public void Apply(NeuralTrainSet set, float learningRate = 0.05f)
{
for (int i = 0; i < Layers.Length; i++)
{
Layers[i].Apply(set.Layers[i], learningRate);
}
}
public NeuralTrainSet(NeuralTrainSet a, NeuralTrainSet b)
{
if (a.NeuralNetwork != b.NeuralNetwork)
{
throw new Exception("Cannot blend sets that belong to a different network.");
}
NeuralNetwork = a.NeuralNetwork;
int layerCount = a.Layers.Length;
Layers = new NeuralTrainLayer[layerCount];
for (int i = 0; i < layerCount; i++)
{
Layers[i] = new NeuralTrainLayer(a.Layers[i], b.Layers[i]);
}
}
public void Apply(IEnumerable <NeuralTrainSet> trainingSets, float learningRate = 0.05f)
{
NeuralTrainSet combination = null;
foreach (NeuralTrainSet set in trainingSets)
{
if (combination == null)
{
combination = set;
}
else
{
combination = new NeuralTrainSet(combination, set);
}
}
Apply(combination, learningRate);
}
public void Train(Sample[] samples, int miniBatchCount, bool parallel, bool shuffle = true)
{
if (shuffle)
{
samples = General.Shuffle(samples);
}
int end = samples.Length;
NeuralTrainSet[] miniBatch;
for (int i = 0; i < end;)
{
if (StopTraining)
{
return;
}
int localEnd = i + miniBatchCount;
if (localEnd > end)
{
localEnd = end;
}
int size = localEnd - i;
miniBatch = new NeuralTrainSet[size];
if (parallel)
{
Parallel.For(0, size, j => {
miniBatch[j] = GetNeuralTrainSet(samples[j + i]);
});
}
else
{
for (int j = 0; j < size; j++)
{
miniBatch[j] = GetNeuralTrainSet(samples[j + i]);
}
}
Apply(miniBatch, LearningRate);
i = localEnd;
}
}
