I adapt neural network instead of train. Any one knows how I can see which stopping criteria causes to stop the network ? (nntraintool command does not work!)

When I use the example of the ANN toolbox with training the network in the matlab, I am able to see number of the Epoch, Time, Performance, Gradient, Mu, and Validation checks with "nntraintool" command and I can figure out which one stops the network. But when I change train to adapt,"nntraintool" does not work and I do not how I can see Epoch, Time, Performance, Gradient, Mu, and Validation checks and the one that stops the network. I was wondering if some one know it?

[Sanaz]

So it means when I use "adapt" and the network stops, is is not clear which criterion stops the network? Does "adapt" have the validation check criteria?

[Greg Heath]

Yes.

No.

Answer 1

[Greg Heath]

If you use the command

 [ net tr Y E Xf Af ] = train(net, x, t, Xi, Ai, EW);

 tr = tr

yields training information on over 30 quantities.

tr.stop yields the stopping criterion.

However, if you use

 [ net Y E Xf Af tr] = adapt(net, x, t, Xi, Ai);

 tr = tr

only yields the training records of epoch and perf.

Don't expect more info than that using the GUI.

Thank you for formally accepting my answer

Greg

P.S. Details obtained using help, doc and type on train and adapt