility, 2)}\n" + return (res, info) + + def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): + shared.state.job = 'extras-upscale' + upscaler = shared.sd_upscalers[scaler_index] + res = upscaler.scaler.upscale(image, resize, upscaler.data_path) + if mode == 1 and crop: + cropped = Image.new("RGB", (resize_w, resize_h)) + cropped.paste(res, box=(resize_w // 2 - res.width // 2, resize_h // 2 - res.height // 2)) + res = cropped + return res + + def run_prepare_crop(image: Image.Image, info: str) -> Tuple[Image.Image, str]: + # Actual crop happens in run_upscalers_blend, this just sets upscaling_resize and adds info text + nonlocal upscaling_resize + if resize_mode == 1: + upscaling_resize = max(upscaling_resize_w/image.width, upscaling_resize_h/image.height) + crop_info = " (crop)" if upscaling_crop else "" + info += f"Resize to: {upscaling_resize_w:g}x{upscaling_resize_h:g}{crop_info}\n" + return (image, info) + + @dataclass + class UpscaleParams: + upscaler_idx: int + blend_alpha: float + + def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]: + blended_result: Image.Image = None + image_hash: str = hash(np.array(image.getdata()).tobytes()) + for upscaler in params: + upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode, + upscaling_resize_w, upscaling_resize_h, upscaling_crop) + cache_key = LruCache.Key(image_hash=image_hash, + info_hash=hash(info), + args_hash=hash(upscale_args)) + cached_entry = cached_images.get(cache_key) + if cached_entry is None: + res = upscale(image, *upscale_args) + info += f"Upscale: {round(upscaling_resize, 3)}, visibility: {upscaler.blend_alpha}, model:{shared.sd_upscalers[upscaler.upscaler_idx].name}\n" + cached_images.put(cache_key, LruCache.Value(image=res, info=info)) + else: + res, info = cached_entry.image, cached_entry.info + + if blended_result is None: + blended_result = res + else: + blended_result = Image.blend(blended_result, res, upscaler.blend_alpha) + return (blended_result, info) + + # Build a list of operations to run + facefix_ops: List[Callable] = [] + facefix_ops += [run_gfpgan] if gfpgan_visibility > 0 else [] + facefix_ops += [run_codeformer] if codeformer_visibility > 0 else [] + + upscale_ops: List[Callable] = [] + upscale_ops += [run_prepare_crop] if resize_mode == 1 else [] + + if upscaling_resize != 0: + step_params: List[UpscaleParams] = [] + step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_1, blend_alpha=1.0)) + if extras_upscaler_2 != 0 and extras_upscaler_2_visibility > 0: + step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_2, blend_alpha=extras_upscaler_2_visibility)) + + upscale_ops.append(partial(run_upscalers_blend, step_params)) + + extras_ops: List[Callable] = (upscale_ops + facefix_ops) if upscale_first else (facefix_ops + upscale_ops) + + for image, image_name in zip(imageArr, imageNameArr): + if image is None: + return outputs, "Please select an input image.", '' + + shared.state.textinfo = f'Processing image {image_name}' + + existing_pnginfo = image.info or {} + + image = image.convert("RGB") + info = "" + # Run each operation on each image + for op in extras_ops: + image, info = op(image, info) + + if opts.use_original_name_batch and image_name is not None: + basename = os.path.splitext(os.path.basename(image_name))[0] + else: + basename = '' + + if opts.enable_pnginfo: # append info before save + image.info = existing_pnginfo + image.info["extras"] = info + + if save_output: + # Add upscaler name as a suffix. + suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" + # Add second upscaler if applicable. + if suffix and extras_upscaler_2 and extras_upscaler_2_visibility: + suffix += f"-{shared.sd_upscalers[extras_upscaler_2].name}" + + images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, + no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) + + if extras_mode != 2 or show_extras_results : + outputs.append(image) + + devices.torch_gc() + + return outputs, plaintext_to_html(info), '' + +def clear_cache(): + cached_images.clear() + + +def run_pnginfo(image): + if image is None: + return '', '', '' + + geninfo, items = images.read_info_from_image(image) + items = {**{'parameters': geninfo}, **items} + + info = '' + for key, text in items.items(): + info += f""" +
    +

    {plaintext_to_html(str(key))}

    +

    {plaintext_to_html(str(text))}

    +
    +""".strip()+"\n" + + if len(info) == 0: + message = "Nothing found in the image." + info = f"

    {message}

    " + + return '', geninfo, info + + +def create_config(ckpt_result, config_source, a, b, c): + def config(x): + res = sd_models.find_checkpoint_config(x) if x else None + return res if res != shared.sd_default_config else None + + if config_source == 0: + cfg = config(a) or config(b) or config(c) + elif config_source == 1: + cfg = config(b) + elif config_source == 2: + cfg = config(c) + else: + cfg = None + + if cfg is None: + return + + filename, _ = os.path.splitext(ckpt_result) + checkpoint_filename = filename + ".yaml" + + print("Copying config:") + print(" from:", cfg) + print(" to:", checkpoint_filename) + shutil.copyfile(cfg, checkpoint_filename) + + +checkpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] + + +def to_half(tensor, enable): + if enable and tensor.dtype == torch.float: + return tensor.half() + + return tensor + + +def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae, discard_weights): + shared.state.begin() + shared.state.job = 'model-merge' + + def fail(message): + shared.state.textinfo = message + shared.state.end() + return [*[gr.update() for _ in range(4)], message] + + def weighted_sum(theta0, theta1, alpha): + return ((1 - alpha) * theta0) + (alpha * theta1) + + def get_difference(theta1, theta2): + return theta1 - theta2 + + def add_difference(theta0, theta1_2_diff, alpha): + return theta0 + (alpha * theta1_2_diff) + + def filename_weighted_sum(): + a = primary_model_info.model_name + b = secondary_model_info.model_name + Ma = round(1 - multiplier, 2) + Mb = round(multiplier, 2) + + return f"{Ma}({a}) + {Mb}({b})" + + def filename_add_difference(): + a = primary_model_info.model_name + b = secondary_model_info.model_name + c = tertiary_model_info.model_name + M = round(multiplier, 2) + + return f"{a} + {M}({b} - {c})" + + def filename_nothing(): + return primary_model_info.model_name + + theta_funcs = { + "Weighted sum": (filename_weighted_sum, None, weighted_sum), + "Add difference": (filename_add_difference, get_difference, add_difference), + "No interpolation": (filename_nothing, None, None), + } + filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] + shared.state.job_count = (1 if theta_func1 else 0) + (1 if theta_func2 else 0) + + if not primary_model_name: + return fail("Failed: Merging requires a primary model.") + + primary_model_info = sd_models.checkpoints_list[primary_model_name] + + if theta_func2 and not secondary_model_name: + return fail("Failed: Merging requires a secondary model.") + + secondary_model_info = sd_models.checkpoints_list[secondary_model_name] if theta_func2 else None + + if theta_func1 and not tertiary_model_name: + return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") + + tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None + + result_is_inpainting_model = False + + if theta_func2: + shared.state.textinfo = f"Loading B" + print(f"Loading {secondary_model_info.filename}...") + theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') + else: + theta_1 = None + + if theta_func1: + shared.state.textinfo = f"Loading C" + print(f"Loading {tertiary_model_info.filename}...") + theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') + + shared.state.textinfo = 'Merging B and C' + shared.state.sampling_steps = len(theta_1.keys()) + for key in tqdm.tqdm(theta_1.keys()): + if key in checkpoint_dict_skip_on_merge: + continue + + if 'model' in key: + if key in theta_2: + t2 = theta_2.get(key, torch.zeros_like(theta_1[key])) + theta_1[key] = theta_func1(theta_1[key], t2) + else: + theta_1[key] = torch.zeros_like(theta_1[key]) + + shared.state.sampling_step += 1 + del theta_2 + + shared.state.nextjob() + + shared.state.textinfo = f"Loading {primary_model_info.filename}..." + print(f"Loading {primary_model_info.filename}...") + theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') + + print("Merging...") + shared.state.textinfo = 'Merging A and B' + shared.state.sampling_steps = len(theta_0.keys()) + for key in tqdm.tqdm(theta_0.keys()): + if theta_1 and 'model' in key and key in theta_1: + + if key in checkpoint_dict_skip_on_merge: + continue + + a = theta_0[key] + b = theta_1[key] + + # this enables merging an inpainting model (A) with another one (B); + # where normal model would have 4 channels, for latenst space, inpainting model would + # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 + if a.shape != b.shape and a.shape[0:1] + a.shape[2:] == b.shape[0:1] + b.shape[2:]: + if a.shape[1] == 4 and b.shape[1] == 9: + raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") + + assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" + + theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) + result_is_inpainting_model = True + else: + theta_0[key] = theta_func2(a, b, multiplier) + + theta_0[key] = to_half(theta_0[key], save_as_half) + + shared.state.sampling_step += 1 + + del theta_1 + + bake_in_vae_filename = sd_vae.vae_dict.get(bake_in_vae, None) + if bake_in_vae_filename is not None: + print(f"Baking in VAE from {bake_in_vae_filename}") + shared.state.textinfo = 'Baking in VAE' + vae_dict = sd_vae.load_vae_dict(bake_in_vae_filename, map_location='cpu') + + for key in vae_dict.keys(): + theta_0_key = 'first_stage_model.' + key + if theta_0_key in theta_0: + theta_0[theta_0_key] = to_half(vae_dict[key], save_as_half) + + del vae_dict + + if save_as_half and not theta_func2: + for key in theta_0.keys(): + theta_0[key] = to_half(theta_0[key], save_as_half) + + if discard_weights: + regex = re.compile(discard_weights) + for key in list(theta_0): + if re.search(regex, key): + theta_0.pop(key, None) + + ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path + + filename = filename_generator() if custom_name == '' else custom_name + filename += ".inpainting" if result_is_inpainting_model else "" + filename += "." + checkpoint_format + + output_modelname = os.path.join(ckpt_dir, filename) + + shared.state.nextjob() + shared.state.textinfo = "Saving" + print(f"Saving to {output_modelname}...") + + _, extension = os.path.splitext(output_modelname) + if extension.lower() == ".safetensors": + safetensors.torch.save_file(theta_0, output_modelname, metadata={"format": "pt"}) + else: + torch.save(theta_0, output_modelname) + + sd_models.list_models() + + create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) + + print(f"Checkpoint saved to {output_modelname}.") + shared.state.textinfo = "Checkpoint saved" + shared.state.end() + + return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] diff --git a/modules/temp b/modules/temp deleted file mode 100644 index 385430dc..00000000 --- a/modules/temp +++ /dev/null @@ -1,466 +0,0 @@ -from __future__ import annotations -import math -import os -import re -import sys -import traceback -import shutil - -import numpy as np -from PIL import Image - -import torch -import tqdm - -from typing import Callable, List, OrderedDict, Tuple -from functools import partial -from dataclasses import dataclass - -from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae -from modules.shared import opts -import modules.gfpgan_model -from modules.ui import plaintext_to_html -import modules.codeformer_model -import gradio as gr -import safetensors.torch - -class LruCache(OrderedDict): - @dataclass(frozen=True) - class Key: - image_hash: int - info_hash: int - args_hash: int - - @dataclass - class Value: - image: Image.Image - info: str - - def __init__(self, max_size: int = 5, *args, **kwargs): - super().__init__(*args, **kwargs) - self._max_size = max_size - - def get(self, key: LruCache.Key) -> LruCache.Value: - ret = super().get(key) - if ret is not None: - self.move_to_end(key) # Move to end of eviction list - return ret - - def put(self, key: LruCache.Key, value: LruCache.Value) -> None: - self[key] = value - while len(self) > self._max_size: - self.popitem(last=False) - - -cached_images: LruCache = LruCache(max_size=5) - - -def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): - devices.torch_gc() - - shared.state.begin() - shared.state.job = 'extras' - - imageArr = [] - # Also keep track of original file names - imageNameArr = [] - outputs = [] - - if extras_mode == 1: - #convert file to pillow image - for img in image_folder: - image = Image.open(img) - imageArr.append(image) - imageNameArr.append(os.path.splitext(img.orig_name)[0]) - elif extras_mode == 2: - assert not shared.cmd_opts.hide_ui_dir_config, '--hide-ui-dir-config option must be disabled' - - if input_dir == '': - return outputs, "Please select an input directory.", '' - image_list = shared.listfiles(input_dir) - for img in image_list: - try: - image = Image.open(img) - except Exception: - continue - imageArr.append(image) - imageNameArr.append(img) - else: - imageArr.append(image) - imageNameArr.append(None) - - if extras_mode == 2 and output_dir != '': - outpath = output_dir - else: - outpath = opts.outdir_samples or opts.outdir_extras_samples - - # Extra operation definitions - - def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-gfpgan' - restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) - res = Image.fromarray(restored_img) - - if gfpgan_visibility < 1.0: - res = Image.blend(image, res, gfpgan_visibility) - - info += f"GFPGAN visibility:{round(gfpgan_visibility, 2)}\n" - return (res, info) - - def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-codeformer' - restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) - res = Image.fromarray(restored_img) - - if codeformer_visibility < 1.0: - res = Image.blend(image, res, codeformer_visibility) - - info += f"CodeFormer w: {round(codeformer_weight, 2)}, CodeFormer visibility:{round(codeformer_visibility, 2)}\n" - return (res, info) - - def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): - shared.state.job = 'extras-upscale' - upscaler = shared.sd_upscalers[scaler_index] - res = upscaler.scaler.upscale(image, resize, upscaler.data_path) - if mode == 1 and crop: - cropped = Image.new("RGB", (resize_w, resize_h)) - cropped.paste(res, box=(resize_w // 2 - res.width // 2, resize_h // 2 - res.height // 2)) - res = cropped - return res - - def run_prepare_crop(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - # Actual crop happens in run_upscalers_blend, this just sets upscaling_resize and adds info text - nonlocal upscaling_resize - if resize_mode == 1: - upscaling_resize = max(upscaling_resize_w/image.width, upscaling_resize_h/image.height) - crop_info = " (crop)" if upscaling_crop else "" - info += f"Resize to: {upscaling_resize_w:g}x{upscaling_resize_h:g}{crop_info}\n" - return (image, info) - - @dataclass - class UpscaleParams: - upscaler_idx: int - blend_alpha: float - - def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]: - blended_result: Image.Image = None - image_hash: str = hash(np.array(image.getdata()).tobytes()) - for upscaler in params: - upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode, - upscaling_resize_w, upscaling_resize_h, upscaling_crop) - cache_key = LruCache.Key(image_hash=image_hash, - info_hash=hash(info), - args_hash=hash(upscale_args)) - cached_entry = cached_images.get(cache_key) - if cached_entry is None: - res = upscale(image, *upscale_args) - info += f"Upscale: {round(upscaling_resize, 3)}, visibility: {upscaler.blend_alpha}, model:{shared.sd_upscalers[upscaler.upscaler_idx].name}\n" - cached_images.put(cache_key, LruCache.Value(image=res, info=info)) - else: - res, info = cached_entry.image, cached_entry.info - - if blended_result is None: - blended_result = res - else: - blended_result = Image.blend(blended_result, res, upscaler.blend_alpha) - return (blended_result, info) - - # Build a list of operations to run - facefix_ops: List[Callable] = [] - facefix_ops += [run_gfpgan] if gfpgan_visibility > 0 else [] - facefix_ops += [run_codeformer] if codeformer_visibility > 0 else [] - - upscale_ops: List[Callable] = [] - upscale_ops += [run_prepare_crop] if resize_mode == 1 else [] - - if upscaling_resize != 0: - step_params: List[UpscaleParams] = [] - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_1, blend_alpha=1.0)) - if extras_upscaler_2 != 0 and extras_upscaler_2_visibility > 0: - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_2, blend_alpha=extras_upscaler_2_visibility)) - - upscale_ops.append(partial(run_upscalers_blend, step_params)) - - extras_ops: List[Callable] = (upscale_ops + facefix_ops) if upscale_first else (facefix_ops + upscale_ops) - - for image, image_name in zip(imageArr, imageNameArr): - if image is None: - return outputs, "Please select an input image.", '' - - shared.state.textinfo = f'Processing image {image_name}' - - existing_pnginfo = image.info or {} - - image = image.convert("RGB") - info = "" - # Run each operation on each image - for op in extras_ops: - image, info = op(image, info) - - if opts.use_original_name_batch and image_name is not None: - basename = os.path.splitext(os.path.basename(image_name))[0] - else: - basename = '' - - if opts.enable_pnginfo: # append info before save - image.info = existing_pnginfo - image.info["extras"] = info - - if save_output: - # Add upscaler name as a suffix. - suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" - # Add second upscaler if applicable. - if suffix and extras_upscaler_2 and extras_upscaler_2_visibility: - suffix += f"-{shared.sd_upscalers[extras_upscaler_2].name}" - - images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, - no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) - - if extras_mode != 2 or show_extras_results : - outputs.append(image) - - devices.torch_gc() - - return outputs, plaintext_to_html(info), '' - -def clear_cache(): - cached_images.clear() - - -def run_pnginfo(image): - if image is None: - return '', '', '' - - geninfo, items = images.read_info_from_image(image) - items = {**{'parameters': geninfo}, **items} - - info = '' - for key, text in items.items(): - info += f""" -
    -

    {plaintext_to_html(str(key))}

    -

    {plaintext_to_html(str(text))}

    -
    -""".strip()+"\n" - - if len(info) == 0: - message = "Nothing found in the image." - info = f"

    {message}

    " - - return '', geninfo, info - - -def create_config(ckpt_result, config_source, a, b, c): - def config(x): - res = sd_models.find_checkpoint_config(x) if x else None - return res if res != shared.sd_default_config else None - - if config_source == 0: - cfg = config(a) or config(b) or config(c) - elif config_source == 1: - cfg = config(b) - elif config_source == 2: - cfg = config(c) - else: - cfg = None - - if cfg is None: - return - - filename, _ = os.path.splitext(ckpt_result) - checkpoint_filename = filename + ".yaml" - - print("Copying config:") - print(" from:", cfg) - print(" to:", checkpoint_filename) - shutil.copyfile(cfg, checkpoint_filename) - - -checkpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] - - -def to_half(tensor, enable): - if enable and tensor.dtype == torch.float: - return tensor.half() - - return tensor - - -def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae, discard_weights): - shared.state.begin() - shared.state.job = 'model-merge' - - def fail(message): - shared.state.textinfo = message - shared.state.end() - return [*[gr.update() for _ in range(4)], message] - - def weighted_sum(theta0, theta1, alpha): - return ((1 - alpha) * theta0) + (alpha * theta1) - - def get_difference(theta1, theta2): - return theta1 - theta2 - - def add_difference(theta0, theta1_2_diff, alpha): - return theta0 + (alpha * theta1_2_diff) - - def filename_weighted_sum(): - a = primary_model_info.model_name - b = secondary_model_info.model_name - Ma = round(1 - multiplier, 2) - Mb = round(multiplier, 2) - - return f"{Ma}({a}) + {Mb}({b})" - - def filename_add_difference(): - a = primary_model_info.model_name - b = secondary_model_info.model_name - c = tertiary_model_info.model_name - M = round(multiplier, 2) - - return f"{a} + {M}({b} - {c})" - - def filename_nothing(): - return primary_model_info.model_name - - theta_funcs = { - "Weighted sum": (filename_weighted_sum, None, weighted_sum), - "Add difference": (filename_add_difference, get_difference, add_difference), - "No interpolation": (filename_nothing, None, None), - } - filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] - shared.state.job_count = (1 if theta_func1 else 0) + (1 if theta_func2 else 0) - - if not primary_model_name: - return fail("Failed: Merging requires a primary model.") - - primary_model_info = sd_models.checkpoints_list[primary_model_name] - - if theta_func2 and not secondary_model_name: - return fail("Failed: Merging requires a secondary model.") - - secondary_model_info = sd_models.checkpoints_list[secondary_model_name] if theta_func2 else None - - if theta_func1 and not tertiary_model_name: - return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") - - tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None - - result_is_inpainting_model = False - - if theta_func2: - shared.state.textinfo = f"Loading B" - print(f"Loading {secondary_model_info.filename}...") - theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') - else: - theta_1 = None - - if theta_func1: - shared.state.textinfo = f"Loading C" - print(f"Loading {tertiary_model_info.filename}...") - theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') - - shared.state.textinfo = 'Merging B and C' - shared.state.sampling_steps = len(theta_1.keys()) - for key in tqdm.tqdm(theta_1.keys()): - if key in checkpoint_dict_skip_on_merge: - continue - - if 'model' in key: - if key in theta_2: - t2 = theta_2.get(key, torch.zeros_like(theta_1[key])) - theta_1[key] = theta_func1(theta_1[key], t2) - else: - theta_1[key] = torch.zeros_like(theta_1[key]) - - shared.state.sampling_step += 1 - del theta_2 - - shared.state.nextjob() - - shared.state.textinfo = f"Loading {primary_model_info.filename}..." - print(f"Loading {primary_model_info.filename}...") - theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') - - print("Merging...") - shared.state.textinfo = 'Merging A and B' - shared.state.sampling_steps = len(theta_0.keys()) - for key in tqdm.tqdm(theta_0.keys()): - if theta_1 and 'model' in key and key in theta_1: - - if key in checkpoint_dict_skip_on_merge: - continue - - a = theta_0[key] - b = theta_1[key] - - # this enables merging an inpainting model (A) with another one (B); - # where normal model would have 4 channels, for latenst space, inpainting model would - # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 - if a.shape != b.shape and a.shape[0:1] + a.shape[2:] == b.shape[0:1] + b.shape[2:]: - if a.shape[1] == 4 and b.shape[1] == 9: - raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") - - assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" - - theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) - result_is_inpainting_model = True - else: - theta_0[key] = theta_func2(a, b, multiplier) - - theta_0[key] = to_half(theta_0[key], save_as_half) - - shared.state.sampling_step += 1 - - del theta_1 - - bake_in_vae_filename = sd_vae.vae_dict.get(bake_in_vae, None) - if bake_in_vae_filename is not None: - print(f"Baking in VAE from {bake_in_vae_filename}") - shared.state.textinfo = 'Baking in VAE' - vae_dict = sd_vae.load_vae_dict(bake_in_vae_filename, map_location='cpu') - - for key in vae_dict.keys(): - theta_0_key = 'first_stage_model.' + key - if theta_0_key in theta_0: - theta_0[theta_0_key] = to_half(vae_dict[key], save_as_half) - - del vae_dict - - if save_as_half and not theta_func2: - for key in theta_0.keys(): - theta_0[key] = to_half(theta_0[key], save_as_half) - - if discard_weights: - regex = re.compile(discard_weights) - for key in list(theta_0): - if re.search(regex, key): - theta_0.pop(key, None) - - ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path - - filename = filename_generator() if custom_name == '' else custom_name - filename += ".inpainting" if result_is_inpainting_model else "" - filename += "." + checkpoint_format - - output_modelname = os.path.join(ckpt_dir, filename) - - shared.state.nextjob() - shared.state.textinfo = "Saving" - print(f"Saving to {output_modelname}...") - - _, extension = os.path.splitext(output_modelname) - if extension.lower() == ".safetensors": - safetensors.torch.save_file(theta_0, output_modelname, metadata={"format": "pt"}) - else: - torch.save(theta_0, output_modelname) - - sd_models.list_models() - - create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) - - print(f"Checkpoint saved to {output_modelname}.") - shared.state.textinfo = "Checkpoint saved" - shared.state.end() - - return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] -- cgit v1.2.3 From 68303c96e5ab31576a8238a24bf5b6191cf16ed1 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sun, 22 Jan 2023 15:38:39 +0300 Subject: split oversize extras.py to postprocessing.py --- modules/extras.py | 217 +------------------------------------- modules/postprocessing.py | 257 +--------------------------------------------- modules/ui.py | 10 +- modules/ui_components.py | 7 ++ 4 files changed, 18 insertions(+), 473 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 385430dc..f04ddfc2 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -1,231 +1,16 @@ -from __future__ import annotations -import math import os import re -import sys -import traceback import shutil -import numpy as np -from PIL import Image import torch import tqdm -from typing import Callable, List, OrderedDict, Tuple -from functools import partial -from dataclasses import dataclass - -from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae -from modules.shared import opts -import modules.gfpgan_model +from modules import shared, images, sd_models, sd_vae from modules.ui import plaintext_to_html -import modules.codeformer_model import gradio as gr import safetensors.torch -class LruCache(OrderedDict): - @dataclass(frozen=True) - class Key: - image_hash: int - info_hash: int - args_hash: int - - @dataclass - class Value: - image: Image.Image - info: str - - def __init__(self, max_size: int = 5, *args, **kwargs): - super().__init__(*args, **kwargs) - self._max_size = max_size - - def get(self, key: LruCache.Key) -> LruCache.Value: - ret = super().get(key) - if ret is not None: - self.move_to_end(key) # Move to end of eviction list - return ret - - def put(self, key: LruCache.Key, value: LruCache.Value) -> None: - self[key] = value - while len(self) > self._max_size: - self.popitem(last=False) - - -cached_images: LruCache = LruCache(max_size=5) - - -def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): - devices.torch_gc() - - shared.state.begin() - shared.state.job = 'extras' - - imageArr = [] - # Also keep track of original file names - imageNameArr = [] - outputs = [] - - if extras_mode == 1: - #convert file to pillow image - for img in image_folder: - image = Image.open(img) - imageArr.append(image) - imageNameArr.append(os.path.splitext(img.orig_name)[0]) - elif extras_mode == 2: - assert not shared.cmd_opts.hide_ui_dir_config, '--hide-ui-dir-config option must be disabled' - - if input_dir == '': - return outputs, "Please select an input directory.", '' - image_list = shared.listfiles(input_dir) - for img in image_list: - try: - image = Image.open(img) - except Exception: - continue - imageArr.append(image) - imageNameArr.append(img) - else: - imageArr.append(image) - imageNameArr.append(None) - - if extras_mode == 2 and output_dir != '': - outpath = output_dir - else: - outpath = opts.outdir_samples or opts.outdir_extras_samples - - # Extra operation definitions - - def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-gfpgan' - restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) - res = Image.fromarray(restored_img) - - if gfpgan_visibility < 1.0: - res = Image.blend(image, res, gfpgan_visibility) - - info += f"GFPGAN visibility:{round(gfpgan_visibility, 2)}\n" - return (res, info) - - def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-codeformer' - restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) - res = Image.fromarray(restored_img) - - if codeformer_visibility < 1.0: - res = Image.blend(image, res, codeformer_visibility) - - info += f"CodeFormer w: {round(codeformer_weight, 2)}, CodeFormer visibility:{round(codeformer_visibility, 2)}\n" - return (res, info) - - def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): - shared.state.job = 'extras-upscale' - upscaler = shared.sd_upscalers[scaler_index] - res = upscaler.scaler.upscale(image, resize, upscaler.data_path) - if mode == 1 and crop: - cropped = Image.new("RGB", (resize_w, resize_h)) - cropped.paste(res, box=(resize_w // 2 - res.width // 2, resize_h // 2 - res.height // 2)) - res = cropped - return res - - def run_prepare_crop(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - # Actual crop happens in run_upscalers_blend, this just sets upscaling_resize and adds info text - nonlocal upscaling_resize - if resize_mode == 1: - upscaling_resize = max(upscaling_resize_w/image.width, upscaling_resize_h/image.height) - crop_info = " (crop)" if upscaling_crop else "" - info += f"Resize to: {upscaling_resize_w:g}x{upscaling_resize_h:g}{crop_info}\n" - return (image, info) - - @dataclass - class UpscaleParams: - upscaler_idx: int - blend_alpha: float - - def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]: - blended_result: Image.Image = None - image_hash: str = hash(np.array(image.getdata()).tobytes()) - for upscaler in params: - upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode, - upscaling_resize_w, upscaling_resize_h, upscaling_crop) - cache_key = LruCache.Key(image_hash=image_hash, - info_hash=hash(info), - args_hash=hash(upscale_args)) - cached_entry = cached_images.get(cache_key) - if cached_entry is None: - res = upscale(image, *upscale_args) - info += f"Upscale: {round(upscaling_resize, 3)}, visibility: {upscaler.blend_alpha}, model:{shared.sd_upscalers[upscaler.upscaler_idx].name}\n" - cached_images.put(cache_key, LruCache.Value(image=res, info=info)) - else: - res, info = cached_entry.image, cached_entry.info - - if blended_result is None: - blended_result = res - else: - blended_result = Image.blend(blended_result, res, upscaler.blend_alpha) - return (blended_result, info) - - # Build a list of operations to run - facefix_ops: List[Callable] = [] - facefix_ops += [run_gfpgan] if gfpgan_visibility > 0 else [] - facefix_ops += [run_codeformer] if codeformer_visibility > 0 else [] - - upscale_ops: List[Callable] = [] - upscale_ops += [run_prepare_crop] if resize_mode == 1 else [] - - if upscaling_resize != 0: - step_params: List[UpscaleParams] = [] - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_1, blend_alpha=1.0)) - if extras_upscaler_2 != 0 and extras_upscaler_2_visibility > 0: - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_2, blend_alpha=extras_upscaler_2_visibility)) - - upscale_ops.append(partial(run_upscalers_blend, step_params)) - - extras_ops: List[Callable] = (upscale_ops + facefix_ops) if upscale_first else (facefix_ops + upscale_ops) - - for image, image_name in zip(imageArr, imageNameArr): - if image is None: - return outputs, "Please select an input image.", '' - - shared.state.textinfo = f'Processing image {image_name}' - - existing_pnginfo = image.info or {} - - image = image.convert("RGB") - info = "" - # Run each operation on each image - for op in extras_ops: - image, info = op(image, info) - - if opts.use_original_name_batch and image_name is not None: - basename = os.path.splitext(os.path.basename(image_name))[0] - else: - basename = '' - - if opts.enable_pnginfo: # append info before save - image.info = existing_pnginfo - image.info["extras"] = info - - if save_output: - # Add upscaler name as a suffix. - suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" - # Add second upscaler if applicable. - if suffix and extras_upscaler_2 and extras_upscaler_2_visibility: - suffix += f"-{shared.sd_upscalers[extras_upscaler_2].name}" - - images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, - no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) - - if extras_mode != 2 or show_extras_results : - outputs.append(image) - - devices.torch_gc() - - return outputs, plaintext_to_html(info), '' - -def clear_cache(): - cached_images.clear() - def run_pnginfo(image): if image is None: diff --git a/modules/postprocessing.py b/modules/postprocessing.py index 385430dc..cb85720b 100644 --- a/modules/postprocessing.py +++ b/modules/postprocessing.py @@ -1,28 +1,18 @@ from __future__ import annotations -import math import os -import re -import sys -import traceback -import shutil import numpy as np from PIL import Image -import torch -import tqdm - from typing import Callable, List, OrderedDict, Tuple from functools import partial from dataclasses import dataclass -from modules import processing, shared, images, devices, sd_models, sd_samplers, sd_vae +from modules import shared, images, devices, ui_components from modules.shared import opts import modules.gfpgan_model -from modules.ui import plaintext_to_html import modules.codeformer_model -import gradio as gr -import safetensors.torch + class LruCache(OrderedDict): @dataclass(frozen=True) @@ -55,7 +45,7 @@ class LruCache(OrderedDict): cached_images: LruCache = LruCache(max_size=5) -def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): +def run_postprocessing(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): devices.torch_gc() shared.state.begin() @@ -221,246 +211,9 @@ def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_ devices.torch_gc() - return outputs, plaintext_to_html(info), '' + return outputs, ui_components.plaintext_to_html(info), '' + def clear_cache(): cached_images.clear() - -def run_pnginfo(image): - if image is None: - return '', '', '' - - geninfo, items = images.read_info_from_image(image) - items = {**{'parameters': geninfo}, **items} - - info = '' - for key, text in items.items(): - info += f""" -
    -

    {plaintext_to_html(str(key))}

    -

    {plaintext_to_html(str(text))}

    -
    -""".strip()+"\n" - - if len(info) == 0: - message = "Nothing found in the image." - info = f"

    {message}

    " - - return '', geninfo, info - - -def create_config(ckpt_result, config_source, a, b, c): - def config(x): - res = sd_models.find_checkpoint_config(x) if x else None - return res if res != shared.sd_default_config else None - - if config_source == 0: - cfg = config(a) or config(b) or config(c) - elif config_source == 1: - cfg = config(b) - elif config_source == 2: - cfg = config(c) - else: - cfg = None - - if cfg is None: - return - - filename, _ = os.path.splitext(ckpt_result) - checkpoint_filename = filename + ".yaml" - - print("Copying config:") - print(" from:", cfg) - print(" to:", checkpoint_filename) - shutil.copyfile(cfg, checkpoint_filename) - - -checkpoint_dict_skip_on_merge = ["cond_stage_model.transformer.text_model.embeddings.position_ids"] - - -def to_half(tensor, enable): - if enable and tensor.dtype == torch.float: - return tensor.half() - - return tensor - - -def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_model_name, interp_method, multiplier, save_as_half, custom_name, checkpoint_format, config_source, bake_in_vae, discard_weights): - shared.state.begin() - shared.state.job = 'model-merge' - - def fail(message): - shared.state.textinfo = message - shared.state.end() - return [*[gr.update() for _ in range(4)], message] - - def weighted_sum(theta0, theta1, alpha): - return ((1 - alpha) * theta0) + (alpha * theta1) - - def get_difference(theta1, theta2): - return theta1 - theta2 - - def add_difference(theta0, theta1_2_diff, alpha): - return theta0 + (alpha * theta1_2_diff) - - def filename_weighted_sum(): - a = primary_model_info.model_name - b = secondary_model_info.model_name - Ma = round(1 - multiplier, 2) - Mb = round(multiplier, 2) - - return f"{Ma}({a}) + {Mb}({b})" - - def filename_add_difference(): - a = primary_model_info.model_name - b = secondary_model_info.model_name - c = tertiary_model_info.model_name - M = round(multiplier, 2) - - return f"{a} + {M}({b} - {c})" - - def filename_nothing(): - return primary_model_info.model_name - - theta_funcs = { - "Weighted sum": (filename_weighted_sum, None, weighted_sum), - "Add difference": (filename_add_difference, get_difference, add_difference), - "No interpolation": (filename_nothing, None, None), - } - filename_generator, theta_func1, theta_func2 = theta_funcs[interp_method] - shared.state.job_count = (1 if theta_func1 else 0) + (1 if theta_func2 else 0) - - if not primary_model_name: - return fail("Failed: Merging requires a primary model.") - - primary_model_info = sd_models.checkpoints_list[primary_model_name] - - if theta_func2 and not secondary_model_name: - return fail("Failed: Merging requires a secondary model.") - - secondary_model_info = sd_models.checkpoints_list[secondary_model_name] if theta_func2 else None - - if theta_func1 and not tertiary_model_name: - return fail(f"Failed: Interpolation method ({interp_method}) requires a tertiary model.") - - tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None - - result_is_inpainting_model = False - - if theta_func2: - shared.state.textinfo = f"Loading B" - print(f"Loading {secondary_model_info.filename}...") - theta_1 = sd_models.read_state_dict(secondary_model_info.filename, map_location='cpu') - else: - theta_1 = None - - if theta_func1: - shared.state.textinfo = f"Loading C" - print(f"Loading {tertiary_model_info.filename}...") - theta_2 = sd_models.read_state_dict(tertiary_model_info.filename, map_location='cpu') - - shared.state.textinfo = 'Merging B and C' - shared.state.sampling_steps = len(theta_1.keys()) - for key in tqdm.tqdm(theta_1.keys()): - if key in checkpoint_dict_skip_on_merge: - continue - - if 'model' in key: - if key in theta_2: - t2 = theta_2.get(key, torch.zeros_like(theta_1[key])) - theta_1[key] = theta_func1(theta_1[key], t2) - else: - theta_1[key] = torch.zeros_like(theta_1[key]) - - shared.state.sampling_step += 1 - del theta_2 - - shared.state.nextjob() - - shared.state.textinfo = f"Loading {primary_model_info.filename}..." - print(f"Loading {primary_model_info.filename}...") - theta_0 = sd_models.read_state_dict(primary_model_info.filename, map_location='cpu') - - print("Merging...") - shared.state.textinfo = 'Merging A and B' - shared.state.sampling_steps = len(theta_0.keys()) - for key in tqdm.tqdm(theta_0.keys()): - if theta_1 and 'model' in key and key in theta_1: - - if key in checkpoint_dict_skip_on_merge: - continue - - a = theta_0[key] - b = theta_1[key] - - # this enables merging an inpainting model (A) with another one (B); - # where normal model would have 4 channels, for latenst space, inpainting model would - # have another 4 channels for unmasked picture's latent space, plus one channel for mask, for a total of 9 - if a.shape != b.shape and a.shape[0:1] + a.shape[2:] == b.shape[0:1] + b.shape[2:]: - if a.shape[1] == 4 and b.shape[1] == 9: - raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") - - assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" - - theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) - result_is_inpainting_model = True - else: - theta_0[key] = theta_func2(a, b, multiplier) - - theta_0[key] = to_half(theta_0[key], save_as_half) - - shared.state.sampling_step += 1 - - del theta_1 - - bake_in_vae_filename = sd_vae.vae_dict.get(bake_in_vae, None) - if bake_in_vae_filename is not None: - print(f"Baking in VAE from {bake_in_vae_filename}") - shared.state.textinfo = 'Baking in VAE' - vae_dict = sd_vae.load_vae_dict(bake_in_vae_filename, map_location='cpu') - - for key in vae_dict.keys(): - theta_0_key = 'first_stage_model.' + key - if theta_0_key in theta_0: - theta_0[theta_0_key] = to_half(vae_dict[key], save_as_half) - - del vae_dict - - if save_as_half and not theta_func2: - for key in theta_0.keys(): - theta_0[key] = to_half(theta_0[key], save_as_half) - - if discard_weights: - regex = re.compile(discard_weights) - for key in list(theta_0): - if re.search(regex, key): - theta_0.pop(key, None) - - ckpt_dir = shared.cmd_opts.ckpt_dir or sd_models.model_path - - filename = filename_generator() if custom_name == '' else custom_name - filename += ".inpainting" if result_is_inpainting_model else "" - filename += "." + checkpoint_format - - output_modelname = os.path.join(ckpt_dir, filename) - - shared.state.nextjob() - shared.state.textinfo = "Saving" - print(f"Saving to {output_modelname}...") - - _, extension = os.path.splitext(output_modelname) - if extension.lower() == ".safetensors": - safetensors.torch.save_file(theta_0, output_modelname, metadata={"format": "pt"}) - else: - torch.save(theta_0, output_modelname) - - sd_models.list_models() - - create_config(output_modelname, config_source, primary_model_info, secondary_model_info, tertiary_model_info) - - print(f"Checkpoint saved to {output_modelname}.") - shared.state.textinfo = "Checkpoint saved" - shared.state.end() - - return [*[gr.Dropdown.update(choices=sd_models.checkpoint_tiles()) for _ in range(4)], "Checkpoint saved to " + output_modelname] diff --git a/modules/ui.py b/modules/ui.py index eb4b7e6b..4116e167 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -20,7 +20,7 @@ import numpy as np from PIL import Image, PngImagePlugin from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, sd_vae, extra_networks +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, sd_vae, extra_networks, postprocessing, ui_components from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML from modules.paths import script_path @@ -95,8 +95,8 @@ extra_networks_symbol = '\U0001F3B4' # 🎴 def plaintext_to_html(text): - text = "

    " + "
    \n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

    " - return text + return ui_components.plaintext_to_html(text) + def send_gradio_gallery_to_image(x): if len(x) == 0: @@ -1152,7 +1152,7 @@ def create_ui(): result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) submit.click( - fn=wrap_gradio_gpu_call(modules.extras.run_extras, extra_outputs=[None, '']), + fn=wrap_gradio_gpu_call(postprocessing.run_postprocessing, extra_outputs=[None, '']), _js="get_extras_tab_index", inputs=[ dummy_component, @@ -1183,7 +1183,7 @@ def create_ui(): parameters_copypaste.add_paste_fields("extras", extras_image, None) extras_image.change( - fn=modules.extras.clear_cache, + fn=postprocessing.clear_cache, inputs=[], outputs=[] ) diff --git a/modules/ui_components.py b/modules/ui_components.py index 46324425..989cc87b 100644 --- a/modules/ui_components.py +++ b/modules/ui_components.py @@ -1,3 +1,5 @@ +import html + import gradio as gr @@ -47,3 +49,8 @@ class FormColorPicker(gr.ColorPicker, gr.components.FormComponent): def get_block_name(self): return "colorpicker" + + +def plaintext_to_html(text): + text = "

    " + "
    \n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

    " + return text -- cgit v1.2.3 From 985c0b8e9abdd67734d638badefb6ea806b1f28b Mon Sep 17 00:00:00 2001 From: Guillermo Moreno Date: Sat, 21 Jan 2023 17:45:36 -0300 Subject: feat(extra-networks): add thumbs view style --- modules/ui_extra_networks.py | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'modules') diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index af2b8071..ce4801b5 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -25,7 +25,7 @@ class ExtraNetworksPage: def refresh(self): pass - def create_html(self, tabname): + def create_html(self, tabname, view = 'cards'): items_html = '' for item in self.list_items(): @@ -36,7 +36,7 @@ class ExtraNetworksPage: items_html = shared.html("extra-networks-no-cards.html").format(dirs=dirs) res = f""" -
    +
    {items_html}
    """ @@ -75,6 +75,7 @@ class ExtraNetworksUi: self.button_save_preview = None self.preview_target_filename = None + self.view_dropdown = None self.tabname = None @@ -110,6 +111,7 @@ def create_ui(container, button, tabname): filter = gr.Textbox('', show_label=False, elem_id=tabname+"_extra_search", placeholder="Search...", visible=False) button_refresh = gr.Button('Refresh', elem_id=tabname+"_extra_refresh") button_close = gr.Button('Close', elem_id=tabname+"_extra_close") + ui.view_dropdown = gr.Dropdown(['cards', 'thumbs'], elem_id=tabname+"_extra_view", label="View as", value='cards') ui.button_save_preview = gr.Button('Save preview', elem_id=tabname+"_save_preview", visible=False) ui.preview_target_filename = gr.Textbox('Preview save filename', elem_id=tabname+"_preview_filename", visible=False) @@ -117,16 +119,17 @@ def create_ui(container, button, tabname): button.click(fn=lambda: gr.update(visible=True), inputs=[], outputs=[container]) button_close.click(fn=lambda: gr.update(visible=False), inputs=[], outputs=[container]) - def refresh(): + def refresh(view='cards'): res = [] for pg in ui.stored_extra_pages: pg.refresh() - res.append(pg.create_html(ui.tabname)) + res.append(pg.create_html(ui.tabname, view)) return res - button_refresh.click(fn=refresh, inputs=[], outputs=ui.pages) + ui.view_dropdown.change(fn=refresh, inputs=[ui.view_dropdown], outputs=ui.pages) + button_refresh.click(fn=refresh, inputs=[ui.view_dropdown], outputs=ui.pages) return ui @@ -139,7 +142,7 @@ def path_is_parent(parent_path, child_path): def setup_ui(ui, gallery): - def save_preview(index, images, filename): + def save_preview(index, images, filename, view='cards'): if len(images) == 0: print("There is no image in gallery to save as a preview.") return [page.create_html(ui.tabname) for page in ui.stored_extra_pages] @@ -161,11 +164,11 @@ def setup_ui(ui, gallery): image.save(filename) - return [page.create_html(ui.tabname) for page in ui.stored_extra_pages] + return [page.create_html(ui.tabname, view) for page in ui.stored_extra_pages] ui.button_save_preview.click( fn=save_preview, - _js="function(x, y, z){console.log(x, y, z); return [selected_gallery_index(), y, z]}", - inputs=[ui.preview_target_filename, gallery, ui.preview_target_filename], + _js="function(x, y, z, a){console.log(x, y, z, a); return [selected_gallery_index(), y, z, a]}", + inputs=[ui.preview_target_filename, gallery, ui.preview_target_filename, ui.view_dropdown], outputs=[*ui.pages] ) -- cgit v1.2.3 From 66eef11ce7f3db108225668c573cb4a763a43fb3 Mon Sep 17 00:00:00 2001 From: Guillermo Moreno Date: Sat, 21 Jan 2023 18:27:57 -0300 Subject: feat(extra-networks): add default view setting --- modules/shared.py | 4 ++++ modules/ui_extra_networks.py | 8 ++++---- 2 files changed, 8 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/shared.py b/modules/shared.py index cd78e50a..e9548864 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -430,6 +430,10 @@ options_templates.update(options_section(('interrogate', "Interrogate Options"), "deepbooru_filter_tags": OptionInfo("", "filter out those tags from deepbooru output (separated by comma)"), })) +options_templates.update(options_section(('extra_networks', "Extra Networks"), { + "extra_networks_default_view": OptionInfo("cards", "Default view for Extra Networks", gr.Dropdown, { "choices": ["cards", "thumbs"] }), +})) + options_templates.update(options_section(('ui', "User interface"), { "return_grid": OptionInfo(True, "Show grid in results for web"), "do_not_show_images": OptionInfo(False, "Do not show any images in results for web"), diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index ce4801b5..179ba47a 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -25,7 +25,7 @@ class ExtraNetworksPage: def refresh(self): pass - def create_html(self, tabname, view = 'cards'): + def create_html(self, tabname, view=shared.opts.extra_networks_default_view): items_html = '' for item in self.list_items(): @@ -111,7 +111,7 @@ def create_ui(container, button, tabname): filter = gr.Textbox('', show_label=False, elem_id=tabname+"_extra_search", placeholder="Search...", visible=False) button_refresh = gr.Button('Refresh', elem_id=tabname+"_extra_refresh") button_close = gr.Button('Close', elem_id=tabname+"_extra_close") - ui.view_dropdown = gr.Dropdown(['cards', 'thumbs'], elem_id=tabname+"_extra_view", label="View as", value='cards') + ui.view_dropdown = gr.Dropdown(['cards', 'thumbs'], elem_id=tabname+"_extra_view", label="View as", value=lambda: shared.opts.extra_networks_default_view) ui.button_save_preview = gr.Button('Save preview', elem_id=tabname+"_save_preview", visible=False) ui.preview_target_filename = gr.Textbox('Preview save filename', elem_id=tabname+"_preview_filename", visible=False) @@ -119,7 +119,7 @@ def create_ui(container, button, tabname): button.click(fn=lambda: gr.update(visible=True), inputs=[], outputs=[container]) button_close.click(fn=lambda: gr.update(visible=False), inputs=[], outputs=[container]) - def refresh(view='cards'): + def refresh(view): res = [] for pg in ui.stored_extra_pages: @@ -142,7 +142,7 @@ def path_is_parent(parent_path, child_path): def setup_ui(ui, gallery): - def save_preview(index, images, filename, view='cards'): + def save_preview(index, images, filename, view): if len(images) == 0: print("There is no image in gallery to save as a preview.") return [page.create_html(ui.tabname) for page in ui.stored_extra_pages] -- cgit v1.2.3 From f80ff3c1e444926879c284be9384a26ca38d4955 Mon Sep 17 00:00:00 2001 From: Guillermo Moreno Date: Sun, 22 Jan 2023 22:01:24 -0300 Subject: feat(extra-networks): remove view dropdown --- modules/ui_extra_networks.py | 20 +++++++++----------- 1 file changed, 9 insertions(+), 11 deletions(-) (limited to 'modules') diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index 179ba47a..2ddac3d8 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -25,7 +25,8 @@ class ExtraNetworksPage: def refresh(self): pass - def create_html(self, tabname, view=shared.opts.extra_networks_default_view): + def create_html(self, tabname): + view = shared.opts.extra_networks_default_view items_html = '' for item in self.list_items(): @@ -75,7 +76,6 @@ class ExtraNetworksUi: self.button_save_preview = None self.preview_target_filename = None - self.view_dropdown = None self.tabname = None @@ -111,7 +111,6 @@ def create_ui(container, button, tabname): filter = gr.Textbox('', show_label=False, elem_id=tabname+"_extra_search", placeholder="Search...", visible=False) button_refresh = gr.Button('Refresh', elem_id=tabname+"_extra_refresh") button_close = gr.Button('Close', elem_id=tabname+"_extra_close") - ui.view_dropdown = gr.Dropdown(['cards', 'thumbs'], elem_id=tabname+"_extra_view", label="View as", value=lambda: shared.opts.extra_networks_default_view) ui.button_save_preview = gr.Button('Save preview', elem_id=tabname+"_save_preview", visible=False) ui.preview_target_filename = gr.Textbox('Preview save filename', elem_id=tabname+"_preview_filename", visible=False) @@ -119,17 +118,16 @@ def create_ui(container, button, tabname): button.click(fn=lambda: gr.update(visible=True), inputs=[], outputs=[container]) button_close.click(fn=lambda: gr.update(visible=False), inputs=[], outputs=[container]) - def refresh(view): + def refresh(): res = [] for pg in ui.stored_extra_pages: pg.refresh() - res.append(pg.create_html(ui.tabname, view)) + res.append(pg.create_html(ui.tabname)) return res - ui.view_dropdown.change(fn=refresh, inputs=[ui.view_dropdown], outputs=ui.pages) - button_refresh.click(fn=refresh, inputs=[ui.view_dropdown], outputs=ui.pages) + button_refresh.click(fn=refresh, inputs=[], outputs=ui.pages) return ui @@ -142,7 +140,7 @@ def path_is_parent(parent_path, child_path): def setup_ui(ui, gallery): - def save_preview(index, images, filename, view): + def save_preview(index, images, filename): if len(images) == 0: print("There is no image in gallery to save as a preview.") return [page.create_html(ui.tabname) for page in ui.stored_extra_pages] @@ -164,11 +162,11 @@ def setup_ui(ui, gallery): image.save(filename) - return [page.create_html(ui.tabname, view) for page in ui.stored_extra_pages] + return [page.create_html(ui.tabname) for page in ui.stored_extra_pages] ui.button_save_preview.click( fn=save_preview, - _js="function(x, y, z, a){console.log(x, y, z, a); return [selected_gallery_index(), y, z, a]}", - inputs=[ui.preview_target_filename, gallery, ui.preview_target_filename, ui.view_dropdown], + _js="function(x, y, z){console.log(x, y, z); return [selected_gallery_index(), y, z]}", + inputs=[ui.preview_target_filename, gallery, ui.preview_target_filename], outputs=[*ui.pages] ) -- cgit v1.2.3 From b5230197a69d36a79fdc4919c59a03e00e872dd3 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 23 Jan 2023 09:24:43 +0300 Subject: rework extras tab to use script system --- modules/api/api.py | 13 +- modules/postprocessing.py | 236 +++++++++------------------------ modules/scripts.py | 28 ++-- modules/scripts_postprocessing.py | 147 +++++++++++++++++++++ modules/shared.py | 5 + modules/ui.py | 265 +------------------------------------- modules/ui_common.py | 202 +++++++++++++++++++++++++++++ modules/ui_components.py | 6 - modules/ui_postprocessing.py | 57 ++++++++ 9 files changed, 500 insertions(+), 459 deletions(-) create mode 100644 modules/scripts_postprocessing.py create mode 100644 modules/ui_common.py create mode 100644 modules/ui_postprocessing.py (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index f2e9e884..5d60fc0a 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -11,10 +11,9 @@ from fastapi.security import HTTPBasic, HTTPBasicCredentials from secrets import compare_digest import modules.shared as shared -from modules import sd_samplers, deepbooru, sd_hijack, images, scripts, ui +from modules import sd_samplers, deepbooru, sd_hijack, images, scripts, ui, postprocessing from modules.api.models import * from modules.processing import StableDiffusionProcessingTxt2Img, StableDiffusionProcessingImg2Img, process_images -from modules.extras import run_extras from modules.textual_inversion.textual_inversion import create_embedding, train_embedding from modules.textual_inversion.preprocess import preprocess from modules.hypernetworks.hypernetwork import create_hypernetwork, train_hypernetwork @@ -45,10 +44,8 @@ def validate_sampler_name(name): def setUpscalers(req: dict): reqDict = vars(req) - reqDict['extras_upscaler_1'] = upscaler_to_index(req.upscaler_1) - reqDict['extras_upscaler_2'] = upscaler_to_index(req.upscaler_2) - reqDict.pop('upscaler_1') - reqDict.pop('upscaler_2') + reqDict['extras_upscaler_1'] = reqDict.pop('upscaler_1', None) + reqDict['extras_upscaler_2'] = reqDict.pop('upscaler_2', None) return reqDict def decode_base64_to_image(encoding): @@ -244,7 +241,7 @@ class Api: reqDict['image'] = decode_base64_to_image(reqDict['image']) with self.queue_lock: - result = run_extras(extras_mode=0, image_folder="", input_dir="", output_dir="", save_output=False, **reqDict) + result = postprocessing.run_extras(extras_mode=0, image_folder="", input_dir="", output_dir="", save_output=False, **reqDict) return ExtrasSingleImageResponse(image=encode_pil_to_base64(result[0][0]), html_info=result[1]) @@ -260,7 +257,7 @@ class Api: reqDict.pop('imageList') with self.queue_lock: - result = run_extras(extras_mode=1, image="", input_dir="", output_dir="", save_output=False, **reqDict) + result = postprocessing.run_extras(extras_mode=1, image="", input_dir="", output_dir="", save_output=False, **reqDict) return ExtrasBatchImagesResponse(images=list(map(encode_pil_to_base64, result[0])), html_info=result[1]) diff --git a/modules/postprocessing.py b/modules/postprocessing.py index cb85720b..8514fea7 100644 --- a/modules/postprocessing.py +++ b/modules/postprocessing.py @@ -1,219 +1,103 @@ -from __future__ import annotations import os -import numpy as np from PIL import Image -from typing import Callable, List, OrderedDict, Tuple -from functools import partial -from dataclasses import dataclass - -from modules import shared, images, devices, ui_components +from modules import shared, images, devices, scripts, scripts_postprocessing, ui_common, generation_parameters_copypaste from modules.shared import opts -import modules.gfpgan_model -import modules.codeformer_model - - -class LruCache(OrderedDict): - @dataclass(frozen=True) - class Key: - image_hash: int - info_hash: int - args_hash: int - - @dataclass - class Value: - image: Image.Image - info: str - - def __init__(self, max_size: int = 5, *args, **kwargs): - super().__init__(*args, **kwargs) - self._max_size = max_size - - def get(self, key: LruCache.Key) -> LruCache.Value: - ret = super().get(key) - if ret is not None: - self.move_to_end(key) # Move to end of eviction list - return ret - - def put(self, key: LruCache.Key, value: LruCache.Value) -> None: - self[key] = value - while len(self) > self._max_size: - self.popitem(last=False) - - -cached_images: LruCache = LruCache(max_size=5) -def run_postprocessing(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): +def run_postprocessing(extras_mode, image, image_folder, input_dir, output_dir, show_extras_results, *args, save_output: bool = True): devices.torch_gc() shared.state.begin() shared.state.job = 'extras' - imageArr = [] - # Also keep track of original file names - imageNameArr = [] + image_data = [] + image_names = [] outputs = [] if extras_mode == 1: - #convert file to pillow image for img in image_folder: image = Image.open(img) - imageArr.append(image) - imageNameArr.append(os.path.splitext(img.orig_name)[0]) + image_data.append(image) + image_names.append(os.path.splitext(img.orig_name)[0]) elif extras_mode == 2: assert not shared.cmd_opts.hide_ui_dir_config, '--hide-ui-dir-config option must be disabled' + assert input_dir, 'input directory not selected' - if input_dir == '': - return outputs, "Please select an input directory.", '' image_list = shared.listfiles(input_dir) - for img in image_list: + for filename in image_list: try: - image = Image.open(img) + image = Image.open(filename) except Exception: continue - imageArr.append(image) - imageNameArr.append(img) + image_data.append(image) + image_names.append(filename) else: - imageArr.append(image) - imageNameArr.append(None) + assert image, 'image not selected' + + image_data.append(image) + image_names.append(None) if extras_mode == 2 and output_dir != '': outpath = output_dir else: outpath = opts.outdir_samples or opts.outdir_extras_samples - # Extra operation definitions - - def run_gfpgan(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-gfpgan' - restored_img = modules.gfpgan_model.gfpgan_fix_faces(np.array(image, dtype=np.uint8)) - res = Image.fromarray(restored_img) - - if gfpgan_visibility < 1.0: - res = Image.blend(image, res, gfpgan_visibility) - - info += f"GFPGAN visibility:{round(gfpgan_visibility, 2)}\n" - return (res, info) - - def run_codeformer(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - shared.state.job = 'extras-codeformer' - restored_img = modules.codeformer_model.codeformer.restore(np.array(image, dtype=np.uint8), w=codeformer_weight) - res = Image.fromarray(restored_img) - - if codeformer_visibility < 1.0: - res = Image.blend(image, res, codeformer_visibility) - - info += f"CodeFormer w: {round(codeformer_weight, 2)}, CodeFormer visibility:{round(codeformer_visibility, 2)}\n" - return (res, info) - - def upscale(image, scaler_index, resize, mode, resize_w, resize_h, crop): - shared.state.job = 'extras-upscale' - upscaler = shared.sd_upscalers[scaler_index] - res = upscaler.scaler.upscale(image, resize, upscaler.data_path) - if mode == 1 and crop: - cropped = Image.new("RGB", (resize_w, resize_h)) - cropped.paste(res, box=(resize_w // 2 - res.width // 2, resize_h // 2 - res.height // 2)) - res = cropped - return res - - def run_prepare_crop(image: Image.Image, info: str) -> Tuple[Image.Image, str]: - # Actual crop happens in run_upscalers_blend, this just sets upscaling_resize and adds info text - nonlocal upscaling_resize - if resize_mode == 1: - upscaling_resize = max(upscaling_resize_w/image.width, upscaling_resize_h/image.height) - crop_info = " (crop)" if upscaling_crop else "" - info += f"Resize to: {upscaling_resize_w:g}x{upscaling_resize_h:g}{crop_info}\n" - return (image, info) - - @dataclass - class UpscaleParams: - upscaler_idx: int - blend_alpha: float - - def run_upscalers_blend(params: List[UpscaleParams], image: Image.Image, info: str) -> Tuple[Image.Image, str]: - blended_result: Image.Image = None - image_hash: str = hash(np.array(image.getdata()).tobytes()) - for upscaler in params: - upscale_args = (upscaler.upscaler_idx, upscaling_resize, resize_mode, - upscaling_resize_w, upscaling_resize_h, upscaling_crop) - cache_key = LruCache.Key(image_hash=image_hash, - info_hash=hash(info), - args_hash=hash(upscale_args)) - cached_entry = cached_images.get(cache_key) - if cached_entry is None: - res = upscale(image, *upscale_args) - info += f"Upscale: {round(upscaling_resize, 3)}, visibility: {upscaler.blend_alpha}, model:{shared.sd_upscalers[upscaler.upscaler_idx].name}\n" - cached_images.put(cache_key, LruCache.Value(image=res, info=info)) - else: - res, info = cached_entry.image, cached_entry.info - - if blended_result is None: - blended_result = res - else: - blended_result = Image.blend(blended_result, res, upscaler.blend_alpha) - return (blended_result, info) - - # Build a list of operations to run - facefix_ops: List[Callable] = [] - facefix_ops += [run_gfpgan] if gfpgan_visibility > 0 else [] - facefix_ops += [run_codeformer] if codeformer_visibility > 0 else [] - - upscale_ops: List[Callable] = [] - upscale_ops += [run_prepare_crop] if resize_mode == 1 else [] - - if upscaling_resize != 0: - step_params: List[UpscaleParams] = [] - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_1, blend_alpha=1.0)) - if extras_upscaler_2 != 0 and extras_upscaler_2_visibility > 0: - step_params.append(UpscaleParams(upscaler_idx=extras_upscaler_2, blend_alpha=extras_upscaler_2_visibility)) - - upscale_ops.append(partial(run_upscalers_blend, step_params)) - - extras_ops: List[Callable] = (upscale_ops + facefix_ops) if upscale_first else (facefix_ops + upscale_ops) - - for image, image_name in zip(imageArr, imageNameArr): - if image is None: - return outputs, "Please select an input image.", '' - - shared.state.textinfo = f'Processing image {image_name}' - + infotext = '' + + for image, name in zip(image_data, image_names): + shared.state.textinfo = name + existing_pnginfo = image.info or {} - image = image.convert("RGB") - info = "" - # Run each operation on each image - for op in extras_ops: - image, info = op(image, info) + pp = scripts_postprocessing.PostprocessedImage(image.convert("RGB")) - if opts.use_original_name_batch and image_name is not None: - basename = os.path.splitext(os.path.basename(image_name))[0] + scripts.scripts_postproc.run(pp, args) + + if opts.use_original_name_batch and name is not None: + basename = os.path.splitext(os.path.basename(name))[0] else: basename = '' - if opts.enable_pnginfo: # append info before save - image.info = existing_pnginfo - image.info["extras"] = info + infotext = ", ".join([k if k == v else f'{k}: {generation_parameters_copypaste.quote(v)}' for k, v in pp.info.items() if v is not None]) - if save_output: - # Add upscaler name as a suffix. - suffix = f"-{shared.sd_upscalers[extras_upscaler_1].name}" if shared.opts.use_upscaler_name_as_suffix else "" - # Add second upscaler if applicable. - if suffix and extras_upscaler_2 and extras_upscaler_2_visibility: - suffix += f"-{shared.sd_upscalers[extras_upscaler_2].name}" + if opts.enable_pnginfo: + pp.image.info = existing_pnginfo + pp.image.info["postprocessing"] = infotext - images.save_image(image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=info, short_filename=True, - no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None, suffix=suffix) + if save_output: + images.save_image(pp.image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=pp.info, short_filename=True, no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None) - if extras_mode != 2 or show_extras_results : - outputs.append(image) + if extras_mode != 2 or show_extras_results: + outputs.append(pp.image) devices.torch_gc() - return outputs, ui_components.plaintext_to_html(info), '' - - -def clear_cache(): - cached_images.clear() - + return outputs, ui_common.plaintext_to_html(infotext), '' + + +def run_extras(extras_mode, resize_mode, image, image_folder, input_dir, output_dir, show_extras_results, gfpgan_visibility, codeformer_visibility, codeformer_weight, upscaling_resize, upscaling_resize_w, upscaling_resize_h, upscaling_crop, extras_upscaler_1, extras_upscaler_2, extras_upscaler_2_visibility, upscale_first: bool, save_output: bool = True): + """old handler for API""" + + args = scripts.scripts_postproc.create_args_for_run({ + "Upscale": { + "upscale_mode": resize_mode, + "upscale_by": upscaling_resize, + "upscale_to_width": upscaling_resize_w, + "upscale_to_height": upscaling_resize_h, + "upscale_crop": upscaling_crop, + "upscaler_1_name": extras_upscaler_1, + "upscaler_2_name": extras_upscaler_2, + "upscaler_2_visibility": extras_upscaler_2_visibility, + }, + "GFPGAN": { + "gfpgan_visibility": gfpgan_visibility, + }, + "CodeFormer": { + "codeformer_visibility": codeformer_visibility, + "codeformer_weight": codeformer_weight, + }, + }) + + return run_postprocessing(extras_mode, image, image_folder, input_dir, output_dir, show_extras_results, *args, save_output=save_output) diff --git a/modules/scripts.py b/modules/scripts.py index 4ffc369b..03907a63 100644 --- a/modules/scripts.py +++ b/modules/scripts.py @@ -7,7 +7,7 @@ from collections import namedtuple import gradio as gr from modules.processing import StableDiffusionProcessing -from modules import shared, paths, script_callbacks, extensions, script_loading +from modules import shared, paths, script_callbacks, extensions, script_loading, scripts_postprocessing AlwaysVisible = object() @@ -150,8 +150,10 @@ def basedir(): return current_basedir -scripts_data = [] ScriptFile = namedtuple("ScriptFile", ["basedir", "filename", "path"]) + +scripts_data = [] +postprocessing_scripts_data = [] ScriptClassData = namedtuple("ScriptClassData", ["script_class", "path", "basedir", "module"]) @@ -190,23 +192,31 @@ def list_files_with_name(filename): def load_scripts(): global current_basedir scripts_data.clear() + postprocessing_scripts_data.clear() script_callbacks.clear_callbacks() scripts_list = list_scripts("scripts", ".py") syspath = sys.path + def register_scripts_from_module(module): + for key, script_class in module.__dict__.items(): + if type(script_class) != type: + continue + + if issubclass(script_class, Script): + scripts_data.append(ScriptClassData(script_class, scriptfile.path, scriptfile.basedir, module)) + elif issubclass(script_class, scripts_postprocessing.ScriptPostprocessing): + postprocessing_scripts_data.append(ScriptClassData(script_class, scriptfile.path, scriptfile.basedir, module)) + for scriptfile in sorted(scripts_list): try: if scriptfile.basedir != paths.script_path: sys.path = [scriptfile.basedir] + sys.path current_basedir = scriptfile.basedir - module = script_loading.load_module(scriptfile.path) - - for key, script_class in module.__dict__.items(): - if type(script_class) == type and issubclass(script_class, Script): - scripts_data.append(ScriptClassData(script_class, scriptfile.path, scriptfile.basedir, module)) + script_module = script_loading.load_module(scriptfile.path) + register_scripts_from_module(script_module) except Exception: print(f"Error loading script: {scriptfile.filename}", file=sys.stderr) @@ -413,6 +423,7 @@ class ScriptRunner: scripts_txt2img = ScriptRunner() scripts_img2img = ScriptRunner() +scripts_postproc = scripts_postprocessing.ScriptPostprocessingRunner() scripts_current: ScriptRunner = None @@ -423,12 +434,13 @@ def reload_script_body_only(): def reload_scripts(): - global scripts_txt2img, scripts_img2img + global scripts_txt2img, scripts_img2img, scripts_postproc load_scripts() scripts_txt2img = ScriptRunner() scripts_img2img = ScriptRunner() + scripts_postproc = scripts_postprocessing.ScriptPostprocessingRunner() def IOComponent_init(self, *args, **kwargs): diff --git a/modules/scripts_postprocessing.py b/modules/scripts_postprocessing.py new file mode 100644 index 00000000..25de02d0 --- /dev/null +++ b/modules/scripts_postprocessing.py @@ -0,0 +1,147 @@ +import os +import gradio as gr + +from modules import errors, shared + + +class PostprocessedImage: + def __init__(self, image): + self.image = image + self.info = {} + + +class ScriptPostprocessing: + filename = None + controls = None + args_from = None + args_to = None + + order = 1000 + """scripts will be ordred by this value in postprocessing UI""" + + name = None + """this function should return the title of the script.""" + + group = None + """A gr.Group component that has all script's UI inside it""" + + def ui(self): + """ + This function should create gradio UI elements. See https://gradio.app/docs/#components + The return value should be a dictionary that maps parameter names to components used in processing. + Values of those components will be passed to process() function. + """ + + pass + + def process(self, pp: PostprocessedImage, **args): + """ + This function is called to postprocess the image. + args contains a dictionary with all values returned by components from ui() + """ + + pass + + def image_changed(self): + pass + + +def wrap_call(func, filename, funcname, *args, default=None, **kwargs): + try: + res = func(*args, **kwargs) + return res + except Exception as e: + errors.display(e, f"calling {filename}/{funcname}") + + return default + + +class ScriptPostprocessingRunner: + def __init__(self): + self.scripts = None + self.ui_created = False + + def initialize_scripts(self, scripts_data): + self.scripts = [] + + for script_class, path, basedir, script_module in scripts_data: + script: ScriptPostprocessing = script_class() + script.filename = path + + self.scripts.append(script) + + def create_script_ui(self, script, inputs): + script.args_from = len(inputs) + script.args_to = len(inputs) + + script.controls = wrap_call(script.ui, script.filename, "ui") + + for control in script.controls.values(): + control.custom_script_source = os.path.basename(script.filename) + + inputs += list(script.controls.values()) + script.args_to = len(inputs) + + def scripts_in_preferred_order(self): + if self.scripts is None: + import modules.scripts + self.initialize_scripts(modules.scripts.postprocessing_scripts_data) + + scripts_order = [x.lower().strip() for x in shared.opts.postprocessing_scipts_order.split(",")] + + def script_score(name): + name = name.lower() + for i, possible_match in enumerate(scripts_order): + if possible_match in name: + return i + + return len(self.scripts) + + script_scores = {script.name: (script_score(script.name), script.order, script.name, original_index) for original_index, script in enumerate(self.scripts)} + + return sorted(self.scripts, key=lambda x: script_scores[x.name]) + + def setup_ui(self): + inputs = [] + + for script in self.scripts_in_preferred_order(): + with gr.Box() as group: + self.create_script_ui(script, inputs) + + script.group = group + + self.ui_created = True + return inputs + + def run(self, pp: PostprocessedImage, args): + for script in self.scripts_in_preferred_order(): + shared.state.job = script.name + + script_args = args[script.args_from:script.args_to] + + process_args = {} + for (name, component), value in zip(script.controls.items(), script_args): + process_args[name] = value + + script.process(pp, **process_args) + + def create_args_for_run(self, scripts_args): + if not self.ui_created: + with gr.Blocks(analytics_enabled=False): + self.setup_ui() + + scripts = self.scripts_in_preferred_order() + args = [None] * max([x.args_to for x in scripts]) + + for script in scripts: + script_args_dict = scripts_args.get(script.name, None) + if script_args_dict is not None: + + for i, name in enumerate(script.controls): + args[script.args_from + i] = script_args_dict.get(name, None) + + return args + + def image_changed(self): + for script in self.scripts_in_preferred_order(): + script.image_changed() diff --git a/modules/shared.py b/modules/shared.py index cd78e50a..cb73bf31 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -474,6 +474,11 @@ options_templates.update(options_section(('sampler-params', "Sampler parameters" 'always_discard_next_to_last_sigma': OptionInfo(False, "Always discard next-to-last sigma"), })) +options_templates.update(options_section(('postprocessing', "Postprocessing"), { + 'postprocessing_scipts_order': OptionInfo("upscale, gfpgan, codeformer", "Postprocessing operation order"), + 'upscaling_max_images_in_cache': OptionInfo(5, "Maximum number of images in upscaling cache", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), +})) + options_templates.update(options_section((None, "Hidden options"), { "disabled_extensions": OptionInfo([], "Disable those extensions"), "sd_checkpoint_hash": OptionInfo("", "SHA256 hash of the current checkpoint"), diff --git a/modules/ui.py b/modules/ui.py index 4116e167..8cb8e613 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -20,7 +20,7 @@ import numpy as np from PIL import Image, PngImagePlugin from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_gradio_call -from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, sd_vae, extra_networks, postprocessing, ui_components +from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, sd_vae, extra_networks, postprocessing, ui_components, ui_common, ui_postprocessing from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML from modules.paths import script_path @@ -86,7 +86,6 @@ css_hide_progressbar = """ random_symbol = '\U0001f3b2\ufe0f' # 🎲️ reuse_symbol = '\u267b\ufe0f' # ♻️ paste_symbol = '\u2199\ufe0f' # ↙ -folder_symbol = '\U0001f4c2' # 📂 refresh_symbol = '\U0001f504' # 🔄 save_style_symbol = '\U0001f4be' # 💾 apply_style_symbol = '\U0001f4cb' # 📋 @@ -95,7 +94,7 @@ extra_networks_symbol = '\U0001F3B4' # 🎴 def plaintext_to_html(text): - return ui_components.plaintext_to_html(text) + return ui_common.plaintext_to_html(text) def send_gradio_gallery_to_image(x): @@ -103,70 +102,6 @@ def send_gradio_gallery_to_image(x): return None return image_from_url_text(x[0]) -def save_files(js_data, images, do_make_zip, index): - import csv - filenames = [] - fullfns = [] - - #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it - class MyObject: - def __init__(self, d=None): - if d is not None: - for key, value in d.items(): - setattr(self, key, value) - - data = json.loads(js_data) - - p = MyObject(data) - path = opts.outdir_save - save_to_dirs = opts.use_save_to_dirs_for_ui - extension: str = opts.samples_format - start_index = 0 - - if index > -1 and opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only - - images = [images[index]] - start_index = index - - os.makedirs(opts.outdir_save, exist_ok=True) - - with open(os.path.join(opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: - at_start = file.tell() == 0 - writer = csv.writer(file) - if at_start: - writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) - - for image_index, filedata in enumerate(images, start_index): - image = image_from_url_text(filedata) - - is_grid = image_index < p.index_of_first_image - i = 0 if is_grid else (image_index - p.index_of_first_image) - - fullfn, txt_fullfn = save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) - - filename = os.path.relpath(fullfn, path) - filenames.append(filename) - fullfns.append(fullfn) - if txt_fullfn: - filenames.append(os.path.basename(txt_fullfn)) - fullfns.append(txt_fullfn) - - writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) - - # Make Zip - if do_make_zip: - zip_filepath = os.path.join(path, "images.zip") - - from zipfile import ZipFile - with ZipFile(zip_filepath, "w") as zip_file: - for i in range(len(fullfns)): - with open(fullfns[i], mode="rb") as f: - zip_file.writestr(filenames[i], f.read()) - fullfns.insert(0, zip_filepath) - - return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") - - def visit(x, func, path=""): if hasattr(x, 'children'): for c in x.children: @@ -444,19 +379,6 @@ def apply_setting(key, value): opts.save(shared.config_filename) return getattr(opts, key) - -def update_generation_info(generation_info, html_info, img_index): - try: - generation_info = json.loads(generation_info) - if img_index < 0 or img_index >= len(generation_info["infotexts"]): - return html_info, gr.update() - return plaintext_to_html(generation_info["infotexts"][img_index]), gr.update() - except Exception: - pass - # if the json parse or anything else fails, just return the old html_info - return html_info, gr.update() - - def create_refresh_button(refresh_component, refresh_method, refreshed_args, elem_id): def refresh(): refresh_method() @@ -477,107 +399,7 @@ def create_refresh_button(refresh_component, refresh_method, refreshed_args, ele def create_output_panel(tabname, outdir): - def open_folder(f): - if not os.path.exists(f): - print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') - return - elif not os.path.isdir(f): - print(f""" -WARNING -An open_folder request was made with an argument that is not a folder. -This could be an error or a malicious attempt to run code on your computer. -Requested path was: {f} -""", file=sys.stderr) - return - - if not shared.cmd_opts.hide_ui_dir_config: - path = os.path.normpath(f) - if platform.system() == "Windows": - os.startfile(path) - elif platform.system() == "Darwin": - sp.Popen(["open", path]) - elif "microsoft-standard-WSL2" in platform.uname().release: - sp.Popen(["wsl-open", path]) - else: - sp.Popen(["xdg-open", path]) - - with gr.Column(variant='panel', elem_id=f"{tabname}_results"): - with gr.Group(elem_id=f"{tabname}_gallery_container"): - result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) - - generation_info = None - with gr.Column(): - with gr.Row(elem_id=f"image_buttons_{tabname}"): - open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') - - if tabname != "extras": - save = gr.Button('Save', elem_id=f'save_{tabname}') - save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') - - buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) - - open_folder_button.click( - fn=lambda: open_folder(opts.outdir_samples or outdir), - inputs=[], - outputs=[], - ) - - if tabname != "extras": - with gr.Row(): - download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') - - with gr.Group(): - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') - if tabname == 'txt2img' or tabname == 'img2img': - generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") - generation_info_button.click( - fn=update_generation_info, - _js="function(x, y, z){ return [x, y, selected_gallery_index()] }", - inputs=[generation_info, html_info, html_info], - outputs=[html_info, html_info], - ) - - save.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ], - show_progress=False, - ) - - save_zip.click( - fn=wrap_gradio_call(save_files), - _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", - inputs=[ - generation_info, - result_gallery, - html_info, - html_info, - ], - outputs=[ - download_files, - html_log, - ] - ) - - else: - html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') - html_info = gr.HTML(elem_id=f'html_info_{tabname}') - html_log = gr.HTML(elem_id=f'html_log_{tabname}') - - parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) - return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log + return ui_common.create_output_panel(tabname, outdir) def create_sampler_and_steps_selection(choices, tabname): @@ -1106,86 +928,7 @@ def create_ui(): modules.scripts.scripts_current = None with gr.Blocks(analytics_enabled=False) as extras_interface: - with gr.Row().style(equal_height=False): - with gr.Column(variant='compact'): - with gr.Tabs(elem_id="mode_extras"): - with gr.TabItem('Single Image', elem_id="extras_single_tab"): - extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") - - with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab"): - image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") - - with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab"): - extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") - extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") - show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") - - submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') - - with gr.Tabs(elem_id="extras_resize_mode"): - with gr.TabItem('Scale by', elem_id="extras_scale_by_tab"): - upscaling_resize = gr.Slider(minimum=1.0, maximum=8.0, step=0.05, label="Resize", value=4, elem_id="extras_upscaling_resize") - with gr.TabItem('Scale to', elem_id="extras_scale_to_tab"): - with gr.Group(): - with gr.Row(): - upscaling_resize_w = gr.Number(label="Width", value=512, precision=0, elem_id="extras_upscaling_resize_w") - upscaling_resize_h = gr.Number(label="Height", value=512, precision=0, elem_id="extras_upscaling_resize_h") - upscaling_crop = gr.Checkbox(label='Crop to fit', value=True, elem_id="extras_upscaling_crop") - - with gr.Group(): - extras_upscaler_1 = gr.Radio(label='Upscaler 1', elem_id="extras_upscaler_1", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - - with gr.Group(): - extras_upscaler_2 = gr.Radio(label='Upscaler 2', elem_id="extras_upscaler_2", choices=[x.name for x in shared.sd_upscalers], value=shared.sd_upscalers[0].name, type="index") - extras_upscaler_2_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="Upscaler 2 visibility", value=1, elem_id="extras_upscaler_2_visibility") - - with gr.Group(): - gfpgan_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="GFPGAN visibility", value=0, interactive=modules.gfpgan_model.have_gfpgan, elem_id="extras_gfpgan_visibility") - - with gr.Group(): - codeformer_visibility = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer visibility", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_visibility") - codeformer_weight = gr.Slider(minimum=0.0, maximum=1.0, step=0.001, label="CodeFormer weight (0 = maximum effect, 1 = minimum effect)", value=0, interactive=modules.codeformer_model.have_codeformer, elem_id="extras_codeformer_weight") - - with gr.Group(): - upscale_before_face_fix = gr.Checkbox(label='Upscale Before Restoring Faces', value=False, elem_id="extras_upscale_before_face_fix") - - result_images, html_info_x, html_info, html_log = create_output_panel("extras", opts.outdir_extras_samples) - - submit.click( - fn=wrap_gradio_gpu_call(postprocessing.run_postprocessing, extra_outputs=[None, '']), - _js="get_extras_tab_index", - inputs=[ - dummy_component, - dummy_component, - extras_image, - image_batch, - extras_batch_input_dir, - extras_batch_output_dir, - show_extras_results, - gfpgan_visibility, - codeformer_visibility, - codeformer_weight, - upscaling_resize, - upscaling_resize_w, - upscaling_resize_h, - upscaling_crop, - extras_upscaler_1, - extras_upscaler_2, - extras_upscaler_2_visibility, - upscale_before_face_fix, - ], - outputs=[ - result_images, - html_info_x, - html_info, - ] - ) - parameters_copypaste.add_paste_fields("extras", extras_image, None) - - extras_image.change( - fn=postprocessing.clear_cache, - inputs=[], outputs=[] - ) + ui_postprocessing.create_ui() with gr.Blocks(analytics_enabled=False) as pnginfo_interface: with gr.Row().style(equal_height=False): diff --git a/modules/ui_common.py b/modules/ui_common.py new file mode 100644 index 00000000..8ce75b8c --- /dev/null +++ b/modules/ui_common.py @@ -0,0 +1,202 @@ +import json +import html +import os +import platform +import sys + +import gradio as gr +import scipy as sp + +from modules import call_queue, shared +from modules.generation_parameters_copypaste import image_from_url_text +import modules.images + +folder_symbol = '\U0001f4c2' # 📂 + + +def update_generation_info(generation_info, html_info, img_index): + try: + generation_info = json.loads(generation_info) + if img_index < 0 or img_index >= len(generation_info["infotexts"]): + return html_info, gr.update() + return plaintext_to_html(generation_info["infotexts"][img_index]), gr.update() + except Exception: + pass + # if the json parse or anything else fails, just return the old html_info + return html_info, gr.update() + + +def plaintext_to_html(text): + text = "

    " + "
    \n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

    " + return text + + +def save_files(js_data, images, do_make_zip, index): + import csv + filenames = [] + fullfns = [] + + #quick dictionary to class object conversion. Its necessary due apply_filename_pattern requiring it + class MyObject: + def __init__(self, d=None): + if d is not None: + for key, value in d.items(): + setattr(self, key, value) + + data = json.loads(js_data) + + p = MyObject(data) + path = shared.opts.outdir_save + save_to_dirs = shared.opts.use_save_to_dirs_for_ui + extension: str = shared.opts.samples_format + start_index = 0 + + if index > -1 and shared.opts.save_selected_only and (index >= data["index_of_first_image"]): # ensures we are looking at a specific non-grid picture, and we have save_selected_only + + images = [images[index]] + start_index = index + + os.makedirs(shared.opts.outdir_save, exist_ok=True) + + with open(os.path.join(shared.opts.outdir_save, "log.csv"), "a", encoding="utf8", newline='') as file: + at_start = file.tell() == 0 + writer = csv.writer(file) + if at_start: + writer.writerow(["prompt", "seed", "width", "height", "sampler", "cfgs", "steps", "filename", "negative_prompt"]) + + for image_index, filedata in enumerate(images, start_index): + image = image_from_url_text(filedata) + + is_grid = image_index < p.index_of_first_image + i = 0 if is_grid else (image_index - p.index_of_first_image) + + fullfn, txt_fullfn = modules.images.save_image(image, path, "", seed=p.all_seeds[i], prompt=p.all_prompts[i], extension=extension, info=p.infotexts[image_index], grid=is_grid, p=p, save_to_dirs=save_to_dirs) + + filename = os.path.relpath(fullfn, path) + filenames.append(filename) + fullfns.append(fullfn) + if txt_fullfn: + filenames.append(os.path.basename(txt_fullfn)) + fullfns.append(txt_fullfn) + + writer.writerow([data["prompt"], data["seed"], data["width"], data["height"], data["sampler_name"], data["cfg_scale"], data["steps"], filenames[0], data["negative_prompt"]]) + + # Make Zip + if do_make_zip: + zip_filepath = os.path.join(path, "images.zip") + + from zipfile import ZipFile + with ZipFile(zip_filepath, "w") as zip_file: + for i in range(len(fullfns)): + with open(fullfns[i], mode="rb") as f: + zip_file.writestr(filenames[i], f.read()) + fullfns.insert(0, zip_filepath) + + return gr.File.update(value=fullfns, visible=True), plaintext_to_html(f"Saved: {filenames[0]}") + + +def create_output_panel(tabname, outdir): + from modules import shared + import modules.generation_parameters_copypaste as parameters_copypaste + + def open_folder(f): + if not os.path.exists(f): + print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.') + return + elif not os.path.isdir(f): + print(f""" +WARNING +An open_folder request was made with an argument that is not a folder. +This could be an error or a malicious attempt to run code on your computer. +Requested path was: {f} +""", file=sys.stderr) + return + + if not shared.cmd_opts.hide_ui_dir_config: + path = os.path.normpath(f) + if platform.system() == "Windows": + os.startfile(path) + elif platform.system() == "Darwin": + sp.Popen(["open", path]) + elif "microsoft-standard-WSL2" in platform.uname().release: + sp.Popen(["wsl-open", path]) + else: + sp.Popen(["xdg-open", path]) + + with gr.Column(variant='panel', elem_id=f"{tabname}_results"): + with gr.Group(elem_id=f"{tabname}_gallery_container"): + result_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"{tabname}_gallery").style(grid=4) + + generation_info = None + with gr.Column(): + with gr.Row(elem_id=f"image_buttons_{tabname}"): + open_folder_button = gr.Button(folder_symbol, elem_id="hidden_element" if shared.cmd_opts.hide_ui_dir_config else f'open_folder_{tabname}') + + if tabname != "extras": + save = gr.Button('Save', elem_id=f'save_{tabname}') + save_zip = gr.Button('Zip', elem_id=f'save_zip_{tabname}') + + buttons = parameters_copypaste.create_buttons(["img2img", "inpaint", "extras"]) + + open_folder_button.click( + fn=lambda: open_folder(shared.opts.outdir_samples or outdir), + inputs=[], + outputs=[], + ) + + if tabname != "extras": + with gr.Row(): + download_files = gr.File(None, file_count="multiple", interactive=False, show_label=False, visible=False, elem_id=f'download_files_{tabname}') + + with gr.Group(): + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') + + generation_info = gr.Textbox(visible=False, elem_id=f'generation_info_{tabname}') + if tabname == 'txt2img' or tabname == 'img2img': + generation_info_button = gr.Button(visible=False, elem_id=f"{tabname}_generation_info_button") + generation_info_button.click( + fn=update_generation_info, + _js="function(x, y, z){ return [x, y, selected_gallery_index()] }", + inputs=[generation_info, html_info, html_info], + outputs=[html_info, html_info], + ) + + save.click( + fn=call_queue.wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, false, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ], + show_progress=False, + ) + + save_zip.click( + fn=call_queue.wrap_gradio_call(save_files), + _js="(x, y, z, w) => [x, y, true, selected_gallery_index()]", + inputs=[ + generation_info, + result_gallery, + html_info, + html_info, + ], + outputs=[ + download_files, + html_log, + ] + ) + + else: + html_info_x = gr.HTML(elem_id=f'html_info_x_{tabname}') + html_info = gr.HTML(elem_id=f'html_info_{tabname}') + html_log = gr.HTML(elem_id=f'html_log_{tabname}') + + parameters_copypaste.bind_buttons(buttons, result_gallery, "txt2img" if tabname == "txt2img" else None) + return result_gallery, generation_info if tabname != "extras" else html_info_x, html_info, html_log diff --git a/modules/ui_components.py b/modules/ui_components.py index 989cc87b..9aec3097 100644 --- a/modules/ui_components.py +++ b/modules/ui_components.py @@ -1,5 +1,3 @@ -import html - import gradio as gr @@ -50,7 +48,3 @@ class FormColorPicker(gr.ColorPicker, gr.components.FormComponent): def get_block_name(self): return "colorpicker" - -def plaintext_to_html(text): - text = "

    " + "
    \n".join([f"{html.escape(x)}" for x in text.split('\n')]) + "

    " - return text diff --git a/modules/ui_postprocessing.py b/modules/ui_postprocessing.py new file mode 100644 index 00000000..b418d955 --- /dev/null +++ b/modules/ui_postprocessing.py @@ -0,0 +1,57 @@ +import gradio as gr +from modules import scripts_postprocessing, scripts, shared, gfpgan_model, codeformer_model, ui_common, postprocessing, call_queue +import modules.generation_parameters_copypaste as parameters_copypaste + + +def create_ui(): + tab_index = gr.State(value=0) + + with gr.Row().style(equal_height=False, variant='compact'): + with gr.Column(variant='compact'): + with gr.Tabs(elem_id="mode_extras"): + with gr.TabItem('Single Image', elem_id="extras_single_tab") as tab_single: + extras_image = gr.Image(label="Source", source="upload", interactive=True, type="pil", elem_id="extras_image") + + with gr.TabItem('Batch Process', elem_id="extras_batch_process_tab") as tab_batch: + image_batch = gr.File(label="Batch Process", file_count="multiple", interactive=True, type="file", elem_id="extras_image_batch") + + with gr.TabItem('Batch from Directory', elem_id="extras_batch_directory_tab") as tab_batch_dir: + extras_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, placeholder="A directory on the same machine where the server is running.", elem_id="extras_batch_input_dir") + extras_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, placeholder="Leave blank to save images to the default path.", elem_id="extras_batch_output_dir") + show_extras_results = gr.Checkbox(label='Show result images', value=True, elem_id="extras_show_extras_results") + + submit = gr.Button('Generate', elem_id="extras_generate", variant='primary') + + script_inputs = scripts.scripts_postproc.setup_ui() + + with gr.Column(): + result_images, html_info_x, html_info, html_log = ui_common.create_output_panel("extras", shared.opts.outdir_extras_samples) + + tab_single.select(fn=lambda: 0, inputs=[], outputs=[tab_index]) + tab_batch.select(fn=lambda: 1, inputs=[], outputs=[tab_index]) + tab_batch_dir.select(fn=lambda: 2, inputs=[], outputs=[tab_index]) + + submit.click( + fn=call_queue.wrap_gradio_gpu_call(postprocessing.run_postprocessing, extra_outputs=[None, '']), + inputs=[ + tab_index, + extras_image, + image_batch, + extras_batch_input_dir, + extras_batch_output_dir, + show_extras_results, + *script_inputs + ], + outputs=[ + result_images, + html_info_x, + html_info, + ] + ) + + parameters_copypaste.add_paste_fields("extras", extras_image, None) + + extras_image.change( + fn=scripts.scripts_postproc.image_changed, + inputs=[], outputs=[] + ) -- cgit v1.2.3 From 669dbd9725b3a285503e093a75c0dfa332073d8a Mon Sep 17 00:00:00 2001 From: Shondoit Date: Mon, 23 Jan 2023 09:54:42 +0100 Subject: Fix dark mode Fixes #7048 Co-Authored-By: J.J. Tolton --- modules/ui.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index eb4b7e6b..43bcb7e5 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -1942,11 +1942,11 @@ def reload_javascript(): if cmd_opts.theme is not None: inline += f"set_theme('{cmd_opts.theme}');" - head += f'\n' - for script in modules.scripts.list_scripts("javascript", ".js"): head += f'\n' + head += f'\n' + def template_response(*args, **kwargs): res = shared.GradioTemplateResponseOriginal(*args, **kwargs) res.body = res.body.replace(b'', f'{head}'.encode("utf8")) -- cgit v1.2.3 From fabdae089e476c66eba3b0562e4e1881891804b2 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 23 Jan 2023 14:42:49 +0300 Subject: add missing import to previous commit --- modules/ui.py | 1 + 1 file changed, 1 insertion(+) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 8cb8e613..6b5dfd61 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -41,6 +41,7 @@ from modules.sd_samplers import samplers, samplers_for_img2img from modules.textual_inversion import textual_inversion import modules.hypernetworks.ui from modules.generation_parameters_copypaste import image_from_url_text +import modules.extras warnings.filterwarnings("default" if opts.show_warnings else "ignore", category=UserWarning) -- cgit v1.2.3 From 41265a026de699cc223ca5b76c69b4e8e74aa7c1 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 23 Jan 2023 14:50:20 +0300 Subject: third time's the charm --- modules/extras.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index f04ddfc2..36123aa5 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -7,7 +7,7 @@ import torch import tqdm from modules import shared, images, sd_models, sd_vae -from modules.ui import plaintext_to_html +from modules.ui_common import plaintext_to_html import gradio as gr import safetensors.torch -- cgit v1.2.3 From 194cbd065e4644e986889b78a5a949e075b610e8 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 23 Jan 2023 15:50:32 +0300 Subject: fix open directory button failing --- modules/ui.py | 1 - modules/ui_common.py | 2 +- 2 files changed, 1 insertion(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 94d4a80a..85ae62c7 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -5,7 +5,6 @@ import mimetypes import os import platform import random -import subprocess as sp import sys import tempfile import time diff --git a/modules/ui_common.py b/modules/ui_common.py index 8ce75b8c..9405ac1f 100644 --- a/modules/ui_common.py +++ b/modules/ui_common.py @@ -5,7 +5,7 @@ import platform import sys import gradio as gr -import scipy as sp +import subprocess as sp from modules import call_queue, shared from modules.generation_parameters_copypaste import image_from_url_text -- cgit v1.2.3 From 59146621e256269b85feb536edeb745da20daf68 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 23 Jan 2023 16:40:20 +0300 Subject: better support for xformers flash attention on older versions of torch --- modules/errors.py | 12 +++++++++++ modules/sd_hijack_optimizations.py | 42 ++++++++++++++++---------------------- 2 files changed, 30 insertions(+), 24 deletions(-) (limited to 'modules') diff --git a/modules/errors.py b/modules/errors.py index a10e8708..f6b80dbb 100644 --- a/modules/errors.py +++ b/modules/errors.py @@ -24,6 +24,18 @@ See https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Features#stable """) +already_displayed = {} + + +def display_once(e: Exception, task): + if task in already_displayed: + return + + display(e, task) + + already_displayed[task] = 1 + + def run(code, task): try: code() diff --git a/modules/sd_hijack_optimizations.py b/modules/sd_hijack_optimizations.py index 9967359b..74452709 100644 --- a/modules/sd_hijack_optimizations.py +++ b/modules/sd_hijack_optimizations.py @@ -9,7 +9,7 @@ from torch import einsum from ldm.util import default from einops import rearrange -from modules import shared +from modules import shared, errors from modules.hypernetworks import hypernetwork from .sub_quadratic_attention import efficient_dot_product_attention @@ -279,6 +279,21 @@ def sub_quad_attention(q, k, v, q_chunk_size=1024, kv_chunk_size=None, kv_chunk_ ) +def get_xformers_flash_attention_op(q, k, v): + if not shared.cmd_opts.xformers_flash_attention: + return None + + try: + flash_attention_op = xformers.ops.MemoryEfficientAttentionFlashAttentionOp + fw, bw = flash_attention_op + if fw.supports(xformers.ops.fmha.Inputs(query=q, key=k, value=v, attn_bias=None)): + return flash_attention_op + except Exception as e: + errors.display_once(e, "enabling flash attention") + + return None + + def xformers_attention_forward(self, x, context=None, mask=None): h = self.heads q_in = self.to_q(x) @@ -291,18 +306,7 @@ def xformers_attention_forward(self, x, context=None, mask=None): q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> b n h d', h=h), (q_in, k_in, v_in)) del q_in, k_in, v_in - if shared.cmd_opts.xformers_flash_attention: - op = xformers.ops.MemoryEfficientAttentionFlashAttentionOp - fw, bw = op - if not fw.supports(xformers.ops.fmha.Inputs(query=q, key=k, value=v, attn_bias=None)): - # print('xformers_attention_forward', q.shape, k.shape, v.shape) - # Flash Attention is not availabe for the input arguments. - # Fallback to default xFormers' backend. - op = None - else: - op = None - - out = xformers.ops.memory_efficient_attention(q, k, v, attn_bias=None, op=op) + out = xformers.ops.memory_efficient_attention(q, k, v, attn_bias=None, op=get_xformers_flash_attention_op(q, k, v)) out = rearrange(out, 'b n h d -> b n (h d)', h=h) return self.to_out(out) @@ -377,17 +381,7 @@ def xformers_attnblock_forward(self, x): q = q.contiguous() k = k.contiguous() v = v.contiguous() - if shared.cmd_opts.xformers_flash_attention: - op = xformers.ops.MemoryEfficientAttentionFlashAttentionOp - fw, bw = op - if not fw.supports(xformers.ops.fmha.Inputs(query=q, key=k, value=v)): - # print('xformers_attnblock_forward', q.shape, k.shape, v.shape) - # Flash Attention is not availabe for the input arguments. - # Fallback to default xFormers' backend. - op = None - else: - op = None - out = xformers.ops.memory_efficient_attention(q, k, v, op=op) + out = xformers.ops.memory_efficient_attention(q, k, v, op=get_xformers_flash_attention_op(q, k, v)) out = rearrange(out, 'b (h w) c -> b c h w', h=h) out = self.proj_out(out) return x + out -- cgit v1.2.3 From 925dd09c91e7338aef72e4ec99d67b8b57280215 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Mon, 23 Jan 2023 09:03:17 -0500 Subject: improve interrogate --- modules/interrogate.py | 29 +++++++++++++++++------------ modules/shared.py | 1 + 2 files changed, 18 insertions(+), 12 deletions(-) (limited to 'modules') diff --git a/modules/interrogate.py b/modules/interrogate.py index 19938cbb..1d1ac572 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -20,6 +20,7 @@ Category = namedtuple("Category", ["name", "topn", "items"]) re_topn = re.compile(r"\.top(\d+)\.") +category_types = ["artists", "flavors", "mediums", "movements"] def download_default_clip_interrogate_categories(content_dir): print("Downloading CLIP categories...") @@ -27,12 +28,8 @@ def download_default_clip_interrogate_categories(content_dir): tmpdir = content_dir + "_tmp" try: os.makedirs(tmpdir) - - torch.hub.download_url_to_file("https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/artists.txt", os.path.join(tmpdir, "artists.txt")) - torch.hub.download_url_to_file("https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/flavors.txt", os.path.join(tmpdir, "flavors.top3.txt")) - torch.hub.download_url_to_file("https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/mediums.txt", os.path.join(tmpdir, "mediums.txt")) - torch.hub.download_url_to_file("https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/movements.txt", os.path.join(tmpdir, "movements.txt")) - + for category_type in category_types: + torch.hub.download_url_to_file(f"https://raw.githubusercontent.com/pharmapsychotic/clip-interrogator/main/clip_interrogator/data/{category_type}.txt", os.path.join(tmpdir, f"{category_type}.txt")) os.rename(tmpdir, content_dir) except Exception as e: @@ -51,12 +48,13 @@ class InterrogateModels: def __init__(self, content_dir): self.loaded_categories = None + self.selected_categories = [] self.content_dir = content_dir self.running_on_cpu = devices.device_interrogate == torch.device("cpu") def categories(self): - if self.loaded_categories is not None: - return self.loaded_categories + if self.loaded_categories is not None and self.selected_categories == shared.opts.interrogate_clip_categories: + return self.loaded_categories self.loaded_categories = [] @@ -64,14 +62,19 @@ class InterrogateModels: download_default_clip_interrogate_categories(self.content_dir) if os.path.exists(self.content_dir): - for filename in os.listdir(self.content_dir): + self.selected_categories = shared.opts.interrogate_clip_categories + for category_type in category_types: + if 'all' not in self.selected_categories and category_type not in self.selected_categories: + continue + filename = os.path.join(self.content_dir, f"{category_type}.txt") + if not os.path.isfile(filename): + continue m = re_topn.search(filename) topn = 1 if m is None else int(m.group(1)) - - with open(os.path.join(self.content_dir, filename), "r", encoding="utf8") as file: + with open(filename, "r", encoding="utf8") as file: lines = [x.strip() for x in file.readlines()] - self.loaded_categories.append(Category(name=filename, topn=topn, items=lines)) + self.loaded_categories.append(Category(name=category_type, topn=topn, items=lines)) return self.loaded_categories @@ -139,6 +142,8 @@ class InterrogateModels: def rank(self, image_features, text_array, top_count=1): import clip + devices.torch_gc() + if shared.opts.interrogate_clip_dict_limit != 0: text_array = text_array[0:int(shared.opts.interrogate_clip_dict_limit)] diff --git a/modules/shared.py b/modules/shared.py index a644c0be..63b236c5 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -424,6 +424,7 @@ options_templates.update(options_section(('interrogate', "Interrogate Options"), "interrogate_clip_min_length": OptionInfo(24, "Interrogate: minimum description length (excluding artists, etc..)", gr.Slider, {"minimum": 1, "maximum": 128, "step": 1}), "interrogate_clip_max_length": OptionInfo(48, "Interrogate: maximum description length", gr.Slider, {"minimum": 1, "maximum": 256, "step": 1}), "interrogate_clip_dict_limit": OptionInfo(1500, "CLIP: maximum number of lines in text file (0 = No limit)"), + "interrogate_clip_categories": OptionInfo(modules.interrogate.category_types, "CLIP: select which categories to inquire", gr.CheckboxGroup, lambda: {"choices": modules.interrogate.category_types}), "interrogate_deepbooru_score_threshold": OptionInfo(0.5, "Interrogate: deepbooru score threshold", gr.Slider, {"minimum": 0, "maximum": 1, "step": 0.01}), "deepbooru_sort_alpha": OptionInfo(True, "Interrogate: deepbooru sort alphabetically"), "deepbooru_use_spaces": OptionInfo(False, "use spaces for tags in deepbooru"), -- cgit v1.2.3 From e8c3d03f7d9966b81458944efb25666b2143153f Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 23 Jan 2023 17:59:58 +0300 Subject: a possible fix for broken image upscaling --- modules/postprocessing.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/postprocessing.py b/modules/postprocessing.py index 8514fea7..09d8e605 100644 --- a/modules/postprocessing.py +++ b/modules/postprocessing.py @@ -67,7 +67,7 @@ def run_postprocessing(extras_mode, image, image_folder, input_dir, output_dir, pp.image.info["postprocessing"] = infotext if save_output: - images.save_image(pp.image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=pp.info, short_filename=True, no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None) + images.save_image(pp.image, path=outpath, basename=basename, seed=None, prompt=None, extension=opts.samples_format, info=infotext, short_filename=True, no_prompt=True, grid=False, pnginfo_section_name="extras", existing_info=existing_pnginfo, forced_filename=None) if extras_mode != 2 or show_extras_results: outputs.append(pp.image) -- cgit v1.2.3 From 6e1b296baf7a2cdc0ee747225f1704bd2d45c118 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Mon, 23 Jan 2023 10:10:59 -0500 Subject: api-image-format --- modules/api/api.py | 34 ++++++++++++++++++++++++---------- 1 file changed, 24 insertions(+), 10 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 5d60fc0a..b1dd14cc 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -22,6 +22,8 @@ from modules.sd_models import checkpoints_list, find_checkpoint_config from modules.realesrgan_model import get_realesrgan_models from modules import devices from typing import List +import piexif +import piexif.helper def upscaler_to_index(name: str): try: @@ -56,18 +58,30 @@ def decode_base64_to_image(encoding): def encode_pil_to_base64(image): with io.BytesIO() as output_bytes: - # Copy any text-only metadata - use_metadata = False - metadata = PngImagePlugin.PngInfo() - for key, value in image.info.items(): - if isinstance(key, str) and isinstance(value, str): - metadata.add_text(key, value) - use_metadata = True + if opts.samples_format.lower() == 'png': + use_metadata = False + metadata = PngImagePlugin.PngInfo() + for key, value in image.info.items(): + if isinstance(key, str) and isinstance(value, str): + metadata.add_text(key, value) + use_metadata = True + image.save(output_bytes, format="PNG", pnginfo=(metadata if use_metadata else None), quality=opts.jpeg_quality) + + elif opts.samples_format.lower() in ("jpg", "jpeg", "webp"): + parameters = image.info.get('parameters', None) + exif_bytes = piexif.dump({ + "Exif": { piexif.ExifIFD.UserComment: piexif.helper.UserComment.dump(parameters or "", encoding="unicode") } + }) + if opts.samples_format.lower() in ("jpg", "jpeg"): + image.save(output_bytes, format="JPEG", exif = exif_bytes, quality=opts.jpeg_quality) + else: + image.save(output_bytes, format="WEBP", exif = exif_bytes, quality=opts.jpeg_quality) + + else: + raise HTTPException(status_code=500, detail="Invalid image format") - image.save( - output_bytes, "PNG", pnginfo=(metadata if use_metadata else None) - ) bytes_data = output_bytes.getvalue() + return base64.b64encode(bytes_data) def api_middleware(app: FastAPI): -- cgit v1.2.3 From 04a561c11c9bf9a00d7f9b50ca3f7962aa59ba6e Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Mon, 23 Jan 2023 12:29:23 -0500 Subject: add option to skip interrogate categories --- modules/interrogate.py | 32 ++++++++++++++++++-------------- modules/shared.py | 2 +- 2 files changed, 19 insertions(+), 15 deletions(-) (limited to 'modules') diff --git a/modules/interrogate.py b/modules/interrogate.py index 1d1ac572..c252b148 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -2,6 +2,7 @@ import os import sys import traceback from collections import namedtuple +from pathlib import Path import re import torch @@ -20,12 +21,16 @@ Category = namedtuple("Category", ["name", "topn", "items"]) re_topn = re.compile(r"\.top(\d+)\.") -category_types = ["artists", "flavors", "mediums", "movements"] +def category_types(): + return [f.stem for f in Path(shared.interrogator.content_dir).glob('*.txt')] + def download_default_clip_interrogate_categories(content_dir): print("Downloading CLIP categories...") tmpdir = content_dir + "_tmp" + category_types = ["artists", "flavors", "mediums", "movements"] + try: os.makedirs(tmpdir) for category_type in category_types: @@ -48,33 +53,32 @@ class InterrogateModels: def __init__(self, content_dir): self.loaded_categories = None - self.selected_categories = [] + self.skip_categories = [] self.content_dir = content_dir self.running_on_cpu = devices.device_interrogate == torch.device("cpu") def categories(self): - if self.loaded_categories is not None and self.selected_categories == shared.opts.interrogate_clip_categories: + if not os.path.exists(self.content_dir): + download_default_clip_interrogate_categories(self.content_dir) + + if self.loaded_categories is not None and self.skip_categories == shared.opts.interrogate_clip_skip_categories: return self.loaded_categories self.loaded_categories = [] - if not os.path.exists(self.content_dir): - download_default_clip_interrogate_categories(self.content_dir) - if os.path.exists(self.content_dir): - self.selected_categories = shared.opts.interrogate_clip_categories - for category_type in category_types: - if 'all' not in self.selected_categories and category_type not in self.selected_categories: - continue - filename = os.path.join(self.content_dir, f"{category_type}.txt") - if not os.path.isfile(filename): + self.skip_categories = shared.opts.interrogate_clip_skip_categories + category_types = [] + for filename in Path(self.content_dir).glob('*.txt'): + category_types.append(filename.stem) + if filename.stem in self.skip_categories: continue - m = re_topn.search(filename) + m = re_topn.search(filename.stem) topn = 1 if m is None else int(m.group(1)) with open(filename, "r", encoding="utf8") as file: lines = [x.strip() for x in file.readlines()] - self.loaded_categories.append(Category(name=category_type, topn=topn, items=lines)) + self.loaded_categories.append(Category(name=filename.stem, topn=topn, items=lines)) return self.loaded_categories diff --git a/modules/shared.py b/modules/shared.py index d7a18f6a..5f713bee 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -424,7 +424,7 @@ options_templates.update(options_section(('interrogate', "Interrogate Options"), "interrogate_clip_min_length": OptionInfo(24, "Interrogate: minimum description length (excluding artists, etc..)", gr.Slider, {"minimum": 1, "maximum": 128, "step": 1}), "interrogate_clip_max_length": OptionInfo(48, "Interrogate: maximum description length", gr.Slider, {"minimum": 1, "maximum": 256, "step": 1}), "interrogate_clip_dict_limit": OptionInfo(1500, "CLIP: maximum number of lines in text file (0 = No limit)"), - "interrogate_clip_categories": OptionInfo(modules.interrogate.category_types, "CLIP: select which categories to inquire", gr.CheckboxGroup, lambda: {"choices": modules.interrogate.category_types}), + "interrogate_clip_skip_categories": OptionInfo([], "CLIP: skip inquire categories", gr.CheckboxGroup, lambda: {"choices": modules.interrogate.category_types()}, refresh=modules.interrogate.category_types), "interrogate_deepbooru_score_threshold": OptionInfo(0.5, "Interrogate: deepbooru score threshold", gr.Slider, {"minimum": 0, "maximum": 1, "step": 0.01}), "deepbooru_sort_alpha": OptionInfo(True, "Interrogate: deepbooru sort alphabetically"), "deepbooru_use_spaces": OptionInfo(False, "use spaces for tags in deepbooru"), -- cgit v1.2.3 From 7b1c7ba87b14da9960d0347269421233f4cb5838 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Mon, 23 Jan 2023 23:11:34 +0300 Subject: add support for apostrophe in extra network names --- modules/ui_extra_networks.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index 2ddac3d8..8b4f97f8 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -3,6 +3,7 @@ import os.path from modules import shared import gradio as gr import json +import html from modules.generation_parameters_copypaste import image_from_url_text @@ -54,12 +55,13 @@ class ExtraNetworksPage: preview = item.get("preview", None) args = { - "preview_html": "style='background-image: url(" + json.dumps(preview) + ")'" if preview else '', + "preview_html": "style='background-image: url(\"" + html.escape(preview) + "\")'" if preview else '', "prompt": item["prompt"], "tabname": json.dumps(tabname), "local_preview": json.dumps(item["local_preview"]), "name": item["name"], - "allow_negative_prompt": "true" if self.allow_negative_prompt else "false", + "card_clicked": '"' + html.escape(f"""return cardClicked({json.dumps(tabname)}, {item["prompt"]}, {"true" if self.allow_negative_prompt else "false"})""") + '"', + "save_card_preview": '"' + html.escape(f"""return saveCardPreview(event, {json.dumps(tabname)}, {json.dumps(item["local_preview"])})""") + '"', } return self.card_page.format(**args) -- cgit v1.2.3 From 5c1cb9263f980641007088a37360fcab01761d37 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 24 Jan 2023 00:24:17 +0300 Subject: fix BLIP failing to import depending on configuration --- modules/interrogate.py | 3 ++- modules/paths.py | 14 ++++++++++++++ 2 files changed, 16 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/interrogate.py b/modules/interrogate.py index c252b148..236e6983 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -83,7 +83,8 @@ class InterrogateModels: return self.loaded_categories def load_blip_model(self): - import models.blip + with paths.Prioritize("BLIP"): + import models.blip files = modelloader.load_models( model_path=os.path.join(paths.models_path, "BLIP"), diff --git a/modules/paths.py b/modules/paths.py index 4dd03a35..20b3e4d8 100644 --- a/modules/paths.py +++ b/modules/paths.py @@ -38,3 +38,17 @@ for d, must_exist, what, options in path_dirs: else: sys.path.append(d) paths[what] = d + + +class Prioritize: + def __init__(self, name): + self.name = name + self.path = None + + def __enter__(self): + self.path = sys.path.copy() + sys.path = [paths[self.name]] + sys.path + + def __exit__(self, exc_type, exc_val, exc_tb): + sys.path = self.path + self.path = None -- cgit v1.2.3 From 45e270dfc853216b2c413f915946f0f2842e57a4 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Mon, 23 Jan 2023 17:11:22 -0500 Subject: add image decod exception handling --- modules/api/api.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index b1dd14cc..e6e31e41 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -53,7 +53,11 @@ def setUpscalers(req: dict): def decode_base64_to_image(encoding): if encoding.startswith("data:image/"): encoding = encoding.split(";")[1].split(",")[1] - return Image.open(BytesIO(base64.b64decode(encoding))) + try: + image = Image.open(BytesIO(base64.b64decode(encoding))) + return image + except Exception as err: + raise HTTPException(status_code=500, detail="Invalid encoded image") def encode_pil_to_base64(image): with io.BytesIO() as output_bytes: -- cgit v1.2.3 From f64af77adcd20fabe00e1e642512db9c6742ed23 Mon Sep 17 00:00:00 2001 From: brkirch Date: Mon, 23 Jan 2023 22:49:20 -0500 Subject: Fix different first gen with Approx NN previews The loading of the model for approx nn live previews can change the internal state of PyTorch, resulting in a different image. This can be avoided by preloading the approx nn model in advance. --- modules/processing.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index bc541e2f..3bd590ba 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -13,7 +13,7 @@ from skimage import exposure from typing import Any, Dict, List, Optional import modules.sd_hijack -from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_parameters_copypaste, script_callbacks, extra_networks +from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_parameters_copypaste, script_callbacks, extra_networks, sd_vae_approx from modules.sd_hijack import model_hijack from modules.shared import opts, cmd_opts, state import modules.shared as shared @@ -568,6 +568,10 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: with devices.autocast(): p.init(p.all_prompts, p.all_seeds, p.all_subseeds) + if shared.opts.live_previews_enable and sd_samplers.approximation_indexes.get(shared.opts.show_progress_type, 0) == 1: + # preload approx nn model before sampling for a more deterministic result + sd_vae_approx.model() + if not p.disable_extra_networks: extra_networks.activate(p, extra_network_data) -- cgit v1.2.3 From 42a70d74771e8920f658e741679768ed145dd76a Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 24 Jan 2023 10:05:45 +0300 Subject: repair sdapi/v1/upscalers returning bogus results --- modules/api/api.py | 16 +++++++++------- modules/api/models.py | 2 +- 2 files changed, 10 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index e6e31e41..da2a5daf 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -375,13 +375,15 @@ class Api: return [{"name": sampler[0], "aliases":sampler[2], "options":sampler[3]} for sampler in sd_samplers.all_samplers] def get_upscalers(self): - upscalers = [] - - for upscaler in shared.sd_upscalers: - u = upscaler.scaler - upscalers.append({"name":u.name, "model_name":u.model_name, "model_path":u.model_path, "model_url":u.model_url}) - - return upscalers + return [ + { + "name": upscaler.name, + "model_name": upscaler.scaler.model_name, + "model_path": upscaler.data_path, + "scale": upscaler.scale, + } + for upscaler in shared.sd_upscalers + ] def get_sd_models(self): return [{"title": x.title, "model_name": x.model_name, "hash": x.shorthash, "sha256": x.sha256, "filename": x.filename, "config": find_checkpoint_config(x)} for x in checkpoints_list.values()] diff --git a/modules/api/models.py b/modules/api/models.py index 1eb1fcf1..e562ab54 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -219,7 +219,7 @@ class UpscalerItem(BaseModel): name: str = Field(title="Name") model_name: Optional[str] = Field(title="Model Name") model_path: Optional[str] = Field(title="Path") - model_url: Optional[str] = Field(title="URL") + scale: Optional[float] = Field(title="Scale") class SDModelItem(BaseModel): title: str = Field(title="Title") -- cgit v1.2.3 From 602a1864b05075ca4283986e6f5c7d5bce864e11 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 24 Jan 2023 10:09:30 +0300 Subject: also return the removed field to sdapi/v1/upscalers because someone might have relied on it existing --- modules/api/api.py | 1 + modules/api/models.py | 1 + 2 files changed, 2 insertions(+) (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index da2a5daf..25c65e57 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -380,6 +380,7 @@ class Api: "name": upscaler.name, "model_name": upscaler.scaler.model_name, "model_path": upscaler.data_path, + "model_url": None, "scale": upscaler.scale, } for upscaler in shared.sd_upscalers diff --git a/modules/api/models.py b/modules/api/models.py index e562ab54..805bd8f7 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -219,6 +219,7 @@ class UpscalerItem(BaseModel): name: str = Field(title="Name") model_name: Optional[str] = Field(title="Model Name") model_path: Optional[str] = Field(title="Path") + model_url: Optional[str] = Field(title="URL") scale: Optional[float] = Field(title="Scale") class SDModelItem(BaseModel): -- cgit v1.2.3 From e46bfa5a9e9b489ae925a9c23880e34fe8d9fffa Mon Sep 17 00:00:00 2001 From: EllangoK Date: Tue, 24 Jan 2023 02:24:32 -0500 Subject: handling sub grids and merging into one --- modules/images.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/images.py b/modules/images.py index 3b1c5f34..0bc3d524 100644 --- a/modules/images.py +++ b/modules/images.py @@ -195,7 +195,7 @@ def draw_grid_annotations(im, width, height, hor_texts, ver_texts): ver_text_heights = [sum([line.size[1] + line_spacing for line in lines]) - line_spacing * len(lines) for lines in ver_texts] - pad_top = max(hor_text_heights) + line_spacing * 2 + pad_top = 0 if sum(hor_text_heights) == 0 else max(hor_text_heights) + line_spacing * 2 result = Image.new("RGB", (im.width + pad_left, im.height + pad_top), "white") result.paste(im, (pad_left, pad_top)) -- cgit v1.2.3 From 28189985e6f56dc725938a3f0e4d2462dad74bc5 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 24 Jan 2023 20:24:27 +0300 Subject: remove fairscale requirement, add fake fairscale to make BLIP not complain about it --- modules/interrogate.py | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/interrogate.py b/modules/interrogate.py index 236e6983..9f063197 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -82,9 +82,16 @@ class InterrogateModels: return self.loaded_categories + def create_fake_fairscale(self): + class FakeFairscale: + def checkpoint_wrapper(self): + pass + + sys.modules["fairscale.nn.checkpoint.checkpoint_activations"] = FakeFairscale + def load_blip_model(self): - with paths.Prioritize("BLIP"): - import models.blip + create_fake_fairscale() + import models.blip files = modelloader.load_models( model_path=os.path.join(paths.models_path, "BLIP"), -- cgit v1.2.3 From 5228ec8bdada50a8d614573e980193ca89192361 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Tue, 24 Jan 2023 20:30:43 +0300 Subject: remove fairscale requirement, add fake fairscale to make BLIP not complain about it mk2 --- modules/interrogate.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/interrogate.py b/modules/interrogate.py index 9f063197..c72ff694 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -90,7 +90,7 @@ class InterrogateModels: sys.modules["fairscale.nn.checkpoint.checkpoint_activations"] = FakeFairscale def load_blip_model(self): - create_fake_fairscale() + self.create_fake_fairscale() import models.blip files = modelloader.load_models( -- cgit v1.2.3 From 84d9ce30cb427759547bc7876ed80ab91787d175 Mon Sep 17 00:00:00 2001 From: brkirch Date: Tue, 24 Jan 2023 23:51:45 -0500 Subject: Add option for float32 sampling with float16 UNet This also handles type casting so that ROCm and MPS torch devices work correctly without --no-half. One cast is required for deepbooru in deepbooru_model.py, some explicit casting is required for img2img and inpainting. depth_model can't be converted to float16 or it won't work correctly on some systems (it's known to have issues on MPS) so in sd_models.py model.depth_model is removed for model.half(). --- modules/deepbooru_model.py | 4 +++- modules/devices.py | 2 ++ modules/processing.py | 15 ++++++++------- modules/sd_hijack_unet.py | 29 +++++++++++++++++++++++++++++ modules/sd_hijack_utils.py | 28 ++++++++++++++++++++++++++++ modules/sd_models.py | 10 ++++++++++ modules/shared.py | 1 + 7 files changed, 81 insertions(+), 8 deletions(-) create mode 100644 modules/sd_hijack_utils.py (limited to 'modules') diff --git a/modules/deepbooru_model.py b/modules/deepbooru_model.py index edd40c81..83d2ff09 100644 --- a/modules/deepbooru_model.py +++ b/modules/deepbooru_model.py @@ -2,6 +2,8 @@ import torch import torch.nn as nn import torch.nn.functional as F +from modules import devices + # see https://github.com/AUTOMATIC1111/TorchDeepDanbooru for more @@ -196,7 +198,7 @@ class DeepDanbooruModel(nn.Module): t_358, = inputs t_359 = t_358.permute(*[0, 3, 1, 2]) t_359_padded = F.pad(t_359, [2, 3, 2, 3], value=0) - t_360 = self.n_Conv_0(t_359_padded) + t_360 = self.n_Conv_0(t_359_padded.to(self.n_Conv_0.bias.dtype) if devices.unet_needs_upcast else t_359_padded) t_361 = F.relu(t_360) t_361 = F.pad(t_361, [0, 1, 0, 1], value=float('-inf')) t_362 = self.n_MaxPool_0(t_361) diff --git a/modules/devices.py b/modules/devices.py index 524ec7af..0981ef80 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -79,6 +79,8 @@ cpu = torch.device("cpu") device = device_interrogate = device_gfpgan = device_esrgan = device_codeformer = None dtype = torch.float16 dtype_vae = torch.float16 +dtype_unet = torch.float16 +unet_needs_upcast = False def randn(seed, shape): diff --git a/modules/processing.py b/modules/processing.py index bc541e2f..2d186ba0 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -172,7 +172,8 @@ class StableDiffusionProcessing: midas_in = torch.from_numpy(transformed["midas_in"][None, ...]).to(device=shared.device) midas_in = repeat(midas_in, "1 ... -> n ...", n=self.batch_size) - conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(source_image)) + conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(source_image.to(devices.dtype_unet) if devices.unet_needs_upcast else source_image)) + conditioning_image = conditioning_image.float() if devices.unet_needs_upcast else conditioning_image conditioning = torch.nn.functional.interpolate( self.sd_model.depth_model(midas_in), size=conditioning_image.shape[2:], @@ -203,7 +204,7 @@ class StableDiffusionProcessing: # Create another latent image, this time with a masked version of the original input. # Smoothly interpolate between the masked and unmasked latent conditioning image using a parameter. - conditioning_mask = conditioning_mask.to(source_image.device).to(source_image.dtype) + conditioning_mask = conditioning_mask.to(device=source_image.device, dtype=source_image.dtype) conditioning_image = torch.lerp( source_image, source_image * (1.0 - conditioning_mask), @@ -211,7 +212,7 @@ class StableDiffusionProcessing: ) # Encode the new masked image using first stage of network. - conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(conditioning_image)) + conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(conditioning_image.to(devices.dtype_unet) if devices.unet_needs_upcast else conditioning_image)) # Create the concatenated conditioning tensor to be fed to `c_concat` conditioning_mask = torch.nn.functional.interpolate(conditioning_mask, size=latent_image.shape[-2:]) @@ -225,10 +226,10 @@ class StableDiffusionProcessing: # HACK: Using introspection as the Depth2Image model doesn't appear to uniquely # identify itself with a field common to all models. The conditioning_key is also hybrid. if isinstance(self.sd_model, LatentDepth2ImageDiffusion): - return self.depth2img_image_conditioning(source_image) + return self.depth2img_image_conditioning(source_image.float() if devices.unet_needs_upcast else source_image) if self.sampler.conditioning_key in {'hybrid', 'concat'}: - return self.inpainting_image_conditioning(source_image, latent_image, image_mask=image_mask) + return self.inpainting_image_conditioning(source_image.float() if devices.unet_needs_upcast else source_image, latent_image, image_mask=image_mask) # Dummy zero conditioning if we're not using inpainting or depth model. return latent_image.new_zeros(latent_image.shape[0], 5, 1, 1) @@ -610,7 +611,7 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if p.n_iter > 1: shared.state.job = f"Batch {n+1} out of {p.n_iter}" - with devices.autocast(): + with devices.autocast(disable=devices.unet_needs_upcast): samples_ddim = p.sample(conditioning=c, unconditional_conditioning=uc, seeds=seeds, subseeds=subseeds, subseed_strength=p.subseed_strength, prompts=prompts) x_samples_ddim = [decode_first_stage(p.sd_model, samples_ddim[i:i+1].to(dtype=devices.dtype_vae))[0].cpu() for i in range(samples_ddim.size(0))] @@ -988,7 +989,7 @@ class StableDiffusionProcessingImg2Img(StableDiffusionProcessing): image = torch.from_numpy(batch_images) image = 2. * image - 1. - image = image.to(shared.device) + image = image.to(device=shared.device, dtype=devices.dtype_unet if devices.unet_needs_upcast else None) self.init_latent = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(image)) diff --git a/modules/sd_hijack_unet.py b/modules/sd_hijack_unet.py index 18daf8c1..88c94e54 100644 --- a/modules/sd_hijack_unet.py +++ b/modules/sd_hijack_unet.py @@ -1,4 +1,8 @@ import torch +from packaging import version + +from modules import devices +from modules.sd_hijack_utils import CondFunc class TorchHijackForUnet: @@ -28,3 +32,28 @@ class TorchHijackForUnet: th = TorchHijackForUnet() + + +# Below are monkey patches to enable upcasting a float16 UNet for float32 sampling +def apply_model(orig_func, self, x_noisy, t, cond, **kwargs): + for y in cond.keys(): + cond[y] = [x.to(devices.dtype_unet) if isinstance(x, torch.Tensor) else x for x in cond[y]] + with devices.autocast(): + return orig_func(self, x_noisy.to(devices.dtype_unet), t.to(devices.dtype_unet), cond, **kwargs).float() + +class GELUHijack(torch.nn.GELU, torch.nn.Module): + def __init__(self, *args, **kwargs): + torch.nn.GELU.__init__(self, *args, **kwargs) + def forward(self, x): + if devices.unet_needs_upcast: + return torch.nn.GELU.forward(self.float(), x.float()).to(devices.dtype_unet) + else: + return torch.nn.GELU.forward(self, x) + +unet_needs_upcast = lambda *args, **kwargs: devices.unet_needs_upcast +CondFunc('ldm.models.diffusion.ddpm.LatentDiffusion.apply_model', apply_model, unet_needs_upcast) +CondFunc('ldm.modules.diffusionmodules.openaimodel.timestep_embedding', lambda orig_func, *args, **kwargs: orig_func(*args, **kwargs).to(devices.dtype_unet), unet_needs_upcast) +if version.parse(torch.__version__) <= version.parse("1.13.1"): + CondFunc('ldm.modules.diffusionmodules.util.GroupNorm32.forward', lambda orig_func, self, *args, **kwargs: orig_func(self.float(), *args, **kwargs), unet_needs_upcast) + CondFunc('ldm.modules.attention.GEGLU.forward', lambda orig_func, self, x: orig_func(self.float(), x.float()).to(devices.dtype_unet), unet_needs_upcast) + CondFunc('open_clip.transformer.ResidualAttentionBlock.__init__', lambda orig_func, *args, **kwargs: kwargs.update({'act_layer': GELUHijack}) and False or orig_func(*args, **kwargs), lambda _, *args, **kwargs: kwargs.get('act_layer') is None or kwargs['act_layer'] == torch.nn.GELU) diff --git a/modules/sd_hijack_utils.py b/modules/sd_hijack_utils.py new file mode 100644 index 00000000..f81b169a --- /dev/null +++ b/modules/sd_hijack_utils.py @@ -0,0 +1,28 @@ +import importlib + +class CondFunc: + def __new__(cls, orig_func, sub_func, cond_func): + self = super(CondFunc, cls).__new__(cls) + if isinstance(orig_func, str): + func_path = orig_func.split('.') + for i in range(len(func_path)-2, -1, -1): + try: + resolved_obj = importlib.import_module('.'.join(func_path[:i])) + break + except ImportError: + pass + for attr_name in func_path[i:-1]: + resolved_obj = getattr(resolved_obj, attr_name) + orig_func = getattr(resolved_obj, func_path[-1]) + setattr(resolved_obj, func_path[-1], lambda *args, **kwargs: self(*args, **kwargs)) + self.__init__(orig_func, sub_func, cond_func) + return lambda *args, **kwargs: self(*args, **kwargs) + def __init__(self, orig_func, sub_func, cond_func): + self.__orig_func = orig_func + self.__sub_func = sub_func + self.__cond_func = cond_func + def __call__(self, *args, **kwargs): + if not self.__cond_func or self.__cond_func(self.__orig_func, *args, **kwargs): + return self.__sub_func(self.__orig_func, *args, **kwargs) + else: + return self.__orig_func(*args, **kwargs) diff --git a/modules/sd_models.py b/modules/sd_models.py index 12083848..7c98991a 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -257,16 +257,24 @@ def load_model_weights(model, checkpoint_info: CheckpointInfo): if not shared.cmd_opts.no_half: vae = model.first_stage_model + depth_model = getattr(model, 'depth_model', None) # with --no-half-vae, remove VAE from model when doing half() to prevent its weights from being converted to float16 if shared.cmd_opts.no_half_vae: model.first_stage_model = None + # with --upcast-sampling, don't convert the depth model weights to float16 + if shared.cmd_opts.upcast_sampling and depth_model: + model.depth_model = None model.half() model.first_stage_model = vae + if depth_model: + model.depth_model = depth_model devices.dtype = torch.float32 if shared.cmd_opts.no_half else torch.float16 devices.dtype_vae = torch.float32 if shared.cmd_opts.no_half or shared.cmd_opts.no_half_vae else torch.float16 + devices.dtype_unet = model.model.diffusion_model.dtype + devices.unet_needs_upcast = shared.cmd_opts.upcast_sampling and devices.dtype == torch.float16 and devices.dtype_unet == torch.float16 model.first_stage_model.to(devices.dtype_vae) @@ -372,6 +380,8 @@ def load_model(checkpoint_info=None): if shared.cmd_opts.no_half: sd_config.model.params.unet_config.params.use_fp16 = False + elif shared.cmd_opts.upcast_sampling: + sd_config.model.params.unet_config.params.use_fp16 = True timer = Timer() diff --git a/modules/shared.py b/modules/shared.py index 5f713bee..4ce1209b 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -45,6 +45,7 @@ parser.add_argument("--lowram", action='store_true', help="load stable diffusion parser.add_argument("--always-batch-cond-uncond", action='store_true', help="disables cond/uncond batching that is enabled to save memory with --medvram or --lowvram") parser.add_argument("--unload-gfpgan", action='store_true', help="does not do anything.") parser.add_argument("--precision", type=str, help="evaluate at this precision", choices=["full", "autocast"], default="autocast") +parser.add_argument("--upcast-sampling", action='store_true', help="upcast sampling. No effect with --no-half. Usually produces similar results to --no-half with better performance while using less memory.") parser.add_argument("--share", action='store_true', help="use share=True for gradio and make the UI accessible through their site") parser.add_argument("--ngrok", type=str, help="ngrok authtoken, alternative to gradio --share", default=None) parser.add_argument("--ngrok-region", type=str, help="The region in which ngrok should start.", default="us") -- cgit v1.2.3 From e3b53fd295aca784253dfc8668ec87b537a72f43 Mon Sep 17 00:00:00 2001 From: brkirch Date: Wed, 25 Jan 2023 00:23:10 -0500 Subject: Add UI setting for upcasting attention to float32 Adds "Upcast cross attention layer to float32" option in Stable Diffusion settings. This allows for generating images using SD 2.1 models without --no-half or xFormers. In order to make upcasting cross attention layer optimizations possible it is necessary to indent several sections of code in sd_hijack_optimizations.py so that a context manager can be used to disable autocast. Also, even though Stable Diffusion (and Diffusers) only upcast q and k, unfortunately my findings were that most of the cross attention layer optimizations could not function unless v is upcast also. --- modules/devices.py | 6 +- modules/processing.py | 2 +- modules/sd_hijack_optimizations.py | 159 +++++++++++++++++++++++-------------- modules/shared.py | 1 + modules/sub_quadratic_attention.py | 4 +- 5 files changed, 108 insertions(+), 64 deletions(-) (limited to 'modules') diff --git a/modules/devices.py b/modules/devices.py index 0981ef80..6b36622c 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -108,6 +108,10 @@ def autocast(disable=False): return torch.autocast("cuda") +def without_autocast(disable=False): + return torch.autocast("cuda", enabled=False) if torch.is_autocast_enabled() and not disable else contextlib.nullcontext() + + class NansException(Exception): pass @@ -125,7 +129,7 @@ def test_for_nans(x, where): message = "A tensor with all NaNs was produced in Unet." if not shared.cmd_opts.no_half: - message += " This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try using --no-half commandline argument to fix this." + message += " This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the \"Upcast cross attention layer to float32\" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this." elif where == "vae": message = "A tensor with all NaNs was produced in VAE." diff --git a/modules/processing.py b/modules/processing.py index 2d186ba0..a850082d 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -611,7 +611,7 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if p.n_iter > 1: shared.state.job = f"Batch {n+1} out of {p.n_iter}" - with devices.autocast(disable=devices.unet_needs_upcast): + with devices.without_autocast() if devices.unet_needs_upcast else devices.autocast(): samples_ddim = p.sample(conditioning=c, unconditional_conditioning=uc, seeds=seeds, subseeds=subseeds, subseed_strength=p.subseed_strength, prompts=prompts) x_samples_ddim = [decode_first_stage(p.sd_model, samples_ddim[i:i+1].to(dtype=devices.dtype_vae))[0].cpu() for i in range(samples_ddim.size(0))] diff --git a/modules/sd_hijack_optimizations.py b/modules/sd_hijack_optimizations.py index 74452709..c02d954c 100644 --- a/modules/sd_hijack_optimizations.py +++ b/modules/sd_hijack_optimizations.py @@ -9,7 +9,7 @@ from torch import einsum from ldm.util import default from einops import rearrange -from modules import shared, errors +from modules import shared, errors, devices from modules.hypernetworks import hypernetwork from .sub_quadratic_attention import efficient_dot_product_attention @@ -52,18 +52,25 @@ def split_cross_attention_forward_v1(self, x, context=None, mask=None): q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> (b h) n d', h=h), (q_in, k_in, v_in)) del q_in, k_in, v_in - r1 = torch.zeros(q.shape[0], q.shape[1], v.shape[2], device=q.device) - for i in range(0, q.shape[0], 2): - end = i + 2 - s1 = einsum('b i d, b j d -> b i j', q[i:end], k[i:end]) - s1 *= self.scale + dtype = q.dtype + if shared.opts.upcast_attn: + q, k, v = q.float(), k.float(), v.float() - s2 = s1.softmax(dim=-1) - del s1 + with devices.without_autocast(disable=not shared.opts.upcast_attn): + r1 = torch.zeros(q.shape[0], q.shape[1], v.shape[2], device=q.device, dtype=q.dtype) + for i in range(0, q.shape[0], 2): + end = i + 2 + s1 = einsum('b i d, b j d -> b i j', q[i:end], k[i:end]) + s1 *= self.scale + + s2 = s1.softmax(dim=-1) + del s1 + + r1[i:end] = einsum('b i j, b j d -> b i d', s2, v[i:end]) + del s2 + del q, k, v - r1[i:end] = einsum('b i j, b j d -> b i d', s2, v[i:end]) - del s2 - del q, k, v + r1 = r1.to(dtype) r2 = rearrange(r1, '(b h) n d -> b n (h d)', h=h) del r1 @@ -82,45 +89,52 @@ def split_cross_attention_forward(self, x, context=None, mask=None): k_in = self.to_k(context_k) v_in = self.to_v(context_v) - k_in *= self.scale - - del context, x - - q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> (b h) n d', h=h), (q_in, k_in, v_in)) - del q_in, k_in, v_in - - r1 = torch.zeros(q.shape[0], q.shape[1], v.shape[2], device=q.device, dtype=q.dtype) - - mem_free_total = get_available_vram() - - gb = 1024 ** 3 - tensor_size = q.shape[0] * q.shape[1] * k.shape[1] * q.element_size() - modifier = 3 if q.element_size() == 2 else 2.5 - mem_required = tensor_size * modifier - steps = 1 - - if mem_required > mem_free_total: - steps = 2 ** (math.ceil(math.log(mem_required / mem_free_total, 2))) - # print(f"Expected tensor size:{tensor_size/gb:0.1f}GB, cuda free:{mem_free_cuda/gb:0.1f}GB " - # f"torch free:{mem_free_torch/gb:0.1f} total:{mem_free_total/gb:0.1f} steps:{steps}") + dtype = q_in.dtype + if shared.opts.upcast_attn: + q_in, k_in, v_in = q_in.float(), k_in.float(), v_in if v_in.device.type == 'mps' else v_in.float() - if steps > 64: - max_res = math.floor(math.sqrt(math.sqrt(mem_free_total / 2.5)) / 8) * 64 - raise RuntimeError(f'Not enough memory, use lower resolution (max approx. {max_res}x{max_res}). ' - f'Need: {mem_required / 64 / gb:0.1f}GB free, Have:{mem_free_total / gb:0.1f}GB free') - - slice_size = q.shape[1] // steps if (q.shape[1] % steps) == 0 else q.shape[1] - for i in range(0, q.shape[1], slice_size): - end = i + slice_size - s1 = einsum('b i d, b j d -> b i j', q[:, i:end], k) - - s2 = s1.softmax(dim=-1, dtype=q.dtype) - del s1 - - r1[:, i:end] = einsum('b i j, b j d -> b i d', s2, v) - del s2 + with devices.without_autocast(disable=not shared.opts.upcast_attn): + k_in = k_in * self.scale + + del context, x + + q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> (b h) n d', h=h), (q_in, k_in, v_in)) + del q_in, k_in, v_in + + r1 = torch.zeros(q.shape[0], q.shape[1], v.shape[2], device=q.device, dtype=q.dtype) + + mem_free_total = get_available_vram() + + gb = 1024 ** 3 + tensor_size = q.shape[0] * q.shape[1] * k.shape[1] * q.element_size() + modifier = 3 if q.element_size() == 2 else 2.5 + mem_required = tensor_size * modifier + steps = 1 + + if mem_required > mem_free_total: + steps = 2 ** (math.ceil(math.log(mem_required / mem_free_total, 2))) + # print(f"Expected tensor size:{tensor_size/gb:0.1f}GB, cuda free:{mem_free_cuda/gb:0.1f}GB " + # f"torch free:{mem_free_torch/gb:0.1f} total:{mem_free_total/gb:0.1f} steps:{steps}") + + if steps > 64: + max_res = math.floor(math.sqrt(math.sqrt(mem_free_total / 2.5)) / 8) * 64 + raise RuntimeError(f'Not enough memory, use lower resolution (max approx. {max_res}x{max_res}). ' + f'Need: {mem_required / 64 / gb:0.1f}GB free, Have:{mem_free_total / gb:0.1f}GB free') + + slice_size = q.shape[1] // steps if (q.shape[1] % steps) == 0 else q.shape[1] + for i in range(0, q.shape[1], slice_size): + end = i + slice_size + s1 = einsum('b i d, b j d -> b i j', q[:, i:end], k) + + s2 = s1.softmax(dim=-1, dtype=q.dtype) + del s1 + + r1[:, i:end] = einsum('b i j, b j d -> b i d', s2, v) + del s2 + + del q, k, v - del q, k, v + r1 = r1.to(dtype) r2 = rearrange(r1, '(b h) n d -> b n (h d)', h=h) del r1 @@ -204,12 +218,20 @@ def split_cross_attention_forward_invokeAI(self, x, context=None, mask=None): context = default(context, x) context_k, context_v = hypernetwork.apply_hypernetworks(shared.loaded_hypernetworks, context) - k = self.to_k(context_k) * self.scale + k = self.to_k(context_k) v = self.to_v(context_v) del context, context_k, context_v, x - q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> (b h) n d', h=h), (q, k, v)) - r = einsum_op(q, k, v) + dtype = q.dtype + if shared.opts.upcast_attn: + q, k, v = q.float(), k.float(), v if v.device.type == 'mps' else v.float() + + with devices.without_autocast(disable=not shared.opts.upcast_attn): + k = k * self.scale + + q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> (b h) n d', h=h), (q, k, v)) + r = einsum_op(q, k, v) + r = r.to(dtype) return self.to_out(rearrange(r, '(b h) n d -> b n (h d)', h=h)) # -- End of code from https://github.com/invoke-ai/InvokeAI -- @@ -234,8 +256,14 @@ def sub_quad_attention_forward(self, x, context=None, mask=None): k = k.unflatten(-1, (h, -1)).transpose(1,2).flatten(end_dim=1) v = v.unflatten(-1, (h, -1)).transpose(1,2).flatten(end_dim=1) + dtype = q.dtype + if shared.opts.upcast_attn: + q, k = q.float(), k.float() + x = sub_quad_attention(q, k, v, q_chunk_size=shared.cmd_opts.sub_quad_q_chunk_size, kv_chunk_size=shared.cmd_opts.sub_quad_kv_chunk_size, chunk_threshold=shared.cmd_opts.sub_quad_chunk_threshold, use_checkpoint=self.training) + x = x.to(dtype) + x = x.unflatten(0, (-1, h)).transpose(1,2).flatten(start_dim=2) out_proj, dropout = self.to_out @@ -268,15 +296,16 @@ def sub_quad_attention(q, k, v, q_chunk_size=1024, kv_chunk_size=None, kv_chunk_ query_chunk_size = q_tokens kv_chunk_size = k_tokens - return efficient_dot_product_attention( - q, - k, - v, - query_chunk_size=q_chunk_size, - kv_chunk_size=kv_chunk_size, - kv_chunk_size_min = kv_chunk_size_min, - use_checkpoint=use_checkpoint, - ) + with devices.without_autocast(disable=q.dtype == v.dtype): + return efficient_dot_product_attention( + q, + k, + v, + query_chunk_size=q_chunk_size, + kv_chunk_size=kv_chunk_size, + kv_chunk_size_min = kv_chunk_size_min, + use_checkpoint=use_checkpoint, + ) def get_xformers_flash_attention_op(q, k, v): @@ -306,8 +335,14 @@ def xformers_attention_forward(self, x, context=None, mask=None): q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> b n h d', h=h), (q_in, k_in, v_in)) del q_in, k_in, v_in + dtype = q.dtype + if shared.opts.upcast_attn: + q, k = q.float(), k.float() + out = xformers.ops.memory_efficient_attention(q, k, v, attn_bias=None, op=get_xformers_flash_attention_op(q, k, v)) + out = out.to(dtype) + out = rearrange(out, 'b n h d -> b n (h d)', h=h) return self.to_out(out) @@ -378,10 +413,14 @@ def xformers_attnblock_forward(self, x): v = self.v(h_) b, c, h, w = q.shape q, k, v = map(lambda t: rearrange(t, 'b c h w -> b (h w) c'), (q, k, v)) + dtype = q.dtype + if shared.opts.upcast_attn: + q, k = q.float(), k.float() q = q.contiguous() k = k.contiguous() v = v.contiguous() out = xformers.ops.memory_efficient_attention(q, k, v, op=get_xformers_flash_attention_op(q, k, v)) + out = out.to(dtype) out = rearrange(out, 'b (h w) c -> b c h w', h=h) out = self.proj_out(out) return x + out diff --git a/modules/shared.py b/modules/shared.py index 4ce1209b..6a0b96cb 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -410,6 +410,7 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "comma_padding_backtrack": OptionInfo(20, "Increase coherency by padding from the last comma within n tokens when using more than 75 tokens", gr.Slider, {"minimum": 0, "maximum": 74, "step": 1 }), "CLIP_stop_at_last_layers": OptionInfo(1, "Clip skip", gr.Slider, {"minimum": 1, "maximum": 12, "step": 1}), "extra_networks_default_multiplier": OptionInfo(1.0, "Multiplier for extra networks", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), + "upcast_attn": OptionInfo(False, "Upcast cross attention layer to float32"), })) options_templates.update(options_section(('compatibility', "Compatibility"), { diff --git a/modules/sub_quadratic_attention.py b/modules/sub_quadratic_attention.py index 55052815..05595323 100644 --- a/modules/sub_quadratic_attention.py +++ b/modules/sub_quadratic_attention.py @@ -67,7 +67,7 @@ def _summarize_chunk( max_score, _ = torch.max(attn_weights, -1, keepdim=True) max_score = max_score.detach() exp_weights = torch.exp(attn_weights - max_score) - exp_values = torch.bmm(exp_weights, value) + exp_values = torch.bmm(exp_weights, value) if query.device.type == 'mps' else torch.bmm(exp_weights, value.to(exp_weights.dtype)).to(value.dtype) max_score = max_score.squeeze(-1) return AttnChunk(exp_values, exp_weights.sum(dim=-1), max_score) @@ -129,7 +129,7 @@ def _get_attention_scores_no_kv_chunking( ) attn_probs = attn_scores.softmax(dim=-1) del attn_scores - hidden_states_slice = torch.bmm(attn_probs, value) + hidden_states_slice = torch.bmm(attn_probs, value) if query.device.type == 'mps' else torch.bmm(attn_probs, value.to(attn_probs.dtype)).to(value.dtype) return hidden_states_slice -- cgit v1.2.3 From ee0a0da3244123cb6d2ba4097a54a1e9caccb687 Mon Sep 17 00:00:00 2001 From: Kyle Date: Wed, 25 Jan 2023 08:53:23 -0500 Subject: Add instruct-pix2pix hijack Allows loading instruct-pix2pix models via same method as inpainting models in sd_models.py and sd_hijack_ip2p.py Adds ddpm_edit.py necessary for instruct-pix2pix --- modules/models/diffusion/ddpm_edit.py | 1459 +++++++++++++++++++++++++++++++++ modules/sd_hijack_ip2p.py | 13 + modules/sd_models.py | 12 +- 3 files changed, 1483 insertions(+), 1 deletion(-) create mode 100644 modules/models/diffusion/ddpm_edit.py create mode 100644 modules/sd_hijack_ip2p.py (limited to 'modules') diff --git a/modules/models/diffusion/ddpm_edit.py b/modules/models/diffusion/ddpm_edit.py new file mode 100644 index 00000000..f3d49c44 --- /dev/null +++ b/modules/models/diffusion/ddpm_edit.py @@ -0,0 +1,1459 @@ +""" +wild mixture of +https://github.com/lucidrains/denoising-diffusion-pytorch/blob/7706bdfc6f527f58d33f84b7b522e61e6e3164b3/denoising_diffusion_pytorch/denoising_diffusion_pytorch.py +https://github.com/openai/improved-diffusion/blob/e94489283bb876ac1477d5dd7709bbbd2d9902ce/improved_diffusion/gaussian_diffusion.py +https://github.com/CompVis/taming-transformers +-- merci +""" + +# File modified by authors of InstructPix2Pix from original (https://github.com/CompVis/stable-diffusion). +# See more details in LICENSE. + +import torch +import torch.nn as nn +import numpy as np +import pytorch_lightning as pl +from torch.optim.lr_scheduler import LambdaLR +from einops import rearrange, repeat +from contextlib import contextmanager +from functools import partial +from tqdm import tqdm +from torchvision.utils import make_grid +from pytorch_lightning.utilities.distributed import rank_zero_only + +from ldm.util import log_txt_as_img, exists, default, ismap, isimage, mean_flat, count_params, instantiate_from_config +from ldm.modules.ema import LitEma +from ldm.modules.distributions.distributions import normal_kl, DiagonalGaussianDistribution +from ldm.models.autoencoder import VQModelInterface, IdentityFirstStage, AutoencoderKL +from ldm.modules.diffusionmodules.util import make_beta_schedule, extract_into_tensor, noise_like +from ldm.models.diffusion.ddim import DDIMSampler + + +__conditioning_keys__ = {'concat': 'c_concat', + 'crossattn': 'c_crossattn', + 'adm': 'y'} + + +def disabled_train(self, mode=True): + """Overwrite model.train with this function to make sure train/eval mode + does not change anymore.""" + return self + + +def uniform_on_device(r1, r2, shape, device): + return (r1 - r2) * torch.rand(*shape, device=device) + r2 + + +class DDPM(pl.LightningModule): + # classic DDPM with Gaussian diffusion, in image space + def __init__(self, + unet_config, + timesteps=1000, + beta_schedule="linear", + loss_type="l2", + ckpt_path=None, + ignore_keys=[], + load_only_unet=False, + monitor="val/loss", + use_ema=True, + first_stage_key="image", + image_size=256, + channels=3, + log_every_t=100, + clip_denoised=True, + linear_start=1e-4, + linear_end=2e-2, + cosine_s=8e-3, + given_betas=None, + original_elbo_weight=0., + v_posterior=0., # weight for choosing posterior variance as sigma = (1-v) * beta_tilde + v * beta + l_simple_weight=1., + conditioning_key=None, + parameterization="eps", # all assuming fixed variance schedules + scheduler_config=None, + use_positional_encodings=False, + learn_logvar=False, + logvar_init=0., + load_ema=True, + ): + super().__init__() + assert parameterization in ["eps", "x0"], 'currently only supporting "eps" and "x0"' + self.parameterization = parameterization + print(f"{self.__class__.__name__}: Running in {self.parameterization}-prediction mode") + self.cond_stage_model = None + self.clip_denoised = clip_denoised + self.log_every_t = log_every_t + self.first_stage_key = first_stage_key + self.image_size = image_size # try conv? + self.channels = channels + self.use_positional_encodings = use_positional_encodings + self.model = DiffusionWrapper(unet_config, conditioning_key) + count_params(self.model, verbose=True) + self.use_ema = use_ema + + self.use_scheduler = scheduler_config is not None + if self.use_scheduler: + self.scheduler_config = scheduler_config + + self.v_posterior = v_posterior + self.original_elbo_weight = original_elbo_weight + self.l_simple_weight = l_simple_weight + + if monitor is not None: + self.monitor = monitor + + if self.use_ema and load_ema: + self.model_ema = LitEma(self.model) + print(f"Keeping EMAs of {len(list(self.model_ema.buffers()))}.") + + if ckpt_path is not None: + self.init_from_ckpt(ckpt_path, ignore_keys=ignore_keys, only_model=load_only_unet) + + # If initialing from EMA-only checkpoint, create EMA model after loading. + if self.use_ema and not load_ema: + self.model_ema = LitEma(self.model) + print(f"Keeping EMAs of {len(list(self.model_ema.buffers()))}.") + + self.register_schedule(given_betas=given_betas, beta_schedule=beta_schedule, timesteps=timesteps, + linear_start=linear_start, linear_end=linear_end, cosine_s=cosine_s) + + self.loss_type = loss_type + + self.learn_logvar = learn_logvar + self.logvar = torch.full(fill_value=logvar_init, size=(self.num_timesteps,)) + if self.learn_logvar: + self.logvar = nn.Parameter(self.logvar, requires_grad=True) + + + def register_schedule(self, given_betas=None, beta_schedule="linear", timesteps=1000, + linear_start=1e-4, linear_end=2e-2, cosine_s=8e-3): + if exists(given_betas): + betas = given_betas + else: + betas = make_beta_schedule(beta_schedule, timesteps, linear_start=linear_start, linear_end=linear_end, + cosine_s=cosine_s) + alphas = 1. - betas + alphas_cumprod = np.cumprod(alphas, axis=0) + alphas_cumprod_prev = np.append(1., alphas_cumprod[:-1]) + + timesteps, = betas.shape + self.num_timesteps = int(timesteps) + self.linear_start = linear_start + self.linear_end = linear_end + assert alphas_cumprod.shape[0] == self.num_timesteps, 'alphas have to be defined for each timestep' + + to_torch = partial(torch.tensor, dtype=torch.float32) + + self.register_buffer('betas', to_torch(betas)) + self.register_buffer('alphas_cumprod', to_torch(alphas_cumprod)) + self.register_buffer('alphas_cumprod_prev', to_torch(alphas_cumprod_prev)) + + # calculations for diffusion q(x_t | x_{t-1}) and others + self.register_buffer('sqrt_alphas_cumprod', to_torch(np.sqrt(alphas_cumprod))) + self.register_buffer('sqrt_one_minus_alphas_cumprod', to_torch(np.sqrt(1. - alphas_cumprod))) + self.register_buffer('log_one_minus_alphas_cumprod', to_torch(np.log(1. - alphas_cumprod))) + self.register_buffer('sqrt_recip_alphas_cumprod', to_torch(np.sqrt(1. / alphas_cumprod))) + self.register_buffer('sqrt_recipm1_alphas_cumprod', to_torch(np.sqrt(1. / alphas_cumprod - 1))) + + # calculations for posterior q(x_{t-1} | x_t, x_0) + posterior_variance = (1 - self.v_posterior) * betas * (1. - alphas_cumprod_prev) / ( + 1. - alphas_cumprod) + self.v_posterior * betas + # above: equal to 1. / (1. / (1. - alpha_cumprod_tm1) + alpha_t / beta_t) + self.register_buffer('posterior_variance', to_torch(posterior_variance)) + # below: log calculation clipped because the posterior variance is 0 at the beginning of the diffusion chain + self.register_buffer('posterior_log_variance_clipped', to_torch(np.log(np.maximum(posterior_variance, 1e-20)))) + self.register_buffer('posterior_mean_coef1', to_torch( + betas * np.sqrt(alphas_cumprod_prev) / (1. - alphas_cumprod))) + self.register_buffer('posterior_mean_coef2', to_torch( + (1. - alphas_cumprod_prev) * np.sqrt(alphas) / (1. - alphas_cumprod))) + + if self.parameterization == "eps": + lvlb_weights = self.betas ** 2 / ( + 2 * self.posterior_variance * to_torch(alphas) * (1 - self.alphas_cumprod)) + elif self.parameterization == "x0": + lvlb_weights = 0.5 * np.sqrt(torch.Tensor(alphas_cumprod)) / (2. * 1 - torch.Tensor(alphas_cumprod)) + else: + raise NotImplementedError("mu not supported") + # TODO how to choose this term + lvlb_weights[0] = lvlb_weights[1] + self.register_buffer('lvlb_weights', lvlb_weights, persistent=False) + assert not torch.isnan(self.lvlb_weights).all() + + @contextmanager + def ema_scope(self, context=None): + if self.use_ema: + self.model_ema.store(self.model.parameters()) + self.model_ema.copy_to(self.model) + if context is not None: + print(f"{context}: Switched to EMA weights") + try: + yield None + finally: + if self.use_ema: + self.model_ema.restore(self.model.parameters()) + if context is not None: + print(f"{context}: Restored training weights") + + def init_from_ckpt(self, path, ignore_keys=list(), only_model=False): + sd = torch.load(path, map_location="cpu") + if "state_dict" in list(sd.keys()): + sd = sd["state_dict"] + keys = list(sd.keys()) + + # Our model adds additional channels to the first layer to condition on an input image. + # For the first layer, copy existing channel weights and initialize new channel weights to zero. + input_keys = [ + "model.diffusion_model.input_blocks.0.0.weight", + "model_ema.diffusion_modelinput_blocks00weight", + ] + + self_sd = self.state_dict() + for input_key in input_keys: + if input_key not in sd or input_key not in self_sd: + continue + + input_weight = self_sd[input_key] + + if input_weight.size() != sd[input_key].size(): + print(f"Manual init: {input_key}") + input_weight.zero_() + input_weight[:, :4, :, :].copy_(sd[input_key]) + ignore_keys.append(input_key) + + for k in keys: + for ik in ignore_keys: + if k.startswith(ik): + print("Deleting key {} from state_dict.".format(k)) + del sd[k] + missing, unexpected = self.load_state_dict(sd, strict=False) if not only_model else self.model.load_state_dict( + sd, strict=False) + print(f"Restored from {path} with {len(missing)} missing and {len(unexpected)} unexpected keys") + if len(missing) > 0: + print(f"Missing Keys: {missing}") + if len(unexpected) > 0: + print(f"Unexpected Keys: {unexpected}") + + def q_mean_variance(self, x_start, t): + """ + Get the distribution q(x_t | x_0). + :param x_start: the [N x C x ...] tensor of noiseless inputs. + :param t: the number of diffusion steps (minus 1). Here, 0 means one step. + :return: A tuple (mean, variance, log_variance), all of x_start's shape. + """ + mean = (extract_into_tensor(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start) + variance = extract_into_tensor(1.0 - self.alphas_cumprod, t, x_start.shape) + log_variance = extract_into_tensor(self.log_one_minus_alphas_cumprod, t, x_start.shape) + return mean, variance, log_variance + + def predict_start_from_noise(self, x_t, t, noise): + return ( + extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x_t.shape) * x_t - + extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x_t.shape) * noise + ) + + def q_posterior(self, x_start, x_t, t): + posterior_mean = ( + extract_into_tensor(self.posterior_mean_coef1, t, x_t.shape) * x_start + + extract_into_tensor(self.posterior_mean_coef2, t, x_t.shape) * x_t + ) + posterior_variance = extract_into_tensor(self.posterior_variance, t, x_t.shape) + posterior_log_variance_clipped = extract_into_tensor(self.posterior_log_variance_clipped, t, x_t.shape) + return posterior_mean, posterior_variance, posterior_log_variance_clipped + + def p_mean_variance(self, x, t, clip_denoised: bool): + model_out = self.model(x, t) + if self.parameterization == "eps": + x_recon = self.predict_start_from_noise(x, t=t, noise=model_out) + elif self.parameterization == "x0": + x_recon = model_out + if clip_denoised: + x_recon.clamp_(-1., 1.) + + model_mean, posterior_variance, posterior_log_variance = self.q_posterior(x_start=x_recon, x_t=x, t=t) + return model_mean, posterior_variance, posterior_log_variance + + @torch.no_grad() + def p_sample(self, x, t, clip_denoised=True, repeat_noise=False): + b, *_, device = *x.shape, x.device + model_mean, _, model_log_variance = self.p_mean_variance(x=x, t=t, clip_denoised=clip_denoised) + noise = noise_like(x.shape, device, repeat_noise) + # no noise when t == 0 + nonzero_mask = (1 - (t == 0).float()).reshape(b, *((1,) * (len(x.shape) - 1))) + return model_mean + nonzero_mask * (0.5 * model_log_variance).exp() * noise + + @torch.no_grad() + def p_sample_loop(self, shape, return_intermediates=False): + device = self.betas.device + b = shape[0] + img = torch.randn(shape, device=device) + intermediates = [img] + for i in tqdm(reversed(range(0, self.num_timesteps)), desc='Sampling t', total=self.num_timesteps): + img = self.p_sample(img, torch.full((b,), i, device=device, dtype=torch.long), + clip_denoised=self.clip_denoised) + if i % self.log_every_t == 0 or i == self.num_timesteps - 1: + intermediates.append(img) + if return_intermediates: + return img, intermediates + return img + + @torch.no_grad() + def sample(self, batch_size=16, return_intermediates=False): + image_size = self.image_size + channels = self.channels + return self.p_sample_loop((batch_size, channels, image_size, image_size), + return_intermediates=return_intermediates) + + def q_sample(self, x_start, t, noise=None): + noise = default(noise, lambda: torch.randn_like(x_start)) + return (extract_into_tensor(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start + + extract_into_tensor(self.sqrt_one_minus_alphas_cumprod, t, x_start.shape) * noise) + + def get_loss(self, pred, target, mean=True): + if self.loss_type == 'l1': + loss = (target - pred).abs() + if mean: + loss = loss.mean() + elif self.loss_type == 'l2': + if mean: + loss = torch.nn.functional.mse_loss(target, pred) + else: + loss = torch.nn.functional.mse_loss(target, pred, reduction='none') + else: + raise NotImplementedError("unknown loss type '{loss_type}'") + + return loss + + def p_losses(self, x_start, t, noise=None): + noise = default(noise, lambda: torch.randn_like(x_start)) + x_noisy = self.q_sample(x_start=x_start, t=t, noise=noise) + model_out = self.model(x_noisy, t) + + loss_dict = {} + if self.parameterization == "eps": + target = noise + elif self.parameterization == "x0": + target = x_start + else: + raise NotImplementedError(f"Paramterization {self.parameterization} not yet supported") + + loss = self.get_loss(model_out, target, mean=False).mean(dim=[1, 2, 3]) + + log_prefix = 'train' if self.training else 'val' + + loss_dict.update({f'{log_prefix}/loss_simple': loss.mean()}) + loss_simple = loss.mean() * self.l_simple_weight + + loss_vlb = (self.lvlb_weights[t] * loss).mean() + loss_dict.update({f'{log_prefix}/loss_vlb': loss_vlb}) + + loss = loss_simple + self.original_elbo_weight * loss_vlb + + loss_dict.update({f'{log_prefix}/loss': loss}) + + return loss, loss_dict + + def forward(self, x, *args, **kwargs): + # b, c, h, w, device, img_size, = *x.shape, x.device, self.image_size + # assert h == img_size and w == img_size, f'height and width of image must be {img_size}' + t = torch.randint(0, self.num_timesteps, (x.shape[0],), device=self.device).long() + return self.p_losses(x, t, *args, **kwargs) + + def get_input(self, batch, k): + return batch[k] + + def shared_step(self, batch): + x = self.get_input(batch, self.first_stage_key) + loss, loss_dict = self(x) + return loss, loss_dict + + def training_step(self, batch, batch_idx): + loss, loss_dict = self.shared_step(batch) + + self.log_dict(loss_dict, prog_bar=True, + logger=True, on_step=True, on_epoch=True) + + self.log("global_step", self.global_step, + prog_bar=True, logger=True, on_step=True, on_epoch=False) + + if self.use_scheduler: + lr = self.optimizers().param_groups[0]['lr'] + self.log('lr_abs', lr, prog_bar=True, logger=True, on_step=True, on_epoch=False) + + return loss + + @torch.no_grad() + def validation_step(self, batch, batch_idx): + _, loss_dict_no_ema = self.shared_step(batch) + with self.ema_scope(): + _, loss_dict_ema = self.shared_step(batch) + loss_dict_ema = {key + '_ema': loss_dict_ema[key] for key in loss_dict_ema} + self.log_dict(loss_dict_no_ema, prog_bar=False, logger=True, on_step=False, on_epoch=True) + self.log_dict(loss_dict_ema, prog_bar=False, logger=True, on_step=False, on_epoch=True) + + def on_train_batch_end(self, *args, **kwargs): + if self.use_ema: + self.model_ema(self.model) + + def _get_rows_from_list(self, samples): + n_imgs_per_row = len(samples) + denoise_grid = rearrange(samples, 'n b c h w -> b n c h w') + denoise_grid = rearrange(denoise_grid, 'b n c h w -> (b n) c h w') + denoise_grid = make_grid(denoise_grid, nrow=n_imgs_per_row) + return denoise_grid + + @torch.no_grad() + def log_images(self, batch, N=8, n_row=2, sample=True, return_keys=None, **kwargs): + log = dict() + x = self.get_input(batch, self.first_stage_key) + N = min(x.shape[0], N) + n_row = min(x.shape[0], n_row) + x = x.to(self.device)[:N] + log["inputs"] = x + + # get diffusion row + diffusion_row = list() + x_start = x[:n_row] + + for t in range(self.num_timesteps): + if t % self.log_every_t == 0 or t == self.num_timesteps - 1: + t = repeat(torch.tensor([t]), '1 -> b', b=n_row) + t = t.to(self.device).long() + noise = torch.randn_like(x_start) + x_noisy = self.q_sample(x_start=x_start, t=t, noise=noise) + diffusion_row.append(x_noisy) + + log["diffusion_row"] = self._get_rows_from_list(diffusion_row) + + if sample: + # get denoise row + with self.ema_scope("Plotting"): + samples, denoise_row = self.sample(batch_size=N, return_intermediates=True) + + log["samples"] = samples + log["denoise_row"] = self._get_rows_from_list(denoise_row) + + if return_keys: + if np.intersect1d(list(log.keys()), return_keys).shape[0] == 0: + return log + else: + return {key: log[key] for key in return_keys} + return log + + def configure_optimizers(self): + lr = self.learning_rate + params = list(self.model.parameters()) + if self.learn_logvar: + params = params + [self.logvar] + opt = torch.optim.AdamW(params, lr=lr) + return opt + + +class LatentDiffusion(DDPM): + """main class""" + def __init__(self, + first_stage_config, + cond_stage_config, + num_timesteps_cond=None, + cond_stage_key="image", + cond_stage_trainable=False, + concat_mode=True, + cond_stage_forward=None, + conditioning_key=None, + scale_factor=1.0, + scale_by_std=False, + load_ema=True, + *args, **kwargs): + self.num_timesteps_cond = default(num_timesteps_cond, 1) + self.scale_by_std = scale_by_std + assert self.num_timesteps_cond <= kwargs['timesteps'] + # for backwards compatibility after implementation of DiffusionWrapper + if conditioning_key is None: + conditioning_key = 'concat' if concat_mode else 'crossattn' + if cond_stage_config == '__is_unconditional__': + conditioning_key = None + ckpt_path = kwargs.pop("ckpt_path", None) + ignore_keys = kwargs.pop("ignore_keys", []) + super().__init__(conditioning_key=conditioning_key, *args, load_ema=load_ema, **kwargs) + self.concat_mode = concat_mode + self.cond_stage_trainable = cond_stage_trainable + self.cond_stage_key = cond_stage_key + try: + self.num_downs = len(first_stage_config.params.ddconfig.ch_mult) - 1 + except: + self.num_downs = 0 + if not scale_by_std: + self.scale_factor = scale_factor + else: + self.register_buffer('scale_factor', torch.tensor(scale_factor)) + self.instantiate_first_stage(first_stage_config) + self.instantiate_cond_stage(cond_stage_config) + self.cond_stage_forward = cond_stage_forward + self.clip_denoised = False + self.bbox_tokenizer = None + + self.restarted_from_ckpt = False + if ckpt_path is not None: + self.init_from_ckpt(ckpt_path, ignore_keys) + self.restarted_from_ckpt = True + + if self.use_ema and not load_ema: + self.model_ema = LitEma(self.model) + print(f"Keeping EMAs of {len(list(self.model_ema.buffers()))}.") + + def make_cond_schedule(self, ): + self.cond_ids = torch.full(size=(self.num_timesteps,), fill_value=self.num_timesteps - 1, dtype=torch.long) + ids = torch.round(torch.linspace(0, self.num_timesteps - 1, self.num_timesteps_cond)).long() + self.cond_ids[:self.num_timesteps_cond] = ids + + @rank_zero_only + @torch.no_grad() + def on_train_batch_start(self, batch, batch_idx, dataloader_idx): + # only for very first batch + if self.scale_by_std and self.current_epoch == 0 and self.global_step == 0 and batch_idx == 0 and not self.restarted_from_ckpt: + assert self.scale_factor == 1., 'rather not use custom rescaling and std-rescaling simultaneously' + # set rescale weight to 1./std of encodings + print("### USING STD-RESCALING ###") + x = super().get_input(batch, self.first_stage_key) + x = x.to(self.device) + encoder_posterior = self.encode_first_stage(x) + z = self.get_first_stage_encoding(encoder_posterior).detach() + del self.scale_factor + self.register_buffer('scale_factor', 1. / z.flatten().std()) + print(f"setting self.scale_factor to {self.scale_factor}") + print("### USING STD-RESCALING ###") + + def register_schedule(self, + given_betas=None, beta_schedule="linear", timesteps=1000, + linear_start=1e-4, linear_end=2e-2, cosine_s=8e-3): + super().register_schedule(given_betas, beta_schedule, timesteps, linear_start, linear_end, cosine_s) + + self.shorten_cond_schedule = self.num_timesteps_cond > 1 + if self.shorten_cond_schedule: + self.make_cond_schedule() + + def instantiate_first_stage(self, config): + model = instantiate_from_config(config) + self.first_stage_model = model.eval() + self.first_stage_model.train = disabled_train + for param in self.first_stage_model.parameters(): + param.requires_grad = False + + def instantiate_cond_stage(self, config): + if not self.cond_stage_trainable: + if config == "__is_first_stage__": + print("Using first stage also as cond stage.") + self.cond_stage_model = self.first_stage_model + elif config == "__is_unconditional__": + print(f"Training {self.__class__.__name__} as an unconditional model.") + self.cond_stage_model = None + # self.be_unconditional = True + else: + model = instantiate_from_config(config) + self.cond_stage_model = model.eval() + self.cond_stage_model.train = disabled_train + for param in self.cond_stage_model.parameters(): + param.requires_grad = False + else: + assert config != '__is_first_stage__' + assert config != '__is_unconditional__' + model = instantiate_from_config(config) + self.cond_stage_model = model + + def _get_denoise_row_from_list(self, samples, desc='', force_no_decoder_quantization=False): + denoise_row = [] + for zd in tqdm(samples, desc=desc): + denoise_row.append(self.decode_first_stage(zd.to(self.device), + force_not_quantize=force_no_decoder_quantization)) + n_imgs_per_row = len(denoise_row) + denoise_row = torch.stack(denoise_row) # n_log_step, n_row, C, H, W + denoise_grid = rearrange(denoise_row, 'n b c h w -> b n c h w') + denoise_grid = rearrange(denoise_grid, 'b n c h w -> (b n) c h w') + denoise_grid = make_grid(denoise_grid, nrow=n_imgs_per_row) + return denoise_grid + + def get_first_stage_encoding(self, encoder_posterior): + if isinstance(encoder_posterior, DiagonalGaussianDistribution): + z = encoder_posterior.sample() + elif isinstance(encoder_posterior, torch.Tensor): + z = encoder_posterior + else: + raise NotImplementedError(f"encoder_posterior of type '{type(encoder_posterior)}' not yet implemented") + return self.scale_factor * z + + def get_learned_conditioning(self, c): + if self.cond_stage_forward is None: + if hasattr(self.cond_stage_model, 'encode') and callable(self.cond_stage_model.encode): + c = self.cond_stage_model.encode(c) + if isinstance(c, DiagonalGaussianDistribution): + c = c.mode() + else: + c = self.cond_stage_model(c) + else: + assert hasattr(self.cond_stage_model, self.cond_stage_forward) + c = getattr(self.cond_stage_model, self.cond_stage_forward)(c) + return c + + def meshgrid(self, h, w): + y = torch.arange(0, h).view(h, 1, 1).repeat(1, w, 1) + x = torch.arange(0, w).view(1, w, 1).repeat(h, 1, 1) + + arr = torch.cat([y, x], dim=-1) + return arr + + def delta_border(self, h, w): + """ + :param h: height + :param w: width + :return: normalized distance to image border, + wtith min distance = 0 at border and max dist = 0.5 at image center + """ + lower_right_corner = torch.tensor([h - 1, w - 1]).view(1, 1, 2) + arr = self.meshgrid(h, w) / lower_right_corner + dist_left_up = torch.min(arr, dim=-1, keepdims=True)[0] + dist_right_down = torch.min(1 - arr, dim=-1, keepdims=True)[0] + edge_dist = torch.min(torch.cat([dist_left_up, dist_right_down], dim=-1), dim=-1)[0] + return edge_dist + + def get_weighting(self, h, w, Ly, Lx, device): + weighting = self.delta_border(h, w) + weighting = torch.clip(weighting, self.split_input_params["clip_min_weight"], + self.split_input_params["clip_max_weight"], ) + weighting = weighting.view(1, h * w, 1).repeat(1, 1, Ly * Lx).to(device) + + if self.split_input_params["tie_braker"]: + L_weighting = self.delta_border(Ly, Lx) + L_weighting = torch.clip(L_weighting, + self.split_input_params["clip_min_tie_weight"], + self.split_input_params["clip_max_tie_weight"]) + + L_weighting = L_weighting.view(1, 1, Ly * Lx).to(device) + weighting = weighting * L_weighting + return weighting + + def get_fold_unfold(self, x, kernel_size, stride, uf=1, df=1): # todo load once not every time, shorten code + """ + :param x: img of size (bs, c, h, w) + :return: n img crops of size (n, bs, c, kernel_size[0], kernel_size[1]) + """ + bs, nc, h, w = x.shape + + # number of crops in image + Ly = (h - kernel_size[0]) // stride[0] + 1 + Lx = (w - kernel_size[1]) // stride[1] + 1 + + if uf == 1 and df == 1: + fold_params = dict(kernel_size=kernel_size, dilation=1, padding=0, stride=stride) + unfold = torch.nn.Unfold(**fold_params) + + fold = torch.nn.Fold(output_size=x.shape[2:], **fold_params) + + weighting = self.get_weighting(kernel_size[0], kernel_size[1], Ly, Lx, x.device).to(x.dtype) + normalization = fold(weighting).view(1, 1, h, w) # normalizes the overlap + weighting = weighting.view((1, 1, kernel_size[0], kernel_size[1], Ly * Lx)) + + elif uf > 1 and df == 1: + fold_params = dict(kernel_size=kernel_size, dilation=1, padding=0, stride=stride) + unfold = torch.nn.Unfold(**fold_params) + + fold_params2 = dict(kernel_size=(kernel_size[0] * uf, kernel_size[0] * uf), + dilation=1, padding=0, + stride=(stride[0] * uf, stride[1] * uf)) + fold = torch.nn.Fold(output_size=(x.shape[2] * uf, x.shape[3] * uf), **fold_params2) + + weighting = self.get_weighting(kernel_size[0] * uf, kernel_size[1] * uf, Ly, Lx, x.device).to(x.dtype) + normalization = fold(weighting).view(1, 1, h * uf, w * uf) # normalizes the overlap + weighting = weighting.view((1, 1, kernel_size[0] * uf, kernel_size[1] * uf, Ly * Lx)) + + elif df > 1 and uf == 1: + fold_params = dict(kernel_size=kernel_size, dilation=1, padding=0, stride=stride) + unfold = torch.nn.Unfold(**fold_params) + + fold_params2 = dict(kernel_size=(kernel_size[0] // df, kernel_size[0] // df), + dilation=1, padding=0, + stride=(stride[0] // df, stride[1] // df)) + fold = torch.nn.Fold(output_size=(x.shape[2] // df, x.shape[3] // df), **fold_params2) + + weighting = self.get_weighting(kernel_size[0] // df, kernel_size[1] // df, Ly, Lx, x.device).to(x.dtype) + normalization = fold(weighting).view(1, 1, h // df, w // df) # normalizes the overlap + weighting = weighting.view((1, 1, kernel_size[0] // df, kernel_size[1] // df, Ly * Lx)) + + else: + raise NotImplementedError + + return fold, unfold, normalization, weighting + + @torch.no_grad() + def get_input(self, batch, k, return_first_stage_outputs=False, force_c_encode=False, + cond_key=None, return_original_cond=False, bs=None, uncond=0.05): + x = super().get_input(batch, k) + if bs is not None: + x = x[:bs] + x = x.to(self.device) + encoder_posterior = self.encode_first_stage(x) + z = self.get_first_stage_encoding(encoder_posterior).detach() + cond_key = cond_key or self.cond_stage_key + xc = super().get_input(batch, cond_key) + if bs is not None: + xc["c_crossattn"] = xc["c_crossattn"][:bs] + xc["c_concat"] = xc["c_concat"][:bs] + cond = {} + + # To support classifier-free guidance, randomly drop out only text conditioning 5%, only image conditioning 5%, and both 5%. + random = torch.rand(x.size(0), device=x.device) + prompt_mask = rearrange(random < 2 * uncond, "n -> n 1 1") + input_mask = 1 - rearrange((random >= uncond).float() * (random < 3 * uncond).float(), "n -> n 1 1 1") + + null_prompt = self.get_learned_conditioning([""]) + cond["c_crossattn"] = [torch.where(prompt_mask, null_prompt, self.get_learned_conditioning(xc["c_crossattn"]).detach())] + cond["c_concat"] = [input_mask * self.encode_first_stage((xc["c_concat"].to(self.device))).mode().detach()] + + out = [z, cond] + if return_first_stage_outputs: + xrec = self.decode_first_stage(z) + out.extend([x, xrec]) + if return_original_cond: + out.append(xc) + return out + + @torch.no_grad() + def decode_first_stage(self, z, predict_cids=False, force_not_quantize=False): + if predict_cids: + if z.dim() == 4: + z = torch.argmax(z.exp(), dim=1).long() + z = self.first_stage_model.quantize.get_codebook_entry(z, shape=None) + z = rearrange(z, 'b h w c -> b c h w').contiguous() + + z = 1. / self.scale_factor * z + + if hasattr(self, "split_input_params"): + if self.split_input_params["patch_distributed_vq"]: + ks = self.split_input_params["ks"] # eg. (128, 128) + stride = self.split_input_params["stride"] # eg. (64, 64) + uf = self.split_input_params["vqf"] + bs, nc, h, w = z.shape + if ks[0] > h or ks[1] > w: + ks = (min(ks[0], h), min(ks[1], w)) + print("reducing Kernel") + + if stride[0] > h or stride[1] > w: + stride = (min(stride[0], h), min(stride[1], w)) + print("reducing stride") + + fold, unfold, normalization, weighting = self.get_fold_unfold(z, ks, stride, uf=uf) + + z = unfold(z) # (bn, nc * prod(**ks), L) + # 1. Reshape to img shape + z = z.view((z.shape[0], -1, ks[0], ks[1], z.shape[-1])) # (bn, nc, ks[0], ks[1], L ) + + # 2. apply model loop over last dim + if isinstance(self.first_stage_model, VQModelInterface): + output_list = [self.first_stage_model.decode(z[:, :, :, :, i], + force_not_quantize=predict_cids or force_not_quantize) + for i in range(z.shape[-1])] + else: + + output_list = [self.first_stage_model.decode(z[:, :, :, :, i]) + for i in range(z.shape[-1])] + + o = torch.stack(output_list, axis=-1) # # (bn, nc, ks[0], ks[1], L) + o = o * weighting + # Reverse 1. reshape to img shape + o = o.view((o.shape[0], -1, o.shape[-1])) # (bn, nc * ks[0] * ks[1], L) + # stitch crops together + decoded = fold(o) + decoded = decoded / normalization # norm is shape (1, 1, h, w) + return decoded + else: + if isinstance(self.first_stage_model, VQModelInterface): + return self.first_stage_model.decode(z, force_not_quantize=predict_cids or force_not_quantize) + else: + return self.first_stage_model.decode(z) + + else: + if isinstance(self.first_stage_model, VQModelInterface): + return self.first_stage_model.decode(z, force_not_quantize=predict_cids or force_not_quantize) + else: + return self.first_stage_model.decode(z) + + # same as above but without decorator + def differentiable_decode_first_stage(self, z, predict_cids=False, force_not_quantize=False): + if predict_cids: + if z.dim() == 4: + z = torch.argmax(z.exp(), dim=1).long() + z = self.first_stage_model.quantize.get_codebook_entry(z, shape=None) + z = rearrange(z, 'b h w c -> b c h w').contiguous() + + z = 1. / self.scale_factor * z + + if hasattr(self, "split_input_params"): + if self.split_input_params["patch_distributed_vq"]: + ks = self.split_input_params["ks"] # eg. (128, 128) + stride = self.split_input_params["stride"] # eg. (64, 64) + uf = self.split_input_params["vqf"] + bs, nc, h, w = z.shape + if ks[0] > h or ks[1] > w: + ks = (min(ks[0], h), min(ks[1], w)) + print("reducing Kernel") + + if stride[0] > h or stride[1] > w: + stride = (min(stride[0], h), min(stride[1], w)) + print("reducing stride") + + fold, unfold, normalization, weighting = self.get_fold_unfold(z, ks, stride, uf=uf) + + z = unfold(z) # (bn, nc * prod(**ks), L) + # 1. Reshape to img shape + z = z.view((z.shape[0], -1, ks[0], ks[1], z.shape[-1])) # (bn, nc, ks[0], ks[1], L ) + + # 2. apply model loop over last dim + if isinstance(self.first_stage_model, VQModelInterface): + output_list = [self.first_stage_model.decode(z[:, :, :, :, i], + force_not_quantize=predict_cids or force_not_quantize) + for i in range(z.shape[-1])] + else: + + output_list = [self.first_stage_model.decode(z[:, :, :, :, i]) + for i in range(z.shape[-1])] + + o = torch.stack(output_list, axis=-1) # # (bn, nc, ks[0], ks[1], L) + o = o * weighting + # Reverse 1. reshape to img shape + o = o.view((o.shape[0], -1, o.shape[-1])) # (bn, nc * ks[0] * ks[1], L) + # stitch crops together + decoded = fold(o) + decoded = decoded / normalization # norm is shape (1, 1, h, w) + return decoded + else: + if isinstance(self.first_stage_model, VQModelInterface): + return self.first_stage_model.decode(z, force_not_quantize=predict_cids or force_not_quantize) + else: + return self.first_stage_model.decode(z) + + else: + if isinstance(self.first_stage_model, VQModelInterface): + return self.first_stage_model.decode(z, force_not_quantize=predict_cids or force_not_quantize) + else: + return self.first_stage_model.decode(z) + + @torch.no_grad() + def encode_first_stage(self, x): + if hasattr(self, "split_input_params"): + if self.split_input_params["patch_distributed_vq"]: + ks = self.split_input_params["ks"] # eg. (128, 128) + stride = self.split_input_params["stride"] # eg. (64, 64) + df = self.split_input_params["vqf"] + self.split_input_params['original_image_size'] = x.shape[-2:] + bs, nc, h, w = x.shape + if ks[0] > h or ks[1] > w: + ks = (min(ks[0], h), min(ks[1], w)) + print("reducing Kernel") + + if stride[0] > h or stride[1] > w: + stride = (min(stride[0], h), min(stride[1], w)) + print("reducing stride") + + fold, unfold, normalization, weighting = self.get_fold_unfold(x, ks, stride, df=df) + z = unfold(x) # (bn, nc * prod(**ks), L) + # Reshape to img shape + z = z.view((z.shape[0], -1, ks[0], ks[1], z.shape[-1])) # (bn, nc, ks[0], ks[1], L ) + + output_list = [self.first_stage_model.encode(z[:, :, :, :, i]) + for i in range(z.shape[-1])] + + o = torch.stack(output_list, axis=-1) + o = o * weighting + + # Reverse reshape to img shape + o = o.view((o.shape[0], -1, o.shape[-1])) # (bn, nc * ks[0] * ks[1], L) + # stitch crops together + decoded = fold(o) + decoded = decoded / normalization + return decoded + + else: + return self.first_stage_model.encode(x) + else: + return self.first_stage_model.encode(x) + + def shared_step(self, batch, **kwargs): + x, c = self.get_input(batch, self.first_stage_key) + loss = self(x, c) + return loss + + def forward(self, x, c, *args, **kwargs): + t = torch.randint(0, self.num_timesteps, (x.shape[0],), device=self.device).long() + if self.model.conditioning_key is not None: + assert c is not None + if self.cond_stage_trainable: + c = self.get_learned_conditioning(c) + if self.shorten_cond_schedule: # TODO: drop this option + tc = self.cond_ids[t].to(self.device) + c = self.q_sample(x_start=c, t=tc, noise=torch.randn_like(c.float())) + return self.p_losses(x, c, t, *args, **kwargs) + + def _rescale_annotations(self, bboxes, crop_coordinates): # TODO: move to dataset + def rescale_bbox(bbox): + x0 = clamp((bbox[0] - crop_coordinates[0]) / crop_coordinates[2]) + y0 = clamp((bbox[1] - crop_coordinates[1]) / crop_coordinates[3]) + w = min(bbox[2] / crop_coordinates[2], 1 - x0) + h = min(bbox[3] / crop_coordinates[3], 1 - y0) + return x0, y0, w, h + + return [rescale_bbox(b) for b in bboxes] + + def apply_model(self, x_noisy, t, cond, return_ids=False): + + if isinstance(cond, dict): + # hybrid case, cond is exptected to be a dict + pass + else: + if not isinstance(cond, list): + cond = [cond] + key = 'c_concat' if self.model.conditioning_key == 'concat' else 'c_crossattn' + cond = {key: cond} + + if hasattr(self, "split_input_params"): + assert len(cond) == 1 # todo can only deal with one conditioning atm + assert not return_ids + ks = self.split_input_params["ks"] # eg. (128, 128) + stride = self.split_input_params["stride"] # eg. (64, 64) + + h, w = x_noisy.shape[-2:] + + fold, unfold, normalization, weighting = self.get_fold_unfold(x_noisy, ks, stride) + + z = unfold(x_noisy) # (bn, nc * prod(**ks), L) + # Reshape to img shape + z = z.view((z.shape[0], -1, ks[0], ks[1], z.shape[-1])) # (bn, nc, ks[0], ks[1], L ) + z_list = [z[:, :, :, :, i] for i in range(z.shape[-1])] + + if self.cond_stage_key in ["image", "LR_image", "segmentation", + 'bbox_img'] and self.model.conditioning_key: # todo check for completeness + c_key = next(iter(cond.keys())) # get key + c = next(iter(cond.values())) # get value + assert (len(c) == 1) # todo extend to list with more than one elem + c = c[0] # get element + + c = unfold(c) + c = c.view((c.shape[0], -1, ks[0], ks[1], c.shape[-1])) # (bn, nc, ks[0], ks[1], L ) + + cond_list = [{c_key: [c[:, :, :, :, i]]} for i in range(c.shape[-1])] + + elif self.cond_stage_key == 'coordinates_bbox': + assert 'original_image_size' in self.split_input_params, 'BoudingBoxRescaling is missing original_image_size' + + # assuming padding of unfold is always 0 and its dilation is always 1 + n_patches_per_row = int((w - ks[0]) / stride[0] + 1) + full_img_h, full_img_w = self.split_input_params['original_image_size'] + # as we are operating on latents, we need the factor from the original image size to the + # spatial latent size to properly rescale the crops for regenerating the bbox annotations + num_downs = self.first_stage_model.encoder.num_resolutions - 1 + rescale_latent = 2 ** (num_downs) + + # get top left postions of patches as conforming for the bbbox tokenizer, therefore we + # need to rescale the tl patch coordinates to be in between (0,1) + tl_patch_coordinates = [(rescale_latent * stride[0] * (patch_nr % n_patches_per_row) / full_img_w, + rescale_latent * stride[1] * (patch_nr // n_patches_per_row) / full_img_h) + for patch_nr in range(z.shape[-1])] + + # patch_limits are tl_coord, width and height coordinates as (x_tl, y_tl, h, w) + patch_limits = [(x_tl, y_tl, + rescale_latent * ks[0] / full_img_w, + rescale_latent * ks[1] / full_img_h) for x_tl, y_tl in tl_patch_coordinates] + # patch_values = [(np.arange(x_tl,min(x_tl+ks, 1.)),np.arange(y_tl,min(y_tl+ks, 1.))) for x_tl, y_tl in tl_patch_coordinates] + + # tokenize crop coordinates for the bounding boxes of the respective patches + patch_limits_tknzd = [torch.LongTensor(self.bbox_tokenizer._crop_encoder(bbox))[None].to(self.device) + for bbox in patch_limits] # list of length l with tensors of shape (1, 2) + print(patch_limits_tknzd[0].shape) + # cut tknzd crop position from conditioning + assert isinstance(cond, dict), 'cond must be dict to be fed into model' + cut_cond = cond['c_crossattn'][0][..., :-2].to(self.device) + print(cut_cond.shape) + + adapted_cond = torch.stack([torch.cat([cut_cond, p], dim=1) for p in patch_limits_tknzd]) + adapted_cond = rearrange(adapted_cond, 'l b n -> (l b) n') + print(adapted_cond.shape) + adapted_cond = self.get_learned_conditioning(adapted_cond) + print(adapted_cond.shape) + adapted_cond = rearrange(adapted_cond, '(l b) n d -> l b n d', l=z.shape[-1]) + print(adapted_cond.shape) + + cond_list = [{'c_crossattn': [e]} for e in adapted_cond] + + else: + cond_list = [cond for i in range(z.shape[-1])] # Todo make this more efficient + + # apply model by loop over crops + output_list = [self.model(z_list[i], t, **cond_list[i]) for i in range(z.shape[-1])] + assert not isinstance(output_list[0], + tuple) # todo cant deal with multiple model outputs check this never happens + + o = torch.stack(output_list, axis=-1) + o = o * weighting + # Reverse reshape to img shape + o = o.view((o.shape[0], -1, o.shape[-1])) # (bn, nc * ks[0] * ks[1], L) + # stitch crops together + x_recon = fold(o) / normalization + + else: + x_recon = self.model(x_noisy, t, **cond) + + if isinstance(x_recon, tuple) and not return_ids: + return x_recon[0] + else: + return x_recon + + def _predict_eps_from_xstart(self, x_t, t, pred_xstart): + return (extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x_t.shape) * x_t - pred_xstart) / \ + extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x_t.shape) + + def _prior_bpd(self, x_start): + """ + Get the prior KL term for the variational lower-bound, measured in + bits-per-dim. + This term can't be optimized, as it only depends on the encoder. + :param x_start: the [N x C x ...] tensor of inputs. + :return: a batch of [N] KL values (in bits), one per batch element. + """ + batch_size = x_start.shape[0] + t = torch.tensor([self.num_timesteps - 1] * batch_size, device=x_start.device) + qt_mean, _, qt_log_variance = self.q_mean_variance(x_start, t) + kl_prior = normal_kl(mean1=qt_mean, logvar1=qt_log_variance, mean2=0.0, logvar2=0.0) + return mean_flat(kl_prior) / np.log(2.0) + + def p_losses(self, x_start, cond, t, noise=None): + noise = default(noise, lambda: torch.randn_like(x_start)) + x_noisy = self.q_sample(x_start=x_start, t=t, noise=noise) + model_output = self.apply_model(x_noisy, t, cond) + + loss_dict = {} + prefix = 'train' if self.training else 'val' + + if self.parameterization == "x0": + target = x_start + elif self.parameterization == "eps": + target = noise + else: + raise NotImplementedError() + + loss_simple = self.get_loss(model_output, target, mean=False).mean([1, 2, 3]) + loss_dict.update({f'{prefix}/loss_simple': loss_simple.mean()}) + + logvar_t = self.logvar[t].to(self.device) + loss = loss_simple / torch.exp(logvar_t) + logvar_t + # loss = loss_simple / torch.exp(self.logvar) + self.logvar + if self.learn_logvar: + loss_dict.update({f'{prefix}/loss_gamma': loss.mean()}) + loss_dict.update({'logvar': self.logvar.data.mean()}) + + loss = self.l_simple_weight * loss.mean() + + loss_vlb = self.get_loss(model_output, target, mean=False).mean(dim=(1, 2, 3)) + loss_vlb = (self.lvlb_weights[t] * loss_vlb).mean() + loss_dict.update({f'{prefix}/loss_vlb': loss_vlb}) + loss += (self.original_elbo_weight * loss_vlb) + loss_dict.update({f'{prefix}/loss': loss}) + + return loss, loss_dict + + def p_mean_variance(self, x, c, t, clip_denoised: bool, return_codebook_ids=False, quantize_denoised=False, + return_x0=False, score_corrector=None, corrector_kwargs=None): + t_in = t + model_out = self.apply_model(x, t_in, c, return_ids=return_codebook_ids) + + if score_corrector is not None: + assert self.parameterization == "eps" + model_out = score_corrector.modify_score(self, model_out, x, t, c, **corrector_kwargs) + + if return_codebook_ids: + model_out, logits = model_out + + if self.parameterization == "eps": + x_recon = self.predict_start_from_noise(x, t=t, noise=model_out) + elif self.parameterization == "x0": + x_recon = model_out + else: + raise NotImplementedError() + + if clip_denoised: + x_recon.clamp_(-1., 1.) + if quantize_denoised: + x_recon, _, [_, _, indices] = self.first_stage_model.quantize(x_recon) + model_mean, posterior_variance, posterior_log_variance = self.q_posterior(x_start=x_recon, x_t=x, t=t) + if return_codebook_ids: + return model_mean, posterior_variance, posterior_log_variance, logits + elif return_x0: + return model_mean, posterior_variance, posterior_log_variance, x_recon + else: + return model_mean, posterior_variance, posterior_log_variance + + @torch.no_grad() + def p_sample(self, x, c, t, clip_denoised=False, repeat_noise=False, + return_codebook_ids=False, quantize_denoised=False, return_x0=False, + temperature=1., noise_dropout=0., score_corrector=None, corrector_kwargs=None): + b, *_, device = *x.shape, x.device + outputs = self.p_mean_variance(x=x, c=c, t=t, clip_denoised=clip_denoised, + return_codebook_ids=return_codebook_ids, + quantize_denoised=quantize_denoised, + return_x0=return_x0, + score_corrector=score_corrector, corrector_kwargs=corrector_kwargs) + if return_codebook_ids: + raise DeprecationWarning("Support dropped.") + model_mean, _, model_log_variance, logits = outputs + elif return_x0: + model_mean, _, model_log_variance, x0 = outputs + else: + model_mean, _, model_log_variance = outputs + + noise = noise_like(x.shape, device, repeat_noise) * temperature + if noise_dropout > 0.: + noise = torch.nn.functional.dropout(noise, p=noise_dropout) + # no noise when t == 0 + nonzero_mask = (1 - (t == 0).float()).reshape(b, *((1,) * (len(x.shape) - 1))) + + if return_codebook_ids: + return model_mean + nonzero_mask * (0.5 * model_log_variance).exp() * noise, logits.argmax(dim=1) + if return_x0: + return model_mean + nonzero_mask * (0.5 * model_log_variance).exp() * noise, x0 + else: + return model_mean + nonzero_mask * (0.5 * model_log_variance).exp() * noise + + @torch.no_grad() + def progressive_denoising(self, cond, shape, verbose=True, callback=None, quantize_denoised=False, + img_callback=None, mask=None, x0=None, temperature=1., noise_dropout=0., + score_corrector=None, corrector_kwargs=None, batch_size=None, x_T=None, start_T=None, + log_every_t=None): + if not log_every_t: + log_every_t = self.log_every_t + timesteps = self.num_timesteps + if batch_size is not None: + b = batch_size if batch_size is not None else shape[0] + shape = [batch_size] + list(shape) + else: + b = batch_size = shape[0] + if x_T is None: + img = torch.randn(shape, device=self.device) + else: + img = x_T + intermediates = [] + if cond is not None: + if isinstance(cond, dict): + cond = {key: cond[key][:batch_size] if not isinstance(cond[key], list) else + list(map(lambda x: x[:batch_size], cond[key])) for key in cond} + else: + cond = [c[:batch_size] for c in cond] if isinstance(cond, list) else cond[:batch_size] + + if start_T is not None: + timesteps = min(timesteps, start_T) + iterator = tqdm(reversed(range(0, timesteps)), desc='Progressive Generation', + total=timesteps) if verbose else reversed( + range(0, timesteps)) + if type(temperature) == float: + temperature = [temperature] * timesteps + + for i in iterator: + ts = torch.full((b,), i, device=self.device, dtype=torch.long) + if self.shorten_cond_schedule: + assert self.model.conditioning_key != 'hybrid' + tc = self.cond_ids[ts].to(cond.device) + cond = self.q_sample(x_start=cond, t=tc, noise=torch.randn_like(cond)) + + img, x0_partial = self.p_sample(img, cond, ts, + clip_denoised=self.clip_denoised, + quantize_denoised=quantize_denoised, return_x0=True, + temperature=temperature[i], noise_dropout=noise_dropout, + score_corrector=score_corrector, corrector_kwargs=corrector_kwargs) + if mask is not None: + assert x0 is not None + img_orig = self.q_sample(x0, ts) + img = img_orig * mask + (1. - mask) * img + + if i % log_every_t == 0 or i == timesteps - 1: + intermediates.append(x0_partial) + if callback: callback(i) + if img_callback: img_callback(img, i) + return img, intermediates + + @torch.no_grad() + def p_sample_loop(self, cond, shape, return_intermediates=False, + x_T=None, verbose=True, callback=None, timesteps=None, quantize_denoised=False, + mask=None, x0=None, img_callback=None, start_T=None, + log_every_t=None): + + if not log_every_t: + log_every_t = self.log_every_t + device = self.betas.device + b = shape[0] + if x_T is None: + img = torch.randn(shape, device=device) + else: + img = x_T + + intermediates = [img] + if timesteps is None: + timesteps = self.num_timesteps + + if start_T is not None: + timesteps = min(timesteps, start_T) + iterator = tqdm(reversed(range(0, timesteps)), desc='Sampling t', total=timesteps) if verbose else reversed( + range(0, timesteps)) + + if mask is not None: + assert x0 is not None + assert x0.shape[2:3] == mask.shape[2:3] # spatial size has to match + + for i in iterator: + ts = torch.full((b,), i, device=device, dtype=torch.long) + if self.shorten_cond_schedule: + assert self.model.conditioning_key != 'hybrid' + tc = self.cond_ids[ts].to(cond.device) + cond = self.q_sample(x_start=cond, t=tc, noise=torch.randn_like(cond)) + + img = self.p_sample(img, cond, ts, + clip_denoised=self.clip_denoised, + quantize_denoised=quantize_denoised) + if mask is not None: + img_orig = self.q_sample(x0, ts) + img = img_orig * mask + (1. - mask) * img + + if i % log_every_t == 0 or i == timesteps - 1: + intermediates.append(img) + if callback: callback(i) + if img_callback: img_callback(img, i) + + if return_intermediates: + return img, intermediates + return img + + @torch.no_grad() + def sample(self, cond, batch_size=16, return_intermediates=False, x_T=None, + verbose=True, timesteps=None, quantize_denoised=False, + mask=None, x0=None, shape=None,**kwargs): + if shape is None: + shape = (batch_size, self.channels, self.image_size, self.image_size) + if cond is not None: + if isinstance(cond, dict): + cond = {key: cond[key][:batch_size] if not isinstance(cond[key], list) else + list(map(lambda x: x[:batch_size], cond[key])) for key in cond} + else: + cond = [c[:batch_size] for c in cond] if isinstance(cond, list) else cond[:batch_size] + return self.p_sample_loop(cond, + shape, + return_intermediates=return_intermediates, x_T=x_T, + verbose=verbose, timesteps=timesteps, quantize_denoised=quantize_denoised, + mask=mask, x0=x0) + + @torch.no_grad() + def sample_log(self,cond,batch_size,ddim, ddim_steps,**kwargs): + + if ddim: + ddim_sampler = DDIMSampler(self) + shape = (self.channels, self.image_size, self.image_size) + samples, intermediates =ddim_sampler.sample(ddim_steps,batch_size, + shape,cond,verbose=False,**kwargs) + + else: + samples, intermediates = self.sample(cond=cond, batch_size=batch_size, + return_intermediates=True,**kwargs) + + return samples, intermediates + + + @torch.no_grad() + def log_images(self, batch, N=4, n_row=4, sample=True, ddim_steps=200, ddim_eta=1., return_keys=None, + quantize_denoised=True, inpaint=False, plot_denoise_rows=False, plot_progressive_rows=False, + plot_diffusion_rows=False, **kwargs): + + use_ddim = False + + log = dict() + z, c, x, xrec, xc = self.get_input(batch, self.first_stage_key, + return_first_stage_outputs=True, + force_c_encode=True, + return_original_cond=True, + bs=N, uncond=0) + N = min(x.shape[0], N) + n_row = min(x.shape[0], n_row) + log["inputs"] = x + log["reals"] = xc["c_concat"] + log["reconstruction"] = xrec + if self.model.conditioning_key is not None: + if hasattr(self.cond_stage_model, "decode"): + xc = self.cond_stage_model.decode(c) + log["conditioning"] = xc + elif self.cond_stage_key in ["caption"]: + xc = log_txt_as_img((x.shape[2], x.shape[3]), batch["caption"]) + log["conditioning"] = xc + elif self.cond_stage_key == 'class_label': + xc = log_txt_as_img((x.shape[2], x.shape[3]), batch["human_label"]) + log['conditioning'] = xc + elif isimage(xc): + log["conditioning"] = xc + if ismap(xc): + log["original_conditioning"] = self.to_rgb(xc) + + if plot_diffusion_rows: + # get diffusion row + diffusion_row = list() + z_start = z[:n_row] + for t in range(self.num_timesteps): + if t % self.log_every_t == 0 or t == self.num_timesteps - 1: + t = repeat(torch.tensor([t]), '1 -> b', b=n_row) + t = t.to(self.device).long() + noise = torch.randn_like(z_start) + z_noisy = self.q_sample(x_start=z_start, t=t, noise=noise) + diffusion_row.append(self.decode_first_stage(z_noisy)) + + diffusion_row = torch.stack(diffusion_row) # n_log_step, n_row, C, H, W + diffusion_grid = rearrange(diffusion_row, 'n b c h w -> b n c h w') + diffusion_grid = rearrange(diffusion_grid, 'b n c h w -> (b n) c h w') + diffusion_grid = make_grid(diffusion_grid, nrow=diffusion_row.shape[0]) + log["diffusion_row"] = diffusion_grid + + if sample: + # get denoise row + with self.ema_scope("Plotting"): + samples, z_denoise_row = self.sample_log(cond=c,batch_size=N,ddim=use_ddim, + ddim_steps=ddim_steps,eta=ddim_eta) + # samples, z_denoise_row = self.sample(cond=c, batch_size=N, return_intermediates=True) + x_samples = self.decode_first_stage(samples) + log["samples"] = x_samples + if plot_denoise_rows: + denoise_grid = self._get_denoise_row_from_list(z_denoise_row) + log["denoise_row"] = denoise_grid + + if quantize_denoised and not isinstance(self.first_stage_model, AutoencoderKL) and not isinstance( + self.first_stage_model, IdentityFirstStage): + # also display when quantizing x0 while sampling + with self.ema_scope("Plotting Quantized Denoised"): + samples, z_denoise_row = self.sample_log(cond=c,batch_size=N,ddim=use_ddim, + ddim_steps=ddim_steps,eta=ddim_eta, + quantize_denoised=True) + # samples, z_denoise_row = self.sample(cond=c, batch_size=N, return_intermediates=True, + # quantize_denoised=True) + x_samples = self.decode_first_stage(samples.to(self.device)) + log["samples_x0_quantized"] = x_samples + + if inpaint: + # make a simple center square + b, h, w = z.shape[0], z.shape[2], z.shape[3] + mask = torch.ones(N, h, w).to(self.device) + # zeros will be filled in + mask[:, h // 4:3 * h // 4, w // 4:3 * w // 4] = 0. + mask = mask[:, None, ...] + with self.ema_scope("Plotting Inpaint"): + + samples, _ = self.sample_log(cond=c,batch_size=N,ddim=use_ddim, eta=ddim_eta, + ddim_steps=ddim_steps, x0=z[:N], mask=mask) + x_samples = self.decode_first_stage(samples.to(self.device)) + log["samples_inpainting"] = x_samples + log["mask"] = mask + + # outpaint + with self.ema_scope("Plotting Outpaint"): + samples, _ = self.sample_log(cond=c, batch_size=N, ddim=use_ddim,eta=ddim_eta, + ddim_steps=ddim_steps, x0=z[:N], mask=mask) + x_samples = self.decode_first_stage(samples.to(self.device)) + log["samples_outpainting"] = x_samples + + if plot_progressive_rows: + with self.ema_scope("Plotting Progressives"): + img, progressives = self.progressive_denoising(c, + shape=(self.channels, self.image_size, self.image_size), + batch_size=N) + prog_row = self._get_denoise_row_from_list(progressives, desc="Progressive Generation") + log["progressive_row"] = prog_row + + if return_keys: + if np.intersect1d(list(log.keys()), return_keys).shape[0] == 0: + return log + else: + return {key: log[key] for key in return_keys} + return log + + def configure_optimizers(self): + lr = self.learning_rate + params = list(self.model.parameters()) + if self.cond_stage_trainable: + print(f"{self.__class__.__name__}: Also optimizing conditioner params!") + params = params + list(self.cond_stage_model.parameters()) + if self.learn_logvar: + print('Diffusion model optimizing logvar') + params.append(self.logvar) + opt = torch.optim.AdamW(params, lr=lr) + if self.use_scheduler: + assert 'target' in self.scheduler_config + scheduler = instantiate_from_config(self.scheduler_config) + + print("Setting up LambdaLR scheduler...") + scheduler = [ + { + 'scheduler': LambdaLR(opt, lr_lambda=scheduler.schedule), + 'interval': 'step', + 'frequency': 1 + }] + return [opt], scheduler + return opt + + @torch.no_grad() + def to_rgb(self, x): + x = x.float() + if not hasattr(self, "colorize"): + self.colorize = torch.randn(3, x.shape[1], 1, 1).to(x) + x = nn.functional.conv2d(x, weight=self.colorize) + x = 2. * (x - x.min()) / (x.max() - x.min()) - 1. + return x + + +class DiffusionWrapper(pl.LightningModule): + def __init__(self, diff_model_config, conditioning_key): + super().__init__() + self.diffusion_model = instantiate_from_config(diff_model_config) + self.conditioning_key = conditioning_key + assert self.conditioning_key in [None, 'concat', 'crossattn', 'hybrid', 'adm'] + + def forward(self, x, t, c_concat: list = None, c_crossattn: list = None): + if self.conditioning_key is None: + out = self.diffusion_model(x, t) + elif self.conditioning_key == 'concat': + xc = torch.cat([x] + c_concat, dim=1) + out = self.diffusion_model(xc, t) + elif self.conditioning_key == 'crossattn': + cc = torch.cat(c_crossattn, 1) + out = self.diffusion_model(x, t, context=cc) + elif self.conditioning_key == 'hybrid': + xc = torch.cat([x] + c_concat, dim=1) + cc = torch.cat(c_crossattn, 1) + out = self.diffusion_model(xc, t, context=cc) + elif self.conditioning_key == 'adm': + cc = c_crossattn[0] + out = self.diffusion_model(x, t, y=cc) + else: + raise NotImplementedError() + + return out + + +class Layout2ImgDiffusion(LatentDiffusion): + # TODO: move all layout-specific hacks to this class + def __init__(self, cond_stage_key, *args, **kwargs): + assert cond_stage_key == 'coordinates_bbox', 'Layout2ImgDiffusion only for cond_stage_key="coordinates_bbox"' + super().__init__(cond_stage_key=cond_stage_key, *args, **kwargs) + + def log_images(self, batch, N=8, *args, **kwargs): + logs = super().log_images(batch=batch, N=N, *args, **kwargs) + + key = 'train' if self.training else 'validation' + dset = self.trainer.datamodule.datasets[key] + mapper = dset.conditional_builders[self.cond_stage_key] + + bbox_imgs = [] + map_fn = lambda catno: dset.get_textual_label(dset.get_category_id(catno)) + for tknzd_bbox in batch[self.cond_stage_key][:N]: + bboximg = mapper.plot(tknzd_bbox.detach().cpu(), map_fn, (256, 256)) + bbox_imgs.append(bboximg) + + cond_img = torch.stack(bbox_imgs, dim=0) + logs['bbox_image'] = cond_img + return logs diff --git a/modules/sd_hijack_ip2p.py b/modules/sd_hijack_ip2p.py new file mode 100644 index 00000000..635f015f --- /dev/null +++ b/modules/sd_hijack_ip2p.py @@ -0,0 +1,13 @@ +import collections +import os.path +import sys +import gc +import time + +def should_hijack_ip2p(checkpoint_info): + from modules import sd_models + + ckpt_basename = os.path.basename(checkpoint_info.filename).lower() + cfg_basename = os.path.basename(sd_models.find_checkpoint_config(checkpoint_info)).lower() + + return "pix2pix" in ckpt_basename and not "pix2pix" in cfg_basename diff --git a/modules/sd_models.py b/modules/sd_models.py index 12083848..cddc2343 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -17,6 +17,7 @@ from ldm.util import instantiate_from_config from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors, hashes from modules.paths import models_path from modules.sd_hijack_inpainting import do_inpainting_hijack, should_hijack_inpainting +from modules.sd_hijack_ip2p import should_hijack_ip2p model_dir = "Stable-diffusion" model_path = os.path.abspath(os.path.join(models_path, model_dir)) @@ -365,6 +366,15 @@ def load_model(checkpoint_info=None): sd_config.model.params.unet_config.params.in_channels = 9 sd_config.model.params.finetune_keys = None + if should_hijack_ip2p(checkpoint_info): + sd_config.model.target = "modules.models.diffusion.ddpm_edit.LatentDiffusion" + sd_config.model.params.conditioning_key = "hybrid" + sd_config.model.params.first_stage_key = "edited" + sd_config.model.params.cond_stage_key = "edit" + sd_config.model.params.image_size = 16 + sd_config.model.params.unet_config.params.in_channels = 8 + sd_config.model.params.unet_config.params.out_channels = 4 + if not hasattr(sd_config.model.params, "use_ema"): sd_config.model.params.use_ema = False @@ -429,7 +439,7 @@ def reload_model_weights(sd_model=None, info=None): checkpoint_config = find_checkpoint_config(current_checkpoint_info) - if current_checkpoint_info is None or checkpoint_config != find_checkpoint_config(checkpoint_info) or should_hijack_inpainting(checkpoint_info) != should_hijack_inpainting(sd_model.sd_checkpoint_info): + if current_checkpoint_info is None or checkpoint_config != find_checkpoint_config(checkpoint_info) or should_hijack_inpainting(checkpoint_info) != should_hijack_inpainting(sd_model.sd_checkpoint_info) or should_hijack_ip2p(checkpoint_info) != should_hijack_ip2p(sd_model.sd_checkpoint_info): del sd_model checkpoints_loaded.clear() load_model(checkpoint_info) -- cgit v1.2.3 From 57c1baa774d07060af0abbd2974c5f36c8cb63ac Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 25 Jan 2023 18:56:23 +0300 Subject: change to code for live preview fix on OSX to be bit more obvious --- modules/processing.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 3bd590ba..57c3db1b 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -568,8 +568,8 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: with devices.autocast(): p.init(p.all_prompts, p.all_seeds, p.all_subseeds) - if shared.opts.live_previews_enable and sd_samplers.approximation_indexes.get(shared.opts.show_progress_type, 0) == 1: - # preload approx nn model before sampling for a more deterministic result + # for OSX, loading the model during sampling changes the generated picture, so it is loaded here + if shared.opts.live_previews_enable and opts.show_progress_type == "Approx NN": sd_vae_approx.model() if not p.disable_extra_networks: -- cgit v1.2.3 From e179b6098ac1b1ce9645fef5bd9fd0bc9b918f30 Mon Sep 17 00:00:00 2001 From: "Alex \"mcmonkey\" Goodwin" Date: Wed, 25 Jan 2023 08:48:40 -0800 Subject: allow symlinks in the textual inversion embeddings folder --- modules/textual_inversion/textual_inversion.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py index 4e90f690..6cf00e65 100644 --- a/modules/textual_inversion/textual_inversion.py +++ b/modules/textual_inversion/textual_inversion.py @@ -194,7 +194,7 @@ class EmbeddingDatabase: if not os.path.isdir(embdir.path): return - for root, dirs, fns in os.walk(embdir.path): + for root, dirs, fns in os.walk(embdir.path, followlinks=True): for fn in fns: try: fullfn = os.path.join(root, fn) -- cgit v1.2.3 From 789d47f832a5c921dbbdd0a657dff9bca7f78d94 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 25 Jan 2023 19:55:31 +0300 Subject: make clicking extra networks button one more time close the extra networks UI --- modules/ui_extra_networks.py | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/ui_extra_networks.py b/modules/ui_extra_networks.py index 8b4f97f8..c6ff889a 100644 --- a/modules/ui_extra_networks.py +++ b/modules/ui_extra_networks.py @@ -117,8 +117,13 @@ def create_ui(container, button, tabname): ui.button_save_preview = gr.Button('Save preview', elem_id=tabname+"_save_preview", visible=False) ui.preview_target_filename = gr.Textbox('Preview save filename', elem_id=tabname+"_preview_filename", visible=False) - button.click(fn=lambda: gr.update(visible=True), inputs=[], outputs=[container]) - button_close.click(fn=lambda: gr.update(visible=False), inputs=[], outputs=[container]) + def toggle_visibility(is_visible): + is_visible = not is_visible + return is_visible, gr.update(visible=is_visible) + + state_visible = gr.State(value=False) + button.click(fn=toggle_visibility, inputs=[state_visible], outputs=[state_visible, container]) + button_close.click(fn=toggle_visibility, inputs=[state_visible], outputs=[state_visible, container]) def refresh(): res = [] -- cgit v1.2.3 From 15e89ef0f6f22f823c19592a401b9e4ee477258c Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 25 Jan 2023 20:11:01 +0300 Subject: fix for unet hijack breaking the train tab --- modules/sd_hijack_unet.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/sd_hijack_unet.py b/modules/sd_hijack_unet.py index 88c94e54..a6ee577c 100644 --- a/modules/sd_hijack_unet.py +++ b/modules/sd_hijack_unet.py @@ -36,8 +36,11 @@ th = TorchHijackForUnet() # Below are monkey patches to enable upcasting a float16 UNet for float32 sampling def apply_model(orig_func, self, x_noisy, t, cond, **kwargs): - for y in cond.keys(): - cond[y] = [x.to(devices.dtype_unet) if isinstance(x, torch.Tensor) else x for x in cond[y]] + + if isinstance(cond, dict): + for y in cond.keys(): + cond[y] = [x.to(devices.dtype_unet) if isinstance(x, torch.Tensor) else x for x in cond[y]] + with devices.autocast(): return orig_func(self, x_noisy.to(devices.dtype_unet), t.to(devices.dtype_unet), cond, **kwargs).float() -- cgit v1.2.3 From d1d6ce29831d1b067801c3206f314258de88f683 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 25 Jan 2023 23:25:25 +0300 Subject: add edit_image_conditioning from my earlier edits in case there's an attempt to inegrate pix2pix properly this allows to use pix2pix model in img2img though it won't work well this way --- modules/processing.py | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 9e5a2f38..cb41288a 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -185,7 +185,12 @@ class StableDiffusionProcessing: conditioning = 2. * (conditioning - depth_min) / (depth_max - depth_min) - 1. return conditioning - def inpainting_image_conditioning(self, source_image, latent_image, image_mask = None): + def edit_image_conditioning(self, source_image): + conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(source_image)) + + return conditioning_image + + def inpainting_image_conditioning(self, source_image, latent_image, image_mask=None): self.is_using_inpainting_conditioning = True # Handle the different mask inputs @@ -228,6 +233,9 @@ class StableDiffusionProcessing: if isinstance(self.sd_model, LatentDepth2ImageDiffusion): return self.depth2img_image_conditioning(source_image.float() if devices.unet_needs_upcast else source_image) + if self.sd_model.cond_stage_key == "edit": + return self.edit_image_conditioning(source_image) + if self.sampler.conditioning_key in {'hybrid', 'concat'}: return self.inpainting_image_conditioning(source_image.float() if devices.unet_needs_upcast else source_image, latent_image, image_mask=image_mask) -- cgit v1.2.3 From 6cff4401824299a983c8e13424018efc347b4a2b Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 25 Jan 2023 23:25:40 +0300 Subject: fix prompt editing break after first batch in img2img --- modules/sd_samplers.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py index 6261d1f7..a7910b56 100644 --- a/modules/sd_samplers.py +++ b/modules/sd_samplers.py @@ -454,7 +454,7 @@ class KDiffusionSampler: def initialize(self, p): self.model_wrap_cfg.mask = p.mask if hasattr(p, 'mask') else None self.model_wrap_cfg.nmask = p.nmask if hasattr(p, 'nmask') else None - self.model_wrap.step = 0 + self.model_wrap_cfg.step = 0 self.eta = p.eta or opts.eta_ancestral k_diffusion.sampling.torch = TorchHijack(self.sampler_noises if self.sampler_noises is not None else []) -- cgit v1.2.3 From 10421f93c3f7f7ce88cb40391b46d4e6664eff74 Mon Sep 17 00:00:00 2001 From: brkirch Date: Thu, 26 Jan 2023 00:34:38 -0500 Subject: Fix full previews, --no-half-vae --- modules/processing.py | 8 ++++---- modules/sd_hijack_utils.py | 2 +- 2 files changed, 5 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index cb41288a..92894d67 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -172,7 +172,7 @@ class StableDiffusionProcessing: midas_in = torch.from_numpy(transformed["midas_in"][None, ...]).to(device=shared.device) midas_in = repeat(midas_in, "1 ... -> n ...", n=self.batch_size) - conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(source_image.to(devices.dtype_unet) if devices.unet_needs_upcast else source_image)) + conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(source_image.to(devices.dtype_vae) if devices.unet_needs_upcast else source_image)) conditioning_image = conditioning_image.float() if devices.unet_needs_upcast else conditioning_image conditioning = torch.nn.functional.interpolate( self.sd_model.depth_model(midas_in), @@ -217,7 +217,7 @@ class StableDiffusionProcessing: ) # Encode the new masked image using first stage of network. - conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(conditioning_image.to(devices.dtype_unet) if devices.unet_needs_upcast else conditioning_image)) + conditioning_image = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(conditioning_image.to(devices.dtype_vae) if devices.unet_needs_upcast else conditioning_image)) # Create the concatenated conditioning tensor to be fed to `c_concat` conditioning_mask = torch.nn.functional.interpolate(conditioning_mask, size=latent_image.shape[-2:]) @@ -417,7 +417,7 @@ def create_random_tensors(shape, seeds, subseeds=None, subseed_strength=0.0, see def decode_first_stage(model, x): with devices.autocast(disable=x.dtype == devices.dtype_vae): - x = model.decode_first_stage(x) + x = model.decode_first_stage(x.to(devices.dtype_vae) if devices.unet_needs_upcast else x) return x @@ -1001,7 +1001,7 @@ class StableDiffusionProcessingImg2Img(StableDiffusionProcessing): image = torch.from_numpy(batch_images) image = 2. * image - 1. - image = image.to(device=shared.device, dtype=devices.dtype_unet if devices.unet_needs_upcast else None) + image = image.to(device=shared.device, dtype=devices.dtype_vae if devices.unet_needs_upcast else None) self.init_latent = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(image)) diff --git a/modules/sd_hijack_utils.py b/modules/sd_hijack_utils.py index f81b169a..f8684475 100644 --- a/modules/sd_hijack_utils.py +++ b/modules/sd_hijack_utils.py @@ -5,7 +5,7 @@ class CondFunc: self = super(CondFunc, cls).__new__(cls) if isinstance(orig_func, str): func_path = orig_func.split('.') - for i in range(len(func_path)-2, -1, -1): + for i in range(len(func_path)-1, -1, -1): try: resolved_obj = importlib.import_module('.'.join(func_path[:i])) break -- cgit v1.2.3 From f4ec411f2c9d6bc6817a2eca8a2c00f255ffb386 Mon Sep 17 00:00:00 2001 From: "ULTRANOX\\Chris" Date: Thu, 26 Jan 2023 03:45:16 -0500 Subject: Allow checkpoint merger to merge pix2pix models in the same way that it currently supports inpainting models. --- modules/extras.py | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 36123aa5..67ffdee3 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -132,6 +132,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None result_is_inpainting_model = False + result_is_pix2pix_model = False if theta_func2: shared.state.textinfo = f"Loading B" @@ -186,13 +187,17 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ if a.shape[1] == 4 and b.shape[1] == 9: raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") - assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" - - theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) - result_is_inpainting_model = True + if a.shape[1] == 8 and b.shape[1] == 4:#If we have an InstructPix2Pix model... + print("Detected possible merge of instruct model with non-instruct model.") + theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier)#Merge only the vectors the models have in common. Otherwise we get an error due to dimension mismatch. + result_is_pix2pix_model = True + else: + assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" + theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) + result_is_inpainting_model = True else: theta_0[key] = theta_func2(a, b, multiplier) - + theta_0[key] = to_half(theta_0[key], save_as_half) shared.state.sampling_step += 1 @@ -226,6 +231,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ filename = filename_generator() if custom_name == '' else custom_name filename += ".inpainting" if result_is_inpainting_model else "" + filename += ".pix2pix" if result_is_pix2pix_model else "" filename += "." + checkpoint_format output_modelname = os.path.join(ckpt_dir, filename) -- cgit v1.2.3 From f90798c6b6cc48e514acb08ce02bdb5874bf74d8 Mon Sep 17 00:00:00 2001 From: "ULTRANOX\\Chris" Date: Thu, 26 Jan 2023 04:38:04 -0500 Subject: Added error check for the rare case a user merges a pix2pix model with a normal model using weighted sum. Also removed bad print message that interfered with merging progress bar. --- modules/extras.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 67ffdee3..badd13c7 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -186,9 +186,10 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ if a.shape != b.shape and a.shape[0:1] + a.shape[2:] == b.shape[0:1] + b.shape[2:]: if a.shape[1] == 4 and b.shape[1] == 9: raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") + if a.shape[1] == 4 and b.shape[1] == 8: + raise RuntimeError("When merging pix2pix model with a normal one, A must be the pix2pix model.") if a.shape[1] == 8 and b.shape[1] == 4:#If we have an InstructPix2Pix model... - print("Detected possible merge of instruct model with non-instruct model.") theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier)#Merge only the vectors the models have in common. Otherwise we get an error due to dimension mismatch. result_is_pix2pix_model = True else: -- cgit v1.2.3 From 9e72dc743480c8b1ca6aeb8ced3af03f3e3243a3 Mon Sep 17 00:00:00 2001 From: "ULTRANOX\\Chris" Date: Thu, 26 Jan 2023 06:05:40 -0500 Subject: Changed all references to "pix2pix" to the more precise name "instruct pix2pix". Also changed extension to instrpix2pix at least for now. --- modules/extras.py | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index badd13c7..2bf0d17e 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -132,7 +132,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ tertiary_model_info = sd_models.checkpoints_list[tertiary_model_name] if theta_func1 else None result_is_inpainting_model = False - result_is_pix2pix_model = False + result_is_instruct_pix2pix_model = False if theta_func2: shared.state.textinfo = f"Loading B" @@ -187,11 +187,11 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ if a.shape[1] == 4 and b.shape[1] == 9: raise RuntimeError("When merging inpainting model with a normal one, A must be the inpainting model.") if a.shape[1] == 4 and b.shape[1] == 8: - raise RuntimeError("When merging pix2pix model with a normal one, A must be the pix2pix model.") + raise RuntimeError("When merging instruct-pix2pix model with a normal one, A must be the instruct-pix2pix model.") - if a.shape[1] == 8 and b.shape[1] == 4:#If we have an InstructPix2Pix model... + if a.shape[1] == 8 and b.shape[1] == 4:#If we have an Instruct-Pix2Pix model... theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier)#Merge only the vectors the models have in common. Otherwise we get an error due to dimension mismatch. - result_is_pix2pix_model = True + result_is_instruct_pix2pix_model = True else: assert a.shape[1] == 9 and b.shape[1] == 4, f"Bad dimensions for merged layer {key}: A={a.shape}, B={b.shape}" theta_0[key][:, 0:4, :, :] = theta_func2(a[:, 0:4, :, :], b, multiplier) @@ -232,7 +232,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ filename = filename_generator() if custom_name == '' else custom_name filename += ".inpainting" if result_is_inpainting_model else "" - filename += ".pix2pix" if result_is_pix2pix_model else "" + filename += ".instrpix2pix" if result_is_instruct_pix2pix_model else "" filename += "." + checkpoint_format output_modelname = os.path.join(ckpt_dir, filename) -- cgit v1.2.3 From cdc2fa209a3efdc71a90643a5e7a1df49869cd5f Mon Sep 17 00:00:00 2001 From: "ULTRANOX\\Chris" Date: Thu, 26 Jan 2023 11:27:07 -0500 Subject: Changed filename addition from "instrpix2pix" to the more readable ".instruct-pix2pix" for newly generated instruct pix2pix models. --- modules/extras.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 2bf0d17e..466ecc15 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -232,7 +232,7 @@ def run_modelmerger(id_task, primary_model_name, secondary_model_name, tertiary_ filename = filename_generator() if custom_name == '' else custom_name filename += ".inpainting" if result_is_inpainting_model else "" - filename += ".instrpix2pix" if result_is_instruct_pix2pix_model else "" + filename += ".instruct-pix2pix" if result_is_instruct_pix2pix_model else "" filename += "." + checkpoint_format output_modelname = os.path.join(ckpt_dir, filename) -- cgit v1.2.3 From 7a14c8ab45da8a681792a6331d48a88dd684a0a9 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 26 Jan 2023 23:29:27 +0300 Subject: add an option to enable sections from extras tab in txt2img/img2img fix some style inconsistenices --- modules/processing.py | 7 +++++- modules/scripts.py | 32 ++++++++++++++++++++++---- modules/scripts_auto_postprocessing.py | 42 ++++++++++++++++++++++++++++++++++ modules/scripts_postprocessing.py | 11 ++++++--- modules/shared.py | 15 ++++-------- modules/shared_items.py | 10 ++++++++ modules/ui_components.py | 8 +++++++ 7 files changed, 107 insertions(+), 18 deletions(-) create mode 100644 modules/scripts_auto_postprocessing.py create mode 100644 modules/shared_items.py (limited to 'modules') diff --git a/modules/processing.py b/modules/processing.py index 92894d67..262806a1 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -13,7 +13,7 @@ from skimage import exposure from typing import Any, Dict, List, Optional import modules.sd_hijack -from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_parameters_copypaste, script_callbacks, extra_networks, sd_vae_approx +from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_parameters_copypaste, script_callbacks, extra_networks, sd_vae_approx, scripts from modules.sd_hijack import model_hijack from modules.shared import opts, cmd_opts, state import modules.shared as shared @@ -658,6 +658,11 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: image = Image.fromarray(x_sample) + if p.scripts is not None: + pp = scripts.PostprocessImageArgs(image) + p.scripts.postprocess_image(p, pp) + image = pp.image + if p.color_corrections is not None and i < len(p.color_corrections): if opts.save and not p.do_not_save_samples and opts.save_images_before_color_correction: image_without_cc = apply_overlay(image, p.paste_to, i, p.overlay_images) diff --git a/modules/scripts.py b/modules/scripts.py index 03907a63..6e9dc0c0 100644 --- a/modules/scripts.py +++ b/modules/scripts.py @@ -6,12 +6,16 @@ from collections import namedtuple import gradio as gr -from modules.processing import StableDiffusionProcessing from modules import shared, paths, script_callbacks, extensions, script_loading, scripts_postprocessing AlwaysVisible = object() +class PostprocessImageArgs: + def __init__(self, image): + self.image = image + + class Script: filename = None args_from = None @@ -65,7 +69,7 @@ class Script: args contains all values returned by components from ui() """ - raise NotImplementedError() + pass def process(self, p, *args): """ @@ -100,6 +104,13 @@ class Script: pass + def postprocess_image(self, p, pp: PostprocessImageArgs, *args): + """ + Called for every image after it has been generated. + """ + + pass + def postprocess(self, p, processed, *args): """ This function is called after processing ends for AlwaysVisible scripts. @@ -247,11 +258,15 @@ class ScriptRunner: self.infotext_fields = [] def initialize_scripts(self, is_img2img): + from modules import scripts_auto_postprocessing + self.scripts.clear() self.alwayson_scripts.clear() self.selectable_scripts.clear() - for script_class, path, basedir, script_module in scripts_data: + auto_processing_scripts = scripts_auto_postprocessing.create_auto_preprocessing_script_data() + + for script_class, path, basedir, script_module in auto_processing_scripts + scripts_data: script = script_class() script.filename = path script.is_txt2img = not is_img2img @@ -332,7 +347,7 @@ class ScriptRunner: return inputs - def run(self, p: StableDiffusionProcessing, *args): + def run(self, p, *args): script_index = args[0] if script_index == 0: @@ -386,6 +401,15 @@ class ScriptRunner: print(f"Error running postprocess_batch: {script.filename}", file=sys.stderr) print(traceback.format_exc(), file=sys.stderr) + def postprocess_image(self, p, pp: PostprocessImageArgs): + for script in self.alwayson_scripts: + try: + script_args = p.script_args[script.args_from:script.args_to] + script.postprocess_image(p, pp, *script_args) + except Exception: + print(f"Error running postprocess_batch: {script.filename}", file=sys.stderr) + print(traceback.format_exc(), file=sys.stderr) + def before_component(self, component, **kwargs): for script in self.scripts: try: diff --git a/modules/scripts_auto_postprocessing.py b/modules/scripts_auto_postprocessing.py new file mode 100644 index 00000000..30d6d658 --- /dev/null +++ b/modules/scripts_auto_postprocessing.py @@ -0,0 +1,42 @@ +from modules import scripts, scripts_postprocessing, shared + + +class ScriptPostprocessingForMainUI(scripts.Script): + def __init__(self, script_postproc): + self.script: scripts_postprocessing.ScriptPostprocessing = script_postproc + self.postprocessing_controls = None + + def title(self): + return self.script.name + + def show(self, is_img2img): + return scripts.AlwaysVisible + + def ui(self, is_img2img): + self.postprocessing_controls = self.script.ui() + return self.postprocessing_controls.values() + + def postprocess_image(self, p, script_pp, *args): + args_dict = {k: v for k, v in zip(self.postprocessing_controls, args)} + + pp = scripts_postprocessing.PostprocessedImage(script_pp.image) + pp.info = {} + self.script.process(pp, **args_dict) + p.extra_generation_params.update(pp.info) + script_pp.image = pp.image + + +def create_auto_preprocessing_script_data(): + from modules import scripts + + res = [] + + for name in shared.opts.postprocessing_enable_in_main_ui: + script = next(iter([x for x in scripts.postprocessing_scripts_data if x.script_class.name == name]), None) + if script is None: + continue + + constructor = lambda s=script: ScriptPostprocessingForMainUI(s.script_class()) + res.append(scripts.ScriptClassData(script_class=constructor, path=script.path, basedir=script.basedir, module=script.module)) + + return res diff --git a/modules/scripts_postprocessing.py b/modules/scripts_postprocessing.py index 25de02d0..ce0ebb61 100644 --- a/modules/scripts_postprocessing.py +++ b/modules/scripts_postprocessing.py @@ -46,6 +46,8 @@ class ScriptPostprocessing: pass + + def wrap_call(func, filename, funcname, *args, default=None, **kwargs): try: res = func(*args, **kwargs) @@ -68,6 +70,9 @@ class ScriptPostprocessingRunner: script: ScriptPostprocessing = script_class() script.filename = path + if script.name == "Simple Upscale": + continue + self.scripts.append(script) def create_script_ui(self, script, inputs): @@ -87,12 +92,11 @@ class ScriptPostprocessingRunner: import modules.scripts self.initialize_scripts(modules.scripts.postprocessing_scripts_data) - scripts_order = [x.lower().strip() for x in shared.opts.postprocessing_scipts_order.split(",")] + scripts_order = shared.opts.postprocessing_operation_order def script_score(name): - name = name.lower() for i, possible_match in enumerate(scripts_order): - if possible_match in name: + if possible_match == name: return i return len(self.scripts) @@ -145,3 +149,4 @@ class ScriptPostprocessingRunner: def image_changed(self): for script in self.scripts_in_preferred_order(): script.image_changed() + diff --git a/modules/shared.py b/modules/shared.py index 6a0b96cb..cdeed55d 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -13,8 +13,8 @@ import modules.interrogate import modules.memmon import modules.styles import modules.devices as devices -from modules import localization, sd_vae, extensions, script_loading, errors, ui_components -from modules.paths import models_path, script_path, sd_path +from modules import localization, sd_vae, extensions, script_loading, errors, ui_components, shared_items +from modules.paths import models_path, script_path demo = None @@ -264,12 +264,6 @@ interrogator = modules.interrogate.InterrogateModels("interrogate") face_restorers = [] - -def realesrgan_models_names(): - import modules.realesrgan_model - return [x.name for x in modules.realesrgan_model.get_realesrgan_models(None)] - - class OptionInfo: def __init__(self, default=None, label="", component=None, component_args=None, onchange=None, section=None, refresh=None): self.default = default @@ -360,7 +354,7 @@ options_templates.update(options_section(('saving-to-dirs', "Saving to a directo options_templates.update(options_section(('upscaling', "Upscaling"), { "ESRGAN_tile": OptionInfo(192, "Tile size for ESRGAN upscalers. 0 = no tiling.", gr.Slider, {"minimum": 0, "maximum": 512, "step": 16}), "ESRGAN_tile_overlap": OptionInfo(8, "Tile overlap, in pixels for ESRGAN upscalers. Low values = visible seam.", gr.Slider, {"minimum": 0, "maximum": 48, "step": 1}), - "realesrgan_enabled_models": OptionInfo(["R-ESRGAN 4x+", "R-ESRGAN 4x+ Anime6B"], "Select which Real-ESRGAN models to show in the web UI. (Requires restart)", gr.CheckboxGroup, lambda: {"choices": realesrgan_models_names()}), + "realesrgan_enabled_models": OptionInfo(["R-ESRGAN 4x+", "R-ESRGAN 4x+ Anime6B"], "Select which Real-ESRGAN models to show in the web UI. (Requires restart)", gr.CheckboxGroup, lambda: {"choices": shared_items.realesrgan_models_names()}), "upscaler_for_img2img": OptionInfo(None, "Upscaler for img2img", gr.Dropdown, lambda: {"choices": [x.name for x in sd_upscalers]}), })) @@ -483,7 +477,8 @@ options_templates.update(options_section(('sampler-params', "Sampler parameters" })) options_templates.update(options_section(('postprocessing', "Postprocessing"), { - 'postprocessing_scipts_order': OptionInfo("upscale, gfpgan, codeformer", "Postprocessing operation order"), + 'postprocessing_enable_in_main_ui': OptionInfo([], "Enable postprocessing operations in txt2img and img2img tabs", ui_components.DropdownMulti, lambda: {"choices": [x.name for x in shared_items.postprocessing_scripts()]}), + 'postprocessing_operation_order': OptionInfo([], "Postprocessing operation order", ui_components.DropdownMulti, lambda: {"choices": [x.name for x in shared_items.postprocessing_scripts()]}), 'upscaling_max_images_in_cache': OptionInfo(5, "Maximum number of images in upscaling cache", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), })) diff --git a/modules/shared_items.py b/modules/shared_items.py new file mode 100644 index 00000000..b5d480c9 --- /dev/null +++ b/modules/shared_items.py @@ -0,0 +1,10 @@ + + +def realesrgan_models_names(): + import modules.realesrgan_model + return [x.name for x in modules.realesrgan_model.get_realesrgan_models(None)] + +def postprocessing_scripts(): + import modules.scripts + + return modules.scripts.scripts_postproc.scripts \ No newline at end of file diff --git a/modules/ui_components.py b/modules/ui_components.py index 9aec3097..284ca0cf 100644 --- a/modules/ui_components.py +++ b/modules/ui_components.py @@ -48,3 +48,11 @@ class FormColorPicker(gr.ColorPicker, gr.components.FormComponent): def get_block_name(self): return "colorpicker" + +class DropdownMulti(gr.Dropdown): + """Same as gr.Dropdown but always multiselect""" + def __init__(self, **kwargs): + super().__init__(multiselect=True, **kwargs) + + def get_block_name(self): + return "dropdown" -- cgit v1.2.3 From d2ac95fa7b2a8d0bcc5361ee16dba9cbb81ff8b2 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 27 Jan 2023 11:28:12 +0300 Subject: remove the need to place configs near models --- modules/api/api.py | 5 +- modules/devices.py | 12 ++- modules/sd_hijack_inpainting.py | 9 -- modules/sd_models.py | 228 ++++++++++++++++++++-------------------- modules/sd_models_config.py | 65 ++++++++++++ modules/shared.py | 7 +- modules/shared_items.py | 15 ++- modules/timer.py | 35 ++++++ 8 files changed, 242 insertions(+), 134 deletions(-) create mode 100644 modules/sd_models_config.py create mode 100644 modules/timer.py (limited to 'modules') diff --git a/modules/api/api.py b/modules/api/api.py index 25c65e57..eb7b1da5 100644 --- a/modules/api/api.py +++ b/modules/api/api.py @@ -18,7 +18,8 @@ from modules.textual_inversion.textual_inversion import create_embedding, train_ from modules.textual_inversion.preprocess import preprocess from modules.hypernetworks.hypernetwork import create_hypernetwork, train_hypernetwork from PIL import PngImagePlugin,Image -from modules.sd_models import checkpoints_list, find_checkpoint_config +from modules.sd_models import checkpoints_list +from modules.sd_models_config import find_checkpoint_config_near_filename from modules.realesrgan_model import get_realesrgan_models from modules import devices from typing import List @@ -387,7 +388,7 @@ class Api: ] def get_sd_models(self): - return [{"title": x.title, "model_name": x.model_name, "hash": x.shorthash, "sha256": x.sha256, "filename": x.filename, "config": find_checkpoint_config(x)} for x in checkpoints_list.values()] + return [{"title": x.title, "model_name": x.model_name, "hash": x.shorthash, "sha256": x.sha256, "filename": x.filename, "config": find_checkpoint_config_near_filename(x)} for x in checkpoints_list.values()] def get_hypernetworks(self): return [{"name": name, "path": shared.hypernetworks[name]} for name in shared.hypernetworks] diff --git a/modules/devices.py b/modules/devices.py index 6b36622c..2d5f797a 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -34,14 +34,18 @@ def get_cuda_device_string(): return "cuda" -def get_optimal_device(): +def get_optimal_device_name(): if torch.cuda.is_available(): - return torch.device(get_cuda_device_string()) + return get_cuda_device_string() if has_mps(): - return torch.device("mps") + return "mps" + + return "cpu" - return cpu + +def get_optimal_device(): + return torch.device(get_optimal_device_name()) def get_device_for(task): diff --git a/modules/sd_hijack_inpainting.py b/modules/sd_hijack_inpainting.py index 31d2c898..478cd499 100644 --- a/modules/sd_hijack_inpainting.py +++ b/modules/sd_hijack_inpainting.py @@ -96,15 +96,6 @@ def p_sample_plms(self, x, c, t, index, repeat_noise=False, use_original_steps=F return x_prev, pred_x0, e_t -def should_hijack_inpainting(checkpoint_info): - from modules import sd_models - - ckpt_basename = os.path.basename(checkpoint_info.filename).lower() - cfg_basename = os.path.basename(sd_models.find_checkpoint_config(checkpoint_info)).lower() - - return "inpainting" in ckpt_basename and not "inpainting" in cfg_basename - - def do_inpainting_hijack(): # p_sample_plms is needed because PLMS can't work with dicts as conditionings diff --git a/modules/sd_models.py b/modules/sd_models.py index 7072eb2e..fa208728 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -2,8 +2,6 @@ import collections import os.path import sys import gc -import time -from collections import namedtuple import torch import re import safetensors.torch @@ -14,10 +12,10 @@ import ldm.modules.midas as midas from ldm.util import instantiate_from_config -from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors, hashes +from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors, hashes, sd_models_config from modules.paths import models_path -from modules.sd_hijack_inpainting import do_inpainting_hijack, should_hijack_inpainting -from modules.sd_hijack_ip2p import should_hijack_ip2p +from modules.sd_hijack_inpainting import do_inpainting_hijack +from modules.timer import Timer model_dir = "Stable-diffusion" model_path = os.path.abspath(os.path.join(models_path, model_dir)) @@ -99,17 +97,6 @@ def checkpoint_tiles(): return sorted([x.title for x in checkpoints_list.values()], key=alphanumeric_key) -def find_checkpoint_config(info): - if info is None: - return shared.cmd_opts.config - - config = os.path.splitext(info.filename)[0] + ".yaml" - if os.path.exists(config): - return config - - return shared.cmd_opts.config - - def list_models(): checkpoints_list.clear() checkpoint_alisases.clear() @@ -215,9 +202,7 @@ def get_state_dict_from_checkpoint(pl_sd): def read_state_dict(checkpoint_file, print_global_state=False, map_location=None): _, extension = os.path.splitext(checkpoint_file) if extension.lower() == ".safetensors": - device = map_location or shared.weight_load_location - if device is None: - device = devices.get_cuda_device_string() if torch.cuda.is_available() else "cpu" + device = map_location or shared.weight_load_location or devices.get_optimal_device_name() pl_sd = safetensors.torch.load_file(checkpoint_file, device=device) else: pl_sd = torch.load(checkpoint_file, map_location=map_location or shared.weight_load_location) @@ -229,60 +214,74 @@ def read_state_dict(checkpoint_file, print_global_state=False, map_location=None return sd -def load_model_weights(model, checkpoint_info: CheckpointInfo): +def get_checkpoint_state_dict(checkpoint_info: CheckpointInfo, timer): + sd_model_hash = checkpoint_info.calculate_shorthash() + timer.record("calculate hash") + + if checkpoint_info in checkpoints_loaded: + # use checkpoint cache + print(f"Loading weights [{sd_model_hash}] from cache") + return checkpoints_loaded[checkpoint_info] + + print(f"Loading weights [{sd_model_hash}] from {checkpoint_info.filename}") + res = read_state_dict(checkpoint_info.filename) + timer.record("load weights from disk") + + return res + + +def load_model_weights(model, checkpoint_info: CheckpointInfo, state_dict, timer): title = checkpoint_info.title sd_model_hash = checkpoint_info.calculate_shorthash() + timer.record("calculate hash") + if checkpoint_info.title != title: shared.opts.data["sd_model_checkpoint"] = checkpoint_info.title - cache_enabled = shared.opts.sd_checkpoint_cache > 0 + if state_dict is None: + state_dict = get_checkpoint_state_dict(checkpoint_info, timer) - if cache_enabled and checkpoint_info in checkpoints_loaded: - # use checkpoint cache - print(f"Loading weights [{sd_model_hash}] from cache") - model.load_state_dict(checkpoints_loaded[checkpoint_info]) - else: - # load from file - print(f"Loading weights [{sd_model_hash}] from {checkpoint_info.filename}") + model.load_state_dict(state_dict, strict=False) + del state_dict + timer.record("apply weights to model") - sd = read_state_dict(checkpoint_info.filename) - model.load_state_dict(sd, strict=False) - del sd - - if cache_enabled: - # cache newly loaded model - checkpoints_loaded[checkpoint_info] = model.state_dict().copy() + if shared.opts.sd_checkpoint_cache > 0: + # cache newly loaded model + checkpoints_loaded[checkpoint_info] = model.state_dict().copy() + + if shared.cmd_opts.opt_channelslast: + model.to(memory_format=torch.channels_last) + timer.record("apply channels_last") - if shared.cmd_opts.opt_channelslast: - model.to(memory_format=torch.channels_last) + if not shared.cmd_opts.no_half: + vae = model.first_stage_model + depth_model = getattr(model, 'depth_model', None) - if not shared.cmd_opts.no_half: - vae = model.first_stage_model - depth_model = getattr(model, 'depth_model', None) + # with --no-half-vae, remove VAE from model when doing half() to prevent its weights from being converted to float16 + if shared.cmd_opts.no_half_vae: + model.first_stage_model = None + # with --upcast-sampling, don't convert the depth model weights to float16 + if shared.cmd_opts.upcast_sampling and depth_model: + model.depth_model = None - # with --no-half-vae, remove VAE from model when doing half() to prevent its weights from being converted to float16 - if shared.cmd_opts.no_half_vae: - model.first_stage_model = None - # with --upcast-sampling, don't convert the depth model weights to float16 - if shared.cmd_opts.upcast_sampling and depth_model: - model.depth_model = None + model.half() + model.first_stage_model = vae + if depth_model: + model.depth_model = depth_model - model.half() - model.first_stage_model = vae - if depth_model: - model.depth_model = depth_model + timer.record("apply half()") - devices.dtype = torch.float32 if shared.cmd_opts.no_half else torch.float16 - devices.dtype_vae = torch.float32 if shared.cmd_opts.no_half or shared.cmd_opts.no_half_vae else torch.float16 - devices.dtype_unet = model.model.diffusion_model.dtype - devices.unet_needs_upcast = shared.cmd_opts.upcast_sampling and devices.dtype == torch.float16 and devices.dtype_unet == torch.float16 + devices.dtype = torch.float32 if shared.cmd_opts.no_half else torch.float16 + devices.dtype_vae = torch.float32 if shared.cmd_opts.no_half or shared.cmd_opts.no_half_vae else torch.float16 + devices.dtype_unet = model.model.diffusion_model.dtype + devices.unet_needs_upcast = shared.cmd_opts.upcast_sampling and devices.dtype == torch.float16 and devices.dtype_unet == torch.float16 - model.first_stage_model.to(devices.dtype_vae) + model.first_stage_model.to(devices.dtype_vae) + timer.record("apply dtype to VAE") # clean up cache if limit is reached - if cache_enabled: - while len(checkpoints_loaded) > shared.opts.sd_checkpoint_cache + 1: # we need to count the current model - checkpoints_loaded.popitem(last=False) # LRU + while len(checkpoints_loaded) > shared.opts.sd_checkpoint_cache: + checkpoints_loaded.popitem(last=False) model.sd_model_hash = sd_model_hash model.sd_model_checkpoint = checkpoint_info.filename @@ -295,6 +294,7 @@ def load_model_weights(model, checkpoint_info: CheckpointInfo): sd_vae.clear_loaded_vae() vae_file, vae_source = sd_vae.resolve_vae(checkpoint_info.filename) sd_vae.load_vae(model, vae_file, vae_source) + timer.record("load VAE") def enable_midas_autodownload(): @@ -340,24 +340,20 @@ def enable_midas_autodownload(): midas.api.load_model = load_model_wrapper -class Timer: - def __init__(self): - self.start = time.time() +def repair_config(sd_config): - def elapsed(self): - end = time.time() - res = end - self.start - self.start = end - return res + if not hasattr(sd_config.model.params, "use_ema"): + sd_config.model.params.use_ema = False + if shared.cmd_opts.no_half: + sd_config.model.params.unet_config.params.use_fp16 = False + elif shared.cmd_opts.upcast_sampling: + sd_config.model.params.unet_config.params.use_fp16 = True -def load_model(checkpoint_info=None): + +def load_model(checkpoint_info=None, already_loaded_state_dict=None, time_taken_to_load_state_dict=None): from modules import lowvram, sd_hijack checkpoint_info = checkpoint_info or select_checkpoint() - checkpoint_config = find_checkpoint_config(checkpoint_info) - - if checkpoint_config != shared.cmd_opts.config: - print(f"Loading config from: {checkpoint_config}") if shared.sd_model: sd_hijack.model_hijack.undo_hijack(shared.sd_model) @@ -365,38 +361,27 @@ def load_model(checkpoint_info=None): gc.collect() devices.torch_gc() - sd_config = OmegaConf.load(checkpoint_config) - - if should_hijack_inpainting(checkpoint_info): - # Hardcoded config for now... - sd_config.model.target = "ldm.models.diffusion.ddpm.LatentInpaintDiffusion" - sd_config.model.params.conditioning_key = "hybrid" - sd_config.model.params.unet_config.params.in_channels = 9 - sd_config.model.params.finetune_keys = None - - if should_hijack_ip2p(checkpoint_info): - sd_config.model.target = "modules.models.diffusion.ddpm_edit.LatentDiffusion" - sd_config.model.params.conditioning_key = "hybrid" - sd_config.model.params.first_stage_key = "edited" - sd_config.model.params.cond_stage_key = "edit" - sd_config.model.params.image_size = 16 - sd_config.model.params.unet_config.params.in_channels = 8 - sd_config.model.params.unet_config.params.out_channels = 4 + do_inpainting_hijack() - if not hasattr(sd_config.model.params, "use_ema"): - sd_config.model.params.use_ema = False + timer = Timer() - do_inpainting_hijack() + if already_loaded_state_dict is not None: + state_dict = already_loaded_state_dict + else: + state_dict = get_checkpoint_state_dict(checkpoint_info, timer) - if shared.cmd_opts.no_half: - sd_config.model.params.unet_config.params.use_fp16 = False - elif shared.cmd_opts.upcast_sampling: - sd_config.model.params.unet_config.params.use_fp16 = True + checkpoint_config = sd_models_config.find_checkpoint_config(state_dict, checkpoint_info) - timer = Timer() + timer.record("find config") - sd_model = None + sd_config = OmegaConf.load(checkpoint_config) + repair_config(sd_config) + + timer.record("load config") + + print(f"Creating model from config: {checkpoint_config}") + sd_model = None try: with sd_disable_initialization.DisableInitialization(): sd_model = instantiate_from_config(sd_config.model) @@ -407,29 +392,35 @@ def load_model(checkpoint_info=None): print('Failed to create model quickly; will retry using slow method.', file=sys.stderr) sd_model = instantiate_from_config(sd_config.model) - elapsed_create = timer.elapsed() + sd_model.used_config = checkpoint_config - load_model_weights(sd_model, checkpoint_info) + timer.record("create model") - elapsed_load_weights = timer.elapsed() + load_model_weights(sd_model, checkpoint_info, state_dict, timer) if shared.cmd_opts.lowvram or shared.cmd_opts.medvram: lowvram.setup_for_low_vram(sd_model, shared.cmd_opts.medvram) else: sd_model.to(shared.device) + timer.record("move model to device") + sd_hijack.model_hijack.hijack(sd_model) + timer.record("hijack") + sd_model.eval() shared.sd_model = sd_model sd_hijack.model_hijack.embedding_db.load_textual_inversion_embeddings(force_reload=True) # Reload embeddings after model load as they may or may not fit the model + timer.record("load textual inversion embeddings") + script_callbacks.model_loaded_callback(sd_model) - elapsed_the_rest = timer.elapsed() + timer.record("scripts callbacks") - print(f"Model loaded in {elapsed_create + elapsed_load_weights + elapsed_the_rest:.1f}s ({elapsed_create:.1f}s create model, {elapsed_load_weights:.1f}s load weights).") + print(f"Model loaded in {timer.summary()}.") return sd_model @@ -440,6 +431,7 @@ def reload_model_weights(sd_model=None, info=None): if not sd_model: sd_model = shared.sd_model + if sd_model is None: # previous model load failed current_checkpoint_info = None else: @@ -447,14 +439,6 @@ def reload_model_weights(sd_model=None, info=None): if sd_model.sd_model_checkpoint == checkpoint_info.filename: return - checkpoint_config = find_checkpoint_config(current_checkpoint_info) - - if current_checkpoint_info is None or checkpoint_config != find_checkpoint_config(checkpoint_info) or should_hijack_inpainting(checkpoint_info) != should_hijack_inpainting(sd_model.sd_checkpoint_info) or should_hijack_ip2p(checkpoint_info) != should_hijack_ip2p(sd_model.sd_checkpoint_info): - del sd_model - checkpoints_loaded.clear() - load_model(checkpoint_info) - return shared.sd_model - if shared.cmd_opts.lowvram or shared.cmd_opts.medvram: lowvram.send_everything_to_cpu() else: @@ -464,21 +448,35 @@ def reload_model_weights(sd_model=None, info=None): timer = Timer() + state_dict = get_checkpoint_state_dict(checkpoint_info, timer) + + checkpoint_config = sd_models_config.find_checkpoint_config(state_dict, checkpoint_info) + + timer.record("find config") + + if sd_model is None or checkpoint_config != sd_model.used_config: + del sd_model + checkpoints_loaded.clear() + load_model(checkpoint_info, already_loaded_state_dict=state_dict, time_taken_to_load_state_dict=timer.records["load weights from disk"]) + return shared.sd_model + try: - load_model_weights(sd_model, checkpoint_info) + load_model_weights(sd_model, checkpoint_info, state_dict, timer) except Exception as e: print("Failed to load checkpoint, restoring previous") - load_model_weights(sd_model, current_checkpoint_info) + load_model_weights(sd_model, current_checkpoint_info, None, timer) raise finally: sd_hijack.model_hijack.hijack(sd_model) + timer.record("hijack") + script_callbacks.model_loaded_callback(sd_model) + timer.record("script callbacks") if not shared.cmd_opts.lowvram and not shared.cmd_opts.medvram: sd_model.to(devices.device) + timer.record("move model to device") - elapsed = timer.elapsed() - - print(f"Weights loaded in {elapsed:.1f}s.") + print(f"Weights loaded in {timer.summary()}.") return sd_model diff --git a/modules/sd_models_config.py b/modules/sd_models_config.py new file mode 100644 index 00000000..ea773a10 --- /dev/null +++ b/modules/sd_models_config.py @@ -0,0 +1,65 @@ +import re +import os + +from modules import shared, paths + +sd_configs_path = shared.sd_configs_path +sd_repo_configs_path = os.path.join(paths.paths['Stable Diffusion'], "configs", "stable-diffusion") + + +config_default = shared.sd_default_config +config_sd2 = os.path.join(sd_repo_configs_path, "v2-inference.yaml") +config_sd2v = os.path.join(sd_repo_configs_path, "v2-inference-v.yaml") +config_inpainting = os.path.join(sd_configs_path, "v1-inpainting-inference.yaml") +config_instruct_pix2pix = os.path.join(sd_configs_path, "instruct-pix2pix.yaml") +config_alt_diffusion = os.path.join(sd_configs_path, "alt-diffusion-inference.yaml") + +re_parametrization_v = re.compile(r'-v\b') + + +def guess_model_config_from_state_dict(sd, filename): + fn = os.path.basename(filename) + + sd2_cond_proj_weight = sd.get('cond_stage_model.model.transformer.resblocks.0.attn.in_proj_weight', None) + diffusion_model_input = sd.get('model.diffusion_model.input_blocks.0.0.weight', None) + roberta_weight = sd.get('cond_stage_model.roberta.embeddings.word_embeddings.weight', None) + + if sd2_cond_proj_weight is not None and sd2_cond_proj_weight.shape[1] == 1024: + if re.search(re_parametrization_v, fn) or "v2-1_768" in fn: + return config_sd2v + else: + return config_sd2 + + if diffusion_model_input is not None: + if diffusion_model_input.shape[1] == 9: + return config_inpainting + if diffusion_model_input.shape[1] == 8: + return config_instruct_pix2pix + + if roberta_weight is not None: + return config_alt_diffusion + + return config_default + + +def find_checkpoint_config(state_dict, info): + if info is None: + return guess_model_config_from_state_dict(state_dict, "") + + config = find_checkpoint_config_near_filename(info) + if config is not None: + return config + + return guess_model_config_from_state_dict(state_dict, info.filename) + + +def find_checkpoint_config_near_filename(info): + if info is None: + return None + + config = os.path.splitext(info.filename)[0] + ".yaml" + if os.path.exists(config): + return config + + return None + diff --git a/modules/shared.py b/modules/shared.py index cdeed55d..14be993d 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -13,13 +13,14 @@ import modules.interrogate import modules.memmon import modules.styles import modules.devices as devices -from modules import localization, sd_vae, extensions, script_loading, errors, ui_components, shared_items +from modules import localization, extensions, script_loading, errors, ui_components, shared_items from modules.paths import models_path, script_path demo = None -sd_default_config = os.path.join(script_path, "configs/v1-inference.yaml") +sd_configs_path = os.path.join(script_path, "configs") +sd_default_config = os.path.join(sd_configs_path, "v1-inference.yaml") sd_model_file = os.path.join(script_path, 'model.ckpt') default_sd_model_file = sd_model_file @@ -391,7 +392,7 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), { "sd_model_checkpoint": OptionInfo(None, "Stable Diffusion checkpoint", gr.Dropdown, lambda: {"choices": list_checkpoint_tiles()}, refresh=refresh_checkpoints), "sd_checkpoint_cache": OptionInfo(0, "Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), "sd_vae_checkpoint_cache": OptionInfo(0, "VAE Checkpoints to cache in RAM", gr.Slider, {"minimum": 0, "maximum": 10, "step": 1}), - "sd_vae": OptionInfo("Automatic", "SD VAE", gr.Dropdown, lambda: {"choices": ["Automatic", "None"] + list(sd_vae.vae_dict)}, refresh=sd_vae.refresh_vae_list), + "sd_vae": OptionInfo("Automatic", "SD VAE", gr.Dropdown, lambda: {"choices": shared_items.sd_vae_items()}, refresh=shared_items.refresh_vae_list), "sd_vae_as_default": OptionInfo(True, "Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them"), "inpainting_mask_weight": OptionInfo(1.0, "Inpainting conditioning mask strength", gr.Slider, {"minimum": 0.0, "maximum": 1.0, "step": 0.01}), "initial_noise_multiplier": OptionInfo(1.0, "Noise multiplier for img2img", gr.Slider, {"minimum": 0.5, "maximum": 1.5, "step": 0.01}), diff --git a/modules/shared_items.py b/modules/shared_items.py index b5d480c9..8b5ec96d 100644 --- a/modules/shared_items.py +++ b/modules/shared_items.py @@ -4,7 +4,20 @@ def realesrgan_models_names(): import modules.realesrgan_model return [x.name for x in modules.realesrgan_model.get_realesrgan_models(None)] + def postprocessing_scripts(): import modules.scripts - return modules.scripts.scripts_postproc.scripts \ No newline at end of file + return modules.scripts.scripts_postproc.scripts + + +def sd_vae_items(): + import modules.sd_vae + + return ["Automatic", "None"] + list(modules.sd_vae.vae_dict) + + +def refresh_vae_list(): + import modules.sd_vae + + return modules.sd_vae.refresh_vae_list diff --git a/modules/timer.py b/modules/timer.py new file mode 100644 index 00000000..57a4f17a --- /dev/null +++ b/modules/timer.py @@ -0,0 +1,35 @@ +import time + + +class Timer: + def __init__(self): + self.start = time.time() + self.records = {} + self.total = 0 + + def elapsed(self): + end = time.time() + res = end - self.start + self.start = end + return res + + def record(self, category, extra_time=0): + e = self.elapsed() + if category not in self.records: + self.records[category] = 0 + + self.records[category] += e + extra_time + self.total += e + extra_time + + def summary(self): + res = f"{self.total:.1f}s" + + additions = [x for x in self.records.items() if x[1] >= 0.1] + if not additions: + return res + + res += " (" + res += ", ".join([f"{category}: {time_taken:.1f}s" for category, time_taken in additions]) + res += ")" + + return res -- cgit v1.2.3 From 6f31d2210c189f8db118e6f95add7ba2a64f0238 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 27 Jan 2023 11:54:19 +0300 Subject: support detecting midas model fix broken api for checkpoint list --- modules/api/models.py | 2 +- modules/sd_models.py | 10 +++++----- modules/sd_models_config.py | 7 +++++-- 3 files changed, 11 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/api/models.py b/modules/api/models.py index 805bd8f7..cba43d3b 100644 --- a/modules/api/models.py +++ b/modules/api/models.py @@ -228,7 +228,7 @@ class SDModelItem(BaseModel): hash: Optional[str] = Field(title="Short hash") sha256: Optional[str] = Field(title="sha256 hash") filename: str = Field(title="Filename") - config: str = Field(title="Config file") + config: Optional[str] = Field(title="Config file") class HypernetworkItem(BaseModel): name: str = Field(title="Name") diff --git a/modules/sd_models.py b/modules/sd_models.py index fa208728..37dad18d 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -439,12 +439,12 @@ def reload_model_weights(sd_model=None, info=None): if sd_model.sd_model_checkpoint == checkpoint_info.filename: return - if shared.cmd_opts.lowvram or shared.cmd_opts.medvram: - lowvram.send_everything_to_cpu() - else: - sd_model.to(devices.cpu) + if shared.cmd_opts.lowvram or shared.cmd_opts.medvram: + lowvram.send_everything_to_cpu() + else: + sd_model.to(devices.cpu) - sd_hijack.model_hijack.undo_hijack(sd_model) + sd_hijack.model_hijack.undo_hijack(sd_model) timer = Timer() diff --git a/modules/sd_models_config.py b/modules/sd_models_config.py index ea773a10..4d1e92e1 100644 --- a/modules/sd_models_config.py +++ b/modules/sd_models_config.py @@ -10,6 +10,7 @@ sd_repo_configs_path = os.path.join(paths.paths['Stable Diffusion'], "configs", config_default = shared.sd_default_config config_sd2 = os.path.join(sd_repo_configs_path, "v2-inference.yaml") config_sd2v = os.path.join(sd_repo_configs_path, "v2-inference-v.yaml") +config_depth_model = os.path.join(sd_repo_configs_path, "v2-midas-inference.yaml") config_inpainting = os.path.join(sd_configs_path, "v1-inpainting-inference.yaml") config_instruct_pix2pix = os.path.join(sd_configs_path, "instruct-pix2pix.yaml") config_alt_diffusion = os.path.join(sd_configs_path, "alt-diffusion-inference.yaml") @@ -22,7 +23,9 @@ def guess_model_config_from_state_dict(sd, filename): sd2_cond_proj_weight = sd.get('cond_stage_model.model.transformer.resblocks.0.attn.in_proj_weight', None) diffusion_model_input = sd.get('model.diffusion_model.input_blocks.0.0.weight', None) - roberta_weight = sd.get('cond_stage_model.roberta.embeddings.word_embeddings.weight', None) + + if sd.get('depth_model.model.pretrained.act_postprocess3.0.project.0.bias', None) is not None: + return config_depth_model if sd2_cond_proj_weight is not None and sd2_cond_proj_weight.shape[1] == 1024: if re.search(re_parametrization_v, fn) or "v2-1_768" in fn: @@ -36,7 +39,7 @@ def guess_model_config_from_state_dict(sd, filename): if diffusion_model_input.shape[1] == 8: return config_instruct_pix2pix - if roberta_weight is not None: + if sd.get('cond_stage_model.roberta.embeddings.word_embeddings.weight', None) is not None: return config_alt_diffusion return config_default -- cgit v1.2.3 From 9beb794e0b0dc1a0f9e89d8e38bd789a8c608397 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 27 Jan 2023 13:08:00 +0300 Subject: clarify the option to disable NaN check. --- modules/devices.py | 2 ++ 1 file changed, 2 insertions(+) (limited to 'modules') diff --git a/modules/devices.py b/modules/devices.py index 2d5f797a..4687944e 100644 --- a/modules/devices.py +++ b/modules/devices.py @@ -143,6 +143,8 @@ def test_for_nans(x, where): else: message = "A tensor with all NaNs was produced." + message += " Use --disable-nan-check commandline argument to disable this check." + raise NansException(message) -- cgit v1.2.3 From 5eee2ac39863f9e44591b50d0710dd2615416a13 Mon Sep 17 00:00:00 2001 From: Max Audron Date: Wed, 25 Jan 2023 17:15:42 +0100 Subject: add data-dir flag and set all user data directories based on it --- modules/extensions.py | 2 +- modules/generation_parameters_copypaste.py | 4 ++-- modules/gfpgan_model.py | 5 ++--- modules/hashes.py | 4 +++- modules/interrogate.py | 2 +- modules/paths.py | 10 +++++++++- modules/processing.py | 3 ++- modules/sd_models.py | 6 +++--- modules/sd_vae.py | 5 ++--- modules/shared.py | 11 ++++++----- modules/textual_inversion/preprocess.py | 5 ++--- modules/ui.py | 6 +++--- modules/ui_extensions.py | 2 +- modules/upscaler.py | 5 ++--- 14 files changed, 39 insertions(+), 31 deletions(-) (limited to 'modules') diff --git a/modules/extensions.py b/modules/extensions.py index b522125c..92ee8144 100644 --- a/modules/extensions.py +++ b/modules/extensions.py @@ -7,7 +7,7 @@ import git from modules import paths, shared extensions = [] -extensions_dir = os.path.join(paths.script_path, "extensions") +extensions_dir = os.path.join(paths.data_path, "extensions") extensions_builtin_dir = os.path.join(paths.script_path, "extensions-builtin") diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 46e12dc6..35f72808 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -6,7 +6,7 @@ import re from pathlib import Path import gradio as gr -from modules.shared import script_path +from modules.paths import data_path, script_path from modules import shared, ui_tempdir, script_callbacks import tempfile from PIL import Image @@ -289,7 +289,7 @@ Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 965400086, Size: 512x512, Model def connect_paste(button, paste_fields, input_comp, jsfunc=None): def paste_func(prompt): if not prompt and not shared.cmd_opts.hide_ui_dir_config: - filename = os.path.join(script_path, "params.txt") + filename = os.path.join(data_path, "params.txt") if os.path.exists(filename): with open(filename, "r", encoding="utf8") as file: prompt = file.read() diff --git a/modules/gfpgan_model.py b/modules/gfpgan_model.py index 1e2dbc32..fbe6215a 100644 --- a/modules/gfpgan_model.py +++ b/modules/gfpgan_model.py @@ -6,12 +6,11 @@ import facexlib import gfpgan import modules.face_restoration -from modules import shared, devices, modelloader -from modules.paths import models_path +from modules import paths, shared, devices, modelloader model_dir = "GFPGAN" user_path = None -model_path = os.path.join(models_path, model_dir) +model_path = os.path.join(paths.models_path, model_dir) model_url = "https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.4.pth" have_gfpgan = False loaded_gfpgan_model = None diff --git a/modules/hashes.py b/modules/hashes.py index b85a7580..819362a3 100644 --- a/modules/hashes.py +++ b/modules/hashes.py @@ -4,8 +4,10 @@ import os.path import filelock +from modules.paths import data_path -cache_filename = "cache.json" + +cache_filename = os.path.join(data_path, "cache.json") cache_data = None diff --git a/modules/interrogate.py b/modules/interrogate.py index c72ff694..cbb80683 100644 --- a/modules/interrogate.py +++ b/modules/interrogate.py @@ -12,7 +12,7 @@ from torchvision import transforms from torchvision.transforms.functional import InterpolationMode import modules.shared as shared -from modules import devices, paths, lowvram, modelloader, errors +from modules import devices, paths, shared, lowvram, modelloader, errors blip_image_eval_size = 384 clip_model_name = 'ViT-L/14' diff --git a/modules/paths.py b/modules/paths.py index 20b3e4d8..08e6f9b9 100644 --- a/modules/paths.py +++ b/modules/paths.py @@ -4,7 +4,15 @@ import sys import modules.safe script_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__))) -models_path = os.path.join(script_path, "models") + +# Parse the --data-dir flag first so we can use it as a base for our other argument default values +parser = argparse.ArgumentParser() +parser.add_argument("--data-dir", type=str, default=os.path.dirname(os.path.dirname(os.path.realpath(__file__))), help="base path where all user data is stored",) +cmd_opts_pre = parser.parse_known_args()[0] +data_path = cmd_opts_pre.data_dir +models_path = os.path.join(data_path, "models") + +# data_path = cmd_opts_pre.data sys.path.insert(0, script_path) # search for directory of stable diffusion in following places diff --git a/modules/processing.py b/modules/processing.py index 262806a1..5072fc40 100644 --- a/modules/processing.py +++ b/modules/processing.py @@ -17,6 +17,7 @@ from modules import devices, prompt_parser, masking, sd_samplers, lowvram, gener from modules.sd_hijack import model_hijack from modules.shared import opts, cmd_opts, state import modules.shared as shared +import modules.paths as paths import modules.face_restoration import modules.images as images import modules.styles @@ -584,7 +585,7 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed: if not p.disable_extra_networks: extra_networks.activate(p, extra_network_data) - with open(os.path.join(shared.script_path, "params.txt"), "w", encoding="utf8") as file: + with open(os.path.join(paths.data_path, "params.txt"), "w", encoding="utf8") as file: processed = Processed(p, [], p.seed, "") file.write(processed.infotext(p, 0)) diff --git a/modules/sd_models.py b/modules/sd_models.py index 37dad18d..b2d48a51 100644 --- a/modules/sd_models.py +++ b/modules/sd_models.py @@ -12,13 +12,13 @@ import ldm.modules.midas as midas from ldm.util import instantiate_from_config -from modules import shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors, hashes, sd_models_config +from modules import paths, shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors, hashes, sd_models_config from modules.paths import models_path from modules.sd_hijack_inpainting import do_inpainting_hijack from modules.timer import Timer model_dir = "Stable-diffusion" -model_path = os.path.abspath(os.path.join(models_path, model_dir)) +model_path = os.path.abspath(os.path.join(paths.models_path, model_dir)) checkpoints_list = {} checkpoint_alisases = {} @@ -307,7 +307,7 @@ def enable_midas_autodownload(): location automatically. """ - midas_path = os.path.join(models_path, 'midas') + midas_path = os.path.join(paths.models_path, 'midas') # stable-diffusion-stability-ai hard-codes the midas model path to # a location that differs from where other scripts using this model look. diff --git a/modules/sd_vae.py b/modules/sd_vae.py index 4ce238b8..9b00f76e 100644 --- a/modules/sd_vae.py +++ b/modules/sd_vae.py @@ -3,13 +3,12 @@ import safetensors.torch import os import collections from collections import namedtuple -from modules import shared, devices, script_callbacks, sd_models -from modules.paths import models_path +from modules import paths, shared, devices, script_callbacks, sd_models import glob from copy import deepcopy -vae_path = os.path.abspath(os.path.join(models_path, "VAE")) +vae_path = os.path.abspath(os.path.join(paths.models_path, "VAE")) vae_ignore_keys = {"model_ema.decay", "model_ema.num_updates"} vae_dict = {} diff --git a/modules/shared.py b/modules/shared.py index 14be993d..474fcc42 100644 --- a/modules/shared.py +++ b/modules/shared.py @@ -14,7 +14,7 @@ import modules.memmon import modules.styles import modules.devices as devices from modules import localization, extensions, script_loading, errors, ui_components, shared_items -from modules.paths import models_path, script_path +from modules.paths import models_path, script_path, data_path demo = None @@ -25,6 +25,7 @@ sd_model_file = os.path.join(script_path, 'model.ckpt') default_sd_model_file = sd_model_file parser = argparse.ArgumentParser() +parser.add_argument("--data-dir", type=str, default=os.path.dirname(os.path.dirname(os.path.realpath(__file__))), help="base path where all user data is stored",) parser.add_argument("--config", type=str, default=sd_default_config, help="path to config which constructs model",) parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; if specified, this checkpoint will be added to the list of checkpoints and loaded",) parser.add_argument("--ckpt-dir", type=str, default=None, help="Path to directory with stable diffusion checkpoints") @@ -35,7 +36,7 @@ parser.add_argument("--no-half", action='store_true', help="do not switch the mo parser.add_argument("--no-half-vae", action='store_true', help="do not switch the VAE model to 16-bit floats") parser.add_argument("--no-progressbar-hiding", action='store_true', help="do not hide progressbar in gradio UI (we hide it because it slows down ML if you have hardware acceleration in browser)") parser.add_argument("--max-batch-count", type=int, default=16, help="maximum batch count value for the UI") -parser.add_argument("--embeddings-dir", type=str, default=os.path.join(script_path, 'embeddings'), help="embeddings directory for textual inversion (default: embeddings)") +parser.add_argument("--embeddings-dir", type=str, default=os.path.join(data_path, 'embeddings'), help="embeddings directory for textual inversion (default: embeddings)") parser.add_argument("--textual-inversion-templates-dir", type=str, default=os.path.join(script_path, 'textual_inversion_templates'), help="directory with textual inversion templates") parser.add_argument("--hypernetwork-dir", type=str, default=os.path.join(models_path, 'hypernetworks'), help="hypernetwork directory") parser.add_argument("--localizations-dir", type=str, default=os.path.join(script_path, 'localizations'), help="localizations directory") @@ -74,16 +75,16 @@ parser.add_argument("--use-cpu", nargs='+', help="use CPU as torch device for sp parser.add_argument("--listen", action='store_true', help="launch gradio with 0.0.0.0 as server name, allowing to respond to network requests") parser.add_argument("--port", type=int, help="launch gradio with given server port, you need root/admin rights for ports < 1024, defaults to 7860 if available", default=None) parser.add_argument("--show-negative-prompt", action='store_true', help="does not do anything", default=False) -parser.add_argument("--ui-config-file", type=str, help="filename to use for ui configuration", default=os.path.join(script_path, 'ui-config.json')) +parser.add_argument("--ui-config-file", type=str, help="filename to use for ui configuration", default=os.path.join(data_path, 'ui-config.json')) parser.add_argument("--hide-ui-dir-config", action='store_true', help="hide directory configuration from webui", default=False) parser.add_argument("--freeze-settings", action='store_true', help="disable editing settings", default=False) -parser.add_argument("--ui-settings-file", type=str, help="filename to use for ui settings", default=os.path.join(script_path, 'config.json')) +parser.add_argument("--ui-settings-file", type=str, help="filename to use for ui settings", default=os.path.join(data_path, 'config.json')) parser.add_argument("--gradio-debug", action='store_true', help="launch gradio with --debug option") parser.add_argument("--gradio-auth", type=str, help='set gradio authentication like "username:password"; or comma-delimit multiple like "u1:p1,u2:p2,u3:p3"', default=None) parser.add_argument("--gradio-img2img-tool", type=str, help='does not do anything') parser.add_argument("--gradio-inpaint-tool", type=str, help="does not do anything") parser.add_argument("--opt-channelslast", action='store_true', help="change memory type for stable diffusion to channels last") -parser.add_argument("--styles-file", type=str, help="filename to use for styles", default=os.path.join(script_path, 'styles.csv')) +parser.add_argument("--styles-file", type=str, help="filename to use for styles", default=os.path.join(data_path, 'styles.csv')) parser.add_argument("--autolaunch", action='store_true', help="open the webui URL in the system's default browser upon launch", default=False) parser.add_argument("--theme", type=str, help="launches the UI with light or dark theme", default=None) parser.add_argument("--use-textbox-seed", action='store_true', help="use textbox for seeds in UI (no up/down, but possible to input long seeds)", default=False) diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py index c0ac11d3..2239cb84 100644 --- a/modules/textual_inversion/preprocess.py +++ b/modules/textual_inversion/preprocess.py @@ -6,8 +6,7 @@ import sys import tqdm import time -from modules import shared, images, deepbooru -from modules.paths import models_path +from modules import paths, shared, images, deepbooru from modules.shared import opts, cmd_opts from modules.textual_inversion import autocrop @@ -199,7 +198,7 @@ def preprocess_work(process_src, process_dst, process_width, process_height, pre dnn_model_path = None try: - dnn_model_path = autocrop.download_and_cache_models(os.path.join(models_path, "opencv")) + dnn_model_path = autocrop.download_and_cache_models(os.path.join(paths.models_path, "opencv")) except Exception as e: print("Unable to load face detection model for auto crop selection. Falling back to lower quality haar method.", e) diff --git a/modules/ui.py b/modules/ui.py index 85ae62c7..0117df3e 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -21,7 +21,7 @@ from modules.call_queue import wrap_gradio_gpu_call, wrap_queued_call, wrap_grad from modules import sd_hijack, sd_models, localization, script_callbacks, ui_extensions, deepbooru, sd_vae, extra_networks, postprocessing, ui_components, ui_common, ui_postprocessing from modules.ui_components import FormRow, FormGroup, ToolButton, FormHTML -from modules.paths import script_path +from modules.paths import script_path, data_path from modules.shared import opts, cmd_opts, restricted_opts @@ -1497,8 +1497,8 @@ def create_ui(): with open(cssfile, "r", encoding="utf8") as file: css += file.read() + "\n" - if os.path.exists(os.path.join(script_path, "user.css")): - with open(os.path.join(script_path, "user.css"), "r", encoding="utf8") as file: + if os.path.exists(os.path.join(data_path, "user.css")): + with open(os.path.join(data_path, "user.css"), "r", encoding="utf8") as file: css += file.read() + "\n" if not cmd_opts.no_progressbar_hiding: diff --git a/modules/ui_extensions.py b/modules/ui_extensions.py index 742e745e..66a41865 100644 --- a/modules/ui_extensions.py +++ b/modules/ui_extensions.py @@ -132,7 +132,7 @@ def install_extension_from_url(dirname, url): normalized_url = normalize_git_url(url) assert len([x for x in extensions.extensions if normalize_git_url(x.remote) == normalized_url]) == 0, 'Extension with this URL is already installed' - tmpdir = os.path.join(paths.script_path, "tmp", dirname) + tmpdir = os.path.join(paths.data_path, "tmp", dirname) try: shutil.rmtree(tmpdir, True) diff --git a/modules/upscaler.py b/modules/upscaler.py index a5bf5acb..e2eaa730 100644 --- a/modules/upscaler.py +++ b/modules/upscaler.py @@ -11,7 +11,6 @@ from modules import modelloader, shared LANCZOS = (Image.Resampling.LANCZOS if hasattr(Image, 'Resampling') else Image.LANCZOS) NEAREST = (Image.Resampling.NEAREST if hasattr(Image, 'Resampling') else Image.NEAREST) -from modules.paths import models_path class Upscaler: @@ -39,7 +38,7 @@ class Upscaler: self.mod_scale = None if self.model_path is None and self.name: - self.model_path = os.path.join(models_path, self.name) + self.model_path = os.path.join(shared.models_path, self.name) if self.model_path and create_dirs: os.makedirs(self.model_path, exist_ok=True) @@ -143,4 +142,4 @@ class UpscalerNearest(Upscaler): def __init__(self, dirname=None): super().__init__(False) self.name = "Nearest" - self.scalers = [UpscalerData("Nearest", None, self)] \ No newline at end of file + self.scalers = [UpscalerData("Nearest", None, self)] -- cgit v1.2.3 From 14c0884fd0948c478db165989cca7aaffc9a0504 Mon Sep 17 00:00:00 2001 From: Max Audron Date: Wed, 25 Jan 2023 17:55:59 +0100 Subject: use python importlib to load and execute extension modules previously module attributes like __file__ where not set correctly, leading to scripts getting the directory of the stable-diffusion repo location instead of their own script. This causes problem when loading user data from an external location using the --data-dir flag, as extensions would look for their own code in the stable-diffusion repo location instead of the data dir location. Using pythons importlib functions sets the modules specs correctly and executes them. But this will break extensions if they build paths based on the previously incorrect __file__ attribute. --- modules/script_loading.py | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'modules') diff --git a/modules/script_loading.py b/modules/script_loading.py index f93f0951..a7d2203f 100644 --- a/modules/script_loading.py +++ b/modules/script_loading.py @@ -1,16 +1,14 @@ import os import sys import traceback +import importlib.util from types import ModuleType def load_module(path): - with open(path, "r", encoding="utf8") as file: - text = file.read() - - compiled = compile(text, path, 'exec') - module = ModuleType(os.path.basename(path)) - exec(compiled, module.__dict__) + module_spec = importlib.util.spec_from_file_location(os.path.basename(path), path) + module = importlib.util.module_from_spec(module_spec) + module_spec.loader.exec_module(module) return module -- cgit v1.2.3 From 6b3981c0685cd1df750df4eb51823f1cfd70c6d5 Mon Sep 17 00:00:00 2001 From: Max Audron Date: Wed, 25 Jan 2023 18:00:09 +0100 Subject: clean up unused script_path imports --- modules/codeformer_model.py | 2 +- modules/generation_parameters_copypaste.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'modules') diff --git a/modules/codeformer_model.py b/modules/codeformer_model.py index ab40d842..01fb7bd8 100644 --- a/modules/codeformer_model.py +++ b/modules/codeformer_model.py @@ -8,7 +8,7 @@ import torch import modules.face_restoration import modules.shared from modules import shared, devices, modelloader -from modules.paths import script_path, models_path +from modules.paths import models_path # codeformer people made a choice to include modified basicsr library to their project which makes # it utterly impossible to use it alongside with other libraries that also use basicsr, like GFPGAN. diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 35f72808..773c5c0e 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -6,7 +6,7 @@ import re from pathlib import Path import gradio as gr -from modules.paths import data_path, script_path +from modules.paths import data_path from modules import shared, ui_tempdir, script_callbacks import tempfile from PIL import Image -- cgit v1.2.3 From 23a9d5e27390846dea0895a02c04aec9583a4d38 Mon Sep 17 00:00:00 2001 From: Max Audron Date: Wed, 25 Jan 2023 18:18:55 +0100 Subject: create user extensions directory if not exists --- modules/extensions.py | 2 ++ 1 file changed, 2 insertions(+) (limited to 'modules') diff --git a/modules/extensions.py b/modules/extensions.py index 92ee8144..5e12b1aa 100644 --- a/modules/extensions.py +++ b/modules/extensions.py @@ -10,6 +10,8 @@ extensions = [] extensions_dir = os.path.join(paths.data_path, "extensions") extensions_builtin_dir = os.path.join(paths.script_path, "extensions-builtin") +if not os.path.exists(extensions_dir): + os.makedirs(extensions_dir) def active(): return [x for x in extensions if x.enabled] -- cgit v1.2.3 From eafaf14167cf574ad0f918c10f60ef86aea9cd20 Mon Sep 17 00:00:00 2001 From: Gazzoo-byte <73721238+Gazzoo-byte@users.noreply.github.com> Date: Fri, 27 Jan 2023 18:34:41 +0000 Subject: Add button to switch width and height Adds a button to switch width and height, allowing quick and easy switching between landscape and portrait. --- modules/ui.py | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 85ae62c7..fb0e4d5c 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -91,6 +91,13 @@ save_style_symbol = '\U0001f4be' # 💾 apply_style_symbol = '\U0001f4cb' # 📋 clear_prompt_symbol = '\U0001F5D1' # 🗑️ extra_networks_symbol = '\U0001F3B4' # 🎴 +switch_values_symbol = '\U000021C5' # ⇅ + +def switch_width_and_height(width, height): + width_temp = width + width = height + height = width_temp + return width, height def plaintext_to_html(text): @@ -466,6 +473,7 @@ def create_ui(): height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="txt2img_height") if opts.dimensions_and_batch_together: + res_switch_btn = ToolButton(value=switch_values_symbol, elem_id="txt2img_res_switch_btn") with gr.Column(elem_id="txt2img_column_batch"): batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="txt2img_batch_count") batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="txt2img_batch_size") @@ -566,6 +574,7 @@ def create_ui(): txt2img_prompt.submit(**txt2img_args) submit.click(**txt2img_args) + res_switch_btn.click(switch_width_and_height, inputs=[width, height], outputs=[width, height]) txt_prompt_img.change( fn=modules.images.image_data, @@ -728,6 +737,7 @@ def create_ui(): height = gr.Slider(minimum=64, maximum=2048, step=8, label="Height", value=512, elem_id="img2img_height") if opts.dimensions_and_batch_together: + res_switch_btn = ToolButton(value=switch_values_symbol, elem_id="img2img_res_switch_btn") with gr.Column(elem_id="img2img_column_batch"): batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1, elem_id="img2img_batch_count") batch_size = gr.Slider(minimum=1, maximum=8, step=1, label='Batch size', value=1, elem_id="img2img_batch_size") @@ -865,6 +875,7 @@ def create_ui(): img2img_prompt.submit(**img2img_args) submit.click(**img2img_args) + res_switch_btn.click(switch_width_and_height, inputs=[width, height], outputs=[width, height]) img2img_interrogate.click( fn=lambda *args: process_interrogate(interrogate, *args), -- cgit v1.2.3 From cc8c9b7474d917888a0bd069fcd59a458c67ae4b Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Fri, 27 Jan 2023 22:43:08 +0300 Subject: fix broken calls to find_checkpoint_config --- modules/extras.py | 4 ++-- modules/sd_hijack_ip2p.py | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/extras.py b/modules/extras.py index 36123aa5..4f842be9 100644 --- a/modules/extras.py +++ b/modules/extras.py @@ -6,7 +6,7 @@ import shutil import torch import tqdm -from modules import shared, images, sd_models, sd_vae +from modules import shared, images, sd_models, sd_vae, sd_models_config from modules.ui_common import plaintext_to_html import gradio as gr import safetensors.torch @@ -37,7 +37,7 @@ def run_pnginfo(image): def create_config(ckpt_result, config_source, a, b, c): def config(x): - res = sd_models.find_checkpoint_config(x) if x else None + res = sd_models_config.find_checkpoint_config_near_filename(x) if x else None return res if res != shared.sd_default_config else None if config_source == 0: diff --git a/modules/sd_hijack_ip2p.py b/modules/sd_hijack_ip2p.py index 635f015f..3c727d3b 100644 --- a/modules/sd_hijack_ip2p.py +++ b/modules/sd_hijack_ip2p.py @@ -5,9 +5,9 @@ import gc import time def should_hijack_ip2p(checkpoint_info): - from modules import sd_models + from modules import sd_models_config ckpt_basename = os.path.basename(checkpoint_info.filename).lower() - cfg_basename = os.path.basename(sd_models.find_checkpoint_config(checkpoint_info)).lower() + cfg_basename = os.path.basename(sd_models_config.find_checkpoint_config_near_filename(checkpoint_info)).lower() return "pix2pix" in ckpt_basename and not "pix2pix" in cfg_basename -- cgit v1.2.3 From 6b82efd737827bbeef202f04ff5a8faec9b64ef8 Mon Sep 17 00:00:00 2001 From: MrCheeze Date: Fri, 27 Jan 2023 20:06:19 -0500 Subject: add v2-inpainting model detection, and broaden v-model detection to include anything with 768 in the name --- modules/sd_models_config.py | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_models_config.py b/modules/sd_models_config.py index 4d1e92e1..73854a45 100644 --- a/modules/sd_models_config.py +++ b/modules/sd_models_config.py @@ -10,6 +10,7 @@ sd_repo_configs_path = os.path.join(paths.paths['Stable Diffusion'], "configs", config_default = shared.sd_default_config config_sd2 = os.path.join(sd_repo_configs_path, "v2-inference.yaml") config_sd2v = os.path.join(sd_repo_configs_path, "v2-inference-v.yaml") +config_sd2_inpainting = os.path.join(sd_repo_configs_path, "v2-inpainting-inference.yaml") config_depth_model = os.path.join(sd_repo_configs_path, "v2-midas-inference.yaml") config_inpainting = os.path.join(sd_configs_path, "v1-inpainting-inference.yaml") config_instruct_pix2pix = os.path.join(sd_configs_path, "instruct-pix2pix.yaml") @@ -28,7 +29,9 @@ def guess_model_config_from_state_dict(sd, filename): return config_depth_model if sd2_cond_proj_weight is not None and sd2_cond_proj_weight.shape[1] == 1024: - if re.search(re_parametrization_v, fn) or "v2-1_768" in fn: + if diffusion_model_input.shape[1] == 9: + return config_sd2_inpainting + elif re.search(re_parametrization_v, fn) or "768" in fn: return config_sd2v else: return config_sd2 -- cgit v1.2.3 From 2aac1d97782b486f3a4a5209cf399dcdcb7bbb4d Mon Sep 17 00:00:00 2001 From: Andrii Skaliuk Date: Fri, 27 Jan 2023 17:32:31 -0800 Subject: Basic inpainting batch support Modifies batch UI to add optional inpainting support --- modules/img2img.py | 20 +++++++++++++++++--- modules/ui.py | 9 ++++++++- 2 files changed, 25 insertions(+), 4 deletions(-) (limited to 'modules') diff --git a/modules/img2img.py b/modules/img2img.py index 2168c8e2..fe9447c7 100644 --- a/modules/img2img.py +++ b/modules/img2img.py @@ -16,11 +16,16 @@ import modules.images as images import modules.scripts -def process_batch(p, input_dir, output_dir, args): +def process_batch(p, input_dir, output_dir, inpaint_mask_dir, args): processing.fix_seed(p) images = shared.listfiles(input_dir) + inpaint_masks = shared.listfiles(inpaint_mask_dir) + is_inpaint_batch = inpaint_mask_dir and len(inpaint_masks) > 0 + if is_inpaint_batch: + print(f"\nInpaint batch is enabled. {len(inpaint_masks)} masks found.") + print(f"Will process {len(images)} images, creating {p.n_iter * p.batch_size} new images for each.") save_normally = output_dir == '' @@ -43,6 +48,15 @@ def process_batch(p, input_dir, output_dir, args): img = ImageOps.exif_transpose(img) p.init_images = [img] * p.batch_size + if is_inpaint_batch: + # try to find corresponding mask for an image using simple filename matching + mask_image_path = os.path.join(inpaint_mask_dir, os.path.basename(image)) + # if not found use first one ("same mask for all images" use-case) + if not mask_image_path in inpaint_masks: + mask_image_path = inpaint_masks[0] + mask_image = Image.open(mask_image_path) + p.image_mask = mask_image + proc = modules.scripts.scripts_img2img.run(p, *args) if proc is None: proc = process_images(p) @@ -59,7 +73,7 @@ def process_batch(p, input_dir, output_dir, args): processed_image.save(os.path.join(output_dir, filename)) -def img2img(id_task: str, mode: int, prompt: str, negative_prompt: str, prompt_styles, init_img, sketch, init_img_with_mask, inpaint_color_sketch, inpaint_color_sketch_orig, init_img_inpaint, init_mask_inpaint, steps: int, sampler_index: int, mask_blur: int, mask_alpha: float, inpainting_fill: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, denoising_strength: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, resize_mode: int, inpaint_full_res: bool, inpaint_full_res_padding: int, inpainting_mask_invert: int, img2img_batch_input_dir: str, img2img_batch_output_dir: str, *args): +def img2img(id_task: str, mode: int, prompt: str, negative_prompt: str, prompt_styles, init_img, sketch, init_img_with_mask, inpaint_color_sketch, inpaint_color_sketch_orig, init_img_inpaint, init_mask_inpaint, steps: int, sampler_index: int, mask_blur: int, mask_alpha: float, inpainting_fill: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, denoising_strength: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, resize_mode: int, inpaint_full_res: bool, inpaint_full_res_padding: int, inpainting_mask_invert: int, img2img_batch_input_dir: str, img2img_batch_output_dir: str, img2img_batch_inpaint_mask_dir: str, *args): is_batch = mode == 5 if mode == 0: # img2img @@ -139,7 +153,7 @@ def img2img(id_task: str, mode: int, prompt: str, negative_prompt: str, prompt_s if is_batch: assert not shared.cmd_opts.hide_ui_dir_config, "Launched with --hide-ui-dir-config, batch img2img disabled" - process_batch(p, img2img_batch_input_dir, img2img_batch_output_dir, args) + process_batch(p, img2img_batch_input_dir, img2img_batch_output_dir, img2img_batch_inpaint_mask_dir, args) processed = Processed(p, [], p.seed, "") else: diff --git a/modules/ui.py b/modules/ui.py index 85ae62c7..fddb9177 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -691,9 +691,15 @@ def create_ui(): with gr.TabItem('Batch', id='batch', elem_id="img2img_batch_tab") as tab_batch: hidden = '
    Disabled when launched with --hide-ui-dir-config.' if shared.cmd_opts.hide_ui_dir_config else '' - gr.HTML(f"

    Process images in a directory on the same machine where the server is running.
    Use an empty output directory to save pictures normally instead of writing to the output directory.{hidden}

    ") + gr.HTML( + f"

    Process images in a directory on the same machine where the server is running." + + f"
    Use an empty output directory to save pictures normally instead of writing to the output directory." + + f"
    Add inpaint batch mask directory to enable inpaint batch processing." + f"{hidden}

    " + ) img2img_batch_input_dir = gr.Textbox(label="Input directory", **shared.hide_dirs, elem_id="img2img_batch_input_dir") img2img_batch_output_dir = gr.Textbox(label="Output directory", **shared.hide_dirs, elem_id="img2img_batch_output_dir") + img2img_batch_inpaint_mask_dir = gr.Textbox(label="Inpaint batch mask directory (required for inpaint batch processing only)", **shared.hide_dirs, elem_id="img2img_batch_inpaint_mask_dir") def copy_image(img): if isinstance(img, dict) and 'image' in img: @@ -838,6 +844,7 @@ def create_ui(): inpainting_mask_invert, img2img_batch_input_dir, img2img_batch_output_dir, + img2img_batch_inpaint_mask_dir ] + custom_inputs, outputs=[ img2img_gallery, -- cgit v1.2.3 From 4c52dfe4ac98c53431ecd267d59f27391d3a63e7 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 28 Jan 2023 08:30:17 +0300 Subject: make the detection for -v models less broad --- modules/sd_models_config.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/sd_models_config.py b/modules/sd_models_config.py index 73854a45..00217990 100644 --- a/modules/sd_models_config.py +++ b/modules/sd_models_config.py @@ -31,7 +31,7 @@ def guess_model_config_from_state_dict(sd, filename): if sd2_cond_proj_weight is not None and sd2_cond_proj_weight.shape[1] == 1024: if diffusion_model_input.shape[1] == 9: return config_sd2_inpainting - elif re.search(re_parametrization_v, fn) or "768" in fn: + elif re.search(re_parametrization_v, fn): return config_sd2v else: return config_sd2 -- cgit v1.2.3 From 0834d4ce374225131e025540220c727e352a3e43 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Sat, 28 Jan 2023 08:41:15 +0300 Subject: simplify #7284 --- modules/ui.py | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) (limited to 'modules') diff --git a/modules/ui.py b/modules/ui.py index 3c0a4050..ca2c1eb6 100644 --- a/modules/ui.py +++ b/modules/ui.py @@ -93,12 +93,6 @@ clear_prompt_symbol = '\U0001F5D1' # 🗑️ extra_networks_symbol = '\U0001F3B4' # 🎴 switch_values_symbol = '\U000021C5' # ⇅ -def switch_width_and_height(width, height): - width_temp = width - width = height - height = width_temp - return width, height - def plaintext_to_html(text): return ui_common.plaintext_to_html(text) @@ -574,7 +568,8 @@ def create_ui(): txt2img_prompt.submit(**txt2img_args) submit.click(**txt2img_args) - res_switch_btn.click(switch_width_and_height, inputs=[width, height], outputs=[width, height]) + + res_switch_btn.click(lambda w, h: (h, w), inputs=[width, height], outputs=[width, height]) txt_prompt_img.change( fn=modules.images.image_data, @@ -882,7 +877,7 @@ def create_ui(): img2img_prompt.submit(**img2img_args) submit.click(**img2img_args) - res_switch_btn.click(switch_width_and_height, inputs=[width, height], outputs=[width, height]) + res_switch_btn.click(lambda w, h: (h, w), inputs=[width, height], outputs=[width, height]) img2img_interrogate.click( fn=lambda *args: process_interrogate(interrogate, *args), -- cgit v1.2.3 From 3752aad23d4be4522f9edf3fe79c1122fa5ad509 Mon Sep 17 00:00:00 2001 From: Mackerel Date: Sat, 28 Jan 2023 02:44:12 -0500 Subject: don't replace regular --help with new paths.py parser help --- modules/paths.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'modules') diff --git a/modules/paths.py b/modules/paths.py index 08e6f9b9..d991cc71 100644 --- a/modules/paths.py +++ b/modules/paths.py @@ -6,7 +6,7 @@ import modules.safe script_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__))) # Parse the --data-dir flag first so we can use it as a base for our other argument default values -parser = argparse.ArgumentParser() +parser = argparse.ArgumentParser(add_help=False) parser.add_argument("--data-dir", type=str, default=os.path.dirname(os.path.dirname(os.path.realpath(__file__))), help="base path where all user data is stored",) cmd_opts_pre = parser.parse_known_args()[0] data_path = cmd_opts_pre.data_dir -- cgit v1.2.3