Question

When training a nn using trainbr like this:

net = newff( inputs_train, outputs_train, layers, { 'tansig', 'purelin' }, 'trainbr' );
net.trainParam.goal = 0.0;
net.trainParam.show = 1;
net.trainParam.epochs = ephocs;
net = train( net, inputs_train, outputs_train );

How do I obtain the number of epohcs/iterations that has been run before max mu has been reached?

Cheers

Was it helpful?

Solution

I would use the newer FEEDFORWARDNET instead of the old NEWFF.

You can get the training record as a second argument from TRAIN:

[net,tr] = train(net,inputs_train,outputs_train);
tr.num_epochs
Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top