Skip to content
/ roop Public
forked from s0md3v/roop

one-click deepfake (face swap)

License

Notifications You must be signed in to change notification settings

walkxcode/roop

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Roop - Face Replacement Software

Roop is a face replacement software that allows you to replace faces in videos with a face of your choice using a single image. With Roop, you can create deepfake videos without the need for a dataset or training. Please note that this software is a hobby project created for learning purposes and should be used responsibly. Before using someone's face, ensure you have their consent and do not mislead others about the nature of the content. The developer of Roop is not responsible for any misuse or malicious behavior by end-users.

Demo

Check out some demos of Roop in action: Demo Videos

Demo GIF

Check out the latest video made by our test CI below:

Demo Video

Disclaimer

Please be aware that there are more advanced deepfake software available. Roop is a personal project created to explore AI technologies. The software includes a built-in check to prevent inappropriate media from being processed.

Installation

To install Roop, there are two types of installations available: basic and GPU-powered.

  • Basic Installation: This option is more likely to work on most computers but may be slower. Follow the instructions for the basic install here.

  • GPU Installation: If you have a powerful GPU and are comfortable troubleshooting potential software issues, you can enable GPU acceleration for faster processing. Start by following the basic installation instructions above, and then refer to the GPU-specific instructions here. Updated documentation coming soon!

Usage

Note: When running Roop for the first time, it will download models of approximately 300MB in size.

  1. Execute the following command to launch Roop:

    python run.py
  2. The Roop graphical user interface (GUI) window will appear as shown below: GUI Demo

  3. Select a face image (the image with the desired face) and a target image/video (the image/video in which you want to replace the face) using the provided options.

  4. Click on the Start button to begin the face replacement process.

  5. Open the file explorer and navigate to the directory where you selected the output to be saved. You will find a directory named <video_title> where you can observe the frames being swapped in real-time. Once the processing is complete, Roop will create the output file.

Caution: Unless you are familiar with the workings of the software, avoid modifying the FPS checkbox.

Additional command-line arguments for Roop are as follows:

Options:
  -h, --help                   Show this help message and exit
  -f SOURCE_IMG,
     --face SOURCE_IMG         Use this face
  -t TARGET_PATH,
     --target TARGET_PATH      Replace this face
  -o OUTPUT_FILE,
     --output OUTPUT_FILE      Save output to this file
  --gpu                        Use GPU
  --keep-fps                   Maintain original FPS
  --keep-frames                Keep frames directory
  --max-memory MAX_MEMORY      Maximum amount of RAM in GB to be used
  --max-cores CORES_COUNT      Number of cores to be used for CPU mode
  --all-faces                  Swap all faces in frame

If you prefer to use the CLI mode, you can use the -f/--face argument to run the program in CLI mode.

Future Plans

The developer has outlined the following future plans for Roop:

  • Improve the quality of faces in the results
  • Enable selective face replacement throughout the video
  • Support for replacing multiple faces

Credits

Roop acknowledges the following projects and libraries for their contributions:

  • FFmpeg: For simplifying video-related operations
  • Deepinsight: For their InsightFace project, providing a well-made library and models
  • All developers behind the libraries utilized in this project

About

one-click deepfake (face swap)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%