I started wondering looking at the code of the Convolutional layer that would it be so easy to add the padding to the descriptor. But after running a few tests and outputting the filters it looks good and also the dimension matches. Would it be so easy breeze?
I started wondering looking at the code of the Convolutional layer that would it be so easy to add the padding to the descriptor. But after running a few tests and outputting the filters it looks good and also the dimension matches. Would it be so easy breeze?