private void applyStep(GradStep step, DerivedLayer layer)
{
for (int i = 0; i < layer.weights.RowCount; i++) //iterate first by column because every column represents a neuron
{
for (int j = 0; j < layer.weights.ColumnCount; j++) //each row represents the jth weight for each neuron
{
layer.weights[i, j] -= learningFactor * step.weightStep[i, j]; //change each weight by its negative gradient
}
layer.biases[i] -= learningFactor * step.biasStep[i]; //change the bias by its negative gradient, one for each neuron in the layer
}
}
/// <summary>
/// Creates a network that takes inputSize neurons
/// </summary>
/// <param name="inputSize"></param>
/// <param name="hiddenSizes"></param>
/// <param name="outputSize"></param>
public Network(int inputSize, int[] hiddenSizes, int outputSize)
{
input = new Layer(new double[inputSize]); //input layer is your data
if (hiddenSizes.Count() > 0)
{
hiddenLayers = new DerivedLayer[hiddenSizes.Count()]; //where the stuff happens
hiddenLayers[0] = new DerivedLayer(input, hiddenSizes[0]); //first hidden layer pulls from input
for (int i = 1; i < hiddenSizes.Count(); i++)
{
hiddenLayers[i] = new DerivedLayer(hiddenLayers[i - 1], hiddenSizes[i]); //then the next hidden layers pull from the hidden layer before them
}
output = new DerivedLayer(hiddenLayers[hiddenLayers.GetUpperBound(0)], outputSize); //the output is derived from the last hidden layer
}
else
{
output = new DerivedLayer(input, outputSize); //directly from input to output (probably sucks)
}
}