github.com

Pytorch implementation for few-shot photorealistic video-to-video translation.

It can be used for generating human motions from poses, synthesizing people talking from edge maps, or turning semantic label maps into photo-realistic videos.