mirror of https://github.com/bmaltais/kohya_ss
Code refactor
bmaltais
parent
315a1cd3d6
commit
b2ffe30a84
|
|
@ -39,6 +39,7 @@ from library.dreambooth_folder_creation_gui import (
|
|||
)
|
||||
from library.utilities import utilities_tab
|
||||
from library.sampler_gui import sample_gradio_config, run_cmd_sample
|
||||
|
||||
# from easygui import msgbox
|
||||
|
||||
folder_symbol = '\U0001f4c2' # 📂
|
||||
|
|
@ -340,39 +341,60 @@ def train_model(
|
|||
wandb_api_key,
|
||||
):
|
||||
headless_bool = True if headless.get('label') == 'True' else False
|
||||
|
||||
|
||||
if pretrained_model_name_or_path == '':
|
||||
output_message(msg='Source model information is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Source model information is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if train_data_dir == '':
|
||||
output_message(msg='Image folder path is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Image folder path is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if not os.path.exists(train_data_dir):
|
||||
output_message(msg='Image folder does not exist', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Image folder does not exist', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if reg_data_dir != '':
|
||||
if not os.path.exists(reg_data_dir):
|
||||
output_message(msg='Regularisation folder does not exist', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Regularisation folder does not exist',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
if output_dir == '':
|
||||
output_message(msg='Output folder path is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Output folder path is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if check_if_model_exist(output_name, output_dir, save_model_as, headless=headless_bool):
|
||||
if check_if_model_exist(
|
||||
output_name, output_dir, save_model_as, headless=headless_bool
|
||||
):
|
||||
return
|
||||
|
||||
if optimizer == 'Adafactor' and lr_warmup != '0':
|
||||
output_message(msg="Warning: lr_scheduler is set to 'Adafactor', so 'LR warmup (% of steps)' will be considered 0.",
|
||||
title='Warning', headless=headless_bool
|
||||
output_message(
|
||||
msg="Warning: lr_scheduler is set to 'Adafactor', so 'LR warmup (% of steps)' will be considered 0.",
|
||||
title='Warning',
|
||||
headless=headless_bool,
|
||||
)
|
||||
lr_warmup = '0'
|
||||
|
||||
if float(noise_offset) > 0 and (multires_noise_iterations > 0 or multires_noise_discount > 0):
|
||||
output_message(msg='noise offset and multires_noise can\'t be set at the same time. Only use one or the other.', title='Error', headless=headless_bool)
|
||||
|
||||
if float(noise_offset) > 0 and (
|
||||
multires_noise_iterations > 0 or multires_noise_discount > 0
|
||||
):
|
||||
output_message(
|
||||
msg="noise offset and multires_noise can't be set at the same time. Only use one or the other.",
|
||||
title='Error',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
# Get a list of all subfolders in train_data_dir, excluding hidden folders
|
||||
|
|
@ -602,7 +624,7 @@ def dreambooth_tab(
|
|||
reg_data_dir=gr.Textbox(),
|
||||
output_dir=gr.Textbox(),
|
||||
logging_dir=gr.Textbox(),
|
||||
headless=False
|
||||
headless=False,
|
||||
):
|
||||
dummy_db_true = gr.Label(value=True, visible=False)
|
||||
dummy_db_false = gr.Label(value=False, visible=False)
|
||||
|
|
@ -753,7 +775,9 @@ def dreambooth_tab(
|
|||
label='VAE',
|
||||
placeholder='(Optiona) path to checkpoint of vae to replace for training',
|
||||
)
|
||||
vae_button = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
vae_button = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
vae_button.click(
|
||||
get_any_file_path,
|
||||
outputs=vae,
|
||||
|
|
@ -950,7 +974,7 @@ def dreambooth_tab(
|
|||
|
||||
def UI(**kwargs):
|
||||
css = ''
|
||||
|
||||
|
||||
headless = kwargs.get('headless', False)
|
||||
print(f'headless: {headless}')
|
||||
|
||||
|
|
@ -978,7 +1002,7 @@ def UI(**kwargs):
|
|||
output_dir_input=output_dir_input,
|
||||
logging_dir_input=logging_dir_input,
|
||||
enable_copy_info_button=True,
|
||||
headless=headless
|
||||
headless=headless,
|
||||
)
|
||||
|
||||
# Show the interface
|
||||
|
|
|
|||
|
|
@ -30,6 +30,7 @@ from library.tensorboard_gui import (
|
|||
)
|
||||
from library.utilities import utilities_tab
|
||||
from library.sampler_gui import sample_gradio_config, run_cmd_sample
|
||||
|
||||
# from easygui import msgbox
|
||||
|
||||
folder_symbol = '\U0001f4c2' # 📂
|
||||
|
|
@ -351,18 +352,27 @@ def train_model(
|
|||
wandb_api_key,
|
||||
):
|
||||
headless_bool = True if headless.get('label') == 'True' else False
|
||||
|
||||
if check_if_model_exist(output_name, output_dir, save_model_as, headless_bool):
|
||||
|
||||
if check_if_model_exist(
|
||||
output_name, output_dir, save_model_as, headless_bool
|
||||
):
|
||||
return
|
||||
|
||||
if float(noise_offset) > 0 and (multires_noise_iterations > 0 or multires_noise_discount > 0):
|
||||
output_message(msg='noise offset and multires_noise can\'t be set at the same time. Only use one or the other.', title='Error', headless=headless_bool)
|
||||
|
||||
if float(noise_offset) > 0 and (
|
||||
multires_noise_iterations > 0 or multires_noise_discount > 0
|
||||
):
|
||||
output_message(
|
||||
msg="noise offset and multires_noise can't be set at the same time. Only use one or the other.",
|
||||
title='Error',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
if optimizer == 'Adafactor' and lr_warmup != '0':
|
||||
output_message(
|
||||
msg="Warning: lr_scheduler is set to 'Adafactor', so 'LR warmup (% of steps)' will be considered 0.",
|
||||
title='Warning', headless=headless_bool
|
||||
title='Warning',
|
||||
headless=headless_bool,
|
||||
)
|
||||
lr_warmup = '0'
|
||||
|
||||
|
|
@ -595,7 +605,9 @@ def finetune_tab(headless=False):
|
|||
placeholder='folder where the training configuration files will be saved',
|
||||
)
|
||||
train_dir_folder = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
train_dir_folder.click(
|
||||
get_folder_path,
|
||||
|
|
@ -608,7 +620,9 @@ def finetune_tab(headless=False):
|
|||
placeholder='folder where the training images are located',
|
||||
)
|
||||
image_folder_input_folder = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
image_folder_input_folder.click(
|
||||
get_folder_path,
|
||||
|
|
@ -621,7 +635,9 @@ def finetune_tab(headless=False):
|
|||
placeholder='folder where the model will be saved',
|
||||
)
|
||||
output_dir_input_folder = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
output_dir_input_folder.click(
|
||||
get_folder_path,
|
||||
|
|
@ -634,7 +650,9 @@ def finetune_tab(headless=False):
|
|||
placeholder='Optional: enable logging and output TensorBoard log to this folder',
|
||||
)
|
||||
logging_dir_input_folder = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
logging_dir_input_folder.click(
|
||||
get_folder_path,
|
||||
|
|
@ -904,7 +922,7 @@ def finetune_tab(headless=False):
|
|||
|
||||
def UI(**kwargs):
|
||||
css = ''
|
||||
|
||||
|
||||
headless = kwargs.get('headless', False)
|
||||
print(f'headless: {headless}')
|
||||
|
||||
|
|
|
|||
|
|
@ -82,7 +82,7 @@ def setup_logging(clean=False):
|
|||
|
||||
def UI(**kwargs):
|
||||
css = ''
|
||||
|
||||
|
||||
headless = kwargs.get('headless', False)
|
||||
print(f'headless: {headless}')
|
||||
|
||||
|
|
@ -116,7 +116,7 @@ def UI(**kwargs):
|
|||
output_dir_input=output_dir_input,
|
||||
logging_dir_input=logging_dir_input,
|
||||
enable_copy_info_button=True,
|
||||
headless=headless
|
||||
headless=headless,
|
||||
)
|
||||
gradio_extract_dylora_tab(headless=headless)
|
||||
gradio_extract_lora_tab(headless=headless)
|
||||
|
|
|
|||
|
|
@ -79,7 +79,9 @@ def gradio_basic_caption_gui_tab(headless=False):
|
|||
placeholder='Directory containing the images to caption',
|
||||
interactive=True,
|
||||
)
|
||||
folder_button = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
folder_button = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
folder_button.click(
|
||||
get_folder_path,
|
||||
outputs=images_dir,
|
||||
|
|
|
|||
|
|
@ -35,11 +35,15 @@ ALL_PRESET_MODELS = V2_BASE_MODELS + V_PARAMETERIZATION_MODELS + V1_MODELS
|
|||
ENV_EXCLUSION = ['COLAB_GPU', 'RUNPOD_POD_ID']
|
||||
|
||||
|
||||
def check_if_model_exist(output_name, output_dir, save_model_as, headless=False):
|
||||
def check_if_model_exist(
|
||||
output_name, output_dir, save_model_as, headless=False
|
||||
):
|
||||
if headless:
|
||||
print('Headless mode, skipping verification if model already exist... if model already exist it will be overwritten...')
|
||||
print(
|
||||
'Headless mode, skipping verification if model already exist... if model already exist it will be overwritten...'
|
||||
)
|
||||
return False
|
||||
|
||||
|
||||
if save_model_as in ['diffusers', 'diffusers_safetendors']:
|
||||
ckpt_folder = os.path.join(output_dir, output_name)
|
||||
if os.path.isdir(ckpt_folder):
|
||||
|
|
@ -66,12 +70,14 @@ def check_if_model_exist(output_name, output_dir, save_model_as, headless=False)
|
|||
|
||||
return False
|
||||
|
||||
|
||||
def output_message(msg='', title='', headless=False):
|
||||
if headless:
|
||||
print(msg)
|
||||
else:
|
||||
msgbox(msg=msg, title=title)
|
||||
|
||||
|
||||
def update_my_data(my_data):
|
||||
# Update the optimizer based on the use_8bit_adam flag
|
||||
use_8bit_adam = my_data.get('use_8bit_adam', False)
|
||||
|
|
@ -95,7 +101,7 @@ def update_my_data(my_data):
|
|||
my_data[key] = int(value)
|
||||
elif not value:
|
||||
my_data[key] = -1
|
||||
|
||||
|
||||
# Convert values to float if they are strings
|
||||
for key in ['noise_offset', 'learning_rate', 'text_encoder_lr', 'unet_lr']:
|
||||
value = my_data.get(key, -1)
|
||||
|
|
@ -347,11 +353,11 @@ def add_pre_postfix(
|
|||
caption_file_path = os.path.join(folder, caption_file_name)
|
||||
|
||||
if not os.path.exists(caption_file_path):
|
||||
with open(caption_file_path, 'w', encoding="utf8") as f:
|
||||
with open(caption_file_path, 'w', encoding='utf8') as f:
|
||||
separator = ' ' if prefix and postfix else ''
|
||||
f.write(f'{prefix}{separator}{postfix}')
|
||||
else:
|
||||
with open(caption_file_path, 'r+', encoding="utf8") as f:
|
||||
with open(caption_file_path, 'r+', encoding='utf8') as f:
|
||||
content = f.read()
|
||||
content = content.rstrip()
|
||||
f.seek(0, 0)
|
||||
|
|
@ -541,8 +547,12 @@ def set_model_list(
|
|||
def gradio_config(headless=False):
|
||||
with gr.Accordion('Configuration file', open=False):
|
||||
with gr.Row():
|
||||
button_open_config = gr.Button('Open 📂', elem_id='open_folder', visible=(not headless))
|
||||
button_save_config = gr.Button('Save 💾', elem_id='open_folder', visible=(not headless))
|
||||
button_open_config = gr.Button(
|
||||
'Open 📂', elem_id='open_folder', visible=(not headless)
|
||||
)
|
||||
button_save_config = gr.Button(
|
||||
'Save 💾', elem_id='open_folder', visible=(not headless)
|
||||
)
|
||||
button_save_as_config = gr.Button(
|
||||
'Save as... 💾', elem_id='open_folder', visible=(not headless)
|
||||
)
|
||||
|
|
@ -583,7 +593,7 @@ def gradio_source_model(
|
|||
'diffusers_safetensors',
|
||||
'safetensors',
|
||||
],
|
||||
headless=False
|
||||
headless=False,
|
||||
):
|
||||
with gr.Tab('Source model'):
|
||||
# Define the input elements
|
||||
|
|
@ -594,7 +604,9 @@ def gradio_source_model(
|
|||
value='runwayml/stable-diffusion-v1-5',
|
||||
)
|
||||
pretrained_model_name_or_path_file = gr.Button(
|
||||
document_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
document_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
pretrained_model_name_or_path_file.click(
|
||||
get_any_file_path,
|
||||
|
|
@ -603,7 +615,9 @@ def gradio_source_model(
|
|||
show_progress=False,
|
||||
)
|
||||
pretrained_model_name_or_path_folder = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
pretrained_model_name_or_path_folder.click(
|
||||
get_folder_path,
|
||||
|
|
@ -758,7 +772,11 @@ def gradio_training(
|
|||
value=lr_scheduler_value,
|
||||
)
|
||||
lr_warmup = gr.Slider(
|
||||
label='LR warmup (% of steps)', value=lr_warmup_value, minimum=0, maximum=100, step=1,
|
||||
label='LR warmup (% of steps)',
|
||||
value=lr_warmup_value,
|
||||
minimum=0,
|
||||
maximum=100,
|
||||
step=1,
|
||||
)
|
||||
optimizer = gr.Dropdown(
|
||||
label='Optimizer',
|
||||
|
|
@ -923,13 +941,28 @@ def gradio_advanced_training(headless=False):
|
|||
)
|
||||
with gr.Row():
|
||||
noise_offset = gr.Slider(
|
||||
label='Noise offset', value=0, minimum=0, maximum=1, step=0.01, info='recommended values are 0.05 - 0.15'
|
||||
label='Noise offset',
|
||||
value=0,
|
||||
minimum=0,
|
||||
maximum=1,
|
||||
step=0.01,
|
||||
info='recommended values are 0.05 - 0.15',
|
||||
)
|
||||
multires_noise_iterations = gr.Slider(
|
||||
label='Multires noise iterations', value=0, minimum=0, maximum=64, step=1, info='enable multires noise (recommended values are 6-10)'
|
||||
label='Multires noise iterations',
|
||||
value=0,
|
||||
minimum=0,
|
||||
maximum=64,
|
||||
step=1,
|
||||
info='enable multires noise (recommended values are 6-10)',
|
||||
)
|
||||
multires_noise_discount = gr.Slider(
|
||||
label='Multires noise discount', value=0, minimum=0, maximum=1, step=0.01, info='recommended values are 0.8. For LoRAs with small datasets, 0.1-0.3'
|
||||
label='Multires noise discount',
|
||||
value=0,
|
||||
minimum=0,
|
||||
maximum=1,
|
||||
step=0.01,
|
||||
info='recommended values are 0.8. For LoRAs with small datasets, 0.1-0.3',
|
||||
)
|
||||
with gr.Row():
|
||||
caption_dropout_every_n_epochs = gr.Number(
|
||||
|
|
@ -947,7 +980,9 @@ def gradio_advanced_training(headless=False):
|
|||
label='Resume from saved training state',
|
||||
placeholder='path to "last-state" state folder to resume from',
|
||||
)
|
||||
resume_button = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
resume_button = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
resume_button.click(
|
||||
get_folder_path,
|
||||
outputs=resume,
|
||||
|
|
@ -1015,69 +1050,50 @@ def run_cmd_advanced_training(**kwargs):
|
|||
f' --max_train_epochs="{kwargs.get("max_train_epochs", "")}"'
|
||||
if kwargs.get('max_train_epochs')
|
||||
else '',
|
||||
|
||||
f' --max_data_loader_n_workers="{kwargs.get("max_data_loader_n_workers", "")}"'
|
||||
if kwargs.get('max_data_loader_n_workers')
|
||||
else '',
|
||||
|
||||
f' --max_token_length={kwargs.get("max_token_length", "")}'
|
||||
if int(kwargs.get('max_token_length', 75)) > 75
|
||||
else '',
|
||||
|
||||
f' --clip_skip={kwargs.get("clip_skip", "")}'
|
||||
if int(kwargs.get('clip_skip', 1)) > 1
|
||||
else '',
|
||||
|
||||
f' --resume="{kwargs.get("resume", "")}"'
|
||||
if kwargs.get('resume')
|
||||
else '',
|
||||
|
||||
f' --keep_tokens="{kwargs.get("keep_tokens", "")}"'
|
||||
if int(kwargs.get('keep_tokens', 0)) > 0
|
||||
else '',
|
||||
|
||||
f' --caption_dropout_every_n_epochs="{int(kwargs.get("caption_dropout_every_n_epochs", 0))}"'
|
||||
if int(kwargs.get('caption_dropout_every_n_epochs', 0)) > 0
|
||||
else '',
|
||||
|
||||
f' --caption_dropout_rate="{float(kwargs.get("caption_dropout_rate", 0))}"'
|
||||
if float(kwargs.get('caption_dropout_rate', 0)) > 0
|
||||
else '',
|
||||
|
||||
f' --vae_batch_size="{kwargs.get("vae_batch_size", 0)}"'
|
||||
if int(kwargs.get('vae_batch_size', 0)) > 0
|
||||
else '',
|
||||
|
||||
f' --bucket_reso_steps={int(kwargs.get("bucket_reso_steps", 1))}'
|
||||
if int(kwargs.get('bucket_reso_steps', 64)) >= 1
|
||||
else '',
|
||||
|
||||
f' --save_every_n_steps="{int(kwargs.get("save_every_n_steps", 0))}"'
|
||||
if int(kwargs.get('save_every_n_steps')) > 0
|
||||
else '',
|
||||
|
||||
f' --save_last_n_steps="{int(kwargs.get("save_last_n_steps", 0))}"'
|
||||
if int(kwargs.get('save_last_n_steps')) > 0
|
||||
else '',
|
||||
|
||||
f' --save_last_n_steps_state="{int(kwargs.get("save_last_n_steps_state", 0))}"'
|
||||
if int(kwargs.get('save_last_n_steps_state')) > 0
|
||||
else '',
|
||||
|
||||
f' --min_snr_gamma={int(kwargs.get("min_snr_gamma", 0))}'
|
||||
if int(kwargs.get('min_snr_gamma', 0)) >= 1
|
||||
else '',
|
||||
|
||||
' --save_state' if kwargs.get('save_state') else '',
|
||||
|
||||
' --mem_eff_attn' if kwargs.get('mem_eff_attn') else '',
|
||||
|
||||
' --color_aug' if kwargs.get('color_aug') else '',
|
||||
|
||||
' --flip_aug' if kwargs.get('flip_aug') else '',
|
||||
|
||||
' --shuffle_caption' if kwargs.get('shuffle_caption') else '',
|
||||
|
||||
' --gradient_checkpointing'
|
||||
if kwargs.get('gradient_checkpointing')
|
||||
else '',
|
||||
|
|
@ -1089,28 +1105,21 @@ def run_cmd_advanced_training(**kwargs):
|
|||
else '',
|
||||
' --bucket_no_upscale' if kwargs.get('bucket_no_upscale') else '',
|
||||
' --random_crop' if kwargs.get('random_crop') else '',
|
||||
|
||||
f' --multires_noise_iterations="{int(kwargs.get("multires_noise_iterations", 0))}"'
|
||||
if kwargs.get('multires_noise_iterations', 0) > 0
|
||||
else '',
|
||||
|
||||
f' --multires_noise_discount="{float(kwargs.get("multires_noise_discount", 0.0))}"'
|
||||
if kwargs.get('multires_noise_discount', 0) > 0
|
||||
else '',
|
||||
|
||||
f' --noise_offset={float(kwargs.get("noise_offset", 0))}'
|
||||
if kwargs.get('noise_offset') > 0
|
||||
else '',
|
||||
|
||||
f' {kwargs.get("additional_parameters", "")}',
|
||||
|
||||
|
||||
' --log_with wandb' if kwargs.get('use_wandb') else '',
|
||||
|
||||
f' --wandb_api_key="{kwargs.get("wandb_api_key", "")}"'
|
||||
if kwargs.get('wandb_api_key')
|
||||
else '',
|
||||
]
|
||||
|
||||
|
||||
run_cmd = ''.join(options)
|
||||
return run_cmd
|
||||
|
|
|
|||
|
|
@ -68,12 +68,12 @@ def convert_model(
|
|||
|
||||
if target_model_type == 'diffuser_safetensors':
|
||||
run_cmd += ' --use_safetensors'
|
||||
|
||||
|
||||
# Fix for stabilityAI diffusers format. When saving v2 models in Diffusers format in training scripts and conversion scripts,
|
||||
# it was found that the U-Net configuration is different from those of Hugging Face's stabilityai models (this repository is
|
||||
# it was found that the U-Net configuration is different from those of Hugging Face's stabilityai models (this repository is
|
||||
# "use_linear_projection": false, stabilityai is true). Please note that the weight shapes are different, so please be careful
|
||||
# when using the weight files directly.
|
||||
|
||||
|
||||
if unet_use_linear_projection:
|
||||
run_cmd += ' --unet_use_linear_projection'
|
||||
|
||||
|
|
@ -176,7 +176,9 @@ def gradio_convert_model_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_source_model_dir = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_source_model_dir.click(
|
||||
get_folder_path,
|
||||
|
|
@ -185,7 +187,9 @@ def gradio_convert_model_tab(headless=False):
|
|||
)
|
||||
|
||||
button_source_model_file = gr.Button(
|
||||
document_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
document_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_source_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -212,7 +216,9 @@ def gradio_convert_model_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_target_model_folder = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_target_model_folder.click(
|
||||
get_folder_path,
|
||||
|
|
@ -239,7 +245,11 @@ def gradio_convert_model_tab(headless=False):
|
|||
choices=['unspecified', 'fp16', 'bf16', 'float'],
|
||||
value='unspecified',
|
||||
)
|
||||
unet_use_linear_projection = gr.Checkbox(label="UNet linear projection", value=False, info="Enable for Hugging Face's stabilityai models")
|
||||
unet_use_linear_projection = gr.Checkbox(
|
||||
label='UNet linear projection',
|
||||
value=False,
|
||||
info="Enable for Hugging Face's stabilityai models",
|
||||
)
|
||||
|
||||
convert_button = gr.Button('Convert model')
|
||||
|
||||
|
|
|
|||
|
|
@ -114,7 +114,7 @@ def gradio_dreambooth_folder_creation_tab(
|
|||
reg_data_dir_input=gr.Textbox(),
|
||||
output_dir_input=gr.Textbox(),
|
||||
logging_dir_input=gr.Textbox(),
|
||||
headless=False
|
||||
headless=False,
|
||||
):
|
||||
with gr.Tab('Dreambooth/LoRA Folder preparation'):
|
||||
gr.Markdown(
|
||||
|
|
|
|||
|
|
@ -67,7 +67,9 @@ def gradio_extract_dylora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -82,7 +84,9 @@ def gradio_extract_dylora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_save_to = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_save_to.click(
|
||||
get_saveasfilename_path,
|
||||
|
|
|
|||
|
|
@ -88,7 +88,9 @@ def gradio_extract_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_model_tuned_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_model_tuned_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -103,7 +105,9 @@ def gradio_extract_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_model_org_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_model_org_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -118,7 +122,9 @@ def gradio_extract_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_save_to = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_save_to.click(
|
||||
get_saveasfilename_path,
|
||||
|
|
|
|||
|
|
@ -138,7 +138,9 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_db_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_db_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -153,7 +155,9 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_base_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_base_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -168,7 +172,9 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_output_name = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_output_name.click(
|
||||
get_saveasfilename_path,
|
||||
|
|
@ -217,7 +223,7 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
value=0.65,
|
||||
step=0.01,
|
||||
interactive=True,
|
||||
info='The higher the value, the smaller the file. Recommended starting value: 0.65'
|
||||
info='The higher the value, the smaller the file. Recommended starting value: 0.65',
|
||||
)
|
||||
conv_threshold = gr.Slider(
|
||||
minimum=0,
|
||||
|
|
@ -226,7 +232,7 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
value=0.65,
|
||||
step=0.01,
|
||||
interactive=True,
|
||||
info='The higher the value, the smaller the file. Recommended starting value: 0.65'
|
||||
info='The higher the value, the smaller the file. Recommended starting value: 0.65',
|
||||
)
|
||||
with gr.Row(visible=False) as ratio:
|
||||
linear_ratio = gr.Slider(
|
||||
|
|
@ -236,7 +242,7 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
value=0.75,
|
||||
step=0.01,
|
||||
interactive=True,
|
||||
info='The higher the value, the smaller the file. Recommended starting value: 0.75'
|
||||
info='The higher the value, the smaller the file. Recommended starting value: 0.75',
|
||||
)
|
||||
conv_ratio = gr.Slider(
|
||||
minimum=0,
|
||||
|
|
@ -245,7 +251,7 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
value=0.75,
|
||||
step=0.01,
|
||||
interactive=True,
|
||||
info='The higher the value, the smaller the file. Recommended starting value: 0.75'
|
||||
info='The higher the value, the smaller the file. Recommended starting value: 0.75',
|
||||
)
|
||||
with gr.Row(visible=False) as quantile:
|
||||
linear_quantile = gr.Slider(
|
||||
|
|
@ -255,7 +261,7 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
value=0.75,
|
||||
step=0.01,
|
||||
interactive=True,
|
||||
info='The higher the value, the larger the file. Recommended starting value: 0.75'
|
||||
info='The higher the value, the larger the file. Recommended starting value: 0.75',
|
||||
)
|
||||
conv_quantile = gr.Slider(
|
||||
minimum=0,
|
||||
|
|
@ -264,7 +270,7 @@ def gradio_extract_lycoris_locon_tab(headless=False):
|
|||
value=0.75,
|
||||
step=0.01,
|
||||
interactive=True,
|
||||
info='The higher the value, the larger the file. Recommended starting value: 0.75'
|
||||
info='The higher the value, the larger the file. Recommended starting value: 0.75',
|
||||
)
|
||||
with gr.Row():
|
||||
use_sparse_bias = gr.Checkbox(
|
||||
|
|
|
|||
|
|
@ -115,7 +115,9 @@ def gradio_merge_lora_tab(headless=False):
|
|||
info='Provide a SD file path IF you want to merge it with LoRA files',
|
||||
)
|
||||
sd_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
sd_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -131,7 +133,9 @@ def gradio_merge_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lora_a_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lora_a_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -146,7 +150,9 @@ def gradio_merge_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lora_b_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lora_b_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -181,7 +187,9 @@ def gradio_merge_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lora_c_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lora_c_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -196,7 +204,9 @@ def gradio_merge_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lora_d_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lora_d_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -231,7 +241,9 @@ def gradio_merge_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_save_to = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_save_to.click(
|
||||
get_saveasfilename_path,
|
||||
|
|
|
|||
|
|
@ -70,7 +70,9 @@ def gradio_merge_lycoris_tab(headless=False):
|
|||
info='Provide a SD file path that you want to merge with the LyCORIS file',
|
||||
)
|
||||
base_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
base_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -86,7 +88,9 @@ def gradio_merge_lycoris_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lycoris_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lycoris_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -112,7 +116,9 @@ def gradio_merge_lycoris_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_output_name = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_output_name.click(
|
||||
get_saveasfilename_path,
|
||||
|
|
|
|||
|
|
@ -94,7 +94,9 @@ def gradio_resize_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lora_a_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lora_a_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -133,7 +135,9 @@ def gradio_resize_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_save_to = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_save_to.click(
|
||||
get_saveasfilename_path,
|
||||
|
|
|
|||
|
|
@ -85,7 +85,9 @@ def gradio_svd_merge_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lora_a_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lora_a_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -100,7 +102,9 @@ def gradio_svd_merge_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lora_b_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lora_b_model_file.click(
|
||||
get_file_path,
|
||||
|
|
@ -141,7 +145,9 @@ def gradio_svd_merge_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_save_to = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_save_to.click(
|
||||
get_saveasfilename_path,
|
||||
|
|
|
|||
|
|
@ -66,7 +66,9 @@ def gradio_verify_lora_tab(headless=False):
|
|||
interactive=True,
|
||||
)
|
||||
button_lora_model_file = gr.Button(
|
||||
folder_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
folder_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
button_lora_model_file.click(
|
||||
get_file_path,
|
||||
|
|
|
|||
111
lora_gui.py
111
lora_gui.py
|
|
@ -4,6 +4,7 @@
|
|||
# v3.1: Adding captionning of images to utilities
|
||||
|
||||
import gradio as gr
|
||||
|
||||
# import easygui
|
||||
import json
|
||||
import math
|
||||
|
|
@ -45,6 +46,7 @@ from library.svd_merge_lora_gui import gradio_svd_merge_lora_tab
|
|||
from library.verify_lora_gui import gradio_verify_lora_tab
|
||||
from library.resize_lora_gui import gradio_resize_lora_tab
|
||||
from library.sampler_gui import sample_gradio_config, run_cmd_sample
|
||||
|
||||
# from easygui import msgbox
|
||||
|
||||
folder_symbol = '\U0001f4c2' # 📂
|
||||
|
|
@ -75,7 +77,8 @@ def save_configuration(
|
|||
save_precision,
|
||||
seed,
|
||||
num_cpu_threads_per_process,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
caption_extension,
|
||||
enable_bucket,
|
||||
gradient_checkpointing,
|
||||
|
|
@ -211,7 +214,8 @@ def open_configuration(
|
|||
save_precision,
|
||||
seed,
|
||||
num_cpu_threads_per_process,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
caption_extension,
|
||||
enable_bucket,
|
||||
gradient_checkpointing,
|
||||
|
|
@ -318,7 +322,6 @@ def open_configuration(
|
|||
return tuple(values)
|
||||
|
||||
|
||||
|
||||
def train_model(
|
||||
headless,
|
||||
print_only,
|
||||
|
|
@ -340,7 +343,8 @@ def train_model(
|
|||
save_precision,
|
||||
seed,
|
||||
num_cpu_threads_per_process,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
caption_extension,
|
||||
enable_bucket,
|
||||
gradient_checkpointing,
|
||||
|
|
@ -414,39 +418,62 @@ def train_model(
|
|||
headless_bool = True if headless.get('label') == 'True' else False
|
||||
|
||||
if pretrained_model_name_or_path == '':
|
||||
output_message(msg='Source model information is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Source model information is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if train_data_dir == '':
|
||||
output_message(msg='Image folder path is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Image folder path is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if not os.path.exists(train_data_dir):
|
||||
output_message(msg='Image folder does not exist', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Image folder does not exist', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if reg_data_dir != '':
|
||||
if not os.path.exists(reg_data_dir):
|
||||
output_message(msg='Regularisation folder does not exist', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Regularisation folder does not exist',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
if output_dir == '':
|
||||
output_message(msg='Output folder path is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Output folder path is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if int(bucket_reso_steps) < 1:
|
||||
output_message(msg='Bucket resolution steps need to be greater than 0', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Bucket resolution steps need to be greater than 0',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
|
||||
if noise_offset == '':
|
||||
noise_offset = 0
|
||||
|
||||
|
||||
if float(noise_offset) > 1 or float(noise_offset) < 0:
|
||||
output_message(msg='Noise offset need to be a value between 0 and 1', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Noise offset need to be a value between 0 and 1',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
if float(noise_offset) > 0 and (multires_noise_iterations > 0 or multires_noise_discount > 0):
|
||||
output_message(msg='noise offset and multires_noise can\'t be set at the same time. Only use one or the other.', title='Error', headless=headless_bool)
|
||||
|
||||
if float(noise_offset) > 0 and (
|
||||
multires_noise_iterations > 0 or multires_noise_discount > 0
|
||||
):
|
||||
output_message(
|
||||
msg="noise offset and multires_noise can't be set at the same time. Only use one or the other.",
|
||||
title='Error',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
if not os.path.exists(output_dir):
|
||||
|
|
@ -454,17 +481,21 @@ def train_model(
|
|||
|
||||
if stop_text_encoder_training_pct > 0:
|
||||
output_message(
|
||||
msg='Output "stop text encoder training" is not yet supported. Ignoring', headless=headless_bool
|
||||
msg='Output "stop text encoder training" is not yet supported. Ignoring',
|
||||
headless=headless_bool,
|
||||
)
|
||||
stop_text_encoder_training_pct = 0
|
||||
|
||||
if check_if_model_exist(output_name, output_dir, save_model_as, headless=headless_bool):
|
||||
if check_if_model_exist(
|
||||
output_name, output_dir, save_model_as, headless=headless_bool
|
||||
):
|
||||
return
|
||||
|
||||
if optimizer == 'Adafactor' and lr_warmup != '0':
|
||||
output_message(
|
||||
msg="Warning: lr_scheduler is set to 'Adafactor', so 'LR warmup (% of steps)' will be considered 0.",
|
||||
title='Warning', headless=headless_bool
|
||||
title='Warning',
|
||||
headless=headless_bool,
|
||||
)
|
||||
lr_warmup = '0'
|
||||
|
||||
|
|
@ -677,7 +708,10 @@ def train_model(
|
|||
run_cmd += f' --network_train_unet_only'
|
||||
else:
|
||||
if float(learning_rate) == 0:
|
||||
output_message(msg='Please input learning rate values.', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Please input learning rate values.',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
run_cmd += f' --network_dim={network_dim}'
|
||||
|
|
@ -800,12 +834,12 @@ def lora_tab(
|
|||
reg_data_dir_input=gr.Textbox(),
|
||||
output_dir_input=gr.Textbox(),
|
||||
logging_dir_input=gr.Textbox(),
|
||||
headless=False
|
||||
headless=False,
|
||||
):
|
||||
dummy_db_true = gr.Label(value=True, visible=False)
|
||||
dummy_db_false = gr.Label(value=False, visible=False)
|
||||
dummy_headless = gr.Label(value=headless, visible=False)
|
||||
|
||||
|
||||
gr.Markdown(
|
||||
'Train a custom model using kohya train network LoRA python code...'
|
||||
)
|
||||
|
|
@ -827,7 +861,8 @@ def lora_tab(
|
|||
save_model_as_choices=[
|
||||
'ckpt',
|
||||
'safetensors',
|
||||
], headless=headless
|
||||
],
|
||||
headless=headless,
|
||||
)
|
||||
|
||||
with gr.Tab('Folders'):
|
||||
|
|
@ -836,7 +871,9 @@ def lora_tab(
|
|||
label='Image folder',
|
||||
placeholder='Folder where the training folders containing the images are located',
|
||||
)
|
||||
train_data_dir_folder = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
train_data_dir_folder = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
train_data_dir_folder.click(
|
||||
get_folder_path,
|
||||
outputs=train_data_dir,
|
||||
|
|
@ -846,7 +883,9 @@ def lora_tab(
|
|||
label='Regularisation folder',
|
||||
placeholder='(Optional) Folder where where the regularization folders containing the images are located',
|
||||
)
|
||||
reg_data_dir_folder = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
reg_data_dir_folder = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
reg_data_dir_folder.click(
|
||||
get_folder_path,
|
||||
outputs=reg_data_dir,
|
||||
|
|
@ -857,7 +896,9 @@ def lora_tab(
|
|||
label='Output folder',
|
||||
placeholder='Folder to output trained model',
|
||||
)
|
||||
output_dir_folder = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
output_dir_folder = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
output_dir_folder.click(
|
||||
get_folder_path,
|
||||
outputs=output_dir,
|
||||
|
|
@ -867,7 +908,9 @@ def lora_tab(
|
|||
label='Logging folder',
|
||||
placeholder='Optional: enable logging and output TensorBoard log to this folder',
|
||||
)
|
||||
logging_dir_folder = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
logging_dir_folder = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
logging_dir_folder.click(
|
||||
get_folder_path,
|
||||
outputs=logging_dir,
|
||||
|
|
@ -924,7 +967,9 @@ def lora_tab(
|
|||
placeholder='{Optional) Path to existing LoRA network weights to resume training',
|
||||
)
|
||||
lora_network_weights_file = gr.Button(
|
||||
document_symbol, elem_id='open_folder_small', visible=(not headless)
|
||||
document_symbol,
|
||||
elem_id='open_folder_small',
|
||||
visible=(not headless),
|
||||
)
|
||||
lora_network_weights_file.click(
|
||||
get_any_file_path,
|
||||
|
|
@ -944,7 +989,8 @@ def lora_tab(
|
|||
num_cpu_threads_per_process,
|
||||
seed,
|
||||
caption_extension,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
optimizer,
|
||||
optimizer_args,
|
||||
) = gradio_training(
|
||||
|
|
@ -1240,7 +1286,8 @@ def lora_tab(
|
|||
save_precision,
|
||||
seed,
|
||||
num_cpu_threads_per_process,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
caption_extension,
|
||||
enable_bucket,
|
||||
gradient_checkpointing,
|
||||
|
|
@ -1361,7 +1408,7 @@ def lora_tab(
|
|||
|
||||
def UI(**kwargs):
|
||||
css = ''
|
||||
|
||||
|
||||
headless = kwargs.get('headless', False)
|
||||
print(f'headless: {headless}')
|
||||
|
||||
|
|
@ -1389,7 +1436,7 @@ def UI(**kwargs):
|
|||
output_dir_input=output_dir_input,
|
||||
logging_dir_input=logging_dir_input,
|
||||
enable_copy_info_button=True,
|
||||
headless=headless
|
||||
headless=headless,
|
||||
)
|
||||
|
||||
# Show the interface
|
||||
|
|
|
|||
|
|
@ -39,6 +39,7 @@ from library.dreambooth_folder_creation_gui import (
|
|||
)
|
||||
from library.utilities import utilities_tab
|
||||
from library.sampler_gui import sample_gradio_config, run_cmd_sample
|
||||
|
||||
# from easygui import msgbox
|
||||
|
||||
folder_symbol = '\U0001f4c2' # 📂
|
||||
|
|
@ -68,7 +69,8 @@ def save_configuration(
|
|||
save_precision,
|
||||
seed,
|
||||
num_cpu_threads_per_process,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
caption_extension,
|
||||
enable_bucket,
|
||||
gradient_checkpointing,
|
||||
|
|
@ -190,7 +192,8 @@ def open_configuration(
|
|||
save_precision,
|
||||
seed,
|
||||
num_cpu_threads_per_process,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
caption_extension,
|
||||
enable_bucket,
|
||||
gradient_checkpointing,
|
||||
|
|
@ -295,7 +298,8 @@ def train_model(
|
|||
save_precision,
|
||||
seed,
|
||||
num_cpu_threads_per_process,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
caption_extension,
|
||||
enable_bucket,
|
||||
gradient_checkpointing,
|
||||
|
|
@ -352,26 +356,37 @@ def train_model(
|
|||
wandb_api_key,
|
||||
):
|
||||
headless_bool = True if headless.get('label') == 'True' else False
|
||||
|
||||
|
||||
if pretrained_model_name_or_path == '':
|
||||
output_message(msg='Source model information is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Source model information is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if train_data_dir == '':
|
||||
output_message(msg='Image folder path is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Image folder path is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if not os.path.exists(train_data_dir):
|
||||
output_message(msg='Image folder does not exist', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Image folder does not exist', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if reg_data_dir != '':
|
||||
if not os.path.exists(reg_data_dir):
|
||||
output_message(msg='Regularisation folder does not exist', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Regularisation folder does not exist',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
if output_dir == '':
|
||||
output_message(msg='Output folder path is missing', headless=headless_bool)
|
||||
output_message(
|
||||
msg='Output folder path is missing', headless=headless_bool
|
||||
)
|
||||
return
|
||||
|
||||
if token_string == '':
|
||||
|
|
@ -385,17 +400,26 @@ def train_model(
|
|||
if not os.path.exists(output_dir):
|
||||
os.makedirs(output_dir)
|
||||
|
||||
if check_if_model_exist(output_name, output_dir, save_model_as, headless_bool):
|
||||
if check_if_model_exist(
|
||||
output_name, output_dir, save_model_as, headless_bool
|
||||
):
|
||||
return
|
||||
|
||||
if float(noise_offset) > 0 and (multires_noise_iterations > 0 or multires_noise_discount > 0):
|
||||
output_message(msg='noise offset and multires_noise can\'t be set at the same time. Only use one or the other.', title='Error', headless=headless_bool)
|
||||
|
||||
if float(noise_offset) > 0 and (
|
||||
multires_noise_iterations > 0 or multires_noise_discount > 0
|
||||
):
|
||||
output_message(
|
||||
msg="noise offset and multires_noise can't be set at the same time. Only use one or the other.",
|
||||
title='Error',
|
||||
headless=headless_bool,
|
||||
)
|
||||
return
|
||||
|
||||
if optimizer == 'Adafactor' and lr_warmup != '0':
|
||||
output_message(
|
||||
msg="Warning: lr_scheduler is set to 'Adafactor', so 'LR warmup (% of steps)' will be considered 0.",
|
||||
title='Warning', headless=headless_bool
|
||||
title='Warning',
|
||||
headless=headless_bool,
|
||||
)
|
||||
lr_warmup = '0'
|
||||
|
||||
|
|
@ -606,7 +630,7 @@ def ti_tab(
|
|||
reg_data_dir=gr.Textbox(),
|
||||
output_dir=gr.Textbox(),
|
||||
logging_dir=gr.Textbox(),
|
||||
headless=False
|
||||
headless=False,
|
||||
):
|
||||
dummy_db_true = gr.Label(value=True, visible=False)
|
||||
dummy_db_false = gr.Label(value=False, visible=False)
|
||||
|
|
@ -630,7 +654,8 @@ def ti_tab(
|
|||
save_model_as_choices=[
|
||||
'ckpt',
|
||||
'safetensors',
|
||||
], headless=headless
|
||||
],
|
||||
headless=headless,
|
||||
)
|
||||
|
||||
with gr.Tab('Folders'):
|
||||
|
|
@ -717,7 +742,9 @@ def ti_tab(
|
|||
label='Resume TI training',
|
||||
placeholder='(Optional) Path to existing TI embeding file to keep training',
|
||||
)
|
||||
weights_file_input = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
weights_file_input = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
weights_file_input.click(
|
||||
get_file_path,
|
||||
outputs=weights,
|
||||
|
|
@ -764,7 +791,8 @@ def ti_tab(
|
|||
num_cpu_threads_per_process,
|
||||
seed,
|
||||
caption_extension,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
optimizer,
|
||||
optimizer_args,
|
||||
) = gradio_training(
|
||||
|
|
@ -802,7 +830,9 @@ def ti_tab(
|
|||
label='VAE',
|
||||
placeholder='(Optiona) path to checkpoint of vae to replace for training',
|
||||
)
|
||||
vae_button = gr.Button('📂', elem_id='open_folder_small', visible=(not headless))
|
||||
vae_button = gr.Button(
|
||||
'📂', elem_id='open_folder_small', visible=(not headless)
|
||||
)
|
||||
vae_button.click(
|
||||
get_any_file_path,
|
||||
outputs=vae,
|
||||
|
|
@ -902,7 +932,8 @@ def ti_tab(
|
|||
save_precision,
|
||||
seed,
|
||||
num_cpu_threads_per_process,
|
||||
cache_latents,cache_latents_to_disk,
|
||||
cache_latents,
|
||||
cache_latents_to_disk,
|
||||
caption_extension,
|
||||
enable_bucket,
|
||||
gradient_checkpointing,
|
||||
|
|
@ -1003,7 +1034,7 @@ def ti_tab(
|
|||
|
||||
def UI(**kwargs):
|
||||
css = ''
|
||||
|
||||
|
||||
headless = kwargs.get('headless', False)
|
||||
print(f'headless: {headless}')
|
||||
|
||||
|
|
@ -1031,7 +1062,7 @@ def UI(**kwargs):
|
|||
output_dir_input=output_dir_input,
|
||||
logging_dir_input=logging_dir_input,
|
||||
enable_copy_info_button=True,
|
||||
headless=headless
|
||||
headless=headless,
|
||||
)
|
||||
|
||||
# Show the interface
|
||||
|
|
|
|||
Loading…
Reference in New Issue