Yixuan Zhu*, Haolin Wang*, Shilin Ma*, Wenliang Zhao, Yansong Tang, $\dagger$ Jie Zhou, $\dagger$

* Equal contribution   $\dagger$ Corresponding author

[Paper]

The repository contains the official implementation for the paper "FADE: Frequency-Aware Diffusion Model Factorization for Video Editing" (CVPR 2025).
FADE, which refers to Frequency-Aware Diffusion Model Factorization for Video Editing, is a training-free yet highly effective video editing approach that fully leverages the inherent priors from pre-trained video diffusion.

• ☑️ Release the code for video editing
• ☑️ Release the paper

We recommend you to use an Anaconda virtual environment. If you have installed Anaconda, run the following commands to create and activate a virtual environment.

conda create --name FADE python=3.10
conda activate FADE
pip install -r requirements.txt

We use Cogvideo-5B as our foundation model, please download this model.

pip install huggingface_hub
huggingface-cli login
hf download zai-org/CogVideoX-5b

We utilize HybridGL to generate mask used for editing, please clone this respository and set up the environment by running the following command.

cd HybridGL
python -m spacy download en_core_web_lg

cd third_party
cd modified_CLIP
pip install -e . --no-build-isolation

cd ..
cd segment-anything
pip install -e .
cd ../..
mkdir checkpoints 
cd checkpoints 
wget https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth

Run editing with CogVideoX-5b: bash edit_pipeline.sh

• Upload video to input folder.

• Modify edit_pipeline.sh, especially init_prompt, edit_prompt, input_video_path, and other related parameters.
  

• Create a config file by following the format of configs/bear.yaml. The trade-off between preservation and editing can be tuned by adjusting self_attn_gs.
  

Run bash edit_pipeline.sh

This project is licensed under the MIT License.

If you use this code for your research, please cite our paper:

@misc{zhu2025fadefrequencyawarediffusionmodel,
      title={FADE: Frequency-Aware Diffusion Model Factorization for Video Editing}, 
      author={Yixuan Zhu and Haolin Wang and Shilin Ma and Wenliang Zhao and Yansong Tang and Lei Chen and Jie Zhou},
      year={2025},
      eprint={2506.05934},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2506.05934}, 
}