Is it possible to view the output of applying a nn.Conv2d with specific weights (so after training) to an image?
I would like to “look” at what the early channels in a CNN look like; as I think their outputs should be relatively “visually” understandable?