Merge branch 'AUTOMATIC1111:master' into hide_ui_tabs

README.md

@@ -104,8 +104,7 @@ Alternatively, use online services (like Google Colab):
 1. Install [Python 3.10.6](https://www.python.org/downloads/windows/), checking "Add Python to PATH"
 2. Install [git](https://git-scm.com/download/win).
 3. Download the stable-diffusion-webui repository, for example by running `git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git`.
-4. Place stable diffusion checkpoint (`model.ckpt`) in the `models/Stable-diffusion` directory (see [dependencies](https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Dependencies) for where to get it).
-5. Run `webui-user.bat` from Windows Explorer as normal, non-administrator, user.
+4. Run `webui-user.bat` from Windows Explorer as normal, non-administrator, user.
 
 ### Automatic Installation on Linux
 1. Install the dependencies:
@@ -121,7 +120,7 @@ sudo pacman -S wget git python3
 ```bash
 bash <(wget -qO- https://raw.githubusercontent.com/AUTOMATIC1111/stable-diffusion-webui/master/webui.sh)
 ```
-
+3. Run `webui.sh`.
 ### Installation on Apple Silicon
 
 Find the instructions [here](https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Installation-on-Apple-Silicon).

javascript/hints.js

@@ -8,8 +8,8 @@ titles = {
 	"DDIM": "Denoising Diffusion Implicit Models - best at inpainting",
 	"DPM adaptive": "Ignores step count - uses a number of steps determined by the CFG and resolution", 
 
-	"Batch count": "How many batches of images to create",
-	"Batch size": "How many image to create in a single batch",
+	"Batch count": "How many batches of images to create (has no impact on generation performance or VRAM usage)",
+	"Batch size": "How many image to create in a single batch (increases generation performance at cost of higher VRAM usage)",
     "CFG Scale": "Classifier Free Guidance Scale - how strongly the image should conform to prompt - lower values produce more creative results",
     "Seed": "A value that determines the output of random number generator - if you create an image with same parameters and seed as another image, you'll get the same result",
     "\u{1f3b2}\ufe0f": "Set seed to -1, which will cause a new random number to be used every time",

launch.py

@@ -242,7 +242,7 @@ def prepare_environment():
 
     sys.argv += shlex.split(commandline_args)
 
-    parser = argparse.ArgumentParser()
+    parser = argparse.ArgumentParser(add_help=False)
     parser.add_argument("--ui-settings-file", type=str, help="filename to use for ui settings", default='config.json')
     args, _ = parser.parse_known_args(sys.argv)
 

modules/api/api.py

@@ -498,7 +498,7 @@ def train_hypernetwork(self, args: dict):
             if not apply_optimizations:
                 sd_hijack.undo_optimizations()
             try:
-                hypernetwork, filename = train_hypernetwork(*args)
+                hypernetwork, filename = train_hypernetwork(**args)
             except Exception as e:
                 error = e
             finally:

modules/esrgan_model_arch.py

@@ -1,5 +1,6 @@
 # this file is adapted from https://github.com/victorca25/iNNfer
 
+from collections import OrderedDict
 import math
 import functools
 import torch

modules/extensions.py

@@ -2,6 +2,7 @@
 import sys
 import traceback
 
+import time
 import git
 
 from modules import paths, shared
@@ -25,6 +26,7 @@ def __init__(self, name, path, enabled=True, is_builtin=False):
         self.status = ''
         self.can_update = False
         self.is_builtin = is_builtin
+        self.version = ''
 
         repo = None
         try:
@@ -40,6 +42,10 @@ def __init__(self, name, path, enabled=True, is_builtin=False):
             try:
                 self.remote = next(repo.remote().urls, None)
                 self.status = 'unknown'
+                head = repo.head.commit
+                ts = time.asctime(time.gmtime(repo.head.commit.committed_date))
+                self.version = f'{head.hexsha[:8]} ({ts})'
+
             except Exception:
                 self.remote = None
 

modules/generation_parameters_copypaste.py

@@ -74,8 +74,8 @@ def image_from_url_text(filedata):
     return image
 
 
-def add_paste_fields(tabname, init_img, fields):
-    paste_fields[tabname] = {"init_img": init_img, "fields": fields}
+def add_paste_fields(tabname, init_img, fields, override_settings_component=None):
+    paste_fields[tabname] = {"init_img": init_img, "fields": fields, "override_settings_component": override_settings_component}
 
     # backwards compatibility for existing extensions
     import modules.ui
@@ -110,6 +110,7 @@ def connect_paste_params_buttons():
     for binding in registered_param_bindings:
         destination_image_component = paste_fields[binding.tabname]["init_img"]
         fields = paste_fields[binding.tabname]["fields"]
+        override_settings_component = binding.override_settings_component or paste_fields[binding.tabname]["override_settings_component"]
 
         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)
@@ -130,7 +131,7 @@ def connect_paste_params_buttons():
             )
 
         if binding.source_text_component is not None and fields is not None:
-            connect_paste(binding.paste_button, fields, binding.source_text_component, binding.override_settings_component, binding.tabname)
+            connect_paste(binding.paste_button, fields, binding.source_text_component, override_settings_component, binding.tabname)
 
         if binding.source_tabname is not None and fields is not None:
             paste_field_names = ['Prompt', 'Negative prompt', 'Steps', 'Face restoration'] + (["Seed"] if shared.opts.send_seed else [])

modules/hypernetworks/hypernetwork.py

@@ -380,8 +380,8 @@ def apply_single_hypernetwork(hypernetwork, context_k, context_v, layer=None):
         layer.hyper_k = hypernetwork_layers[0]
         layer.hyper_v = hypernetwork_layers[1]
 
-    context_k = hypernetwork_layers[0](context_k)
-    context_v = hypernetwork_layers[1](context_v)
+    context_k = devices.cond_cast_unet(hypernetwork_layers[0](devices.cond_cast_float(context_k)))
+    context_v = devices.cond_cast_unet(hypernetwork_layers[1](devices.cond_cast_float(context_v)))
     return context_k, context_v
 
 
@@ -496,7 +496,7 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None,
     shared.reload_hypernetworks()
 
 
-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):
+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, use_weight, 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
 
@@ -554,7 +554,7 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi
 
     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, varsize=varsize)
+    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, use_weight=use_weight)
 
     if shared.opts.save_training_settings_to_txt:
         saved_params = dict(
@@ -640,13 +640,19 @@ def train_hypernetwork(id_task, hypernetwork_name, learn_rate, batch_size, gradi
 
                 with devices.autocast():
                     x = batch.latent_sample.to(devices.device, non_blocking=pin_memory)
+                    if use_weight:
+                        w = batch.weight.to(devices.device, non_blocking=pin_memory)
                     if tag_drop_out != 0 or shuffle_tags:
                         shared.sd_model.cond_stage_model.to(devices.device)
                         c = shared.sd_model.cond_stage_model(batch.cond_text).to(devices.device, non_blocking=pin_memory)
                         shared.sd_model.cond_stage_model.to(devices.cpu)
                     else:
                         c = stack_conds(batch.cond).to(devices.device, non_blocking=pin_memory)
-                    loss = shared.sd_model(x, c)[0] / gradient_step
+                    if use_weight:
+                        loss = shared.sd_model.weighted_forward(x, c, w)[0] / gradient_step
+                        del w
+                    else:
+                        loss = shared.sd_model.forward(x, c)[0] / gradient_step
                     del x
                     del c
 

modules/images.py

@@ -18,7 +18,7 @@
 import json
 import hashlib
 
-from modules import sd_samplers, shared, script_callbacks
+from modules import sd_samplers, shared, script_callbacks, errors
 from modules.shared import opts, cmd_opts
 
 LANCZOS = (Image.Resampling.LANCZOS if hasattr(Image, 'Resampling') else Image.LANCZOS)
@@ -553,6 +553,8 @@ def _atomically_save_image(image_to_save, filename_without_extension, extension)
         elif extension.lower() in (".jpg", ".jpeg", ".webp"):
             if image_to_save.mode == 'RGBA':
                 image_to_save = image_to_save.convert("RGB")
+            elif image_to_save.mode == 'I;16':
+                image_to_save = image_to_save.point(lambda p: p * 0.0038910505836576).convert("RGB" if extension.lower() == ".webp" else "L")
 
             image_to_save.save(temp_file_path, format=image_format, quality=opts.jpeg_quality)
 
@@ -575,17 +577,19 @@ def _atomically_save_image(image_to_save, filename_without_extension, extension)
 
     image.already_saved_as = fullfn
 
-    target_side_length = 4000
-    oversize = image.width > target_side_length or image.height > target_side_length
-    if opts.export_for_4chan and (oversize or os.stat(fullfn).st_size > 4 * 1024 * 1024):
+    oversize = image.width > opts.target_side_length or image.height > opts.target_side_length
+    if opts.export_for_4chan and (oversize or os.stat(fullfn).st_size > opts.img_downscale_threshold * 1024 * 1024):
         ratio = image.width / image.height
 
         if oversize and ratio > 1:
-            image = image.resize((target_side_length, image.height * target_side_length // image.width), LANCZOS)
+            image = image.resize((opts.target_side_length, image.height * opts.target_side_length // image.width), LANCZOS)
         elif oversize:
-            image = image.resize((image.width * target_side_length // image.height, target_side_length), LANCZOS)
+            image = image.resize((image.width * opts.target_side_length // image.height, opts.target_side_length), LANCZOS)
 
-        _atomically_save_image(image, fullfn_without_extension, ".jpg")
+        try:
+            _atomically_save_image(image, fullfn_without_extension, ".jpg")
+        except Exception as e:
+            errors.display(e, "saving image as downscaled JPG")
 
     if opts.save_txt and info is not None:
         txt_fullfn = f"{fullfn_without_extension}.txt"

modules/img2img.py

@@ -73,6 +73,8 @@ def process_batch(p, input_dir, output_dir, inpaint_mask_dir, args):
 
             if not save_normally:
                 os.makedirs(output_dir, exist_ok=True)
+                if processed_image.mode == 'RGBA':
+                    processed_image = processed_image.convert("RGB")
                 processed_image.save(os.path.join(output_dir, filename))
 
 

modules/processing.py

@@ -543,8 +543,6 @@ def infotext(iteration=0, position_in_batch=0):
     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)
 
@@ -582,13 +580,6 @@ def get_conds_with_caching(function, required_prompts, steps, cache):
             if shared.opts.live_previews_enable and opts.show_progress_type == "Approx NN":
                 sd_vae_approx.model()
 
-            if not p.disable_extra_networks:
-                extra_networks.activate(p, extra_network_data)
-
-        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))
-
         if state.job_count == -1:
             state.job_count = p.n_iter
 
@@ -609,11 +600,24 @@ def get_conds_with_caching(function, required_prompts, steps, cache):
             if len(prompts) == 0:
                 break
 
-            prompts, _ = extra_networks.parse_prompts(prompts)
+            prompts, extra_network_data = extra_networks.parse_prompts(prompts)
+
+            if not p.disable_extra_networks:
+                with devices.autocast():
+                    extra_networks.activate(p, extra_network_data)
 
             if p.scripts is not None:
                 p.scripts.process_batch(p, batch_number=n, prompts=prompts, seeds=seeds, subseeds=subseeds)
 
+            # params.txt should be saved after scripts.process_batch, since the
+            # infotext could be modified by that callback
+            # Example: a wildcard processed by process_batch sets an extra model
+            # strength, which is saved as "Model Strength: 1.0" in the infotext
+            if n == 0:
+                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))
+
             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)
 

modules/script_callbacks.py

@@ -46,6 +46,18 @@ def __init__(self, x, image_cond, sigma, sampling_step, total_sampling_steps):
         """Total number of sampling steps planned"""
 
 
+class CFGDenoisedParams:
+    def __init__(self, x, sampling_step, total_sampling_steps):
+        self.x = x
+        """Latent image representation in the process of being denoised"""
+
+        self.sampling_step = sampling_step
+        """Current Sampling step number"""
+
+        self.total_sampling_steps = total_sampling_steps
+        """Total number of sampling steps planned"""
+
+
 class UiTrainTabParams:
     def __init__(self, txt2img_preview_params):
         self.txt2img_preview_params = txt2img_preview_params
@@ -68,6 +80,7 @@ def __init__(self, imgs, cols, rows):
     callbacks_before_image_saved=[],
     callbacks_image_saved=[],
     callbacks_cfg_denoiser=[],
+    callbacks_cfg_denoised=[],
     callbacks_before_component=[],
     callbacks_after_component=[],
     callbacks_image_grid=[],
@@ -150,6 +163,14 @@ def cfg_denoiser_callback(params: CFGDenoiserParams):
             report_exception(c, 'cfg_denoiser_callback')
 
 
+def cfg_denoised_callback(params: CFGDenoisedParams):
+    for c in callback_map['callbacks_cfg_denoised']:
+        try:
+            c.callback(params)
+        except Exception:
+            report_exception(c, 'cfg_denoised_callback')
+
+
 def before_component_callback(component, **kwargs):
     for c in callback_map['callbacks_before_component']:
         try:
@@ -283,6 +304,14 @@ def on_cfg_denoiser(callback):
     add_callback(callback_map['callbacks_cfg_denoiser'], callback)
 
 
+def on_cfg_denoised(callback):
+    """register a function to be called in the kdiffussion cfg_denoiser method after building the inner model inputs.
+    The callback is called with one argument:
+        - params: CFGDenoisedParams - parameters to be passed to the inner model and sampling state details.
+    """
+    add_callback(callback_map['callbacks_cfg_denoised'], callback)
+
+
 def on_before_component(callback):
     """register a function to be called before a component is created.
     The callback is called with arguments: