search

mcahny / Deep-Video-Inpainting

Official pytorch implementation for "Deep Video Inpainting" (CVPR 2019)
508 stars 95 forks source link

Confusion about previous completed frame in the training phase. #11

Closed facetohard closed 5 years ago

facetohard commented 5 years ago

How can I get the previous completed frame or just use previous gt frame?

facetohard commented 5 years ago

The answer is that both two training phase use video sequence to train network, so that in each sequence, the previous frame of the first frame is set zero value, and then use the outputs of network as the input of next iteration. I had made a mistake that I thought two training phase was different.