camir-aes2014: toolboxes/FullBNT-1.0.7/netlab3.3/nethess.m annotate

first hg version after svn

rev	line source
wolffd@0	1 function [h, varargout] = nethess(w, net, x, t, varargin)
wolffd@0	2 %NETHESS Evaluate network Hessian
wolffd@0	3 %
wolffd@0	4 % Description
wolffd@0	5 %
wolffd@0	6 % H = NETHESS(W, NET, X, T) takes a weight vector W and a network data
wolffd@0	7 % structure NET, together with the matrix X of input vectors and the
wolffd@0	8 % matrix T of target vectors, and returns the value of the Hessian
wolffd@0	9 % evaluated at W.
wolffd@0	10 %
wolffd@0	11 % [E, VARARGOUT] = NETHESS(W, NET, X, T, VARARGIN) also returns any
wolffd@0	12 % additional return values from the network Hessian function, and
wolffd@0	13 % passes additional arguments to that function.
wolffd@0	14 %
wolffd@0	15 % See also
wolffd@0	16 % NETERR, NETGRAD, NETOPT
wolffd@0	17 %
wolffd@0	18
wolffd@0	19 % Copyright (c) Ian T Nabney (1996-2001)
wolffd@0	20
wolffd@0	21 hess_str = [net.type, 'hess'];
wolffd@0	22
wolffd@0	23 net = netunpak(net, w);
wolffd@0	24
wolffd@0	25 [s{1:nargout}] = feval(hess_str, net, x, t, varargin{:});
wolffd@0	26 h = s{1};
wolffd@0	27 for i = 2:nargout
wolffd@0	28 varargout{i-1} = s{i};
wolffd@0	29 end

Mercurial > hg > camir-aes2014