Implemented dynamic learning

Layer.cpp

@@ -53,7 +53,7 @@ void Layer::connectTo(const Layer & nextLayer)
 
 void Layer::updateInputWeights(Layer & prevLayer)
 {
-	static const double trainingRate = 0.5;
+	static const double trainingRate = 0.2;
 
 	for (size_t currentLayerIndex = 0; currentLayerIndex < sizeWithoutBiasNeuron(); ++currentLayerIndex)
 	{

Net.cpp

@@ -2,9 +2,9 @@
 
 Net::Net(std::initializer_list<size_t> layerSizes)
 {
-	if (layerSizes.size() < 3)
+	if (layerSizes.size() < 2)
 	{
-		throw std::exception("A net needs at least 3 layers");
+		throw std::exception("A net needs at least 2 layers");
 	}
 
 	for (size_t numNeurons : layerSizes)

Neuro.cpp

@@ -9,22 +9,27 @@ int main()
 	{
 		std::cout << "Neuro running" << std::endl;
 
-		std::vector<double> inputValues = { 0.1, 0.2, 0.8 };
+		Net myNet({ 3, 2, 1 });
-		std::vector<double> targetValues = { 0.8 };
 
-		Net myNet({ inputValues.size(), 4, targetValues.size() });
+		for (int i = 0; i < 100000; ++i)
-
-		for (int i = 0; i < 200; ++i)
 		{
+			std::vector<double> inputValues =
+			{
+				std::rand() / (double)RAND_MAX,
+				std::rand() / (double)RAND_MAX,
+				std::rand() / (double)RAND_MAX
+			};
+
+			std::vector<double> targetValues = { inputValues[2] };
+
 			myNet.feedForward(inputValues);
 
 			std::vector<double> outputValues = myNet.getOutput();
 
-			std::cout << "Result: ";
+			double error = outputValues[0] - targetValues[0];
-			for (double &value : outputValues)
+
-			{
+			std::cout << "Error: ";
-				std::cout << value << " ";
+			std::cout << std::abs(error);
-			}
 			std::cout << std::endl;
 
 			myNet.backProp(targetValues);