Pytorch implementation for high-resolution (e.g., 2048x1024) photorealistic video-to-video translation. It can be used for turning semantic label maps into photo-realistic videos, synthesizing people ...