Sefi Kligler
Sefi Kligler
Sorry I do not have access to the results since I am no longer in the institute. If necessary we can try to get them. Note: 1. PSNR code calculations...
Thanks for the compliments. That was a challenging issue to fit our representation of kernel to SRMD's. It is not a simple crop (as far as I remember) My colleague...
I change the dimensions from `1,3,H,W` to `3,1,H,W` because this way the first layer is `1,Cout,k,k rather then `3,Cout,k,k`. The reason this is important is because I want a single...
It either clips or compresses the values accordingly. I do not remember. Any, I think it is not significant at all
This is a basic convolutional layer. A convolutional layer with stride 2 resizes the image from 64x64 to 32x32. If you use 64 filters of size 3 you get an...
Then I might have a mistake in the figure. It seems the stride is the last layer. Sorry for that
Pytorch's default pads the image to keep the resolution. In order to ignore boundaries artifacts I discard them (as far as I recall - I use a mask to discard...
Yes, G is always **exactly** equivalent to a convolution with a single kernel. Since there are no non-linear activations, the sequence of convolutions (which are LINEAR!!!) can be replaced with...
I am sorry but I don't have the time to thoroughly check it. Your script seems right. I verified correctness when the model was developed, but since it was a...
Great news @jnoylinc. thanks!