yhw-yhw / TalkSHOW

This is the official repository for TalkSHOW: Generating Holistic 3D Human Motion from Speech [CVPR2023].
301 stars 26 forks source link

Generalization Ability To Chinese #13

Open FishWoWater opened 1 year ago

FishWoWater commented 1 year ago

Really excellent project! I am wondering whether TalkSHOW could generalise to chinese speech? If not, how can I make this possible? Perhaps I should create paired chineseSpeech-smplx labels(fitted from videos using pipeline in your paper).

Thank you very much.