annotate toolboxes/FullBNT-1.0.7/nethelp3.3/netinit.htm @ 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 <html>
wolffd@0 2 <head>
wolffd@0 3 <title>
wolffd@0 4 Netlab Reference Manual netinit
wolffd@0 5 </title>
wolffd@0 6 </head>
wolffd@0 7 <body>
wolffd@0 8 <H1> netinit
wolffd@0 9 </H1>
wolffd@0 10 <h2>
wolffd@0 11 Purpose
wolffd@0 12 </h2>
wolffd@0 13 Initialise the weights in a network.
wolffd@0 14
wolffd@0 15 <p><h2>
wolffd@0 16 Synopsis
wolffd@0 17 </h2>
wolffd@0 18 <PRE>
wolffd@0 19 net = netinit(net, prior)
wolffd@0 20 </PRE>
wolffd@0 21
wolffd@0 22
wolffd@0 23 <p><h2>
wolffd@0 24 Description
wolffd@0 25 </h2>
wolffd@0 26
wolffd@0 27 <p><CODE>net = netinit(net, prior)</CODE> takes a network data structure
wolffd@0 28 <CODE>net</CODE> and sets the weights and biases by sampling from a Gaussian
wolffd@0 29 distribution. If <CODE>prior</CODE> is a scalar, then all of the parameters
wolffd@0 30 (weights and biases) are sampled from a single isotropic Gaussian with
wolffd@0 31 inverse variance equal to <CODE>prior</CODE>. If <CODE>prior</CODE> is a data
wolffd@0 32 structure of the kind generated by <CODE>mlpprior</CODE>, then the parameters
wolffd@0 33 are sampled from multiple Gaussians according to their groupings
wolffd@0 34 (defined by the <CODE>index</CODE> field) with corresponding variances
wolffd@0 35 (defined by the <CODE>alpha</CODE> field).
wolffd@0 36
wolffd@0 37 <p><h2>
wolffd@0 38 See Also
wolffd@0 39 </h2>
wolffd@0 40 <CODE><a href="mlpprior.htm">mlpprior</a></CODE>, <CODE><a href="netunpak.htm">netunpak</a></CODE>, <CODE><a href="rbfprior.htm">rbfprior</a></CODE><hr>
wolffd@0 41 <b>Pages:</b>
wolffd@0 42 <a href="index.htm">Index</a>
wolffd@0 43 <hr>
wolffd@0 44 <p>Copyright (c) Ian T Nabney (1996-9)
wolffd@0 45
wolffd@0 46
wolffd@0 47 </body>
wolffd@0 48 </html>