From 70931652a4289e28d83869b6d10cf11e80a70345 Mon Sep 17 00:00:00 2001 From: RnDMonkey Date: Fri, 30 Sep 2022 18:02:46 -0700 Subject: [PATCH 01/60] [xy_grid] made -1 seed fixing apply to Var. seed too --- scripts/xy_grid.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/xy_grid.py b/scripts/xy_grid.py index 146663b0..9c078888 100644 --- a/scripts/xy_grid.py +++ b/scripts/xy_grid.py @@ -218,7 +218,7 @@ class Script(scripts.Script): ys = process_axis(y_opt, y_values) def fix_axis_seeds(axis_opt, axis_list): - if axis_opt.label == 'Seed': + if axis_opt.label == 'Seed' or 'Var. seed': return [int(random.randrange(4294967294)) if val is None or val == '' or val == -1 else val for val in axis_list] else: return axis_list From cf141157e7b49b0b3a6e57dc7aa0d1345158b4c8 Mon Sep 17 00:00:00 2001 From: RnDMonkey Date: Fri, 30 Sep 2022 22:02:29 -0700 Subject: [PATCH 02/60] Added X/Y plot parameters to extra_generation_params --- scripts/xy_grid.py | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/scripts/xy_grid.py b/scripts/xy_grid.py index 9c078888..d9f8d55b 100644 --- a/scripts/xy_grid.py +++ b/scripts/xy_grid.py @@ -244,6 +244,14 @@ class Script(scripts.Script): return process_images(pc) + if not x_opt.label == 'Nothing': + p.extra_generation_params["X/Y Plot X Type"] = x_opt.label + p.extra_generation_params["X Values"] = '{' + ", ".join([f'{x}' for x in xs]) + '}' + + if not y_opt.label == 'Nothing': + p.extra_generation_params["X/Y Plot Y Type"] = y_opt.label + p.extra_generation_params["Y Values"] = '{' + ", ".join([f'{y}' for y in ys]) + '}' + processed = draw_xy_grid( p, xs=xs, From eba0c29dbc3bad8c4e32f1fa3a03dc6f9caf1f5a Mon Sep 17 00:00:00 2001 From: RnDMonkey Date: Sat, 1 Oct 2022 13:56:29 -0700 Subject: [PATCH 03/60] Updated xy_grid infotext formatting, parser regex --- modules/generation_parameters_copypaste.py | 2 +- scripts/xy_grid.py | 12 ++++++++---- 2 files changed, 9 insertions(+), 5 deletions(-) diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index ac1ba7f4..39d67d94 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -1,7 +1,7 @@ import re import gradio as gr -re_param_code = r"\s*([\w ]+):\s*([^,]+)(?:,|$)" +re_param_code = r"\s*([\w ]+):\s*((?:{[^}]+})|(?:[^,]+))(?:,|$)" re_param = re.compile(re_param_code) re_params = re.compile(r"^(?:" + re_param_code + "){3,}$") re_imagesize = re.compile(r"^(\d+)x(\d+)$") diff --git a/scripts/xy_grid.py b/scripts/xy_grid.py index d9f8d55b..f87c6c1f 100644 --- a/scripts/xy_grid.py +++ b/scripts/xy_grid.py @@ -245,12 +245,16 @@ class Script(scripts.Script): return process_images(pc) if not x_opt.label == 'Nothing': - p.extra_generation_params["X/Y Plot X Type"] = x_opt.label - p.extra_generation_params["X Values"] = '{' + ", ".join([f'{x}' for x in xs]) + '}' + p.extra_generation_params["XY Plot X Type"] = x_opt.label + p.extra_generation_params["X Values"] = '{' + x_values + '}' + if x_opt.label in ["Seed","Var. seed"] and not no_fixed_seeds: + p.extra_generation_params["Fixed X Values"] = '{' + ", ".join([str(x) for x in xs])+ '}' if not y_opt.label == 'Nothing': - p.extra_generation_params["X/Y Plot Y Type"] = y_opt.label - p.extra_generation_params["Y Values"] = '{' + ", ".join([f'{y}' for y in ys]) + '}' + p.extra_generation_params["XY Plot Y Type"] = y_opt.label + p.extra_generation_params["Y Values"] = '{' + y_values + '}' + if y_opt.label in ["Seed","Var. seed"] and not no_fixed_seeds: + p.extra_generation_params["Fixed Y Values"] = '{' + ", ".join([str(y) for y in ys])+ '}' processed = draw_xy_grid( p, From b99a4f769f11ed74df0344a23069d3858613fbef Mon Sep 17 00:00:00 2001 From: RnDMonkey Date: Sat, 1 Oct 2022 14:26:12 -0700 Subject: [PATCH 04/60] fixed expression error in condition --- scripts/xy_grid.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/xy_grid.py b/scripts/xy_grid.py index f87c6c1f..f1f54d9c 100644 --- a/scripts/xy_grid.py +++ b/scripts/xy_grid.py @@ -218,7 +218,7 @@ class Script(scripts.Script): ys = process_axis(y_opt, y_values) def fix_axis_seeds(axis_opt, axis_list): - if axis_opt.label == 'Seed' or 'Var. seed': + if axis_opt.label in ["Seed","Var. seed"]: return [int(random.randrange(4294967294)) if val is None or val == '' or val == -1 else val for val in axis_list] else: return axis_list From f6a97868e57e44fba6c4283769fedd30ee11cacf Mon Sep 17 00:00:00 2001 From: RnDMonkey Date: Sat, 1 Oct 2022 14:36:09 -0700 Subject: [PATCH 05/60] fix to allow empty {} values --- modules/generation_parameters_copypaste.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/modules/generation_parameters_copypaste.py b/modules/generation_parameters_copypaste.py index 39d67d94..27d58dfd 100644 --- a/modules/generation_parameters_copypaste.py +++ b/modules/generation_parameters_copypaste.py @@ -1,7 +1,7 @@ import re import gradio as gr -re_param_code = r"\s*([\w ]+):\s*((?:{[^}]+})|(?:[^,]+))(?:,|$)" +re_param_code = r"\s*([\w ]+):\s*((?:{[^}]*})|(?:[^,]+))(?:,|$)" re_param = re.compile(re_param_code) re_params = re.compile(r"^(?:" + re_param_code + "){3,}$") re_imagesize = re.compile(r"^(\d+)x(\d+)$") From fe6e2362e8fa5d739de6997ab155a26686d20a49 Mon Sep 17 00:00:00 2001 From: RnDMonkey Date: Sun, 2 Oct 2022 22:04:28 -0700 Subject: [PATCH 06/60] Update xy_grid.py Changed XY Plot infotext value keys to not be so generic. --- scripts/xy_grid.py | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/scripts/xy_grid.py b/scripts/xy_grid.py index f1f54d9c..ae011a17 100644 --- a/scripts/xy_grid.py +++ b/scripts/xy_grid.py @@ -246,15 +246,15 @@ class Script(scripts.Script): if not x_opt.label == 'Nothing': p.extra_generation_params["XY Plot X Type"] = x_opt.label - p.extra_generation_params["X Values"] = '{' + x_values + '}' + p.extra_generation_params["XY Plot X Values"] = '{' + x_values + '}' if x_opt.label in ["Seed","Var. seed"] and not no_fixed_seeds: - p.extra_generation_params["Fixed X Values"] = '{' + ", ".join([str(x) for x in xs])+ '}' + p.extra_generation_params["XY Plot Fixed X Values"] = '{' + ", ".join([str(x) for x in xs])+ '}' if not y_opt.label == 'Nothing': p.extra_generation_params["XY Plot Y Type"] = y_opt.label - p.extra_generation_params["Y Values"] = '{' + y_values + '}' + p.extra_generation_params["XY Plot Y Values"] = '{' + y_values + '}' if y_opt.label in ["Seed","Var. seed"] and not no_fixed_seeds: - p.extra_generation_params["Fixed Y Values"] = '{' + ", ".join([str(y) for y in ys])+ '}' + p.extra_generation_params["XY Plot Fixed Y Values"] = '{' + ", ".join([str(y) for y in ys])+ '}' processed = draw_xy_grid( p, From 14c1c2b9351f16d43ba4e6b6c9062edad44a6bec Mon Sep 17 00:00:00 2001 From: Alexandre Simard Date: Wed, 19 Oct 2022 13:53:52 -0400 Subject: [PATCH 07/60] 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(-) 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) From 4fbdbddc18b21f712acae58bf41740d27023285f Mon Sep 17 00:00:00 2001 From: Alexandre Simard Date: Wed, 19 Oct 2022 15:21:36 -0400 Subject: [PATCH 08/60] Remove pad spaces from progress bar text --- javascript/progressbar.js | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/javascript/progressbar.js b/javascript/progressbar.js index 7a05726e..24ab4795 100644 --- a/javascript/progressbar.js +++ b/javascript/progressbar.js @@ -10,7 +10,7 @@ function check_progressbar(id_part, id_progressbar, id_progressbar_span, id_skip if(opts.show_progress_in_title && progressbar && progressbar.offsetParent){ if(progressbar.innerText){ - let newtitle = 'Stable Diffusion - ' + progressbar.innerText + let newtitle = 'Stable Diffusion - ' + progressbar.innerText.slice(2) if(document.title != newtitle){ document.title = newtitle; } From c4b5ca5778340b21288d84dfb8fe1d5773c886a8 Mon Sep 17 00:00:00 2001 From: Yuta Hayashibe Date: Thu, 27 Oct 2022 22:00:28 +0900 Subject: [PATCH 09/60] Truncate too long filename --- modules/images.py | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) 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: From 2a25729623717cc499e873752d9f4ebebd1e1078 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Fri, 28 Oct 2022 09:44:56 +0700 Subject: [PATCH 10/60] Gradient clipping in train tab --- modules/hypernetworks/hypernetwork.py | 10 +++++++++- modules/ui.py | 7 +++++++ 2 files changed, 16 insertions(+), 1 deletion(-) 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, From a133042c669f666763f5da0f4440abdc839db653 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Fri, 28 Oct 2022 10:01:46 +0700 Subject: [PATCH 11/60] Forgot to remove this from train_embedding --- modules/ui.py | 2 -- 1 file changed, 2 deletions(-) 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, From 1618df41bad092e068c61bf510b1e20856821ad5 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Fri, 28 Oct 2022 10:31:27 +0700 Subject: [PATCH 12/60] Gradient clipping for textual embedding --- modules/textual_inversion/textual_inversion.py | 11 ++++++++++- modules/ui.py | 2 ++ 2 files changed, 12 insertions(+), 1 deletion(-) 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, From 16451ca573220e49f2eaaab97580b6b91287c8c4 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Fri, 28 Oct 2022 17:16:23 +0700 Subject: [PATCH 13/60] 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(-) 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) From 840307f23738c38f7ac3ad636e53ccec66e71f8b Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Mon, 31 Oct 2022 13:49:24 +0700 Subject: [PATCH 14/60] 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(-) 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") From 4123be632a98f70cda06e14c2f556f7ad38cd436 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Mon, 31 Oct 2022 13:53:22 +0700 Subject: [PATCH 15/60] Fix merge conflicts --- modules/hypernetworks/hypernetwork.py | 17 ++++++----------- 1 file changed, 6 insertions(+), 11 deletions(-) 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: From d5ea878b2aa117588d85287cbd8983aa52177df5 Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Mon, 31 Oct 2022 13:54:40 +0700 Subject: [PATCH 16/60] Fix merge conflicts --- modules/hypernetworks/hypernetwork.py | 5 ----- 1 file changed, 5 deletions(-) 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 From cffc240a7327ae60671ff533469fc4ed4bf605de Mon Sep 17 00:00:00 2001 From: Nerogar Date: Sun, 23 Oct 2022 14:05:25 +0200 Subject: [PATCH 17/60] fixed textual inversion training with inpainting models --- .../textual_inversion/textual_inversion.py | 27 ++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) 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() From d624cb82a7c65a1ea04e4b6e23f0164a3ba25e25 Mon Sep 17 00:00:00 2001 From: Ikko Ashimine Date: Thu, 3 Nov 2022 01:05:00 +0900 Subject: [PATCH 18/60] Fix typo in ui.js interation -> interaction --- javascript/ui.js | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/javascript/ui.js b/javascript/ui.js index 7e116465..0308dce3 100644 --- a/javascript/ui.js +++ b/javascript/ui.js @@ -1,4 +1,4 @@ -// various functions for interation with ui.py not large enough to warrant putting them in separate files +// various functions for interaction with ui.py not large enough to warrant putting them in separate files function set_theme(theme){ gradioURL = window.location.href From bb832d7725187f8a8ab44faa6ee1b38cb5f600aa Mon Sep 17 00:00:00 2001 From: Muhammad Rizqi Nur Date: Sat, 5 Nov 2022 11:48:38 +0700 Subject: [PATCH 19/60] Simplify grad clip --- modules/hypernetworks/hypernetwork.py | 16 +++++++--------- modules/textual_inversion/textual_inversion.py | 16 +++++++--------- 2 files changed, 14 insertions(+), 18 deletions(-) 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() From 9fd457e21d6c809a69a1318f03d75f7b3e09b865 Mon Sep 17 00:00:00 2001 From: camenduru <54370274+camenduru@users.noreply.github.com> Date: Thu, 15 Dec 2022 21:57:48 +0300 Subject: [PATCH 20/60] allow_credentials and allow_headers for api from https://fastapi.tiangolo.com/tutorial/cors/ --- webui.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/webui.py b/webui.py index c2d0c6be..13a4d14a 100644 --- a/webui.py +++ b/webui.py @@ -90,11 +90,11 @@ def initialize(): def setup_cors(app): if cmd_opts.cors_allow_origins and cmd_opts.cors_allow_origins_regex: - app.add_middleware(CORSMiddleware, allow_origins=cmd_opts.cors_allow_origins.split(','), allow_origin_regex=cmd_opts.cors_allow_origins_regex, allow_methods=['*']) + app.add_middleware(CORSMiddleware, allow_origins=cmd_opts.cors_allow_origins.split(','), allow_origin_regex=cmd_opts.cors_allow_origins_regex, allow_methods=['*'], allow_credentials=True, allow_headers=['*']) elif cmd_opts.cors_allow_origins: - app.add_middleware(CORSMiddleware, allow_origins=cmd_opts.cors_allow_origins.split(','), allow_methods=['*']) + app.add_middleware(CORSMiddleware, allow_origins=cmd_opts.cors_allow_origins.split(','), allow_methods=['*'], allow_credentials=True, allow_headers=['*']) elif cmd_opts.cors_allow_origins_regex: - app.add_middleware(CORSMiddleware, allow_origin_regex=cmd_opts.cors_allow_origins_regex, allow_methods=['*']) + app.add_middleware(CORSMiddleware, allow_origin_regex=cmd_opts.cors_allow_origins_regex, allow_methods=['*'], allow_credentials=True, allow_headers=['*']) def create_api(app): From f23a822f1c9cb3bd2e8772c75af429e06515eaef Mon Sep 17 00:00:00 2001 From: Philpax Date: Sat, 24 Dec 2022 20:45:16 +1100 Subject: [PATCH 21/60] feat(api): include job_timestamp in progress --- modules/shared.py | 1 + 1 file changed, 1 insertion(+) 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, From 5ba04f9ec050a66e918571f07e8863f157f05b44 Mon Sep 17 00:00:00 2001 From: Nicolas Patry Date: Wed, 21 Dec 2022 13:45:58 +0100 Subject: [PATCH 22/60] Attempting to solve slow loads for `safetensors`. Fixes #5893 --- modules/sd_models.py | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) 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) From 5a523d13050a5ede43c473767f29dfe2e391136a Mon Sep 17 00:00:00 2001 From: Nicolas Patry Date: Tue, 27 Dec 2022 11:27:40 +0100 Subject: [PATCH 23/60] Version 0.2.7 Fixes Windows SAFETENSORS_FAST_GPU path. --- requirements_versions.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/requirements_versions.txt b/requirements_versions.txt index c126c8c4..52e98818 100644 --- a/requirements_versions.txt +++ b/requirements_versions.txt @@ -26,5 +26,5 @@ lark==1.1.2 inflection==0.5.1 GitPython==3.1.27 torchsde==0.2.5 -safetensors==0.2.5 +safetensors==0.2.7 httpcore<=0.15 From 9a3b0ee960b0c61c4f60e3081ae6f2098533d393 Mon Sep 17 00:00:00 2001 From: hithereai <121192995+hithereai@users.noreply.github.com> Date: Tue, 3 Jan 2023 11:22:06 +0200 Subject: [PATCH 24/60] update req.txt The old 'opencv-python' package is very limiting in terms of optical flow - so I propose a package change to 'opencv-contrib-python', which has more cv2.optflow methods. These are needed for optical flow trickery in auto1111 and its extensions, and it cannot be installed by an extension as only a single package of opencv needs to be installed for optical flow to work properly. Change of the main one is Inevitable. --- requirements.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/requirements.txt b/requirements.txt index e2c3876b..4f09385f 100644 --- a/requirements.txt +++ b/requirements.txt @@ -9,7 +9,7 @@ gradio==3.15.0 invisible-watermark numpy omegaconf -opencv-python +opencv-contrib-python requests piexif Pillow From bddebe09edeb6a18f2c06986d5658a7be3a563ea Mon Sep 17 00:00:00 2001 From: Shondoit Date: Tue, 3 Jan 2023 10:26:37 +0100 Subject: [PATCH 25/60] 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 +- .../textual_inversion/textual_inversion.py | 40 +++++++++++++++---- 2 files changed, 33 insertions(+), 9 deletions(-) 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 From aaa4c2aacbb6523077334093c81bd475d757f7a1 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 09:45:16 -0500 Subject: [PATCH 26/60] add api logging --- modules/api/api.py | 24 +++++++++++++++++++++++- modules/shared.py | 1 + 2 files changed, 24 insertions(+), 1 deletion(-) 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) From 1d9dc48efda2e8da6d13fc62e65500198a9b041c Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 10:21:51 -0500 Subject: [PATCH 27/60] init job and add info to model merge --- modules/extras.py | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) 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)] From 192ddc04d6de0d780f73aa5fbaa8c66cd4642e1c Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 10:34:51 -0500 Subject: [PATCH 28/60] 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(-) 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 From cec209981ee988536c2521297baf9bc1b256005f Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 10:58:52 -0500 Subject: [PATCH 29/60] log only sdapi --- modules/api/api.py | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) 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 From d8d206c1685d1e7027d4af82ed18d106f41d1cc4 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Tue, 3 Jan 2023 11:01:04 -0500 Subject: [PATCH 30/60] add state to interrogate --- modules/interrogate.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) 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 From 7c89f3718f9f078113833a88a86f02d3205855b4 Mon Sep 17 00:00:00 2001 From: MMaker Date: Tue, 3 Jan 2023 12:46:48 -0500 Subject: [PATCH 31/60] Add image paste fallback Fixes Firefox pasting support (and possibly other browsers) --- javascript/dragdrop.js | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/javascript/dragdrop.js b/javascript/dragdrop.js index 3ed1cb3c..fe008924 100644 --- a/javascript/dragdrop.js +++ b/javascript/dragdrop.js @@ -9,11 +9,19 @@ function dropReplaceImage( imgWrap, files ) { return; } + const tmpFile = files[0]; + imgWrap.querySelector('.modify-upload button + button, .touch-none + div button + button')?.click(); const callback = () => { const fileInput = imgWrap.querySelector('input[type="file"]'); if ( fileInput ) { - fileInput.files = files; + if ( files.length === 0 ) { + files = new DataTransfer(); + files.items.add(tmpFile); + fileInput.files = files.files; + } else { + fileInput.files = files; + } fileInput.dispatchEvent(new Event('change')); } }; From 917b5bd8d0cd47c9dc241c1852ccd440a8c61668 Mon Sep 17 00:00:00 2001 From: Max Weber Date: Tue, 3 Jan 2023 18:19:56 -0700 Subject: [PATCH 32/60] ui: save dropdown sampling method to the ui-config --- modules/ui.py | 1 + 1 file changed, 1 insertion(+) 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}"): From 4fc81542077af73610279ad7b6b26e38718a0f81 Mon Sep 17 00:00:00 2001 From: Gerschel Date: Tue, 3 Jan 2023 23:25:34 -0800 Subject: [PATCH 33/60] better targetting, class tabs was autoassigned I moved a preset manager into quicksettings, this function was targeting my component instead of the tabs. This is because class tabs is autoassigned, while element id #tabs is not, this allows a tabbed component to live in the quicksettings. --- script.js | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/script.js b/script.js index 9748ec90..0e117d06 100644 --- a/script.js +++ b/script.js @@ -4,7 +4,7 @@ function gradioApp() { } function get_uiCurrentTab() { - return gradioApp().querySelector('.tabs button:not(.border-transparent)') + return gradioApp().querySelector('#tabs button:not(.border-transparent)') } function get_uiCurrentTabContent() { From e5b7ee910e7bb88f08e8876b5732cb034c6fe529 Mon Sep 17 00:00:00 2001 From: MMaker Date: Wed, 4 Jan 2023 04:22:01 -0500 Subject: [PATCH 34/60] fix: Save full res of intermediate step --- modules/processing.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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") From 02d7abf5141431b9a3a8a189bb3136c71abd5e79 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 12:35:07 +0300 Subject: [PATCH 35/60] 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 | 24 +++++++++++++++++------- modules/shared.py | 9 +++++++-- webui.py | 12 ++++++++++-- 4 files changed, 57 insertions(+), 13 deletions(-) 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) + 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) - 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) + 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 diff --git a/webui.py b/webui.py index c7d55a97..13375e71 100644 --- a/webui.py +++ b/webui.py @@ -9,7 +9,7 @@ from fastapi import FastAPI from fastapi.middleware.cors import CORSMiddleware from fastapi.middleware.gzip import GZipMiddleware -from modules import import_hook +from modules import import_hook, errors from modules.call_queue import wrap_queued_call, queue_lock, wrap_gradio_gpu_call from modules.paths import script_path @@ -61,7 +61,15 @@ def initialize(): modelloader.load_upscalers() modules.sd_vae.refresh_vae_list() - modules.sd_models.load_model() + + try: + modules.sd_models.load_model() + except Exception as e: + errors.display(e, "loading stable diffusion model") + print("", file=sys.stderr) + print("Stable diffusion model failed to load, exiting", file=sys.stderr) + exit(1) + shared.opts.onchange("sd_model_checkpoint", wrap_queued_call(lambda: modules.sd_models.reload_model_weights())) shared.opts.onchange("sd_vae", wrap_queued_call(lambda: modules.sd_vae.reload_vae_weights()), call=False) shared.opts.onchange("sd_vae_as_default", wrap_queued_call(lambda: modules.sd_vae.reload_vae_weights()), call=False) From 8d8a05a3bbb50fdfeab51679a919d2487bd97976 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 12:47:42 +0300 Subject: [PATCH 36/60] 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(-) 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) From 96cf15bedecbed97ef9b70b8413d543a9aee5adf Mon Sep 17 00:00:00 2001 From: MMaker Date: Wed, 4 Jan 2023 05:12:06 -0500 Subject: [PATCH 37/60] Add new latent upscale modes --- modules/shared.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) 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 = [] From 15fd0b8bc4734ea85bca1acfb12b51465ab9817d Mon Sep 17 00:00:00 2001 From: MMaker Date: Wed, 4 Jan 2023 05:12:54 -0500 Subject: [PATCH 38/60] Update processing.py --- modules/processing.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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. From 4ec6470a1a2d9430b91266426f995e48f59564e1 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 13:26:23 +0300 Subject: [PATCH 39/60] fix checkpoint list API --- modules/api/api.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) 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] From b2151b934fe0a3613570c6abd7615d3788fd1c8f Mon Sep 17 00:00:00 2001 From: MMaker Date: Wed, 4 Jan 2023 05:36:18 -0500 Subject: [PATCH 40/60] 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(-) 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}, } From 3bd737767b071878ea980e94b8705f603bcf545e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 14:20:32 +0300 Subject: [PATCH 41/60] disable broken API logging --- modules/api/api.py | 1 - 1 file changed, 1 deletion(-) 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) From 0cd6399b8b1699b8b7acad6f0ad2988111fe618e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 14:29:13 +0300 Subject: [PATCH 42/60] fix broken inpainting model --- modules/sd_models.py | 3 --- 1 file changed, 3 deletions(-) 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 From 11b8160a086c434d5baf4971edda46e6d2126800 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 4 Jan 2023 06:36:57 -0500 Subject: [PATCH 43/60] fix typo --- modules/api/api.py | 1 + 1 file changed, 1 insertion(+) 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) From 642142556d8ecdea9beb86d7618b628b1803ab98 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 15:09:53 +0300 Subject: [PATCH 44/60] 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(-) 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) From 21ee77db314ede7ccbb18787962347c09a4df0c7 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 4 Jan 2023 08:04:38 -0500 Subject: [PATCH 45/60] add cross-attention info --- modules/sd_hijack.py | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) 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 From 1cfd8aec4ae5a6ca1afd67b44cb4ef6dd14d8c34 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 16:05:42 +0300 Subject: [PATCH 46/60] 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(-) 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')) From 79c682ad4f2d982b26fa1a15044582d1005134f9 Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 4 Jan 2023 08:20:42 -0500 Subject: [PATCH 47/60] fix jpeg --- modules/extras.py | 2 -- modules/images.py | 2 ++ requirements_versions.txt | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) 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: diff --git a/requirements_versions.txt b/requirements_versions.txt index 975102d9..7ae118cb 100644 --- a/requirements_versions.txt +++ b/requirements_versions.txt @@ -5,7 +5,7 @@ basicsr==1.4.2 gfpgan==1.3.8 gradio==3.15.0 numpy==1.23.3 -Pillow==9.2.0 +Pillow==9.3.0 realesrgan==0.3.0 torch omegaconf==2.2.3 From 4d66bf2c0d27702cc83b9cc57ebb1f359d18d938 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 17:24:46 +0300 Subject: [PATCH 48/60] add infotext to "-before-highres-fix" images --- modules/processing.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) 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]): From 184e670126f5fc50ba56fa0fedcf0cf60e45ed7e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 17:45:01 +0300 Subject: [PATCH 49/60] fix the merge --- modules/textual_inversion/textual_inversion.py | 14 +++++--------- 1 file changed, 5 insertions(+), 9 deletions(-) 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 From 590c5ae016ae494f4873ca20079b30684ea3060c Mon Sep 17 00:00:00 2001 From: Vladimir Mandic Date: Wed, 4 Jan 2023 09:48:54 -0500 Subject: [PATCH 50/60] update pillow --- modules/images.py | 2 -- requirements_versions.txt | 2 +- 2 files changed, 1 insertion(+), 3 deletions(-) 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: diff --git a/requirements_versions.txt b/requirements_versions.txt index 7ae118cb..d2899292 100644 --- a/requirements_versions.txt +++ b/requirements_versions.txt @@ -5,7 +5,7 @@ basicsr==1.4.2 gfpgan==1.3.8 gradio==3.15.0 numpy==1.23.3 -Pillow==9.3.0 +Pillow==9.4.0 realesrgan==0.3.0 torch omegaconf==2.2.3 From 525cea924562afd676f55470095268a0f6fca59e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 17:58:07 +0300 Subject: [PATCH 51/60] use shared function from processing for creating dummy mask when training inpainting model --- modules/processing.py | 39 ++++++++++--------- .../textual_inversion/textual_inversion.py | 33 +++++----------- 2 files changed, 29 insertions(+), 43 deletions(-) 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 From a8eb9e3bf814f72293e474c11e9ff0098859a942 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 18:20:38 +0300 Subject: [PATCH 52/60] 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(-) 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: From 3dae545a03f5102ba5d9c3f27bb6241824c5a916 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 18:42:51 +0300 Subject: [PATCH 53/60] rename weirdly named variables from #3176 --- modules/ui.py | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) 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) From 097a90b88bb92878cf435c513b4757b5b82ae299 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 19:19:11 +0300 Subject: [PATCH 54/60] add XY plot parameters to grid image and do not add them to individual images --- modules/processing.py | 2 +- scripts/xy_grid.py | 38 ++++++++++++++++++++++++-------------- 2 files changed, 25 insertions(+), 15 deletions(-) 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) diff --git a/scripts/xy_grid.py b/scripts/xy_grid.py index 59907f0b..78ff12c5 100644 --- a/scripts/xy_grid.py +++ b/scripts/xy_grid.py @@ -10,7 +10,7 @@ import numpy as np import modules.scripts as scripts import gradio as gr -from modules import images, paths, sd_samplers +from modules import images, paths, sd_samplers, processing from modules.hypernetworks import hypernetwork from modules.processing import process_images, Processed, StableDiffusionProcessingTxt2Img from modules.shared import opts, cmd_opts, state @@ -285,6 +285,7 @@ re_range_float = re.compile(r"\s*([+-]?\s*\d+(?:.\d*)?)\s*-\s*([+-]?\s*\d+(?:.\d re_range_count = re.compile(r"\s*([+-]?\s*\d+)\s*-\s*([+-]?\s*\d+)(?:\s*\[(\d+)\s*\])?\s*") re_range_count_float = re.compile(r"\s*([+-]?\s*\d+(?:.\d*)?)\s*-\s*([+-]?\s*\d+(?:.\d*)?)(?:\s*\[(\d+(?:.\d*)?)\s*\])?\s*") + class Script(scripts.Script): def title(self): return "X/Y plot" @@ -381,7 +382,7 @@ class Script(scripts.Script): ys = process_axis(y_opt, y_values) def fix_axis_seeds(axis_opt, axis_list): - if axis_opt.label in ['Seed','Var. seed']: + if axis_opt.label in ['Seed', 'Var. seed']: return [int(random.randrange(4294967294)) if val is None or val == '' or val == -1 else val for val in axis_list] else: return axis_list @@ -403,24 +404,33 @@ class Script(scripts.Script): print(f"X/Y plot will create {len(xs) * len(ys) * p.n_iter} images on a {len(xs)}x{len(ys)} grid. (Total steps to process: {total_steps * p.n_iter})") shared.total_tqdm.updateTotal(total_steps * p.n_iter) + grid_infotext = [None] + def cell(x, y): pc = copy(p) x_opt.apply(pc, x, xs) y_opt.apply(pc, y, ys) - return process_images(pc) + res = process_images(pc) - if not x_opt.label == 'Nothing': - p.extra_generation_params["XY Plot X Type"] = x_opt.label - p.extra_generation_params["XY Plot X Values"] = '{' + x_values + '}' - if x_opt.label in ["Seed","Var. seed"] and not no_fixed_seeds: - p.extra_generation_params["XY Plot Fixed X Values"] = '{' + ", ".join([str(x) for x in xs])+ '}' + if grid_infotext[0] is None: + pc.extra_generation_params = copy(pc.extra_generation_params) - if not y_opt.label == 'Nothing': - p.extra_generation_params["XY Plot Y Type"] = y_opt.label - p.extra_generation_params["XY Plot Y Values"] = '{' + y_values + '}' - if y_opt.label in ["Seed","Var. seed"] and not no_fixed_seeds: - p.extra_generation_params["XY Plot Fixed Y Values"] = '{' + ", ".join([str(y) for y in ys])+ '}' + if x_opt.label != 'Nothing': + pc.extra_generation_params["X Type"] = x_opt.label + pc.extra_generation_params["X Values"] = x_values + if x_opt.label in ["Seed", "Var. seed"] and not no_fixed_seeds: + pc.extra_generation_params["Fixed X Values"] = ", ".join([str(x) for x in xs]) + + if y_opt.label != 'Nothing': + pc.extra_generation_params["Y Type"] = y_opt.label + pc.extra_generation_params["Y Values"] = y_values + if y_opt.label in ["Seed", "Var. seed"] and not no_fixed_seeds: + pc.extra_generation_params["Fixed Y Values"] = ", ".join([str(y) for y in ys]) + + grid_infotext[0] = processing.create_infotext(pc, pc.all_prompts, pc.all_seeds, pc.all_subseeds) + + return res with SharedSettingsStackHelper(): processed = draw_xy_grid( @@ -435,6 +445,6 @@ class Script(scripts.Script): ) if opts.grid_save: - images.save_image(processed.images[0], p.outpath_grids, "xy_grid", extension=opts.grid_format, prompt=p.prompt, seed=processed.seed, grid=True, p=p) + images.save_image(processed.images[0], p.outpath_grids, "xy_grid", info=grid_infotext[0], extension=opts.grid_format, prompt=p.prompt, seed=processed.seed, grid=True, p=p) return processed From 24d4a0841d3cc0e5908b098f65a9caa3fa889af8 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 20:10:40 +0300 Subject: [PATCH 55/60] train tab visual updates allow setting train tab values from ui-config.json --- modules/ui.py | 35 +++++++++++++++++++++-------------- style.css | 2 +- 2 files changed, 22 insertions(+), 15 deletions(-) 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: diff --git a/style.css b/style.css index 2116ec3c..09ee540b 100644 --- a/style.css +++ b/style.css @@ -611,7 +611,7 @@ img2maskimg, #img2maskimg > .h-60, #img2maskimg > .h-60 > div, #img2maskimg > .h padding-top: 0.9em; } -#img2img_settings div.gr-form .gr-form, #txt2img_settings div.gr-form .gr-form{ +#img2img_settings div.gr-form .gr-form, #txt2img_settings div.gr-form .gr-form, #train_tabs div.gr-form .gr-form{ border: none; padding-bottom: 0.5em; } From 81490780949fffed77493b4bd741e96ec737fe27 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 22:04:40 +0300 Subject: [PATCH 56/60] added the option to specify target resolution with possibility of truncating for hires fix; also sampling steps --- javascript/hints.js | 11 +++-- modules/generation_parameters_copypaste.py | 9 ++-- modules/processing.py | 51 +++++++++++++++++++--- modules/txt2img.py | 5 ++- modules/ui.py | 24 +++++++--- 5 files changed, 81 insertions(+), 19 deletions(-) diff --git a/javascript/hints.js b/javascript/hints.js index 63e17e05..dda66e09 100644 --- a/javascript/hints.js +++ b/javascript/hints.js @@ -81,9 +81,6 @@ titles = { "vram": "Torch active: Peak amount of VRAM used by Torch during generation, excluding cached data.\nTorch reserved: Peak amount of VRAM allocated by Torch, including all active and cached data.\nSys VRAM: Peak amount of VRAM allocation across all applications / total GPU VRAM (peak utilization%).", - "Highres. fix": "Use a two step process to partially create an image at smaller resolution, upscale, and then improve details in it without changing composition", - "Scale latent": "Uscale the image in latent space. Alternative is to produce the full image from latent representation, upscale that, and then move it back to latent space.", - "Eta noise seed delta": "If this values is non-zero, it will be added to seed and used to initialize RNG for noises when using samplers with Eta. You can use this to produce even more variation of images, or you can use this to match images of other software if you know what you are doing.", "Do not add watermark to images": "If this option is enabled, watermark will not be added to created images. Warning: if you do not add watermark, you may be behaving in an unethical manner.", @@ -100,7 +97,13 @@ titles = { "Clip skip": "Early stopping parameter for CLIP model; 1 is stop at last layer as usual, 2 is stop at penultimate layer, etc.", "Approx NN": "Cheap neural network approximation. Very fast compared to VAE, but produces pictures with 4 times smaller horizontal/vertical resoluton and lower quality.", - "Approx cheap": "Very cheap approximation. Very fast compared to VAE, but produces pictures with 8 times smaller horizontal/vertical resoluton and extremely low quality." + "Approx cheap": "Very cheap approximation. Very fast compared to VAE, but produces pictures with 8 times smaller horizontal/vertical resoluton and extremely low quality.", + + "Hires. fix": "Use a two step process to partially create an image at smaller resolution, upscale, and then improve details in it without changing composition", + "Hires steps": "Number of sampling steps for upscaled picture. If 0, uses same as for original.", + "Upscale by": "Adjusts the size of the image by multiplying the original width and height by the selected value. Ignored if either Resize width to or Resize height to are non-zero.", + "Resize width to": "Resizes image to this width. If 0, width is inferred from either of two nearby sliders.", + "Resize height to": "Resizes image to this height. If 0, height is inferred from either of two nearby sliders." } 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) From bc43293c640aef65df3136de9e5bd8b7e79eb3e0 Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Wed, 4 Jan 2023 23:56:43 +0300 Subject: [PATCH 57/60] 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(-) 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: From b663ee2cff6831354e1b5326800c8d1bf300cafe Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 00:36:10 +0300 Subject: [PATCH 58/60] fix fullscreen view showing wrong image on firefox --- javascript/imageviewer.js | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/javascript/imageviewer.js b/javascript/imageviewer.js index 67916536..97f56c07 100644 --- a/javascript/imageviewer.js +++ b/javascript/imageviewer.js @@ -148,7 +148,7 @@ function showGalleryImage() { if(e && e.parentElement.tagName == 'DIV'){ e.style.cursor='pointer' e.style.userSelect='none' - e.addEventListener('click', function (evt) { + e.addEventListener('mousedown', function (evt) { if(!opts.js_modal_lightbox) return; modalZoomSet(gradioApp().getElementById('modalImage'), opts.js_modal_lightbox_initially_zoomed) showModal(evt) From 99b67cff0b48c4a1ad6e14d9cc591b11db6e293c Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 01:25:52 +0300 Subject: [PATCH 59/60] 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(-) 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 From 5f4fa942b8ec3ed3b15a352903489d6f9e6eb46e Mon Sep 17 00:00:00 2001 From: AUTOMATIC <16777216c@gmail.com> Date: Thu, 5 Jan 2023 02:38:52 +0300 Subject: [PATCH 60/60] do not show full window image preview when right mouse button is used --- javascript/imageviewer.js | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/javascript/imageviewer.js b/javascript/imageviewer.js index 97f56c07..b7bc2fe1 100644 --- a/javascript/imageviewer.js +++ b/javascript/imageviewer.js @@ -149,7 +149,7 @@ function showGalleryImage() { e.style.cursor='pointer' e.style.userSelect='none' e.addEventListener('mousedown', function (evt) { - if(!opts.js_modal_lightbox) return; + if(!opts.js_modal_lightbox || evt.button != 0) return; modalZoomSet(gradioApp().getElementById('modalImage'), opts.js_modal_lightbox_initially_zoomed) showModal(evt) }, true);