fix: 0 faces found error

tencent-ailab / V-Express

V-Express aims to generate a talking head video under the control of a reference image, an audio, and a sequence of V-Kps images.

2.2k stars 277 forks source link

fix: 0 faces found error #36

Closed KMiNT21 closed 3 months ago

KMiNT21 commented 3 months ago

The resolution setting in insightface.app/FaceAnalysis > app.prepare () doesn't control the input image resolution. It's recommended to use a maximum resolution of 640x640 for optimal results.

FurkanGozukara commented 3 months ago

@tiankuan93 @Tenvence please check and merge if possible thank you so much