Class CrossEntropy

source code

The cross-entropy criterion is well suited for targets vector that are normalized between 0 and 1, used along with a final Sigmoid activation module.

It has been experimentally demonstrated that an AutoEncoder trains faster with a cross-entropy criterion.

The cross-entropy loss can be written as follows :

L_CE =  − 1 ⁄ N ∑^N_k = 1(y⋅log(ŷ) + (1 − y)⋅log(1 − ŷ))