search

EvelynFan / FaceFormer

[CVPR 2022] FaceFormer: Speech-Driven 3D Facial Animation with Transformers
MIT License
761 stars 135 forks source link

my dataset #69

Open zmj1125 opened 1 year ago

zmj1125 commented 1 year ago

when using my own dataset, it occurs a problem. How can I solve it? @SEAY`}T A5ZAO 8 J8 Z74

ShaojunBian commented 1 year ago

Hiya, the audio you used for test, should be less then 20s, if your audio is longer than 20s, you need to adjust the code as #48.

zmj1125 commented 1 year ago

I have one more question, how did you get the 50? QQ图片20230310173841

Shirley-0708 commented 1 year ago

May I ask how you create your own dataset? Can you be specific? Thank you！ @zmj1125 @ShaojunBian

zhennanh commented 12 months ago

I have one more question, how did you get the 50? QQ图片20230310173841

Excuse me，have you got the answer for this question?