Open super-wcg opened 4 years ago
I don't understand how to update the self.softmax() when training, because the acc don't backward. How does it work?
@jefflai108
I don't understand how to update the self.softmax() when training, because the acc don't backward. How does it work?