Mercurial > hg > camir-aes2014
annotate toolboxes/FullBNT-1.0.7/netlab3.3/netpak.m @ 0:e9a9cd732c1e tip
first hg version after svn
| author | wolffd |
|---|---|
| date | Tue, 10 Feb 2015 15:05:51 +0000 |
| parents | |
| children |
| rev | line source |
|---|---|
| wolffd@0 | 1 function w = netpak(net) |
| wolffd@0 | 2 %NETPAK Combines weights and biases into one weights vector. |
| wolffd@0 | 3 % |
| wolffd@0 | 4 % Description |
| wolffd@0 | 5 % W = NETPAK(NET) takes a network data structure NET and combines the |
| wolffd@0 | 6 % component weight matrices into a single row vector W. The facility |
| wolffd@0 | 7 % to switch between these two representations for the network |
| wolffd@0 | 8 % parameters is useful, for example, in training a network by error |
| wolffd@0 | 9 % function minimization, since a single vector of parameters can be |
| wolffd@0 | 10 % handled by general-purpose optimization routines. This function also |
| wolffd@0 | 11 % takes into account a MASK defined as a field in NET by removing any |
| wolffd@0 | 12 % weights that correspond to entries of 0 in the mask. |
| wolffd@0 | 13 % |
| wolffd@0 | 14 % See also |
| wolffd@0 | 15 % NET, NETUNPAK, NETFWD, NETERR, NETGRAD |
| wolffd@0 | 16 % |
| wolffd@0 | 17 |
| wolffd@0 | 18 % Copyright (c) Ian T Nabney (1996-2001) |
| wolffd@0 | 19 |
| wolffd@0 | 20 pakstr = [net.type, 'pak']; |
| wolffd@0 | 21 w = feval(pakstr, net); |
| wolffd@0 | 22 % Return masked subset of weights |
| wolffd@0 | 23 if (isfield(net, 'mask')) |
| wolffd@0 | 24 w = w(logical(net.mask)); |
| wolffd@0 | 25 end |