Closed barapa closed 11 years ago
Currently, in the NN code, it should getused, but it doesnt in the rbmtrain code. Also need to scale the gradient with respect to the actual minibatch size.
Currently, in the NN code, it should getused, but it doesnt in the rbmtrain code. Also need to scale the gradient with respect to the actual minibatch size.