How to replace faces of videos using Roop from GitHub on Windows?
1. First download the code from roop GitHub repository. Open a dos prompt, in the directory where you want to install, type the command:
>git clone https://github.com/s0md3v/roop.git
2. Go to Visual studio download page. Select community and click Free download. Run the installer. Select three elements to install: Python development, desktop c++, and visual studio extension development.
3. Go to ffmpeg download page and download the full_build.zip file for windows. Unzip it and put on your local drive. Add its bin path to your environment variables.
4. Download inswapper_128.onnx from huggingface or Google drive. Put it under “roop\models” directory.
5. You need to configure a virtual environment to run the code. If you haven’t installed Anaconda3, go to install Anaconda3.
6. Setup a conda environment for roop with this instruction. If you run into problems when you setup, you can download a ready-for-use roop_env at Gumroad (compatible with CUDA 11.8). Unzip it and put it at your anaconda3 installation under “envs” directory.
7. Prepare your input files. The first one is a short video mp4 file with a face movement. The dimension is 512 x 512. The second file is an image with a face that you want to use for your new video. The dimension is 512 x 512.
8. Open an Anaconda Prompt. Run command:
>conda activate roop_env
9. Still in the Anaconda prompt, go to the directory “roop” and run command:
>python run.py
If you have setup PyTorch GPU on Windows, you can run this command:
>python run.py –execution-provider cuda
10. A roop window pop up. Click “Select a face” button to upload the face image. Click “Select a target” to load the video file. Click “Start” button.
11. When the run finishes, the new video is called “output.mp4” under the same directory with input files.