3 ms·
Lots of linear functions with \x -> max(0, x) thrown in between. (That's literally what neural nets with relu as activation unit do.)
by eru 4y ago
Lots of linear functions with \x -> max(0, x) thrown in between.
(That's literally what neural nets with relu as activation unit do.)