mirror of
https://github.com/modelscope/DiffSynth-Studio.git
synced 2026-03-24 10:18:12 +00:00
37 lines
2.1 KiB
Bash
37 lines
2.1 KiB
Bash
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "wanvideo/Wan2.1-I2V-14B-480P/*" --local_dir ./data/diffsynth_example_dataset
|
|
|
|
accelerate launch examples/wanvideo/model_training/train.py \
|
|
--dataset_base_path data/diffsynth_example_dataset/wanvideo/Wan2.1-I2V-14B-480P \
|
|
--dataset_metadata_path data/diffsynth_example_dataset/wanvideo/Wan2.1-I2V-14B-480P/metadata.csv \
|
|
--height 480 \
|
|
--width 832 \
|
|
--dataset_repeat 1 \
|
|
--model_id_with_origin_paths "Wan-AI/Wan2.1-I2V-14B-480P:diffusion_pytorch_model*.safetensors,Wan-AI/Wan2.1-I2V-14B-480P:models_t5_umt5-xxl-enc-bf16.pth,Wan-AI/Wan2.1-I2V-14B-480P:Wan2.1_VAE.pth,Wan-AI/Wan2.1-I2V-14B-480P:models_clip_open-clip-xlm-roberta-large-vit-huge-14.pth" \
|
|
--learning_rate 1e-4 \
|
|
--num_epochs 5 \
|
|
--remove_prefix_in_ckpt "pipe.dit." \
|
|
--output_path "./models/train/Wan2.1-I2V-14B-480P_lora_split_cache" \
|
|
--lora_base_model "dit" \
|
|
--lora_target_modules "q,k,v,o,ffn.0,ffn.2" \
|
|
--lora_rank 32 \
|
|
--extra_inputs "input_image" \
|
|
--task "sft:data_process" \
|
|
--offload_models "Wan-AI/Wan2.1-I2V-14B-480P:diffusion_pytorch_model*.safetensors"
|
|
|
|
accelerate launch examples/wanvideo/model_training/train.py \
|
|
--dataset_base_path "./models/train/Wan2.1-I2V-14B-480P_lora_split_cache" \
|
|
--height 480 \
|
|
--width 832 \
|
|
--dataset_repeat 100 \
|
|
--model_id_with_origin_paths "Wan-AI/Wan2.1-I2V-14B-480P:diffusion_pytorch_model*.safetensors,Wan-AI/Wan2.1-I2V-14B-480P:models_t5_umt5-xxl-enc-bf16.pth,Wan-AI/Wan2.1-I2V-14B-480P:Wan2.1_VAE.pth,Wan-AI/Wan2.1-I2V-14B-480P:models_clip_open-clip-xlm-roberta-large-vit-huge-14.pth" \
|
|
--learning_rate 1e-4 \
|
|
--num_epochs 5 \
|
|
--remove_prefix_in_ckpt "pipe.dit." \
|
|
--output_path "./models/train/Wan2.1-I2V-14B-480P_lora_split" \
|
|
--lora_base_model "dit" \
|
|
--lora_target_modules "q,k,v,o,ffn.0,ffn.2" \
|
|
--lora_rank 32 \
|
|
--extra_inputs "input_image" \
|
|
--task "sft:train" \
|
|
--offload_models "Wan-AI/Wan2.1-I2V-14B-480P:models_t5_umt5-xxl-enc-bf16.pth,Wan-AI/Wan2.1-I2V-14B-480P:Wan2.1_VAE.pth,Wan-AI/Wan2.1-I2V-14B-480P:models_clip_open-clip-xlm-roberta-large-vit-huge-14.pth"
|