1020 B
1020 B
(wanvideo)=
WanVideo
Inference T2V with WanVideo
First, download the model:
python scripts/huggingface/download_hf.py --repo_id=Wan-AI/Wan2.1-T2V-1.3B-Diffusers --local_dir=YOUR_LOCAL_DIR --repo_type=model
or
python scripts/huggingface/download_hf.py --repo_id=Wan-AI/Wan2.1-T2V-14B-Diffusers --local_dir=YOUR_LOCAL_DIR --repo_type=model
Then run the inference using:
sh scripts/inference/v1_inference_wan.sh
Remember to set MODEL_BASE and num_gpus accordingly.
Inference I2V with WanVideo
First, download the model:
python scripts/huggingface/download_hf.py --repo_id=Wan-AI/Wan2.1-I2V-14B-480P-Diffusers --local_dir=YOUR_LOCAL_DIR --repo_type=model
or
python scripts/huggingface/download_hf.py --repo_id=Wan-AI/Wan2.1-I2V-14B-720P-Diffusers --local_dir=YOUR_LOCAL_DIR --repo_type=model
Then run the inference using:
sh scripts/inference/v1_inference_wan_i2v.sh
Remember to set MODEL_BASE and num_gpus accordingly.