| import subprocess |
|
|
| pretrained_model = "/path/to/directorychilloutmix_NiPrunedFp32Fix.safetensors" |
| train_data_dir = "/path/to/directory" |
| reg_data_dir = "/path/to/directory" |
|
|
| resolution = "768,768" |
| batch_size = 2 |
| max_train_epochs = 10 |
| save_every_n_epochs = 1 |
| network_dim = 64 |
| network_alpha = 32 |
| clip_skip = 2 |
| train_unet_only = 0 |
| train_text_encoder_only = 0 |
|
|
| lr = "5e-5" |
| unet_lr = "5e-5" |
| text_encoder_lr = "6e-6" |
| lr_scheduler = "cosine_with_restarts" |
| lr_warmup_steps = 50 |
| lr_restart_cycles = 1 |
|
|
| output_name = "yujinive_v2" |
| save_model_as = "safetensors" |
|
|
| network_weights = "" |
| min_bucket_reso = 256 |
| max_bucket_reso = 1024 |
| persistent_data_loader_workers = 0 |
|
|
| subprocess.run(["source", "venv/bin/activate"], shell=True) |
|
|
| subprocess.run(["export", "HF_HOME=huggingface"], shell=True) |
|
|
| subprocess.run([ |
| "python", "-m", "accelerate", "launch", "--num_processes=1", "--num_workers=8", "--use_env", |
| "./sd-scripts/train_network.py", |
| "--enable_bucket", |
| f"--pretrained_model_name_or_path={pretrained_model}", |
| f"--train_data_dir={train_data_dir}", |
| "--output_dir=./output", |
| "--logging_dir=./logs", |
| f"--resolution={resolution}", |
| "--network_module=networks.lora", |
| f"--max_train_epochs={max_train_epochs}", |
| f"--learning_rate={lr}", |
| f"--unet_lr={unet_lr}", |
| f"--text_encoder_lr={text_encoder_lr}", |
| f"--lr_scheduler={lr_scheduler}", |
| f"--lr_warmup_steps={lr_warmup_steps}", |
| f"--lr_scheduler_num_cycles={lr_restart_cycles}", |
| f"--network_dim={network_dim}", |
| f"--network_alpha={network_alpha}", |
| f"--output_name={output_name}", |
| f"--train_batch_size={batch_size}", |
| f"--save_every_n_epochs={save_every_n_epochs}", |
| "--mixed_precision=fp16", |
| "--save_precision=fp16", |
| "--seed=1337", |
| "--cache_latents", |
| f"--clip_skip={clip_skip}", |
| "--prior_loss_weight=1", |
| "--max_token_length=225", |
| "--caption_extension=.txt", |
| f"--save_model_as={save_model_as}", |
| f"--min_bucket_reso={min_bucket_reso}", |
| f"--max_bucket_reso={max_bucket_reso}", |
| "--xformers", |
| "--shuffle_caption" |
| ]) |
|
|
| print("Train finished") |
| input("Press any key to continue...") |
|
|
