atiyo/deep_image_prior

Why 124 instead of 128 in U-Net deconv

vis-opt opened this issue · 1 comments

In https://github.com/atiyo/deep_image_prior/blob/master/deep_image_prior.py#L188 you have 124 instead of 128.

Can you explain why?

I see why the channel sizes were chosen this way after printing.