NVlabs/few-shot-vid2vid
A PyTorch implementation of a few-shot neural network for photorealistic video-to-video translation that generates human motions from poses and synthesizes talking head videos.
Velocity · 7d
+0.7
★ / day
Trend
→steady
star history
This repository provides a deep learning model for few-shot video-to-video synthesis using PyTorch. The system can translate semantic label maps, pose sequences, and edge maps into photorealistic videos. It leverages image-to-image translation techniques and was published as a NeurIPS 2019 paper by NVIDIA researchers.