search

NVIDIA / vid2vid

Pytorch implementation of our method for high-resolution (e.g. 2048x1024) photorealistic video-to-video translation.
Other
8.62k stars 1.2k forks source link

I got the bad result about edge to face. #109

Open tjusxh opened 5 years ago

tjusxh commented 5 years ago

Thanks for your sharing. I have several questions.

  1. If I set fg， the labels_nc should be stetted 35, what's the meaning of the 35 classes? or which dataset was used?
  2. I try to different parameters for edge to face,( --no_canny_edge --fg), but I got the bad result. As the number of frames increases， the identity changed more and more. Which detail I didn't notice?

image

linkAmy commented 5 years ago

well, I got the same problem.

primejava commented 5 years ago

I got the same problem.

jakeelwes commented 4 years ago

Did you work this out @tjusxh? Thanks :)