Daniel@0: function g = netgrad(w, net, x, t)
Daniel@0: %NETGRAD Evaluate network error gradient for generic optimizers
Daniel@0: %
Daniel@0: %	Description
Daniel@0: %
Daniel@0: %	G = NETGRAD(W, NET, X, T) takes a weight vector W and a network data
Daniel@0: %	structure NET, together with the matrix X of input vectors and the
Daniel@0: %	matrix T of target vectors, and returns the gradient of the error
Daniel@0: %	function evaluated at W.
Daniel@0: %
Daniel@0: %	See also
Daniel@0: %	MLP, NETERR, NETOPT
Daniel@0: %
Daniel@0: 
Daniel@0: %	Copyright (c) Ian T Nabney (1996-2001)
Daniel@0: 
Daniel@0: gradstr = [net.type, 'grad'];
Daniel@0: 
Daniel@0: net = netunpak(net, w);
Daniel@0: 
Daniel@0: g = feval(gradstr, net, x, t);