"Face Renderer" takes almost an hour to complete even for 15 sec audio clip

OpenTalker / SadTalker

[CVPR 2023] SadTalker：Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation

https://sadtalker.github.io/

Other

11.8k stars 2.2k forks source link

"Face Renderer" takes almost an hour to complete even for 15 sec audio clip #510

Open NBNew opened 1 year ago

NBNew commented 1 year ago

Hello there,

I have trying to make SadTalker work with my SD WebUI however having challenges with Face Renderer step. It takes almost an hour to complete for a 15 sec audio itself which I strongly believe is not expected.

I tried running in WebUI extension as we;l as directly launching through py script, both shows almost same result

Any help would be greatly appreciated !!

Sadtalker Face Renderer Issue

Simmer1234569 commented 1 year ago

Did you ever find a fix?