| accelerate launch --num_cpu_threads_per_process=2 "./sdxl_train_network.py" --enable_bucket --min_bucket_reso=256 --max_bucket_reso=2048 --pretrained_model_name_or_path="/workspace/model.safetensors" --train_data_dir="/workspace/kohya_ss/dataset/train_data" --resolution="1024,1024" --output_dir="/workspace/kohya_ss/dataset/result-4000" --logging_dir="/workspace/logs" --network_alpha="16" --save_model_as=safetensors --network_module=lycoris.kohya --text_encoder_lr=0.00005 --unet_lr=0.0003 --network_dim=32 --output_name="<REPLACE>" --no_half_vae --learning_rate="0.0003" --lr_scheduler="cosine_with_restarts" --train_batch_size="3" --max_train_steps="4000" --save_every_n_epochs="3" --mixed_precision="bf16" --save_precision="bf16" --seed="12345" --caption_extension=".txt" --cache_latents --cache_latents_to_disk --optimizer_type="AdamW" --max_data_loader_n_workers="0" --bucket_reso_steps=64 --xformers --max_grad_norm=0 --optimizer_args="decouple=True" --optimizer_args="weight_decay=0.01" --optimizer_args="betas=0.9,0.99" --lr_warmup_steps=0 --lr_scheduler_num_cycles="1" --shuffle_caption --keep_tokens=1 |