May 2020
Beginner to intermediate
430 pages
10h 39m
English
Normally, in a convolution, we move the kernel by one step, apply a convolution to that step, and so on. The stride allows us to skip a step. Let's take a look:

Here, each 3 x 3 window shows the results of skipping one step. The result of a stride is to shrink the dimension since we are skipping over potential x, y locations.
Read now
Unlock full access