Toggle navigation
chex
Archive
Categories
Pages
Tags
Submit
Categories
pages
1
pages
Why do we rotate weights when computing the gradients in a convolution layer of a convolution network?