public Dataset Generate(int length)
{
var dataArray = new Data[length];
for (int i = 0; i < length; i++)
{
var input = new ShapedArray <double>(InputShape, () => GRandom.Uniform() < 0.5 ? 0.0 : 1.0);
int oneCount = 0;
input.ForEach(X => oneCount += X == 1.0 ? 1 : 0);
bool isEven = oneCount % 2 == 0;
double[] outArr;
if (IsBinary)
{
outArr = isEven ? new double[] { 1.0, 0.0 } : new double[] { 0.0, 1.0 };
}
else
{
outArr = isEven ? new double[] { 1.0 } : new double[] { 0.0 };
}
dataArray[i] = new Data(input, ShapedArray <double> .FromRef(TargetShape, outArr));
}
return(Dataset.FromRef(dataArray));
}
public override void Connect(Layer inLayer)
{
if (inLayer.Shape != InputShape)
{
throw new ShapeMismatchException(nameof(inLayer));
}
ShapedArray <Neuron> padded = Padder.PadArray(inLayer.Neurons.ToShape(Shape), Paddings, () => new Neuron()
{
OutVal = PadVal
});
var inConnections = new ShapedArray <List <Synapse> >(PaddedShape, () => new List <Synapse>());
IndexGen.ByStrides(PaddedShape, Strides, KernelShape).ForEach((idxKernel, i) =>
{
Neuron outN = base.Neurons[i];
outN.InSynapses = IndexGen.ByStart(KernelShape, Array <int> .FromRef(idxKernel)).Select((idx, j) =>
{
var S = new Synapse(padded[idx], outN);
inConnections[idx].Add(S);
return(S);
});
});
padded.ForEach((N, i) => N.OutSynapses = Array <Synapse> .FromRef(inConnections[i].ToArray()));
}
public Dataset Generate(int length)
{
var dataArray = new Data[length];
for (int i = 0; i < length; i++)
{
var arr = new ShapedArray <double>(InputShape, () => GRandom.Uniform() < 0.5 ? 0.0 : 1.0);
dataArray[i] = new Data(arr, arr);
}
return(Dataset.FromRef(dataArray));
}
public static ShapedArray <T> PadArray <T>(ShapedArray <T> array, Array <int> paddings, Func <T> padVal)
{
if (array.Shape.Rank != paddings.Length)
{
throw new RankException(nameof(array));
}
Shape paddedShape = array.Shape.Pad(paddings);
var internalArray = new T[paddedShape.Volume];
IndexGen.ByStart(array.Shape, paddings).ForEach((idx, i) => internalArray[paddedShape.FlattenIndices(idx)] = array[i]);
for (int i = 0; i < internalArray.Length; i++)
{
internalArray[i] ??= padVal();
}
return(ShapedArray <T> .FromRef(paddedShape, internalArray));
}
public override void Connect(Layer inLayer)
{
if (inLayer.Shape != InputShape)
{
throw new ShapeMismatchException(nameof(inLayer));
}
ShapedArray <Neuron> padded = Padder.PadArray(inLayer.Neurons.ToShape(Shape), Paddings, () => new Neuron()
{
OutVal = PadVal
});
var inConnections = new ShapedArray <List <Synapse> >(PaddedShape, () => new List <Synapse>());
Kernels.ForEach((kernel, i) =>
{
int offset = i * Shape.Volume / KernelsNum;
IndexGen.ByStrides(PaddedShape, Strides, KernelShape).ForEach((idxKernel, j) =>
{
var outN = (CNeuron)Neurons[offset + j];
outN.KernelBias = kernel.Bias;
outN.InSynapses = IndexGen.ByStart(KernelShape, Array <int> .FromRef(idxKernel)).Select((idx, k) =>
{
var S = new CSynapse(padded[idx], outN)
{
KernelWeight = kernel.Weights[k]
};
inConnections[idx].Add(S);
return((Synapse)S);
});
});
});
padded.ForEach((N, i) => N.OutSynapses = Array <Synapse> .FromRef(inConnections[i].ToArray()));
}
protected override void CalcGrads(ILoss loss, ShapedArray <double> targets)
{
OutputNode.CalcGrads(loss, targets);
}
protected override void Forward(ShapedArray <double> inputs, bool isTraining)
{
InputNode.Forward(inputs, isTraining);
}
