theluyuan/DiffSynth-Studio
1
0
Fork 0
mirror of https://github.com/modelscope/DiffSynth-Studio.git synced 2026-03-23 17:38:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #1034 from modelscope/video_as_prompt

Browse Source 
Video as prompt
This commit is contained in:
Zhongjie Duan
2025-11-04 17:32:50 +08:00
committed by GitHub
parent 401d7d74a5
commit 8332ecebb7
13 changed files with 635 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
examples/wanvideo/model_training/lora/Video-As-Prompt-Wan2.1-14B.sh Normal file
View File

@@ -0,0 +1,18 @@
accelerate launch examples/wanvideo/model_training/train.py \
--dataset_base_path data/example_video_dataset \
--dataset_metadata_path data/example_video_dataset/metadata_vap.csv \
--data_file_keys "video,vap_video" \
--height 480 \
--width 832 \
--num_frames 49 \
--dataset_repeat 10 \
--model_id_with_origin_paths "ByteDance/Video-As-Prompt-Wan2.1-14B:transformer/diffusion_pytorch_model*.safetensors,Wan-AI/Wan2.1-I2V-14B-720P:models_t5_umt5-xxl-enc-bf16.pth,Wan-AI/Wan2.1-I2V-14B-720P:Wan2.1_VAE.pth,Wan-AI/Wan2.1-I2V-14B-720P: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/Video-As-Prompt-Wan2.1-14B_lora" \
--lora_base_model "dit" \
--lora_target_modules "q,k,v,o,ffn.0,ffn.2" \
--lora_rank 32 \
--extra_inputs "vap_video,input_image" \
--use_gradient_checkpointing_offload