annotate toolboxes/FullBNT-1.0.7/netlab3.3/netpak.m @ 0:cc4b1211e677
tip
initial commit to HG from
Changeset:
646 (e263d8a21543) added further path and more save "camirversion.m"
| author |
Daniel Wolff |
| date |
Fri, 19 Aug 2016 13:07:06 +0200 |
| parents |
|
| children |
|
| rev |
line source |
|
Daniel@0
|
1 function w = netpak(net)
|
|
Daniel@0
|
2 %NETPAK Combines weights and biases into one weights vector.
|
|
Daniel@0
|
3 %
|
|
Daniel@0
|
4 % Description
|
|
Daniel@0
|
5 % W = NETPAK(NET) takes a network data structure NET and combines the
|
|
Daniel@0
|
6 % component weight matrices into a single row vector W. The facility
|
|
Daniel@0
|
7 % to switch between these two representations for the network
|
|
Daniel@0
|
8 % parameters is useful, for example, in training a network by error
|
|
Daniel@0
|
9 % function minimization, since a single vector of parameters can be
|
|
Daniel@0
|
10 % handled by general-purpose optimization routines. This function also
|
|
Daniel@0
|
11 % takes into account a MASK defined as a field in NET by removing any
|
|
Daniel@0
|
12 % weights that correspond to entries of 0 in the mask.
|
|
Daniel@0
|
13 %
|
|
Daniel@0
|
14 % See also
|
|
Daniel@0
|
15 % NET, NETUNPAK, NETFWD, NETERR, NETGRAD
|
|
Daniel@0
|
16 %
|
|
Daniel@0
|
17
|
|
Daniel@0
|
18 % Copyright (c) Ian T Nabney (1996-2001)
|
|
Daniel@0
|
19
|
|
Daniel@0
|
20 pakstr = [net.type, 'pak'];
|
|
Daniel@0
|
21 w = feval(pakstr, net);
|
|
Daniel@0
|
22 % Return masked subset of weights
|
|
Daniel@0
|
23 if (isfield(net, 'mask'))
|
|
Daniel@0
|
24 w = w(logical(net.mask));
|
|
Daniel@0
|
25 end |
