You can swap your face from a webcam or the face in the video using trained face models. Here is a list of available ready-to-use public face models. These persons do not exists. Similarities with real people are accidental. Except Keanu Reeves. He exists, and he's breathtaking! | |||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||
If you want a higher quality or better face match, you can train your own face model using DeepFaceLab
Here is an example of Arnold Schwarzneggar trained on a particular face and used in a video call. Read the FAQ for more information. |
You can swap your face from a webcam or the face in the video using your own single photo. |
There is also a Face Animator module in DeepFaceLive app. You can control a static face picture using video or your own face from the camera. The quality is not the best, and requires fine face matching and tuning parameters for every face pair, but enough for funny videos and memes or real-time streaming at 25 fps using 35 TFLOPS GPU. Here is a mini video showing the process of setting up the Face Animator for Obama controlling Kim Chen's face. |
any DirectX12 compatible graphics card (Recommended RTX 2070+ / Radeon RX 5700 XT+ ) Modern CPU with AVX instructions 4GB RAM, 32GB+ paging file Windows 10 | |
Windows | |
Linux | Build info |
Frequently asked questions | for User |
Contains stand-alone zero-dependency all-in-one ready-to-use portable self-extracting folder! You don't need to install anything other than video drivers.
DirectX12 build : NVIDIA, AMD, Intel videocards. NVIDIA build : NVIDIA cards only, GT730 and higher. Works faster than DX12. FaceMerger can work also on AMD/Intel. | |
Discord | Official discord channel. English / Russian. |
QQ群124500433 | 中文交流QQ群,商务合作找群主 |
Train your own face model by following the recommendations in the FAQ section and share it on Discord. If the model fits the quality, it will be added to the public library. | |
Register github account and push "Star" button. | |
Donate via Yoomoney | |
bitcoin:bc1qewl062v70rszulml3f0mjdjrys8uxdydw3v6rq | |