Cyra

> Hi! I also cannot get the evaluation(PSNR, SSIM, LPIPS) method as original paper, have you finally get the result similar to table 1?

> Hey, > > I ran the inference on the 29 Huang annotated sequences from DAVIS 2017. > > ``` > srun python video_completion.py \ > --mode object_removal \ >...

> I have tried to re-implement your work, but the result of the pretrained model can not correspond to that of the paper. Specifically, the warping error is smaller and...

> @phoenix104104 Hey. During training, how are loss functions expected to behave, especially temporal losses? Hi, when I was trainning this network, temporal loss didn't converge, how about your loss...