Endless iterations with the sample code...

Hello,

Yes, I'm a neural network / encog newbie...

But I think I should be able to implement the most basic example from the XOR sample code.

Here's what I wrote:

private double[][] input = { { 0.0, 0.0 }, { 1.0, 0.0 }, { 0.0, 1.0 }, { 1.0, 1.0 } };
private double[][] ideal = { { 0.0 }, { 1.0 }, { 1.0 }, { 0.0 } };

BasicNetwork network = new BasicNetwork();
network.addLayer(new BasicLayer(2));
network.addLayer(new BasicLayer(2));
network.addLayer(new BasicLayer(1));

network.getStructure().finalizeStructure();
network.reset();

NeuralDataSet trainingSet = new BasicNeuralDataSet(input, ideal);

Train train = new ResilientPropagation(network, trainingSet);
int epoch = 1;
Logging.allConsoleLogging();
do
{
train.iteration();
System.out.println("Epoch #" + epoch + " Error:" + Format.formatPercent(train.getError()));
epoch++;
}
while(train.getError() > 0.02);

Using Encog 2.4 this do loop sometimes runs endless.

OK, not endless, but it runs until I stop it. The (truncated) output is:
...
Epoch #350402 Error:40,824829%
Epoch #350403 Error:40,824829%
Epoch #350404 Error:40,824829%
Epoch #350405 Error:40,824829%
Epoch #350406 Error:40,824829%
Epoch #350407 Error:40,824829%
Epoch #350408 Error:40,824829%
Epoch #350409 Error:40,824829%
Epoch #350410 Error:40,824829%
Epoch #350411 Error:40,824829%
Epoch #350412 Error:40,824829%
Epoch #350413 Error:40,824829%
...

Most likely the problem is sitting in front of the monitor and I'm sure I'm missing something simple...

Some other time the training succeds after a few hundred iterations.

And if I use my target data: An imput layer of 185 neurons, it also seems to be stuck...

André