private int InternalCallback(global::System.IntPtr netPtr, global::System.IntPtr dataPtr, uint max_epochs, uint epochs_between_reports, float desired_error, uint epochs, global::System.IntPtr user_data)
{
NeuralNet callbackNet = new NeuralNet(new neural_net(netPtr, false));
TrainingData callbackData = new TrainingData(new training_data(dataPtr, false));
GCHandle handle = (GCHandle)user_data;
return Callback(callbackNet, callbackData, max_epochs, epochs_between_reports, desired_error, epochs, handle.Target as Object);
}
/* Constructor: NeuralNet
Creates a copy the other NeuralNet.
*/
public NeuralNet(NeuralNet other)
{
net = new neural_net(other.Net.to_fann());
Outputs = other.Outputs;
}
static void Main()
{
const float desired_error = 0.0F;
uint max_neurons = 30;
uint neurons_between_reports = 1;
uint bit_fail_train, bit_fail_test;
float mse_train, mse_test;
DataType[] output;
DataType[] steepness = new DataType[1];
int multi = 0;
ActivationFunction[] activation = new ActivationFunction[1];
TrainingAlgorithm training_algorithm = TrainingAlgorithm.TRAIN_RPROP;
Console.WriteLine("Reading data.");
using (TrainingData trainData = new TrainingData("..\\..\\..\\datasets\\parity8.train"))
using (TrainingData testData = new TrainingData("..\\..\\..\\datasets\\parity8.test"))
{
trainData.ScaleTrainData(-1, 1);
testData.ScaleTrainData(-1, 1);
Console.WriteLine("Creating network.");
using (NeuralNet net = new NeuralNet(NetworkType.SHORTCUT, 2, trainData.InputCount, trainData.OutputCount))
{
net.TrainingAlgorithm = training_algorithm;
net.ActivationFunctionHidden = ActivationFunction.SIGMOID_SYMMETRIC;
net.ActivationFunctionOutput = ActivationFunction.LINEAR;
net.TrainErrorFunction = ErrorFunction.ERRORFUNC_LINEAR;
if (multi == 0)
{
steepness[0] = 1;
net.CascadeActivationSteepnesses = steepness;
activation[0] = ActivationFunction.SIGMOID_SYMMETRIC;
net.CascadeActivationFunctions = activation;
net.CascadeCandidateGroupsCount = 8;
}
if (training_algorithm == TrainingAlgorithm.TRAIN_QUICKPROP)
{
net.LearningRate = 0.35F;
net.RandomizeWeights(-2.0F, 2.0F);
}
net.BitFailLimit = (DataType)0.9;
net.TrainStopFunction = StopFunction.STOPFUNC_BIT;
net.PrintParameters();
net.Save("..\\..\\..\\examples\\cascade_train2.net");
Console.WriteLine("Training network.");
net.CascadetrainOnData(trainData, max_neurons, neurons_between_reports, desired_error);
net.PrintConnections();
mse_train = net.TestData(trainData);
bit_fail_train = net.BitFail;
mse_test = net.TestData(testData);
bit_fail_test = net.BitFail;
Console.WriteLine("\nTrain error: {0}, Train bit-fail: {1}, Test error: {2}, Test bit-fail: {3}\n",
mse_train, bit_fail_train, mse_test, bit_fail_test);
for (int i = 0; i < trainData.TrainDataLength; i++)
{
output = net.Run(trainData.GetTrainInput((uint)i));
if ((trainData.GetTrainOutput((uint)i)[0] >= 0 && output[0] <= 0) ||
(trainData.GetTrainOutput((uint)i)[0] <= 0 && output[0] >= 0))
{
Console.WriteLine("ERROR: {0} does not match {1}", trainData.GetTrainOutput((uint)i)[0], output[0]);
}
}
Console.WriteLine("Saving network.");
net.Save("..\\..\\..\\examples\\cascade_train.net");
Console.ReadKey();
}
}
}
