| public GetLayerInput ( double input ) : double[][] | ||
| input | double | The batch of input data. |
| return | double[][] | |
public void ExampleTest1()
{
Accord.Math.Tools.SetupGenerator(0);
// We'll use a simple XOR function as input.
double[][] inputs =
{
new double[] { 0, 0 }, // 0 xor 0
new double[] { 0, 1 }, // 0 xor 1
new double[] { 1, 0 }, // 1 xor 0
new double[] { 1, 1 }, // 1 xor 1
};
// XOR output, corresponding to the input.
double[][] outputs =
{
new double[] { 0 }, // 0 xor 0 = 0
new double[] { 1 }, // 0 xor 1 = 1
new double[] { 1 }, // 1 xor 0 = 1
new double[] { 0 }, // 1 xor 1 = 0
};
// Setup the deep belief network (2 inputs, 3 hidden, 1 output)
DeepBeliefNetwork network = new DeepBeliefNetwork(2, 3, 1);
// Initialize the network with Gaussian weights
new GaussianWeights(network, 0.1).Randomize();
// Update the visible layer with the new weights
network.UpdateVisibleWeights();
// Setup the learning algorithm.
DeepBeliefNetworkLearning teacher = new DeepBeliefNetworkLearning(network)
{
Algorithm = (h, v, i) => new ContrastiveDivergenceLearning(h, v)
{
LearningRate = 0.1,
Momentum = 0.5,
Decay = 0.001,
}
};
// Unsupervised learning on each hidden layer, except for the output.
for (int i = 0; i < network.Layers.Length - 1; i++)
{
teacher.LayerIndex = i;
// Compute the learning data with should be used
var layerInput = teacher.GetLayerInput(inputs);
// Train the layer iteratively
for (int j = 0; j < 5000; j++)
teacher.RunEpoch(layerInput);
}
// Supervised learning on entire network, to provide output classification.
var backpropagation = new BackPropagationLearning(network)
{
LearningRate = 0.1,
Momentum = 0.5
};
// Run supervised learning.
for (int i = 0; i < 5000; i++)
backpropagation.RunEpoch(inputs, outputs);
// Test the resulting accuracy.
int correct = 0;
for (int i = 0; i < inputs.Length; i++)
{
double[] outputValues = network.Compute(inputs[i]);
double outputResult = outputValues.First() >= 0.5 ? 1 : 0;
if (outputResult == outputs[i].First())
{
correct++;
}
}
Assert.AreEqual(4, correct);
}
public void Train(double[][] i, double[][] o = null, int outputLength = 10, int hiddenLayer = -1)
{
if (n == null)
{
if (File.Exists(p)) n = DeepBeliefNetwork.Load(p);
else
{
outputLength = (o == null) ? outputLength : o[0].Length;
hiddenLayer = (hiddenLayer == -1) ? (int)Math.Log(i[0].Length, outputLength) : hiddenLayer;
List<int> layers = new List<int>();
for (int j = 0; j < hiddenLayer; j++) layers.Add(i[0].Length);
layers.Add(outputLength);
n = new DeepBeliefNetwork(new BernoulliFunction(), i[0].Length, layers.ToArray());
new GaussianWeights(n).Randomize();
}
}
dynamic t;
if (o == null)
{
t = new DeepBeliefNetworkLearning(n) { Algorithm = (h, v, j) => new ContrastiveDivergenceLearning(h, v), LayerIndex = n.Machines.Count - 1, };
while (true) e = t.RunEpoch(t.GetLayerInput(i));
}
else
{
t = new DeepNeuralNetworkLearning(n) { Algorithm = (ann, j) => new ParallelResilientBackpropagationLearning(ann), LayerIndex = n.Machines.Count - 1, };
while (true) e = t.RunEpoch(t.GetLayerInput(i), o);
}
}
private static DeepBeliefNetwork createNetwork(double[][] inputs)
{
DeepBeliefNetwork network = new DeepBeliefNetwork(6, 2, 1);
network.Machines[0].Hidden.Neurons[0].Weights[0] = 0.00461421;
network.Machines[0].Hidden.Neurons[0].Weights[1] = 0.04337112;
network.Machines[0].Hidden.Neurons[0].Weights[2] = -0.10839599;
network.Machines[0].Hidden.Neurons[0].Weights[3] = -0.06234004;
network.Machines[0].Hidden.Neurons[0].Weights[4] = -0.03017057;
network.Machines[0].Hidden.Neurons[0].Weights[5] = 0.09520391;
network.Machines[0].Hidden.Neurons[0].Threshold = 0;
network.Machines[0].Hidden.Neurons[1].Weights[0] = 0.08263872;
network.Machines[0].Hidden.Neurons[1].Weights[1] = -0.118437;
network.Machines[0].Hidden.Neurons[1].Weights[2] = -0.21710971;
network.Machines[0].Hidden.Neurons[1].Weights[3] = 0.02332903;
network.Machines[0].Hidden.Neurons[1].Weights[4] = 0.00953116;
network.Machines[0].Hidden.Neurons[1].Weights[5] = 0.09870652;
network.Machines[0].Hidden.Neurons[1].Threshold = 0;
network.Machines[0].Visible.Neurons[0].Threshold = 0;
network.Machines[0].Visible.Neurons[1].Threshold = 0;
network.Machines[0].Visible.Neurons[2].Threshold = 0;
network.Machines[0].Visible.Neurons[3].Threshold = 0;
network.Machines[0].Visible.Neurons[4].Threshold = 0;
network.Machines[0].Visible.Neurons[5].Threshold = 0;
network.UpdateVisibleWeights();
DeepBeliefNetworkLearning target = new DeepBeliefNetworkLearning(network)
{
Algorithm = (h, v, i) => new ContrastiveDivergenceLearning(h, v)
};
for (int layer = 0; layer < 2; layer++)
{
target.LayerIndex = layer;
double[][] layerInputs = target.GetLayerInput(inputs);
int iterations = 5000;
double[] errors = new double[iterations];
for (int i = 0; i < iterations; i++)
errors[i] = target.RunEpoch(layerInputs);
}
return network;
}
static void Main(string[] args)
{
double[][] inputs;
double[][] outputs;
double[][] testInputs;
double[][] testOutputs;
// Load ascii digits dataset.
inputs = DataManager.Load(@"../../../data/data.txt", out outputs);
// The first 500 data rows will be for training. The rest will be for testing.
testInputs = inputs.Skip(500).ToArray();
testOutputs = outputs.Skip(500).ToArray();
inputs = inputs.Take(500).ToArray();
outputs = outputs.Take(500).ToArray();
// Setup the deep belief network and initialize with random weights.
DeepBeliefNetwork network = new DeepBeliefNetwork(inputs.First().Length, 10, 10);
new GaussianWeights(network, 0.1).Randomize();
network.UpdateVisibleWeights();
// Setup the learning algorithm.
DeepBeliefNetworkLearning teacher = new DeepBeliefNetworkLearning(network)
{
Algorithm = (h, v, i) => new ContrastiveDivergenceLearning(h, v)
{
LearningRate = 0.1,
Momentum = 0.5,
Decay = 0.001,
}
};
// Setup batches of input for learning.
int batchCount = Math.Max(1, inputs.Length / 100);
// Create mini-batches to speed learning.
int[] groups = Accord.Statistics.Tools.RandomGroups(inputs.Length, batchCount);
double[][][] batches = inputs.Subgroups(groups);
// Learning data for the specified layer.
double[][][] layerData;
// Unsupervised learning on each hidden layer, except for the output layer.
for (int layerIndex = 0; layerIndex < network.Machines.Count - 1; layerIndex++)
{
teacher.LayerIndex = layerIndex;
layerData = teacher.GetLayerInput(batches);
for (int i = 0; i < 200; i++)
{
double error = teacher.RunEpoch(layerData) / inputs.Length;
if (i % 10 == 0)
{
Console.WriteLine(i + ", Error = " + error);
}
}
}
// Supervised learning on entire network, to provide output classification.
var teacher2 = new BackPropagationLearning(network)
{
LearningRate = 0.1,
Momentum = 0.5
};
// Run supervised learning.
for (int i = 0; i < 500; i++)
{
double error = teacher2.RunEpoch(inputs, outputs) / inputs.Length;
if (i % 10 == 0)
{
Console.WriteLine(i + ", Error = " + error);
}
}
// Test the resulting accuracy.
int correct = 0;
for (int i = 0; i < inputs.Length; i++)
{
double[] outputValues = network.Compute(testInputs[i]);
if (DataManager.FormatOutputResult(outputValues) == DataManager.FormatOutputResult(testOutputs[i]))
{
correct++;
}
}
Console.WriteLine("Correct " + correct + "/" + inputs.Length + ", " + Math.Round(((double)correct / (double)inputs.Length * 100), 2) + "%");
Console.Write("Press any key to quit ..");
Console.ReadKey();
}
private void learnLayerUnsupervised()
{
if (!Main.CanGenerate) return;
Dispatcher dispatcher = Dispatcher.CurrentDispatcher;
new Task(() =>
{
DeepBeliefNetworkLearning teacher = new DeepBeliefNetworkLearning(Main.Network)
{
Algorithm = (h, v, i) => new ContrastiveDivergenceLearning(h, v)
{
LearningRate = LearningRate,
Momentum = 0.5,
Decay = WeightDecay,
},
LayerIndex = SelectedLayerIndex - 1,
};
double[][] inputs;
Main.Database.Training.GetInstances(out inputs);
int batchCount = Math.Max(1, inputs.Length / BatchSize);
// Create mini-batches to speed learning
int[] groups = Accord.Statistics.Tools
.RandomGroups(inputs.Length, batchCount);
double[][][] batches = inputs.Subgroups(groups);
// Gather learning data for the layer
double[][][] layerData = teacher.GetLayerInput(batches);
var cd = teacher.GetLayerAlgorithm(teacher.LayerIndex) as ContrastiveDivergenceLearning;
// Start running the learning procedure
for (int i = 0; i < Epochs && !shouldStop; i++)
{
double error = teacher.RunEpoch(layerData) / inputs.Length;
dispatcher.BeginInvoke((Action<int, double>)updateError,
DispatcherPriority.ContextIdle, i + 1, error);
if (i == 10)
cd.Momentum = Momentum;
}
IsLearning = false;
}).Start();
}