From e14b586d0494d6c5cc3cbc45b5fa00c03d052443 Mon Sep 17 00:00:00 2001
From: Sakura-Luna <53183413+Sakura-Luna@users.noreply.github.com>
Date: Sun, 14 May 2023 12:42:44 +0800
Subject: Add Tiny AE live preview

---
 modules/sd_samplers_common.py | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)

(limited to 'modules/sd_samplers_common.py')

diff --git a/modules/sd_samplers_common.py b/modules/sd_samplers_common.py
index bc074238..d3dc130c 100644
--- a/modules/sd_samplers_common.py
+++ b/modules/sd_samplers_common.py
@@ -2,7 +2,7 @@ from collections import namedtuple
 import numpy as np
 import torch
 from PIL import Image
-from modules import devices, processing, images, sd_vae_approx
+from modules import devices, processing, images, sd_vae_approx, sd_vae_taesd
 
 from modules.shared import opts, state
 import modules.shared as shared
@@ -22,21 +22,26 @@ def setup_img2img_steps(p, steps=None):
     return steps, t_enc
 
 
-approximation_indexes = {"Full": 0, "Approx NN": 1, "Approx cheap": 2}
+approximation_indexes = {"Full": 0, "Tiny AE": 1, "Approx NN": 2, "Approx cheap": 3}
 
 
 def single_sample_to_image(sample, approximation=None):
     if approximation is None:
         approximation = approximation_indexes.get(opts.show_progress_type, 0)
 
-    if approximation == 2:
-        x_sample = sd_vae_approx.cheap_approximation(sample)
-    elif approximation == 1:
-        x_sample = sd_vae_approx.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
+    if approximation == 1:
+        x_sample = sd_vae_taesd.decode()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
+        x_sample = sd_vae_taesd.TAESD.unscale_latents(x_sample)
+        x_sample = torch.clamp((x_sample * 0.25) + 0.5, 0, 1)
     else:
-        x_sample = processing.decode_first_stage(shared.sd_model, sample.unsqueeze(0))[0]
+        if approximation == 3:
+            x_sample = sd_vae_approx.cheap_approximation(sample)
+        elif approximation == 2:
+            x_sample = sd_vae_approx.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
+        else:
+            x_sample = processing.decode_first_stage(shared.sd_model, sample.unsqueeze(0))[0]
+        x_sample = torch.clamp((x_sample + 1.0) / 2.0, min=0.0, max=1.0)
 
-    x_sample = torch.clamp((x_sample + 1.0) / 2.0, min=0.0, max=1.0)
     x_sample = 255. * np.moveaxis(x_sample.cpu().numpy(), 0, 2)
     x_sample = x_sample.astype(np.uint8)
     return Image.fromarray(x_sample)
-- 
cgit v1.2.3


From bd9b9d425a355e151b43047a5df5fcead2fcdc52 Mon Sep 17 00:00:00 2001
From: Sakura-Luna <53183413+Sakura-Luna@users.noreply.github.com>
Date: Sun, 14 May 2023 13:19:11 +0800
Subject: Add live preview mode check

---
 modules/sd_samplers_common.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'modules/sd_samplers_common.py')

diff --git a/modules/sd_samplers_common.py b/modules/sd_samplers_common.py
index d3dc130c..b1e8a780 100644
--- a/modules/sd_samplers_common.py
+++ b/modules/sd_samplers_common.py
@@ -26,8 +26,8 @@ approximation_indexes = {"Full": 0, "Tiny AE": 1, "Approx NN": 2, "Approx cheap"
 
 
 def single_sample_to_image(sample, approximation=None):
-    if approximation is None:
-        approximation = approximation_indexes.get(opts.show_progress_type, 0)
+    if approximation is None or approximation not in approximation_indexes.keys():
+        approximation = approximation_indexes.get(opts.show_progress_type, 1)
 
     if approximation == 1:
         x_sample = sd_vae_taesd.decode()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
-- 
cgit v1.2.3


From cdac5ace1456ba779d5a0171ff8757f31955bfee Mon Sep 17 00:00:00 2001
From: AUTOMATIC <16777216c@gmail.com>
Date: Tue, 16 May 2023 11:54:02 +0300
Subject: suppress ENSD infotext for samplers that don't use it

---
 modules/processing.py             | 11 +++++++----
 modules/sd_samplers.py            |  8 +++++++-
 modules/sd_samplers_common.py     | 21 ++++++++++++++++++++-
 modules/sd_samplers_compvis.py    |  8 ++++++--
 modules/sd_samplers_kdiffusion.py | 16 ++++++++--------
 5 files changed, 48 insertions(+), 16 deletions(-)

(limited to 'modules/sd_samplers_common.py')

diff --git a/modules/processing.py b/modules/processing.py
index 94fe2625..15806f78 100644
--- a/modules/processing.py
+++ b/modules/processing.py
@@ -13,7 +13,7 @@ from skimage import exposure
 from typing import Any, Dict, List
 
 import modules.sd_hijack
-from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_parameters_copypaste, extra_networks, sd_vae_approx, scripts
+from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_parameters_copypaste, extra_networks, sd_vae_approx, scripts, sd_samplers_common
 from modules.sd_hijack import model_hijack
 from modules.shared import opts, cmd_opts, state
 import modules.shared as shared
@@ -480,6 +480,10 @@ def create_infotext(p, all_prompts, all_seeds, all_subseeds, comments=None, iter
     clip_skip = getattr(p, 'clip_skip', opts.CLIP_stop_at_last_layers)
     enable_hr = getattr(p, 'enable_hr', False)
 
+    uses_ensd = opts.eta_noise_seed_delta != 0
+    if uses_ensd:
+        uses_ensd = sd_samplers_common.is_sampler_using_eta_noise_seed_delta(p)
+
     generation_params = {
         "Steps": p.steps,
         "Sampler": p.sampler_name,
@@ -496,17 +500,16 @@ def create_infotext(p, all_prompts, all_seeds, all_subseeds, comments=None, iter
         "Denoising strength": getattr(p, 'denoising_strength', None),
         "Conditional mask weight": getattr(p, "inpainting_mask_weight", shared.opts.inpainting_mask_weight) if p.is_using_inpainting_conditioning else None,
         "Clip skip": None if clip_skip <= 1 else clip_skip,
-        "ENSD": None if opts.eta_noise_seed_delta == 0 else opts.eta_noise_seed_delta,
+        "ENSD": opts.eta_noise_seed_delta if uses_ensd else None,
         "Token merging ratio": None if opts.token_merging_ratio == 0 else opts.token_merging_ratio,
         "Token merging ratio hr": None if not enable_hr or opts.token_merging_ratio_hr == 0 else opts.token_merging_ratio_hr,
         "Init image hash": getattr(p, 'init_img_hash', None),
         "RNG": opts.randn_source if opts.randn_source != "GPU" else None,
         "NGMS": None if p.s_min_uncond == 0 else p.s_min_uncond,
+        **p.extra_generation_params,
         "Version": program_version() if opts.add_version_to_infotext else None,
     }
 
-    generation_params.update(p.extra_generation_params)
-
     generation_params_text = ", ".join([k if k == v else f'{k}: {generation_parameters_copypaste.quote(v)}' for k, v in generation_params.items() if v is not None])
 
     negative_prompt_text = f"\nNegative prompt: {p.all_negative_prompts[index]}" if p.all_negative_prompts[index] else ""
diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py
index 4f1bf21d..f22aad8f 100644
--- a/modules/sd_samplers.py
+++ b/modules/sd_samplers.py
@@ -14,12 +14,18 @@ samplers_for_img2img = []
 samplers_map = {}
 
 
-def create_sampler(name, model):
+def find_sampler_config(name):
     if name is not None:
         config = all_samplers_map.get(name, None)
     else:
         config = all_samplers[0]
 
+    return config
+
+
+def create_sampler(name, model):
+    config = find_sampler_config(name)
+
     assert config is not None, f'bad sampler name: {name}'
 
     sampler = config.constructor(model)
diff --git a/modules/sd_samplers_common.py b/modules/sd_samplers_common.py
index bc074238..92880caf 100644
--- a/modules/sd_samplers_common.py
+++ b/modules/sd_samplers_common.py
@@ -2,7 +2,7 @@ from collections import namedtuple
 import numpy as np
 import torch
 from PIL import Image
-from modules import devices, processing, images, sd_vae_approx
+from modules import devices, processing, images, sd_vae_approx, sd_samplers
 
 from modules.shared import opts, state
 import modules.shared as shared
@@ -58,6 +58,25 @@ def store_latent(decoded):
             shared.state.assign_current_image(sample_to_image(decoded))
 
 
+def is_sampler_using_eta_noise_seed_delta(p):
+    """returns whether sampler from config will use eta noise seed delta for image creation"""
+
+    sampler_config = sd_samplers.find_sampler_config(p.sampler_name)
+
+    eta = p.eta
+
+    if eta is None and p.sampler is not None:
+        eta = p.sampler.eta
+
+    if eta is None and sampler_config is not None:
+        eta = 0 if sampler_config.options.get("default_eta_is_0", False) else 1.0
+
+    if eta == 0:
+        return False
+
+    return sampler_config.options.get("uses_ensd", False)
+
+
 class InterruptedException(BaseException):
     pass
 
diff --git a/modules/sd_samplers_compvis.py b/modules/sd_samplers_compvis.py
index b1ee3be7..bdae8b40 100644
--- a/modules/sd_samplers_compvis.py
+++ b/modules/sd_samplers_compvis.py
@@ -11,7 +11,7 @@ import modules.models.diffusion.uni_pc
 
 
 samplers_data_compvis = [
-    sd_samplers_common.SamplerData('DDIM', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.ddim.DDIMSampler, model), [], {}),
+    sd_samplers_common.SamplerData('DDIM', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.ddim.DDIMSampler, model), [], {"default_eta_is_0": True, "uses_ensd": True}),
     sd_samplers_common.SamplerData('PLMS', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.plms.PLMSSampler, model), [], {}),
     sd_samplers_common.SamplerData('UniPC', lambda model: VanillaStableDiffusionSampler(modules.models.diffusion.uni_pc.UniPCSampler, model), [], {}),
 ]
@@ -134,7 +134,11 @@ class VanillaStableDiffusionSampler:
         self.update_step(x)
 
     def initialize(self, p):
-        self.eta = p.eta if p.eta is not None else shared.opts.eta_ddim
+        if self.is_ddim:
+            self.eta = p.eta if p.eta is not None else shared.opts.eta_ddim
+        else:
+            self.eta = 0.0
+
         if self.eta != 0.0:
             p.extra_generation_params["Eta DDIM"] = self.eta
 
diff --git a/modules/sd_samplers_kdiffusion.py b/modules/sd_samplers_kdiffusion.py
index 61f23ad7..5455561a 100644
--- a/modules/sd_samplers_kdiffusion.py
+++ b/modules/sd_samplers_kdiffusion.py
@@ -11,21 +11,21 @@ from modules.script_callbacks import CFGDenoisedParams, cfg_denoised_callback
 from modules.script_callbacks import AfterCFGCallbackParams, cfg_after_cfg_callback
 
 samplers_k_diffusion = [
-    ('Euler a', 'sample_euler_ancestral', ['k_euler_a', 'k_euler_ancestral'], {}),
+    ('Euler a', 'sample_euler_ancestral', ['k_euler_a', 'k_euler_ancestral'], {"uses_ensd": True}),
     ('Euler', 'sample_euler', ['k_euler'], {}),
     ('LMS', 'sample_lms', ['k_lms'], {}),
     ('Heun', 'sample_heun', ['k_heun'], {}),
     ('DPM2', 'sample_dpm_2', ['k_dpm_2'], {'discard_next_to_last_sigma': True}),
-    ('DPM2 a', 'sample_dpm_2_ancestral', ['k_dpm_2_a'], {'discard_next_to_last_sigma': True}),
-    ('DPM++ 2S a', 'sample_dpmpp_2s_ancestral', ['k_dpmpp_2s_a'], {}),
+    ('DPM2 a', 'sample_dpm_2_ancestral', ['k_dpm_2_a'], {'discard_next_to_last_sigma': True, "uses_ensd": True}),
+    ('DPM++ 2S a', 'sample_dpmpp_2s_ancestral', ['k_dpmpp_2s_a'], {"uses_ensd": True}),
     ('DPM++ 2M', 'sample_dpmpp_2m', ['k_dpmpp_2m'], {}),
     ('DPM++ SDE', 'sample_dpmpp_sde', ['k_dpmpp_sde'], {}),
-    ('DPM fast', 'sample_dpm_fast', ['k_dpm_fast'], {}),
-    ('DPM adaptive', 'sample_dpm_adaptive', ['k_dpm_ad'], {}),
+    ('DPM fast', 'sample_dpm_fast', ['k_dpm_fast'], {"uses_ensd": True}),
+    ('DPM adaptive', 'sample_dpm_adaptive', ['k_dpm_ad'], {"uses_ensd": True}),
     ('LMS Karras', 'sample_lms', ['k_lms_ka'], {'scheduler': 'karras'}),
-    ('DPM2 Karras', 'sample_dpm_2', ['k_dpm_2_ka'], {'scheduler': 'karras', 'discard_next_to_last_sigma': True}),
-    ('DPM2 a Karras', 'sample_dpm_2_ancestral', ['k_dpm_2_a_ka'], {'scheduler': 'karras', 'discard_next_to_last_sigma': True}),
-    ('DPM++ 2S a Karras', 'sample_dpmpp_2s_ancestral', ['k_dpmpp_2s_a_ka'], {'scheduler': 'karras'}),
+    ('DPM2 Karras', 'sample_dpm_2', ['k_dpm_2_ka'], {'scheduler': 'karras', 'discard_next_to_last_sigma': True, "uses_ensd": True}),
+    ('DPM2 a Karras', 'sample_dpm_2_ancestral', ['k_dpm_2_a_ka'], {'scheduler': 'karras', 'discard_next_to_last_sigma': True, "uses_ensd": True}),
+    ('DPM++ 2S a Karras', 'sample_dpmpp_2s_ancestral', ['k_dpmpp_2s_a_ka'], {'scheduler': 'karras', "uses_ensd": True}),
     ('DPM++ 2M Karras', 'sample_dpmpp_2m', ['k_dpmpp_2m_ka'], {'scheduler': 'karras'}),
     ('DPM++ SDE Karras', 'sample_dpmpp_sde', ['k_dpmpp_sde_ka'], {'scheduler': 'karras'}),
 ]
-- 
cgit v1.2.3


From 56a2672831751480f94a018f861f0143a8234ae8 Mon Sep 17 00:00:00 2001
From: AUTOMATIC <16777216c@gmail.com>
Date: Wed, 17 May 2023 09:24:01 +0300
Subject: return live preview defaults to how they were only download TAESD
 model when it's needed return calculations in single_sample_to_image to just
 if/elif/elif blocks keep taesd model in its own directory

---
 modules/sd_samplers_common.py | 29 +++++++++++++++--------------
 modules/sd_vae_taesd.py       | 18 +++++++++++++++---
 modules/shared.py             |  2 +-
 webui.py                      | 11 -----------
 4 files changed, 31 insertions(+), 29 deletions(-)

(limited to 'modules/sd_samplers_common.py')

diff --git a/modules/sd_samplers_common.py b/modules/sd_samplers_common.py
index b1e8a780..20a9af20 100644
--- a/modules/sd_samplers_common.py
+++ b/modules/sd_samplers_common.py
@@ -22,28 +22,29 @@ def setup_img2img_steps(p, steps=None):
     return steps, t_enc
 
 
-approximation_indexes = {"Full": 0, "Tiny AE": 1, "Approx NN": 2, "Approx cheap": 3}
+approximation_indexes = {"Full": 0, "Approx NN": 1, "Approx cheap": 2, "TAESD": 3}
 
 
 def single_sample_to_image(sample, approximation=None):
-    if approximation is None or approximation not in approximation_indexes.keys():
-        approximation = approximation_indexes.get(opts.show_progress_type, 1)
 
-    if approximation == 1:
-        x_sample = sd_vae_taesd.decode()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
-        x_sample = sd_vae_taesd.TAESD.unscale_latents(x_sample)
-        x_sample = torch.clamp((x_sample * 0.25) + 0.5, 0, 1)
+    if approximation is None:
+        approximation = approximation_indexes.get(opts.show_progress_type, 0)
+
+    if approximation == 2:
+        x_sample = sd_vae_approx.cheap_approximation(sample)
+    elif approximation == 1:
+        x_sample = sd_vae_approx.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
+    elif approximation == 3:
+        x_sample = sd_vae_taesd.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
+        x_sample = sd_vae_taesd.TAESD.unscale_latents(x_sample)  # returns value in [-2, 2]
+        x_sample = x_sample * 0.5
     else:
-        if approximation == 3:
-            x_sample = sd_vae_approx.cheap_approximation(sample)
-        elif approximation == 2:
-            x_sample = sd_vae_approx.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
-        else:
-            x_sample = processing.decode_first_stage(shared.sd_model, sample.unsqueeze(0))[0]
-        x_sample = torch.clamp((x_sample + 1.0) / 2.0, min=0.0, max=1.0)
+        x_sample = processing.decode_first_stage(shared.sd_model, sample.unsqueeze(0))[0]
 
+    x_sample = torch.clamp((x_sample + 1.0) / 2.0, min=0.0, max=1.0)
     x_sample = 255. * np.moveaxis(x_sample.cpu().numpy(), 0, 2)
     x_sample = x_sample.astype(np.uint8)
+
     return Image.fromarray(x_sample)
 
 
diff --git a/modules/sd_vae_taesd.py b/modules/sd_vae_taesd.py
index 927a7298..d23812ef 100644
--- a/modules/sd_vae_taesd.py
+++ b/modules/sd_vae_taesd.py
@@ -61,16 +61,28 @@ class TAESD(nn.Module):
         return x.sub(TAESD.latent_shift).mul(2 * TAESD.latent_magnitude)
 
 
-def decode():
+def download_model(model_path):
+    model_url = 'https://github.com/madebyollin/taesd/raw/main/taesd_decoder.pth'
+
+    if not os.path.exists(model_path):
+        os.makedirs(os.path.dirname(model_path), exist_ok=True)
+
+        print(f'Downloading TAESD decoder to: {model_path}')
+        torch.hub.download_url_to_file(model_url, model_path)
+
+
+def model():
     global sd_vae_taesd
 
     if sd_vae_taesd is None:
-        model_path = os.path.join(paths_internal.models_path, "VAE-approx", "taesd_decoder.pth")
+        model_path = os.path.join(paths_internal.models_path, "VAE-taesd", "taesd_decoder.pth")
+        download_model(model_path)
+
         if os.path.exists(model_path):
             sd_vae_taesd = TAESD(model_path)
             sd_vae_taesd.eval()
             sd_vae_taesd.to(devices.device, devices.dtype)
         else:
-            raise FileNotFoundError('Tiny AE model not found')
+            raise FileNotFoundError('TAESD model not found')
 
     return sd_vae_taesd.decoder
diff --git a/modules/shared.py b/modules/shared.py
index 6760a900..96036d38 100644
--- a/modules/shared.py
+++ b/modules/shared.py
@@ -425,7 +425,7 @@ 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("Tiny AE", "Image creation progress preview mode", gr.Radio, {"choices": ["Full", "Tiny AE", "Approx NN", "Approx cheap"]}),
+    "show_progress_type": OptionInfo("Approx NN", "Image creation progress preview mode", gr.Radio, {"choices": ["Full", "Approx NN", "Approx cheap", "TAESD"]}),
     "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")
 }))
diff --git a/webui.py b/webui.py
index 0aa03ea8..727ebd31 100644
--- a/webui.py
+++ b/webui.py
@@ -144,21 +144,10 @@ Use --skip-version-check commandline argument to disable this check.
             """.strip())
 
 
-def check_taesd():
-    from modules.paths_internal import models_path
-
-    model_url = 'https://github.com/madebyollin/taesd/raw/main/taesd_decoder.pth'
-    model_path = os.path.join(models_path, "VAE-approx", "taesd_decoder.pth")
-    if not os.path.exists(model_path):
-        print('From taesd repo download decoder model')
-        torch.hub.download_url_to_file(model_url, model_path)
-
-
 def initialize():
     fix_asyncio_event_loop_policy()
 
     check_versions()
-    check_taesd()
 
     extensions.list_extensions()
     localization.list_localizations(cmd_opts.localizations_dir)
-- 
cgit v1.2.3


From 7a13a3f4ba86dc44fcf7d9944b179018744862f5 Mon Sep 17 00:00:00 2001
From: Sakura-Luna <53183413+Sakura-Luna@users.noreply.github.com>
Date: Wed, 17 May 2023 17:39:07 +0800
Subject: TAESD fix

---
 modules/sd_samplers_common.py | 9 +++++----
 modules/sd_vae_taesd.py       | 2 +-
 2 files changed, 6 insertions(+), 5 deletions(-)

(limited to 'modules/sd_samplers_common.py')

diff --git a/modules/sd_samplers_common.py b/modules/sd_samplers_common.py
index ceda6a35..d99c933d 100644
--- a/modules/sd_samplers_common.py
+++ b/modules/sd_samplers_common.py
@@ -35,13 +35,14 @@ def single_sample_to_image(sample, approximation=None):
     elif approximation == 1:
         x_sample = sd_vae_approx.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
     elif approximation == 3:
-        x_sample = sd_vae_taesd.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
-        x_sample = sd_vae_taesd.TAESD.unscale_latents(x_sample)  # returns value in [-2, 2]
-        x_sample = x_sample * 0.5
+        x_sample = sample * 1.5
+        x_sample = sd_vae_taesd.model()(x_sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
     else:
         x_sample = processing.decode_first_stage(shared.sd_model, sample.unsqueeze(0))[0]
 
-    x_sample = torch.clamp((x_sample + 1.0) / 2.0, min=0.0, max=1.0)
+    if approximation != 3:
+        x_sample = (x_sample + 1.0) / 2.0
+    x_sample = torch.clamp(x_sample, min=0.0, max=1.0)
     x_sample = 255. * np.moveaxis(x_sample.cpu().numpy(), 0, 2)
     x_sample = x_sample.astype(np.uint8)
 
diff --git a/modules/sd_vae_taesd.py b/modules/sd_vae_taesd.py
index d23812ef..5e8496e8 100644
--- a/modules/sd_vae_taesd.py
+++ b/modules/sd_vae_taesd.py
@@ -45,7 +45,7 @@ def decoder():
 
 
 class TAESD(nn.Module):
-    latent_magnitude = 2
+    latent_magnitude = 3
     latent_shift = 0.5
 
     def __init__(self, decoder_path="taesd_decoder.pth"):
-- 
cgit v1.2.3


From 1210548cba9dbd78378a710d75601922addefca2 Mon Sep 17 00:00:00 2001
From: AUTOMATIC <16777216c@gmail.com>
Date: Wed, 17 May 2023 14:53:39 +0300
Subject: simplify single_sample_to_image

---
 modules/sd_samplers_common.py | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

(limited to 'modules/sd_samplers_common.py')

diff --git a/modules/sd_samplers_common.py b/modules/sd_samplers_common.py
index d99c933d..763829f1 100644
--- a/modules/sd_samplers_common.py
+++ b/modules/sd_samplers_common.py
@@ -26,22 +26,19 @@ approximation_indexes = {"Full": 0, "Approx NN": 1, "Approx cheap": 2, "TAESD":
 
 
 def single_sample_to_image(sample, approximation=None):
-
     if approximation is None:
         approximation = approximation_indexes.get(opts.show_progress_type, 0)
 
     if approximation == 2:
-        x_sample = sd_vae_approx.cheap_approximation(sample)
+        x_sample = sd_vae_approx.cheap_approximation(sample) * 0.5 + 0.5
     elif approximation == 1:
-        x_sample = sd_vae_approx.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
+        x_sample = sd_vae_approx.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach() * 0.5 + 0.5
     elif approximation == 3:
         x_sample = sample * 1.5
         x_sample = sd_vae_taesd.model()(x_sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()
     else:
-        x_sample = processing.decode_first_stage(shared.sd_model, sample.unsqueeze(0))[0]
+        x_sample = processing.decode_first_stage(shared.sd_model, sample.unsqueeze(0))[0] * 0.5 + 0.5
 
-    if approximation != 3:
-        x_sample = (x_sample + 1.0) / 2.0
     x_sample = torch.clamp(x_sample, min=0.0, max=1.0)
     x_sample = 255. * np.moveaxis(x_sample.cpu().numpy(), 0, 2)
     x_sample = x_sample.astype(np.uint8)
-- 
cgit v1.2.3