Pavel

Results 1 issues of Pavel

neural-fortran in it's current state effectively supports supervised learning, where gradients are computed in the network%backward() method of each layer based on the derivative of the loss function with respect...