Useful for calc gradients in backpropagation step.
Ref.:
http://cs231n.github.io/optimization-2/
https://wiseodd.github.io/techblog/2016/08/12/lstm-backprop/
Useful for calc gradients in backpropagation step.
Ref.:
http://cs231n.github.io/optimization-2/
https://wiseodd.github.io/techblog/2016/08/12/lstm-backprop/