/// <summary>
/// Adds two matrices together
/// </summary>
/// <param name="inputs1">Matrix 1</param>
/// <param name="inputs2">Matrix 2</param>
public void Calculate(List <double[]> inputs1, List <double[]> inputs2)
{
ZVals = new List <double[]>();
if (inputs1.Count != inputs2.Count)
{
throw new Exception("List sizes do not match");
}
for (int b = 0; b < NN.BatchSize; b++)
{
if (inputs1[b].Length != inputs2[b].Length)
{
throw new Exception("Array sizes do not match");
}
double[] output = new double[inputs1[b].Length];
for (int i = 0; i < inputs1[b].Length; i++)
{
output[i] = inputs1[b][i] + inputs2[b][i];
}
ZVals.Add(output);
}
//If normalizing, do so, but only if it won't return an all-zero matrix
if (NN.NormOutputs && ZVals[0].Length > 1)
{
ZVals = Maths.Normalize(ZVals);
}
//Use the specified type of activation function
if (ActivationFunction == 0)
{
Values = Maths.Tanh(ZVals); return;
}
if (ActivationFunction == 1)
{
Values = Maths.ReLu(ZVals); return;
}
else
{
Values = ZVals;
}
}
public override void Calculate(List <double[]> inputs, bool output)
{
ZVals = new List <double[]>();
for (int i = 0; i < NN.BatchSize; i++)
{
ZVals.Add(Maths.Convert(Pool(Maths.Convert(inputs[i]), output)));
}
//If normalizing, do so, but only if it won't return an all-zero matrix
if (NN.NormOutputs && ZVals[0].Length > 1)
{
ZVals = Maths.Normalize(ZVals);
}
//Use the specified type of activation function
if (ActivationFunction == 0)
{
Values = Maths.Tanh(ZVals); return;
}
if (ActivationFunction == 1)
{
Values = Maths.ReLu(ZVals); return;
}
Values = ZVals;
}
/// <summary>
/// Calculates the dot product of the kernel and input matrix.
/// Matrices should be size [x, y] and [y], respectively, where x is the output size and y is the latent space's size
/// </summary>
/// <param name="inputs">The input matrix</param>
/// <param name="isoutput">Whether to use hyperbolic tangent on the output</param>
/// <returns></returns>
public override void Calculate(List <double[]> inputs, bool isoutput)
{
ZVals = new List <double[]>();
for (int b = 0; b < NN.BatchSize; b++)
{
ZVals.Add(Maths.Convert(DownOrUp ? Convolve(Weights, Pad(Maths.Convert(inputs[b]))) : FullConvolve(Weights, Pad(Maths.Convert(inputs[b])))));
}
//If normalizing, do so, but only if it won't return an all-zero matrix
if (NN.NormOutputs && ZVals[0].Length > 1)
{
ZVals = Maths.Normalize(ZVals);
}
//Use the specified type of activation function
if (ActivationFunction == 0)
{
Values = Maths.Tanh(ZVals); return;
}
if (ActivationFunction == 1)
{
Values = Maths.ReLu(ZVals); return;
}
Values = ZVals;
}
public override void Calculate(List <double[]> inputs, bool output)
{
ZVals = new List <double[]>();
for (int b = 0; b < NN.BatchSize; b++)
{
var vals = new double[Length];
for (int k = 0; k < Length; k++)
{
//Values = (weights * inputs) + biases
for (int j = 0; j < InputLength; j++)
{
vals[k] += Weights[k, j] * inputs[b][j];
}
//Output layers don't use biases
if (!output)
{
vals[k] += Biases[k];
}
}
ZVals.Add(vals);
}
//If normalizing, do so, but only if it won't return an all-zero matrix
if (NN.NormOutputs && ZVals[0].Length > 1)
{
ZVals = Maths.Normalize(ZVals);
}
//Use the specified type of activation function
if (ActivationFunction == 0)
{
Values = Maths.Tanh(ZVals); return;
}
if (ActivationFunction == 1)
{
Values = Maths.ReLu(ZVals); return;
}
Values = ZVals;
}
