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

first hg version after svn

rev	line source
wolffd@0	1 function w = mlppak(net)
wolffd@0	2 %MLPPAK Combines weights and biases into one weights vector.
wolffd@0	3 %
wolffd@0	4 % Description
wolffd@0	5 % W = MLPPAK(NET) takes a network data structure NET and combines the
wolffd@0	6 % component weight matrices bias vectors into a single row vector W.
wolffd@0	7 % The facility to switch between these two representations for the
wolffd@0	8 % network parameters is useful, for example, in training a network by
wolffd@0	9 % error function minimization, since a single vector of parameters can
wolffd@0	10 % be handled by general-purpose optimization routines.
wolffd@0	11 %
wolffd@0	12 % The ordering of the paramters in W is defined by
wolffd@0	13 % w = [net.w1(:)', net.b1, net.w2(:)', net.b2];
wolffd@0	14 % where W1 is the first-layer weight matrix, B1 is the first-layer
wolffd@0	15 % bias vector, W2 is the second-layer weight matrix, and B2 is the
wolffd@0	16 % second-layer bias vector.
wolffd@0	17 %
wolffd@0	18 % See also
wolffd@0	19 % MLP, MLPUNPAK, MLPFWD, MLPERR, MLPBKP, MLPGRAD
wolffd@0	20 %
wolffd@0	21
wolffd@0	22 % Copyright (c) Ian T Nabney (1996-2001)
wolffd@0	23
wolffd@0	24 % Check arguments for consistency
wolffd@0	25 errstring = consist(net, 'mlp');
wolffd@0	26 if ~isempty(errstring);
wolffd@0	27 error(errstring);
wolffd@0	28 end
wolffd@0	29
wolffd@0	30 w = [net.w1(:)', net.b1, net.w2(:)', net.b2];
wolffd@0	31

Mercurial > hg > camir-aes2014