public void TestLayersInitializers()
{
var layer = new Layer(3, new SigmoidFunction(), new UnbiasedUtility());
var network = new Network(2);
network.AddLayer(layer);
}
public void TestGetOutputTransform()
{
// var layer1 = new Layer(2, new SigmoidFunction(), new BiasedUtility());
var layer2 = new Layer(3, new SigmoidFunction(), new BiasedUtility());
var layer3 = new Layer(4, new SigmoidFunction(), new BiasedUtility());
var network = new Network(4);
// network.AddLayer(layer1);
network.AddLayer(layer2);
network.AddLayer(layer3);
// var input = new double[,] { { 2, 3, 1 }, { 4, 2, 1 } };
//layer1.WeightMatrix = Matrix<double>.Build.DenseOfArray(input);
//layer2.WeightMatrix = Matrix<double>.Build.DenseOfArray(new double[,] { { 1, 1, 1 } });
var trainingElems = new List <TrainingElement>();
trainingElems.Add(new TrainingElement(new double[, ] {
{ 1 }, { 0 }, { 0 }, { 0 }
}, new double[, ] {
{ 1 }, { 0 }, { 0 }, { 0 }
}));
trainingElems.Add(new TrainingElement(new double[, ] {
{ 0 }, { 1 }, { 0 }, { 0 }
}, new double[, ] {
{ 0 }, { 1 }, { 0 }, { 0 }
}));
trainingElems.Add(new TrainingElement(new double[, ] {
{ 0 }, { 0 }, { 1 }, { 0 }
}, new double[, ] {
{ 0 }, { 0 }, { 1 }, { 0 }
}));
trainingElems.Add(new TrainingElement(new double[, ] {
{ 0 }, { 0 }, { 0 }, { 1 }
}, new double[, ] {
{ 0 }, { 0 }, { 0 }, { 1 }
}));
network.Train(0.5, 1000, 0.85, trainingElems);
var kupa0 = network.ForwardPropagation(Matrix <double> .Build.DenseOfArray(new double[, ] {
{ 1 }, { 0 }, { 0 }, { 0 }
}));
Console.WriteLine($"5 300 1 1 {kupa0}");
var kupa1 = network.ForwardPropagation(Matrix <double> .Build.DenseOfArray(new double[, ] {
{ 0 }, { 1 }, { 0 }, { 0 }
}));
Console.WriteLine($"0 1 {kupa1}");
var kup2a = network.ForwardPropagation(Matrix <double> .Build.DenseOfArray(new double[, ] {
{ 0 }, { 0 }, { 1 }, { 0 }
}));
Console.WriteLine($"1 0 {kup2a}");
var kupa3 = network.ForwardPropagation(Matrix <double> .Build.DenseOfArray(new double[, ] {
{ 0 }, { 0 }, { 0 }, { 1 }
}));
Console.WriteLine($"0 0 {kupa3}");
}
