search

OpenTalker / SadTalker

[CVPR 2023] SadTalker：Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation
https://sadtalker.github.io/
Other
11.62k stars 2.17k forks source link

How do I train my auido2exp model #801

Open xyh19960622 opened 7 months ago

xyh19960622 commented 7 months ago

The lip synthesis effect of auido2exp_00300-model.pth given by the author is not satisfied, can I use the wav2lip model trained by myself, can you tell me what I should do?

StephanPan commented 6 months ago

The lip synthesis effect of auido2exp_00300-model.pth given by the author is not satisfied, can I use the wav2lip model trained by myself, can you tell me what I should do?

do you reproduct the training code ?