search

ZiqiaoPeng / SyncTalk

[CVPR 2024] This is the official source for our paper "SyncTalk: The Devil is in the Synchronization for Talking Head Synthesis"
https://ziqiaopeng.github.io/synctalk/
Other
1.35k stars 161 forks source link

我想知道是所有人物只需要训练一次就足够还是，每换过一个目标人物就要训练某个部分 #20

Closed zckcola closed 8 months ago

benchrus commented 8 months ago

I want to know if there's a way to control head pose.

ZiqiaoPeng commented 8 months ago

和ER-NeRF、GeneFace等方法一致，需要针对特定人物进行训练。