From 3110f895b2718a3a25aae419fdf5c87c177ec9f4 Mon Sep 17 00:00:00 2001
From: alg-wiki <alg.4chan@gmail.com>
Date: Mon, 10 Oct 2022 17:07:46 +0900
Subject: Textual Inversion: Added custom training image size and number of
 repeats per input image in a single epoch

---
 modules/textual_inversion/textual_inversion.py | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index cd9f3498..e34dc2e8 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -6,6 +6,7 @@ import torch
 import tqdm
 import html
 import datetime
+import math
 
 
 from modules import shared, devices, sd_hijack, processing, sd_models
@@ -156,7 +157,7 @@ def create_embedding(name, num_vectors_per_token, init_text='*'):
     return fn
 
 
-def train_embedding(embedding_name, learn_rate, data_root, log_directory, steps, create_image_every, save_embedding_every, template_file):
+def train_embedding(embedding_name, learn_rate, data_root, log_directory, training_size, steps, num_repeats, create_image_every, save_embedding_every, template_file):
     assert embedding_name, 'embedding not selected'
 
     shared.state.textinfo = "Initializing textual inversion training..."
@@ -182,7 +183,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, steps,
 
     shared.state.textinfo = f"Preparing dataset from {html.escape(data_root)}..."
     with torch.autocast("cuda"):
-        ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, size=512, placeholder_token=embedding_name, model=shared.sd_model, device=devices.device, template_file=template_file)
+        ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, size=training_size, repeats=num_repeats, placeholder_token=embedding_name, model=shared.sd_model, device=devices.device, template_file=template_file)
 
     hijack = sd_hijack.model_hijack
 
@@ -200,6 +201,9 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, steps,
     if ititial_step > steps:
         return embedding, filename
 
+    tr_img_len = len([os.path.join(data_root, file_path) for file_path in os.listdir(data_root)])
+    epoch_len = (tr_img_len * num_repeats) + tr_img_len
+
     pbar = tqdm.tqdm(enumerate(ds), total=steps-ititial_step)
     for i, (x, text) in pbar:
         embedding.step = i + ititial_step
@@ -223,7 +227,10 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, steps,
             loss.backward()
             optimizer.step()
 
-        pbar.set_description(f"loss: {losses.mean():.7f}")
+        epoch_num = math.floor(embedding.step / epoch_len)
+        epoch_step = embedding.step - (epoch_num * epoch_len)
+
+        pbar.set_description(f"[Epoch {epoch_num}: {epoch_step}/{epoch_len}]loss: {losses.mean():.7f}")
 
         if embedding.step > 0 and embedding_dir is not None and embedding.step % save_embedding_every == 0:
             last_saved_file = os.path.join(embedding_dir, f'{embedding_name}-{embedding.step}.pt')
@@ -236,6 +243,8 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, steps,
                 sd_model=shared.sd_model,
                 prompt=text,
                 steps=20,
+                height=training_size,
+                width=training_size,
                 do_not_save_grid=True,
                 do_not_save_samples=True,
             )
-- 
cgit v1.2.3


From 4ee7519fc2e459ce8eff1f61f1655afba393357c Mon Sep 17 00:00:00 2001
From: alg-wiki <alg.4chan@gmail.com>
Date: Mon, 10 Oct 2022 17:31:33 +0900
Subject: Fixed progress bar output for epoch

---
 modules/textual_inversion/textual_inversion.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index e34dc2e8..769682ea 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -228,7 +228,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
             optimizer.step()
 
         epoch_num = math.floor(embedding.step / epoch_len)
-        epoch_step = embedding.step - (epoch_num * epoch_len)
+        epoch_step = embedding.step - (epoch_num * epoch_len) + 1
 
         pbar.set_description(f"[Epoch {epoch_num}: {epoch_step}/{epoch_len}]loss: {losses.mean():.7f}")
 
-- 
cgit v1.2.3


From 04c745ea4f81518999927fee5f78500560c25e29 Mon Sep 17 00:00:00 2001
From: alg-wiki <alg.4chan@gmail.com>
Date: Mon, 10 Oct 2022 22:35:35 +0900
Subject: Custom Width and Height

---
 modules/textual_inversion/dataset.py           |  7 +++----
 modules/textual_inversion/preprocess.py        | 19 ++++++++++---------
 modules/textual_inversion/textual_inversion.py | 11 +++++------
 modules/ui.py                                  | 12 ++++++++----
 4 files changed, 26 insertions(+), 23 deletions(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py
index acc4ce59..bcf772d2 100644
--- a/modules/textual_inversion/dataset.py
+++ b/modules/textual_inversion/dataset.py
@@ -15,13 +15,12 @@ re_tag = re.compile(r"[a-zA-Z][_\w\d()]+")
 
 
 class PersonalizedBase(Dataset):
-    def __init__(self, data_root, size, repeats, flip_p=0.5, placeholder_token="*", model=None, device=None, template_file=None):
+    def __init__(self, data_root, width, height, repeats, flip_p=0.5, placeholder_token="*", model=None, device=None, template_file=None):
 
         self.placeholder_token = placeholder_token
 
-        self.size = size
-        self.width = size
-        self.height = size
+        self.width = width
+        self.height = height
         self.flip = transforms.RandomHorizontalFlip(p=flip_p)
 
         self.dataset = []
diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py
index b3de6fd7..d7efdef2 100644
--- a/modules/textual_inversion/preprocess.py
+++ b/modules/textual_inversion/preprocess.py
@@ -7,8 +7,9 @@ import tqdm
 from modules import shared, images
 
 
-def preprocess(process_src, process_dst, process_size, process_flip, process_split, process_caption):
-    size = process_size
+def preprocess(process_src, process_dst, process_width, process_height, process_flip, process_split, process_caption):
+    width = process_width
+    height = process_height
     src = os.path.abspath(process_src)
     dst = os.path.abspath(process_dst)
 
@@ -55,23 +56,23 @@ def preprocess(process_src, process_dst, process_size, process_flip, process_spl
         is_wide = ratio < 1 / 1.35
 
         if process_split and is_tall:
-            img = img.resize((size, size * img.height // img.width))
+            img = img.resize((width, height * img.height // img.width))
 
-            top = img.crop((0, 0, size, size))
+            top = img.crop((0, 0, width, height))
             save_pic(top, index)
 
-            bot = img.crop((0, img.height - size, size, img.height))
+            bot = img.crop((0, img.height - height, width, img.height))
             save_pic(bot, index)
         elif process_split and is_wide:
-            img = img.resize((size * img.width // img.height, size))
+            img = img.resize((width * img.width // img.height, height))
 
-            left = img.crop((0, 0, size, size))
+            left = img.crop((0, 0, width, height))
             save_pic(left, index)
 
-            right = img.crop((img.width - size, 0, img.width, size))
+            right = img.crop((img.width - width, 0, img.width, height))
             save_pic(right, index)
         else:
-            img = images.resize_image(1, img, size, size)
+            img = images.resize_image(1, img, width, height)
             save_pic(img, index)
 
         shared.state.nextjob()
diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index 769682ea..5965c5a0 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -6,7 +6,6 @@ import torch
 import tqdm
 import html
 import datetime
-import math
 
 
 from modules import shared, devices, sd_hijack, processing, sd_models
@@ -157,7 +156,7 @@ def create_embedding(name, num_vectors_per_token, init_text='*'):
     return fn
 
 
-def train_embedding(embedding_name, learn_rate, data_root, log_directory, training_size, steps, num_repeats, create_image_every, save_embedding_every, template_file):
+def train_embedding(embedding_name, learn_rate, data_root, log_directory, training_width, training_height, steps, num_repeats, create_image_every, save_embedding_every, template_file):
     assert embedding_name, 'embedding not selected'
 
     shared.state.textinfo = "Initializing textual inversion training..."
@@ -183,7 +182,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
 
     shared.state.textinfo = f"Preparing dataset from {html.escape(data_root)}..."
     with torch.autocast("cuda"):
-        ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, size=training_size, repeats=num_repeats, placeholder_token=embedding_name, model=shared.sd_model, device=devices.device, template_file=template_file)
+        ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=num_repeats, placeholder_token=embedding_name, model=shared.sd_model, device=devices.device, template_file=template_file)
 
     hijack = sd_hijack.model_hijack
 
@@ -227,7 +226,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
             loss.backward()
             optimizer.step()
 
-        epoch_num = math.floor(embedding.step / epoch_len)
+        epoch_num = embedding.step // epoch_len
         epoch_step = embedding.step - (epoch_num * epoch_len) + 1
 
         pbar.set_description(f"[Epoch {epoch_num}: {epoch_step}/{epoch_len}]loss: {losses.mean():.7f}")
@@ -243,8 +242,8 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
                 sd_model=shared.sd_model,
                 prompt=text,
                 steps=20,
-                height=training_size,
-                width=training_size,
+				height=training_height,
+				width=training_width,
                 do_not_save_grid=True,
                 do_not_save_samples=True,
             )
diff --git a/modules/ui.py b/modules/ui.py
index f821fd8d..8c06ad7c 100644
--- a/modules/ui.py
+++ b/modules/ui.py
@@ -1029,7 +1029,8 @@ def create_ui(wrap_gradio_gpu_call):
 
                     process_src = gr.Textbox(label='Source directory')
                     process_dst = gr.Textbox(label='Destination directory')
-                    process_size = gr.Slider(minimum=64, maximum=2048, step=64, label="Size (width and height)", value=512)
+                    process_width = gr.Slider(minimum=64, maximum=2048, step=64, label="Width", value=512)
+                    process_height = gr.Slider(minimum=64, maximum=2048, step=64, label="Height", value=512)
 
                     with gr.Row():
                         process_flip = gr.Checkbox(label='Create flipped copies')
@@ -1050,7 +1051,8 @@ def create_ui(wrap_gradio_gpu_call):
                     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_size = gr.Slider(minimum=64, maximum=2048, step=64, label="Size (width and height)", value=512)
+                    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)
                     num_repeats = gr.Number(label='Number of repeats for a single input image per epoch', value=100, precision=0)
                     create_image_every = gr.Number(label='Save an image to log directory every N steps, 0 to disable', value=500, precision=0)
@@ -1095,7 +1097,8 @@ def create_ui(wrap_gradio_gpu_call):
             inputs=[
                 process_src,
                 process_dst,
-                process_size,
+                process_width,
+                process_height,
                 process_flip,
                 process_split,
                 process_caption,
@@ -1114,7 +1117,8 @@ def create_ui(wrap_gradio_gpu_call):
                 learn_rate,
                 dataset_directory,
                 log_directory,
-                training_size,
+                training_width,
+                training_height,
                 steps,
                 num_repeats,
                 create_image_every,
-- 
cgit v1.2.3


From bc3e183b739913e7be91213a256f038b10eb71e9 Mon Sep 17 00:00:00 2001
From: alg-wiki <alg.4chan@gmail.com>
Date: Tue, 11 Oct 2022 04:30:13 +0900
Subject: Textual Inversion: Preprocess and Training will only pick-up image
 files

---
 modules/textual_inversion/dataset.py           | 3 ++-
 modules/textual_inversion/preprocess.py        | 3 ++-
 modules/textual_inversion/textual_inversion.py | 3 ++-
 3 files changed, 6 insertions(+), 3 deletions(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py
index bcf772d2..d4baf066 100644
--- a/modules/textual_inversion/dataset.py
+++ b/modules/textual_inversion/dataset.py
@@ -22,6 +22,7 @@ class PersonalizedBase(Dataset):
         self.width = width
         self.height = height
         self.flip = transforms.RandomHorizontalFlip(p=flip_p)
+        self.extns = [".jpg",".jpeg",".png"]
 
         self.dataset = []
 
@@ -32,7 +33,7 @@ class PersonalizedBase(Dataset):
 
         assert data_root, 'dataset directory not specified'
 
-        self.image_paths = [os.path.join(data_root, file_path) for file_path in os.listdir(data_root)]
+        self.image_paths = [os.path.join(data_root, file_path) for file_path in os.listdir(data_root) if os.path.splitext(file_path.casefold())[1] in self.extns]
         print("Preparing dataset...")
         for path in tqdm.tqdm(self.image_paths):
             image = Image.open(path)
diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py
index d7efdef2..b6c78cf8 100644
--- a/modules/textual_inversion/preprocess.py
+++ b/modules/textual_inversion/preprocess.py
@@ -12,12 +12,13 @@ def preprocess(process_src, process_dst, process_width, process_height, process_
     height = process_height
     src = os.path.abspath(process_src)
     dst = os.path.abspath(process_dst)
+    extns = [".jpg",".jpeg",".png"]
 
     assert src != dst, 'same directory specified as source and destination'
 
     os.makedirs(dst, exist_ok=True)
 
-    files = os.listdir(src)
+    files = [i for i in os.listdir(src) if os.path.splitext(i.casefold())[1] in extns]
 
     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 5965c5a0..45397be9 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -161,6 +161,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
 
     shared.state.textinfo = "Initializing textual inversion training..."
     shared.state.job_count = steps
+    extns = [".jpg",".jpeg",".png"]
 
     filename = os.path.join(shared.cmd_opts.embeddings_dir, f'{embedding_name}.pt')
 
@@ -200,7 +201,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
     if ititial_step > steps:
         return embedding, filename
 
-    tr_img_len = len([os.path.join(data_root, file_path) for file_path in os.listdir(data_root)])
+    tr_img_len = len([os.path.join(data_root, file_path) for file_path in os.listdir(data_root) if os.path.splitext(file_path.casefold())[1] in extns])
     epoch_len = (tr_img_len * num_repeats) + tr_img_len
 
     pbar = tqdm.tqdm(enumerate(ds), total=steps-ititial_step)
-- 
cgit v1.2.3


From 2536ecbb1790da2af0d61b6a26f38732cba665cd Mon Sep 17 00:00:00 2001
From: Fampai <>
Date: Mon, 10 Oct 2022 17:10:29 -0400
Subject: Refactored learning rate code

---
 modules/textual_inversion/textual_inversion.py | 51 ++++++++++++++++++++++++--
 modules/ui.py                                  |  2 +-
 2 files changed, 48 insertions(+), 5 deletions(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index 5965c5a0..c64a4598 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -189,8 +189,6 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
     embedding = hijack.embedding_db.word_embeddings[embedding_name]
     embedding.vec.requires_grad = True
 
-    optimizer = torch.optim.AdamW([embedding.vec], lr=learn_rate)
-
     losses = torch.zeros((32,))
 
     last_saved_file = "<none>"
@@ -203,12 +201,24 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
     tr_img_len = len([os.path.join(data_root, file_path) for file_path in os.listdir(data_root)])
     epoch_len = (tr_img_len * num_repeats) + tr_img_len
 
+    scheduleIter = iter(LearnSchedule(learn_rate, steps, ititial_step))
+    (learn_rate, end_step) = next(scheduleIter)
+    print(f'Training at rate of {learn_rate} until step {end_step}')
+
+    optimizer = torch.optim.AdamW([embedding.vec], lr=learn_rate)
+
     pbar = tqdm.tqdm(enumerate(ds), total=steps-ititial_step)
     for i, (x, text) in pbar:
         embedding.step = i + ititial_step
 
-        if embedding.step > steps:
-            break
+        if embedding.step > end_step:
+            try:
+                (learn_rate, end_step) = next(scheduleIter)
+            except:
+                break
+            tqdm.tqdm.write(f'Training at rate of {learn_rate} until step {end_step}')
+            for pg in optimizer.param_groups:
+                pg['lr'] = learn_rate
 
         if shared.state.interrupted:
             break
@@ -277,3 +287,36 @@ Last saved image: {html.escape(last_saved_image)}<br/>
 
     return embedding, filename
 
+class LearnSchedule:
+    def __init__(self, learn_rate, max_steps, cur_step=0):
+        pairs = learn_rate.split(',')
+        self.rates = []
+        self.it = 0
+        self.maxit = 0
+        for i, pair in enumerate(pairs):
+            tmp = pair.split(':')
+            if len(tmp) == 2:
+                step = int(tmp[1])
+                if step > cur_step:
+                    self.rates.append((float(tmp[0]), min(step, max_steps)))
+                    self.maxit += 1
+                    if step > max_steps:
+                        return
+                elif step == -1:
+                    self.rates.append((float(tmp[0]), max_steps))
+                    self.maxit += 1
+                    return
+            else:
+                self.rates.append((float(tmp[0]), max_steps))
+                self.maxit += 1
+                return
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        if self.it < self.maxit:
+            self.it += 1
+            return self.rates[self.it - 1]
+        else:
+            raise StopIteration
diff --git a/modules/ui.py b/modules/ui.py
index 8c06ad7c..c9e8355b 100644
--- a/modules/ui.py
+++ b/modules/ui.py
@@ -1047,7 +1047,7 @@ def create_ui(wrap_gradio_gpu_call):
                 with gr.Group():
                     gr.HTML(value="<p style='margin-bottom: 0.7em'>Train an embedding; must specify a directory with a set of 1:1 ratio images</p>")
                     train_embedding_name = gr.Dropdown(label='Embedding', choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys()))
-                    learn_rate = gr.Number(label='Learning rate', value=5.0e-03)
+                    learn_rate = gr.Textbox(label='Learning rate', placeholder="Learning rate", value = "5.0e-03")
                     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"))
-- 
cgit v1.2.3


From 907a88b2d0be320575c2129d8d6a1d4f3a68f9eb Mon Sep 17 00:00:00 2001
From: alg-wiki <alg.4chan@gmail.com>
Date: Tue, 11 Oct 2022 06:33:08 +0900
Subject: Added .webp .bmp

---
 modules/textual_inversion/dataset.py           | 2 +-
 modules/textual_inversion/preprocess.py        | 2 +-
 modules/textual_inversion/textual_inversion.py | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py
index d4baf066..0dc54fb7 100644
--- a/modules/textual_inversion/dataset.py
+++ b/modules/textual_inversion/dataset.py
@@ -22,7 +22,7 @@ class PersonalizedBase(Dataset):
         self.width = width
         self.height = height
         self.flip = transforms.RandomHorizontalFlip(p=flip_p)
-        self.extns = [".jpg",".jpeg",".png"]
+        self.extns = [".jpg",".jpeg",".png",".webp",".bmp"]
 
         self.dataset = []
 
diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py
index b6c78cf8..8290abe8 100644
--- a/modules/textual_inversion/preprocess.py
+++ b/modules/textual_inversion/preprocess.py
@@ -12,7 +12,7 @@ def preprocess(process_src, process_dst, process_width, process_height, process_
     height = process_height
     src = os.path.abspath(process_src)
     dst = os.path.abspath(process_dst)
-    extns = [".jpg",".jpeg",".png"]
+    extns = [".jpg",".jpeg",".png",".webp",".bmp"]
 
     assert src != dst, 'same directory specified as source and destination'
 
diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index a03b299c..33c923d1 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -161,7 +161,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
 
     shared.state.textinfo = "Initializing textual inversion training..."
     shared.state.job_count = steps
-    extns = [".jpg",".jpeg",".png"]
+    extns = [".jpg",".jpeg",".png",".webp",".bmp"]
 
     filename = os.path.join(shared.cmd_opts.embeddings_dir, f'{embedding_name}.pt')
 
-- 
cgit v1.2.3


From b2368a3bce663f19a7209d9cb38617e635ca6e3c Mon Sep 17 00:00:00 2001
From: alg-wiki <alg.4chan@gmail.com>
Date: Tue, 11 Oct 2022 17:32:46 +0900
Subject: Switched to exception handling

---
 modules/textual_inversion/dataset.py           | 10 +++++-----
 modules/textual_inversion/preprocess.py        |  8 +++++---
 modules/textual_inversion/textual_inversion.py | 18 ++++++++----------
 3 files changed, 18 insertions(+), 18 deletions(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py
index 0dc54fb7..4d006366 100644
--- a/modules/textual_inversion/dataset.py
+++ b/modules/textual_inversion/dataset.py
@@ -22,7 +22,6 @@ class PersonalizedBase(Dataset):
         self.width = width
         self.height = height
         self.flip = transforms.RandomHorizontalFlip(p=flip_p)
-        self.extns = [".jpg",".jpeg",".png",".webp",".bmp"]
 
         self.dataset = []
 
@@ -33,12 +32,13 @@ class PersonalizedBase(Dataset):
 
         assert data_root, 'dataset directory not specified'
 
-        self.image_paths = [os.path.join(data_root, file_path) for file_path in os.listdir(data_root) if os.path.splitext(file_path.casefold())[1] in self.extns]
+        self.image_paths = [os.path.join(data_root, file_path) for file_path in os.listdir(data_root)]
         print("Preparing dataset...")
         for path in tqdm.tqdm(self.image_paths):
-            image = Image.open(path)
-            image = image.convert('RGB')
-            image = image.resize((self.width, self.height), PIL.Image.BICUBIC)
+            try:
+                image = Image.open(path).convert('RGB').resize((self.width, self.height), PIL.Image.BICUBIC)
+            except Exception:
+                continue
 
             filename = os.path.basename(path)
             filename_tokens = os.path.splitext(filename)[0]
diff --git a/modules/textual_inversion/preprocess.py b/modules/textual_inversion/preprocess.py
index 8290abe8..1a672725 100644
--- a/modules/textual_inversion/preprocess.py
+++ b/modules/textual_inversion/preprocess.py
@@ -12,13 +12,12 @@ def preprocess(process_src, process_dst, process_width, process_height, process_
     height = process_height
     src = os.path.abspath(process_src)
     dst = os.path.abspath(process_dst)
-    extns = [".jpg",".jpeg",".png",".webp",".bmp"]
 
     assert src != dst, 'same directory specified as source and destination'
 
     os.makedirs(dst, exist_ok=True)
 
-    files = [i for i in os.listdir(src) if os.path.splitext(i.casefold())[1] in extns]
+    files = os.listdir(src)
 
     shared.state.textinfo = "Preprocessing..."
     shared.state.job_count = len(files)
@@ -47,7 +46,10 @@ def preprocess(process_src, process_dst, process_width, process_height, process_
     for index, imagefile in enumerate(tqdm.tqdm(files)):
         subindex = [0]
         filename = os.path.join(src, imagefile)
-        img = Image.open(filename).convert("RGB")
+        try:
+            img = Image.open(filename).convert("RGB")
+        except Exception:
+            continue
 
         if shared.state.interrupted:
             break
diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index 33c923d1..91cde04b 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -161,7 +161,6 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
 
     shared.state.textinfo = "Initializing textual inversion training..."
     shared.state.job_count = steps
-    extns = [".jpg",".jpeg",".png",".webp",".bmp"]
 
     filename = os.path.join(shared.cmd_opts.embeddings_dir, f'{embedding_name}.pt')
 
@@ -201,10 +200,6 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
     if ititial_step > steps:
         return embedding, filename
 
-    tr_img_len = len([os.path.join(data_root, file_path) for file_path in os.listdir(data_root) if os.path.splitext(file_path.casefold())[1] in extns])
-
-    epoch_len = (tr_img_len * num_repeats) + tr_img_len
-
     pbar = tqdm.tqdm(enumerate(ds), total=steps-ititial_step)
     for i, (x, text) in pbar:
         embedding.step = i + ititial_step
@@ -228,10 +223,10 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
             loss.backward()
             optimizer.step()
 
-        epoch_num = embedding.step // epoch_len
-        epoch_step = embedding.step - (epoch_num * epoch_len) + 1
+        epoch_num = embedding.step // len(ds)
+        epoch_step = embedding.step - (epoch_num * len(ds)) + 1
 
-        pbar.set_description(f"[Epoch {epoch_num}: {epoch_step}/{epoch_len}]loss: {losses.mean():.7f}")
+        pbar.set_description(f"[Epoch {epoch_num}: {epoch_step}/{len(ds)}]loss: {losses.mean():.7f}")
 
         if embedding.step > 0 and embedding_dir is not None and embedding.step % save_embedding_every == 0:
             last_saved_file = os.path.join(embedding_dir, f'{embedding_name}-{embedding.step}.pt')
@@ -243,9 +238,12 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
             p = processing.StableDiffusionProcessingTxt2Img(
                 sd_model=shared.sd_model,
                 prompt=text,
-                steps=20,
-				height=training_height,
+                steps=28,
+				height=768,
 				width=training_width,
+                negative_prompt="lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts,signature, watermark, username, blurry, artist name",
+                cfg_scale=7.0,
+                sampler_index=0,
                 do_not_save_grid=True,
                 do_not_save_samples=True,
             )
-- 
cgit v1.2.3


From 8bacbca0a1ab9aabcb0ad0cbf070e0006991e98a Mon Sep 17 00:00:00 2001
From: alg-wiki <alg.4chan@gmail.com>
Date: Tue, 11 Oct 2022 17:35:09 +0900
Subject: Removed my local edits to checkpoint image generation

---
 modules/textual_inversion/textual_inversion.py | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index 91cde04b..e9ff80c2 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -238,12 +238,9 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
             p = processing.StableDiffusionProcessingTxt2Img(
                 sd_model=shared.sd_model,
                 prompt=text,
-                steps=28,
-				height=768,
+                steps=20,
+				height=training_height,
 				width=training_width,
-                negative_prompt="lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts,signature, watermark, username, blurry, artist name",
-                cfg_scale=7.0,
-                sampler_index=0,
                 do_not_save_grid=True,
                 do_not_save_samples=True,
             )
-- 
cgit v1.2.3


From d4ea5f4d8631f778d11efcde397e4a5b8801d43b Mon Sep 17 00:00:00 2001
From: AUTOMATIC <16777216c@gmail.com>
Date: Tue, 11 Oct 2022 19:03:08 +0300
Subject: add an option to unload models during hypernetwork training to save
 VRAM

---
 modules/hypernetworks/hypernetwork.py          | 25 +++++++++++++++-------
 modules/hypernetworks/ui.py                    |  4 +++-
 modules/shared.py                              |  4 ++++
 modules/textual_inversion/dataset.py           | 29 ++++++++++++++++++--------
 modules/textual_inversion/textual_inversion.py |  2 +-
 5 files changed, 46 insertions(+), 18 deletions(-)

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py
index b081f14e..4700e1ec 100644
--- a/modules/hypernetworks/hypernetwork.py
+++ b/modules/hypernetworks/hypernetwork.py
@@ -175,6 +175,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, data_root, log_directory,
     filename = os.path.join(shared.cmd_opts.hypernetwork_dir, f'{hypernetwork_name}.pt')
 
     log_directory = os.path.join(log_directory, datetime.datetime.now().strftime("%Y-%m-%d"), hypernetwork_name)
+    unload = shared.opts.unload_models_when_training
 
     if save_hypernetwork_every > 0:
         hypernetwork_dir = os.path.join(log_directory, "hypernetworks")
@@ -188,11 +189,13 @@ def train_hypernetwork(hypernetwork_name, learn_rate, data_root, log_directory,
     else:
         images_dir = None
 
-    cond_model = shared.sd_model.cond_stage_model
-
     shared.state.textinfo = f"Preparing dataset from {html.escape(data_root)}..."
     with torch.autocast("cuda"):
-        ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=512, height=512, repeats=1, placeholder_token=hypernetwork_name, model=shared.sd_model, device=devices.device, template_file=template_file)
+        ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=512, height=512, repeats=1, placeholder_token=hypernetwork_name, model=shared.sd_model, device=devices.device, template_file=template_file, include_cond=True)
+
+    if unload:
+        shared.sd_model.cond_stage_model.to(devices.cpu)
+        shared.sd_model.first_stage_model.to(devices.cpu)
 
     hypernetwork = shared.loaded_hypernetwork
     weights = hypernetwork.weights()
@@ -211,7 +214,7 @@ def train_hypernetwork(hypernetwork_name, learn_rate, data_root, log_directory,
         return hypernetwork, filename
 
     pbar = tqdm.tqdm(enumerate(ds), total=steps - ititial_step)
-    for i, (x, text) in pbar:
+    for i, (x, text, cond) in pbar:
         hypernetwork.step = i + ititial_step
 
         if hypernetwork.step > steps:
@@ -221,11 +224,11 @@ def train_hypernetwork(hypernetwork_name, learn_rate, data_root, log_directory,
             break
 
         with torch.autocast("cuda"):
-            c = cond_model([text])
-
+            cond = cond.to(devices.device)
             x = x.to(devices.device)
-            loss = shared.sd_model(x.unsqueeze(0), c)[0]
+            loss = shared.sd_model(x.unsqueeze(0), cond)[0]
             del x
+            del cond
 
             losses[hypernetwork.step % losses.shape[0]] = loss.item()
 
@@ -244,6 +247,10 @@ def train_hypernetwork(hypernetwork_name, learn_rate, data_root, log_directory,
 
             preview_text = text if preview_image_prompt == "" else preview_image_prompt
 
+            optimizer.zero_grad()
+            shared.sd_model.cond_stage_model.to(devices.device)
+            shared.sd_model.first_stage_model.to(devices.device)
+
             p = processing.StableDiffusionProcessingTxt2Img(
                 sd_model=shared.sd_model,
                 prompt=preview_text,
@@ -255,6 +262,10 @@ def train_hypernetwork(hypernetwork_name, learn_rate, data_root, log_directory,
             processed = processing.process_images(p)
             image = processed.images[0]
 
+            if unload:
+                shared.sd_model.cond_stage_model.to(devices.cpu)
+                shared.sd_model.first_stage_model.to(devices.cpu)
+
             shared.state.current_image = image
             image.save(last_saved_image)
 
diff --git a/modules/hypernetworks/ui.py b/modules/hypernetworks/ui.py
index 3541a388..c67facbb 100644
--- a/modules/hypernetworks/ui.py
+++ b/modules/hypernetworks/ui.py
@@ -5,7 +5,7 @@ import gradio as gr
 
 import modules.textual_inversion.textual_inversion
 import modules.textual_inversion.preprocess
-from modules import sd_hijack, shared
+from modules import sd_hijack, shared, devices
 from modules.hypernetworks import hypernetwork
 
 
@@ -41,5 +41,7 @@ Hypernetwork saved to {html.escape(filename)}
         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/shared.py b/modules/shared.py
index 20b45f23..c1092ff7 100644
--- a/modules/shared.py
+++ b/modules/shared.py
@@ -228,6 +228,10 @@ options_templates.update(options_section(('system', "System"), {
     "multiple_tqdm": OptionInfo(True, "Add a second progress bar to the console that shows progress for an entire job."),
 }))
 
+options_templates.update(options_section(('training', "Training"), {
+    "unload_models_when_training": OptionInfo(False, "Unload VAE and CLIP form VRAM when training"),
+}))
+
 options_templates.update(options_section(('sd', "Stable Diffusion"), {
     "sd_model_checkpoint": OptionInfo(None, "Stable Diffusion checkpoint", gr.Dropdown, lambda: {"choices": modules.sd_models.checkpoint_tiles()}, show_on_main_page=True),
     "sd_hypernetwork": OptionInfo("None", "Stable Diffusion finetune hypernetwork", gr.Dropdown, lambda: {"choices": ["None"] + [x for x in hypernetworks.keys()]}),
diff --git a/modules/textual_inversion/dataset.py b/modules/textual_inversion/dataset.py
index 4d006366..f61f40d3 100644
--- a/modules/textual_inversion/dataset.py
+++ b/modules/textual_inversion/dataset.py
@@ -8,14 +8,14 @@ from torchvision import transforms
 
 import random
 import tqdm
-from modules import devices
+from modules import devices, shared
 import re
 
 re_tag = re.compile(r"[a-zA-Z][_\w\d()]+")
 
 
 class PersonalizedBase(Dataset):
-    def __init__(self, data_root, width, height, repeats, flip_p=0.5, placeholder_token="*", model=None, device=None, template_file=None):
+    def __init__(self, data_root, width, height, repeats, flip_p=0.5, placeholder_token="*", model=None, device=None, template_file=None, include_cond=False):
 
         self.placeholder_token = placeholder_token
 
@@ -32,6 +32,8 @@ class PersonalizedBase(Dataset):
 
         assert data_root, 'dataset directory not specified'
 
+        cond_model = shared.sd_model.cond_stage_model
+
         self.image_paths = [os.path.join(data_root, file_path) for file_path in os.listdir(data_root)]
         print("Preparing dataset...")
         for path in tqdm.tqdm(self.image_paths):
@@ -53,7 +55,13 @@ class PersonalizedBase(Dataset):
             init_latent = model.get_first_stage_encoding(model.encode_first_stage(torchdata.unsqueeze(dim=0))).squeeze()
             init_latent = init_latent.to(devices.cpu)
 
-            self.dataset.append((init_latent, filename_tokens))
+            if include_cond:
+                text = self.create_text(filename_tokens)
+                cond = cond_model([text]).to(devices.cpu)
+            else:
+                cond = None
+
+            self.dataset.append((init_latent, filename_tokens, cond))
 
         self.length = len(self.dataset) * repeats
 
@@ -64,6 +72,12 @@ class PersonalizedBase(Dataset):
     def shuffle(self):
         self.indexes = self.initial_indexes[torch.randperm(self.initial_indexes.shape[0])]
 
+    def create_text(self, filename_tokens):
+        text = random.choice(self.lines)
+        text = text.replace("[name]", self.placeholder_token)
+        text = text.replace("[filewords]", ' '.join(filename_tokens))
+        return text
+
     def __len__(self):
         return self.length
 
@@ -72,10 +86,7 @@ class PersonalizedBase(Dataset):
             self.shuffle()
 
         index = self.indexes[i % len(self.indexes)]
-        x, filename_tokens = self.dataset[index]
-
-        text = random.choice(self.lines)
-        text = text.replace("[name]", self.placeholder_token)
-        text = text.replace("[filewords]", ' '.join(filename_tokens))
+        x, filename_tokens, cond = self.dataset[index]
 
-        return x, text
+        text = self.create_text(filename_tokens)
+        return x, text, cond
diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index bb05cdc6..35f4bd9e 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -201,7 +201,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
         return embedding, filename
 
     pbar = tqdm.tqdm(enumerate(ds), total=steps-ititial_step)
-    for i, (x, text) in pbar:
+    for i, (x, text, _) in pbar:
         embedding.step = i + ititial_step
 
         if embedding.step > steps:
-- 
cgit v1.2.3


From d6fcc6b87bc00fcdecea276fe5b7c7945f7a8b14 Mon Sep 17 00:00:00 2001
From: AUTOMATIC <16777216c@gmail.com>
Date: Tue, 11 Oct 2022 22:03:05 +0300
Subject: apply lr schedule to hypernets

---
 modules/hypernetworks/hypernetwork.py          | 19 ++++++++---
 modules/textual_inversion/learn_schedule.py    | 34 ++++++++++++++++++++
 modules/textual_inversion/textual_inversion.py | 44 +++-----------------------
 modules/ui.py                                  |  2 +-
 4 files changed, 54 insertions(+), 45 deletions(-)
 create mode 100644 modules/textual_inversion/learn_schedule.py

(limited to 'modules/textual_inversion/textual_inversion.py')

diff --git a/modules/hypernetworks/hypernetwork.py b/modules/hypernetworks/hypernetwork.py
index 5608e799..470659df 100644
--- a/modules/hypernetworks/hypernetwork.py
+++ b/modules/hypernetworks/hypernetwork.py
@@ -14,6 +14,7 @@ import torch
 from torch import einsum
 from einops import rearrange, repeat
 import modules.textual_inversion.dataset
+from modules.textual_inversion.learn_schedule import LearnSchedule
 
 
 class HypernetworkModule(torch.nn.Module):
@@ -202,8 +203,6 @@ def train_hypernetwork(hypernetwork_name, learn_rate, data_root, log_directory,
     for weight in weights:
         weight.requires_grad = True
 
-    optimizer = torch.optim.AdamW(weights, lr=learn_rate)
-
     losses = torch.zeros((32,))
 
     last_saved_file = "<none>"
@@ -213,12 +212,24 @@ def train_hypernetwork(hypernetwork_name, learn_rate, data_root, log_directory,
     if ititial_step > steps:
         return hypernetwork, filename
 
+    schedules = iter(LearnSchedule(learn_rate, steps, ititial_step))
+    (learn_rate, end_step) = next(schedules)
+    print(f'Training at rate of {learn_rate} until step {end_step}')
+
+    optimizer = torch.optim.AdamW(weights, lr=learn_rate)
+
     pbar = tqdm.tqdm(enumerate(ds), total=steps - ititial_step)
     for i, (x, text, cond) in pbar:
         hypernetwork.step = i + ititial_step
 
-        if hypernetwork.step > steps:
-            break
+        if hypernetwork.step > end_step:
+            try:
+                (learn_rate, end_step) = next(schedules)
+            except Exception:
+                break
+            tqdm.tqdm.write(f'Training at rate of {learn_rate} until step {end_step}')
+            for pg in optimizer.param_groups:
+                pg['lr'] = learn_rate
 
         if shared.state.interrupted:
             break
diff --git a/modules/textual_inversion/learn_schedule.py b/modules/textual_inversion/learn_schedule.py
new file mode 100644
index 00000000..db720271
--- /dev/null
+++ b/modules/textual_inversion/learn_schedule.py
@@ -0,0 +1,34 @@
+
+class LearnSchedule:
+    def __init__(self, learn_rate, max_steps, cur_step=0):
+        pairs = learn_rate.split(',')
+        self.rates = []
+        self.it = 0
+        self.maxit = 0
+        for i, pair in enumerate(pairs):
+            tmp = pair.split(':')
+            if len(tmp) == 2:
+                step = int(tmp[1])
+                if step > cur_step:
+                    self.rates.append((float(tmp[0]), min(step, max_steps)))
+                    self.maxit += 1
+                    if step > max_steps:
+                        return
+                elif step == -1:
+                    self.rates.append((float(tmp[0]), max_steps))
+                    self.maxit += 1
+                    return
+            else:
+                self.rates.append((float(tmp[0]), max_steps))
+                self.maxit += 1
+                return
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        if self.it < self.maxit:
+            self.it += 1
+            return self.rates[self.it - 1]
+        else:
+            raise StopIteration
diff --git a/modules/textual_inversion/textual_inversion.py b/modules/textual_inversion/textual_inversion.py
index 47a27faf..7717837d 100644
--- a/modules/textual_inversion/textual_inversion.py
+++ b/modules/textual_inversion/textual_inversion.py
@@ -10,6 +10,7 @@ import datetime
 
 from modules import shared, devices, sd_hijack, processing, sd_models
 import modules.textual_inversion.dataset
+from modules.textual_inversion.learn_schedule import LearnSchedule
 
 
 class Embedding:
@@ -198,11 +199,8 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
     if ititial_step > steps:
         return embedding, filename
 
-    tr_img_len = len([os.path.join(data_root, file_path) for file_path in os.listdir(data_root)])
-    epoch_len = (tr_img_len * num_repeats) + tr_img_len
-
-    scheduleIter = iter(LearnSchedule(learn_rate, steps, ititial_step))
-    (learn_rate, end_step) = next(scheduleIter)
+    schedules = iter(LearnSchedule(learn_rate, steps, ititial_step))
+    (learn_rate, end_step) = next(schedules)
     print(f'Training at rate of {learn_rate} until step {end_step}')
 
     optimizer = torch.optim.AdamW([embedding.vec], lr=learn_rate)
@@ -213,7 +211,7 @@ def train_embedding(embedding_name, learn_rate, data_root, log_directory, traini
 
         if embedding.step > end_step:
             try:
-                (learn_rate, end_step) = next(scheduleIter)
+                (learn_rate, end_step) = next(schedules)
             except:
                 break
             tqdm.tqdm.write(f'Training at rate of {learn_rate} until step {end_step}')
@@ -288,37 +286,3 @@ Last saved image: {html.escape(last_saved_image)}<br/>
     embedding.save(filename)
 
     return embedding, filename
-
-class LearnSchedule:
-    def __init__(self, learn_rate, max_steps, cur_step=0):
-        pairs = learn_rate.split(',')
-        self.rates = []
-        self.it = 0
-        self.maxit = 0
-        for i, pair in enumerate(pairs):
-            tmp = pair.split(':')
-            if len(tmp) == 2:
-                step = int(tmp[1])
-                if step > cur_step:
-                    self.rates.append((float(tmp[0]), min(step, max_steps)))
-                    self.maxit += 1
-                    if step > max_steps:
-                        return
-                elif step == -1:
-                    self.rates.append((float(tmp[0]), max_steps))
-                    self.maxit += 1
-                    return
-            else:
-                self.rates.append((float(tmp[0]), max_steps))
-                self.maxit += 1
-                return
-
-    def __iter__(self):
-        return self
-
-    def __next__(self):
-        if self.it < self.maxit:
-            self.it += 1
-            return self.rates[self.it - 1]
-        else:
-            raise StopIteration
diff --git a/modules/ui.py b/modules/ui.py
index 2b688e32..1204eef7 100644
--- a/modules/ui.py
+++ b/modules/ui.py
@@ -1070,7 +1070,7 @@ def create_ui(wrap_gradio_gpu_call):
                     gr.HTML(value="<p style='margin-bottom: 0.7em'>Train an embedding; must specify a directory with a set of 1:1 ratio images</p>")
                     train_embedding_name = gr.Dropdown(label='Embedding', choices=sorted(sd_hijack.model_hijack.embedding_db.word_embeddings.keys()))
                     train_hypernetwork_name = gr.Dropdown(label='Hypernetwork', choices=[x for x in shared.hypernetworks.keys()])
-                    learn_rate = gr.Textbox(label='Learning rate', placeholder="Learning rate", value = "5.0e-03")
+                    learn_rate = gr.Textbox(label='Learning rate', placeholder="Learning rate", value="0.005")
                     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"))
-- 
cgit v1.2.3