Clean up ti, add same behavior to hypernetwork.

This commit is contained in:
timntorres 2023-01-05 08:14:38 -08:00
parent eea8fc40e1
commit b85c2b5cf4
3 changed files with 40 additions and 7 deletions

View File

@ -401,7 +401,33 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None,
hypernet.save(fn)
shared.reload_hypernetworks()
# Note: textual_inversion.py has a nearly identical function of the same name.
def save_settings_to_file(initial_step, num_of_dataset_images, hypernetwork_name, layer_structure, activation_func, weight_init, add_layer_norm, use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height):
checkpoint = sd_models.select_checkpoint()
model_name = checkpoint.model_name
model_hash = '[{}]'.format(checkpoint.hash)
# Starting index of preview-related arguments.
border_index = 19
# Get a list of the argument names, excluding default argument.
sig = inspect.signature(save_settings_to_file)
arg_names = [p.name for p in sig.parameters.values() if p.default == p.empty]
# Create a list of the argument names to include in the settings string.
names = arg_names[:border_index] # Include all arguments up until the preview-related ones.
# Include preview-related arguments if applicable.
if preview_from_txt2img:
names.extend(arg_names[border_index:])
# Build the settings string.
settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n"
for name in names:
value = locals()[name]
settings_str += f"{name}: {value}\n"
with open(os.path.join(log_directory, 'settings.txt'), "a+") as fout:
fout.write(settings_str + "\n\n")
def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step, data_root, log_directory, training_width, training_height, steps, clip_grad_mode, clip_grad_value, shuffle_tags, tag_drop_out, latent_sampling_method, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height):
# images allows training previews to have infotext. Importing it at the top causes a circular import problem.
@ -458,6 +484,9 @@ def train_hypernetwork(hypernetwork_name, learn_rate, batch_size, gradient_step,
ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=hypernetwork_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, include_cond=True, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method)
if shared.opts.save_training_settings_to_txt:
save_settings_to_file(initial_step, len(ds), hypernetwork_name, hypernetwork.layer_structure, hypernetwork.activation_func, hypernetwork.weight_init, hypernetwork.add_layer_norm, hypernetwork.use_dropout, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_hypernetwork_every, template_file, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height)
latent_sampling_method = ds.latent_sampling_method
dl = modules.textual_inversion.dataset.PersonalizedDataLoader(ds, latent_sampling_method=latent_sampling_method, batch_size=ds.batch_size, pin_memory=pin_memory)

View File

@ -362,7 +362,7 @@ options_templates.update(options_section(('training', "Training"), {
"unload_models_when_training": OptionInfo(False, "Move VAE and CLIP to RAM when training if possible. Saves VRAM."),
"pin_memory": OptionInfo(False, "Turn on pin_memory for DataLoader. Makes training slightly faster but can increase memory usage."),
"save_optimizer_state": OptionInfo(False, "Saves Optimizer state as separate *.optim file. Training of embedding or HN can be resumed with the matching optim file."),
"save_train_settings_to_txt": OptionInfo(False, "Save textual inversion and hypernet settings to a text file when training starts."),
"save_training_settings_to_txt": OptionInfo(False, "Save textual inversion and hypernet settings to a text file whenever training starts."),
"dataset_filename_word_regex": OptionInfo("", "Filename word regex"),
"dataset_filename_join_string": OptionInfo(" ", "Filename join string"),
"training_image_repeats_per_epoch": OptionInfo(1, "Number of repeats for a single input image per epoch; used only for displaying epoch number", gr.Number, {"precision": 0}),

View File

@ -230,18 +230,20 @@ def write_loss(log_directory, filename, step, epoch_len, values):
**values,
})
# Note: hypernetwork.py has a nearly identical function of the same name.
def save_settings_to_file(initial_step, num_of_dataset_images, embedding_name, vectors_per_token, learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height):
checkpoint = sd_models.select_checkpoint()
model_name = checkpoint.model_name
model_hash = '[{}]'.format(checkpoint.hash)
# Starting index of preview-related arguments.
border_index = 16
# Get a list of the argument names.
arg_names = inspect.getfullargspec(save_settings_to_file).args
# Create a list of the argument names to include in the settings string.
names = arg_names[:16] # Include all arguments up until the preview-related ones.
names = arg_names[:border_index] # Include all arguments up until the preview-related ones.
if preview_from_txt2img:
names.extend(arg_names[16:]) # Include all remaining arguments if `preview_from_txt2img` is True.
names.extend(arg_names[border_index:]) # Include all remaining arguments if `preview_from_txt2img` is True.
# Build the settings string.
settings_str = "datetime : " + datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S") + "\n"
@ -329,8 +331,10 @@ def train_embedding(embedding_name, learn_rate, batch_size, gradient_step, data_
pin_memory = shared.opts.pin_memory
ds = modules.textual_inversion.dataset.PersonalizedBase(data_root=data_root, width=training_width, height=training_height, repeats=shared.opts.training_image_repeats_per_epoch, placeholder_token=embedding_name, model=shared.sd_model, cond_model=shared.sd_model.cond_stage_model, device=devices.device, template_file=template_file, batch_size=batch_size, gradient_step=gradient_step, shuffle_tags=shuffle_tags, tag_drop_out=tag_drop_out, latent_sampling_method=latent_sampling_method)
if shared.opts.save_train_settings_to_txt:
save_settings_to_file(initial_step , len(ds) , embedding_name, len(embedding.vec) , learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height)
if shared.opts.save_training_settings_to_txt:
save_settings_to_file(initial_step, len(ds), embedding_name, len(embedding.vec), learn_rate, batch_size, data_root, log_directory, training_width, training_height, steps, create_image_every, save_embedding_every, template_file, save_image_with_stored_embedding, preview_from_txt2img, preview_prompt, preview_negative_prompt, preview_steps, preview_sampler_index, preview_cfg_scale, preview_seed, preview_width, preview_height)
latent_sampling_method = ds.latent_sampling_method
dl = modules.textual_inversion.dataset.PersonalizedDataLoader(ds, latent_sampling_method=latent_sampling_method, batch_size=ds.batch_size, pin_memory=pin_memory)