Notice

• This is a diffusers gguf ，not a comfyUI gguf ，注意，这是基于diffuser的管线量化的gguf测试模型

pipeline

from diffusers import  GGUFQuantizationConfig,WanTransformer3DModel,WanVideoToVideoPipeline
from transformers import UMT5EncoderModel
from diffusers.models import AutoencoderKLWan

ggufpath="https://huggingface.co/smthem/SteadyDancer-14B-gguf/blob/main/SteadyDancer-14B-Q80.gguf"
model_id="Wan-AI/Wan2.1-I2V-14B-720P-Diffusers"

transformer = WanTransformer3DModel.fromsinglefile(
        gguf_path,
        config=model_id,
        quantizationconfig=GGUFQuantizationConfig(computedtype=torch.bfloat16),
        torch_dtype=torch.bfloat16,
        )

vae=AutoencoderKLWan.frompretrained(modelid, torch_dtype=torch.bfloat16,)
textencoder=UMT5EncoderModel.frompretrained(modelid, torchdtype=torch.bfloat16,)

pipe = WanVideoToVideoPipeline.frompretrained(modelid, vae=vae,transformer=transformer,textencoder=textencoder, torch_dtype=torch.bfloat16)

run infer
...

📚 Citation

If you find our paper or this codebase useful for your research, please cite us.

@misc{zhang2025steadydancer,
      title={SteadyDancer: Harmonized and Coherent Human Image Animation with First-Frame Preservation}, 
      author={Jiaming Zhang and Shengming Cao and Rui Li and Xiaotong Zhao and Yutao Cui and Xinglin Hou and Gangshan Wu and Haolan Chen and Yu Xu and Limin Wang and Kai Ma},
      year={2025},
      eprint={2511.19320},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2511.19320}, 
}