/avsync-ai

implementation of Out of time: automated lip sync in the wild

Primary LanguagePython

This is a pytorch implementation of 《 Out of time: automated lip sync in the wild》 When I use multi frames to sync audio and video, I get accuracy of 91.6% with offset=0.08s

The code is absolutely ok to run

This is the dependency CUDA10.0+RTX 2080 Ti+python2.7+pytorch1.0