@Override
public void handleLearningEvent(LearningEvent event) {
BackPropagation bp = (BackPropagation) event.getSource();
LOG.error("Current iteration: " + bp.getCurrentIteration());
LOG.error("Error: " + bp.getTotalNetworkError());
LOG.error("Calculation time: " + (System.currentTimeMillis() - start) / 1000.0);
// neuralNetwork.save(bp.getCurrentIteration() + "CNN_MNIST" + bp.getCurrentIteration() +
// ".nnet");
start = System.currentTimeMillis();
// NeuralNetworkEvaluationService.completeEvaluation(neuralNetwork, testSet);
}
public static void main(String[] args) {
try {
DataSet trainSet =
MNISTDataSet.createFromFile(
MNISTDataSet.TRAIN_LABEL_NAME, MNISTDataSet.TRAIN_IMAGE_NAME, 60000);
DataSet testSet =
MNISTDataSet.createFromFile(
MNISTDataSet.TEST_LABEL_NAME, MNISTDataSet.TEST_IMAGE_NAME, 10000);
Layer2D.Dimensions inputDimension = new Layer2D.Dimensions(32, 32);
Kernel convolutionKernel = new Kernel(5, 5);
Kernel poolingKernel = new Kernel(2, 2);
ConvolutionalNetwork convolutionNetwork =
new ConvolutionalNetwork.Builder(inputDimension, 1)
.withConvolutionLayer(convolutionKernel, 10)
.withPoolingLayer(poolingKernel)
.withConvolutionLayer(convolutionKernel, 1)
.withPoolingLayer(poolingKernel)
.withConvolutionLayer(convolutionKernel, 1)
.withFullConnectedLayer(10)
.createNetwork();
BackPropagation backPropagation = new MomentumBackpropagation();
backPropagation.setLearningRate(0.0001);
backPropagation.setMaxError(0.00001);
backPropagation.setMaxIterations(500);
backPropagation.addListener(new LearningListener());
backPropagation.setErrorFunction(new MeanSquaredError());
convolutionNetwork.setLearningRule(backPropagation);
backPropagation.addListener(new LearningListener());
System.out.println("Started training...");
convolutionNetwork.learn(testSet);
System.out.println("Done training!");
CrossValidation crossValidation = new CrossValidation(convolutionNetwork, testSet, 6);
crossValidation.run();
// ClassificationMetrics validationResult =
// crossValidation.computeErrorEstimate(convolutionNetwork, trainSet);
// Evaluation.runFullEvaluation(convolutionNetwork, testSet);
convolutionNetwork.save("/mnist.nnet");
} catch (IOException e) {
e.printStackTrace();
}
}