public FlattenLayerBackprop(FlattenLayer unit)
{
_unit = unit;
_outgoing = new Matrix <float> [unit.Z];
for (int i = 0; i < _outgoing.Length; i++)
{
_outgoing[i] = Matrix <float> .Build.Dense(unit.X, unit.Y);
}
}
public Vector <float> Compute(StatePair input, bool training)
{
// Forward propagate.
var img = InputLayer.Compute(input.Spatial);
for (int i = 0; i < ConvolutionalLayers.Length; i++)
{
img = SubSampleLayers[i].Compute(ConvolutionalLayers[i].Compute(img));
}
_vecp.left = FlattenLayer.Compute(img);
_vecp.right = LinearHiddenLayer.Compute(VectorInput.Compute(input.Linear));
IsOutputFromTraining = false;
var res = OutputLayer.Compute(CombinationLayer.Compute(_vecp));
if (ValuesComputed != null)
{
ValuesComputed(res, training);
}
return(res);
}
public ConvolutionalNetwork(int matsize, int vecsize, int depth, int labels, params CNNArgs[] args)
{
_matsize = matsize;
_vecsize = vecsize;
_depth = depth;
_labels = labels;
_args = args;
InputLayer = new SpatialLayer(matsize, depth);
ConvolutionalLayers = new ConvolutionalLayer[args.Length];
SubSampleLayers = new MeanPoolLayer[args.Length];
ConvolutionalLayers[0] = new ConvolutionalLayer(args[0].FilterSize, args[0].FilterCount, args[0].Stride, InputLayer, Functions.Rectifier2D);
SubSampleLayers[0] = new MeanPoolLayer(args[0].PoolLayerSize, ConvolutionalLayers[0]);
for (int i = 1; i < args.Length; i++)
{
ConvolutionalLayers[i] = new ConvolutionalLayer(args[i].FilterSize, args[i].FilterCount, args[i].Stride, SubSampleLayers[i - 1], Functions.Rectifier2D);
SubSampleLayers[i] = new MeanPoolLayer(args[i].PoolLayerSize, ConvolutionalLayers[i]);
}
FlattenLayer = new FlattenLayer(SubSampleLayers[SubSampleLayers.Length - 1]);
VectorInput = new InputLayer(vecsize);
LinearHiddenLayer = new DenseLayer(vecsize, VectorInput, Functions.Sigmoid);
CombinationLayer = new TreeLayer(FlattenLayer.Size(), vecsize);
OutputLayer = new DenseLayer(labels, CombinationLayer, Functions.Identity);
}