backprop: momentums + decay, quickprop: renaming
This commit is contained in:
@@ -9,9 +9,11 @@ void NeuralNetwork::Learning::BackPropagation::teach(const std::vector<float> &i
|
||||
|
||||
resize();
|
||||
|
||||
computeDeltas(expectation);
|
||||
computeSlopes(expectation);
|
||||
|
||||
updateWeights(input);
|
||||
|
||||
std::swap(deltas,lastDeltas);
|
||||
}
|
||||
|
||||
|
||||
@@ -28,21 +30,25 @@ void NeuralNetwork::Learning::BackPropagation::updateWeights(const std::vector<f
|
||||
|
||||
float delta =slopes[layerIndex][j]*learningCoefficient;
|
||||
|
||||
layer[j].weight(0)+=delta;
|
||||
//momentum
|
||||
delta += momentumWeight * lastDeltas[layerIndex][j];
|
||||
|
||||
deltas[layerIndex][j]=delta;
|
||||
|
||||
layer[j].weight(0)+=delta - weightDecay *layer[j].weight(0);
|
||||
|
||||
for(std::size_t k=1;k<prevLayerSize;k++) {
|
||||
if(layerIndex==1) {
|
||||
layer[j].weight(k)+=delta*input[k-1];
|
||||
layer[j].weight(k)+=delta*input[k-1] - weightDecay * layer[j].weight(k);
|
||||
} else {
|
||||
layer[j].weight(k)+=delta*prevLayer[k].output();
|
||||
layer[j].weight(k)+=delta*prevLayer[k].output() - weightDecay * layer[j].weight(k);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
void NeuralNetwork::Learning::BackPropagation::computeDeltas(const std::vector<float> &expectation) {
|
||||
void NeuralNetwork::Learning::BackPropagation::computeSlopes(const std::vector<float> &expectation) {
|
||||
auto& outputLayer=network[network.size()-1];
|
||||
for(std::size_t j=1;j<outputLayer.size();j++) {
|
||||
auto& neuron = outputLayer[j];
|
||||
|
||||
@@ -18,17 +18,17 @@ void NeuralNetwork::Learning::QuickPropagation::updateWeights(const std::vector<
|
||||
|
||||
float newChange=0;
|
||||
|
||||
if(fabs (lastWeightChange[layerIndex][j])> 0.0001) {
|
||||
if(std::signbit(lastWeightChange[layerIndex][j]) == std::signbit(slopes[layerIndex][j])) {
|
||||
if(fabs (deltas[layerIndex][j])> 0.0001) {
|
||||
if(std::signbit(deltas[layerIndex][j]) == std::signbit(slopes[layerIndex][j])) {
|
||||
newChange+= slopes[layerIndex][j]*_epsilon;
|
||||
|
||||
if(fabs(slopes[layerIndex][j]) > fabs(shrinkFactor * previousSlopes[layerIndex][j])) {
|
||||
newChange += _maxChange * lastWeightChange[layerIndex][j];
|
||||
newChange += _maxChange * deltas[layerIndex][j];
|
||||
}else {
|
||||
newChange+=slopes[layerIndex][j]/(previousSlopes[layerIndex][j]-slopes[layerIndex][j]) * lastWeightChange[layerIndex][j];
|
||||
newChange+=slopes[layerIndex][j]/(previousSlopes[layerIndex][j]-slopes[layerIndex][j]) * deltas[layerIndex][j];
|
||||
}
|
||||
} else {
|
||||
newChange+=slopes[layerIndex][j]/(previousSlopes[layerIndex][j]-slopes[layerIndex][j]) * lastWeightChange[layerIndex][j];
|
||||
newChange+=slopes[layerIndex][j]/(previousSlopes[layerIndex][j]-slopes[layerIndex][j]) * deltas[layerIndex][j];
|
||||
}
|
||||
} else {
|
||||
newChange+= slopes[layerIndex][j]*_epsilon;
|
||||
@@ -49,5 +49,5 @@ void NeuralNetwork::Learning::QuickPropagation::updateWeights(const std::vector<
|
||||
}
|
||||
|
||||
slopes.swap(previousSlopes);
|
||||
weightChange.swap(lastWeightChange);
|
||||
weightChange.swap(deltas);
|
||||
}
|
||||
Reference in New Issue
Block a user