Refacer, a simple tool that allows you to create deepfakes with multiple faces with just one click! This project was inspired by Roop and is powered by the excellent Insightface. Refacer requires no training - just one photo and you're ready to go.

⚠️ Please, before using the code from this repository, make sure to read the disclaimer.

Demonstration

System Compatibility

Refacer has been thoroughly tested on the following operating systems:

Operating System	CPU Support	GPU Support
MacOSX	✅	⚠️
Windows	✅	✅
Linux	✅	✅

The application is compatible with both CPU and GPU (Nvidia CUDA) environments, and MacOSX(CoreML)

⚠️ Please note, we do not recommend using onnxruntime-silicon on MacOSX due to an apparent issue with memory management. If you manage to compile onnxruntime for Silicon, the program is prepared to use CoreML.

Prerequisites

Ensure that you have ffmpeg installed and correctly configured. There are many guides available on the internet to help with this. Here are a few (note: I did not create these guides):

How to Install FFmpeg

Installation

Refacer has been tested and is known to work with Python 3.10.9, but it is likely to work with other Python versions as well. It is recommended to use a virtual environment for setting up and running the project to avoid potential conflicts with other Python packages you may have installed.

Follow these steps to install Refacer:

Clone the repository:

git clone https://github.com/xaviviro/refacer.git
cd refacer

Download the Insightface model: You can manually download the model created by Insightface from this link and add it to the project folder. Alternatively, if you have wget installed, you can use the following command:
```
wget --content-disposition https://huggingface.co/deepinsight/inswapper/resolve/main/inswapper_128.onnx
```
Install dependencies:
- For CPU (compatible with Windows, MacOSX, and Linux):
```
pip install -r requirements.txt
```
- For GPU (compatible with Windows and Linux only, requires a NVIDIA GPU with CUDA and its libraries):
```
pip install -r requirements-GPU.txt
```
- For CoreML (compatible with MacOSX, requires Silicon architecture):
```
pip install -r requirements-COREML.txt
```
For more information on installing the CUDA necessary to use onnxruntime-gpu, please refer directly to the official ONNX Runtime repository.

For more details on using the Insightface model, you can refer to their example.

Usage

Once you have successfully installed Refacer and its dependencies, you can run the application using the following command:

python app.py

Then, open your web browser and navigate to the following address:

http://127.0.0.1:7680

Questions?

If you have any questions or issues, feel free to open an issue or submit a pull request.

Recognition Module

The recognition folder in this repository is derived from Insightface's GitHub repository. You can find the original source code here: Insightface Recognition Source Code

This module is used for recognizing and handling face data within the Refacer application, enabling its powerful deepfake capabilities. We are grateful to Insightface for their work and for making their code available.

License

Note: This project uses a Custom MIT License. See LICENSE for full terms.

The generated content does not represent the views, beliefs, or attitudes of the authors of this Software. Please use the Software and its outputs responsibly, ethically, and with respect toward others.