This project is under development. There may be compatibility issues in future upgrades. If you encounter any problems, please create an issue, thanks.

Installation

Install custom node

git clone https://github.com/jeffy5/comfyui-faceless-node
pip install -r requirements.txt

Download model

All models is same as facefusion which can be found in facefusion assets.

Now it will use the following model by default. In the future, dynamic model selection will be supported. Please download the model based on the following directory structure.

Put all into directory CmofyUI/models/faceless.

|-- face_detector
|   `-- yoloface_8n.onnx
|-- face_landmarker
|   |-- 2dfan4.onnx
|   `-- face_landmarker_68_5.onnx
|-- face_recognizer
|   `-- arcface_w600k_r50.onnx
├── face_restoration
│   └── gfpgan_1.4.onnx
|-- face_swapper
|   └── inswapper_128.onnx
|-- gender_age.onnx
`-- rmbg.pth

You can download models by running download_models.py script under directory CmofyUI/custom_nodes/comfyui-faceless-node.

# Install required models
python download_models.py

# Install all models
python download_models.py --all

Example workflows

Image face swap

Video face swap

Load frames to preview image

It doesn't support preview video for now. All video will be save in directory output/faceless.

Thanks

Thanks to Facefusion. This project is based on facefusion to implenment a special version for ComfyUI.