Feedforward Closedloop Learning
|
Derived classes of the FeedforwardClosedloopLearning class for special functionality. More...
#include <fcl_util.h>
Public Member Functions | |
FeedforwardClosedloopLearningWithFilterbank (int num_of_inputs, int *num_of_neurons_per_layer_array, int num_layers, int num_filtersInput, double minT, double maxT) | |
FeedforwardClosedloopLearning with Filterbank at each input. More... | |
~FeedforwardClosedloopLearningWithFilterbank () | |
Destructor. | |
void | doStep (double *input, double *error) |
Performs the simulation step. More... | |
void | doStep (double *input, int n1, double *error, int n2) |
Python wrapper function. More... | |
double | getFilterOutput (int inputIdx, int filterIdx) |
int | getNFiltersPerInput () |
![]() | |
FeedforwardClosedloopLearning (int num_of_inputs, int *num_of_neurons_per_layer_array, int _num_layers) | |
Constructor: FCL without any filters. More... | |
~FeedforwardClosedloopLearning () | |
Destructor De-allocated any memory. More... | |
void | doStep (double *input, double *error) |
Performs the simulation step. More... | |
void | doStep (double *input, int n1, double *error, int n2) |
Python wrapper function. More... | |
double | getOutput (int index) |
Gets the output from one of the output neurons. More... | |
void | setLearningRate (double learningRate) |
Sets globally the learning rate. More... | |
void | setLearningRateDiscountFactor (double _learningRateDiscountFactor) |
Sets how the learnign rate increases or decreases from layer to layer. More... | |
void | setDecay (double decay) |
Sets a typical weight decay scaled with the learning rate. More... | |
void | setMomentum (double momentum) |
Sets the global momentum for all layers. More... | |
void | setActivationFunction (Neuron::ActivationFunction _activationFunction) |
Sets the activation function of the Neuron. More... | |
void | initWeights (double max=0.001, int initBias=1, Neuron::WeightInitMethod weightInitMethod=Neuron::MAX_OUTPUT_RANDOM) |
Inits the weights in all layers. More... | |
void | seedRandom (int s) |
Seeds the random number generator. More... | |
void | setBias (double _bias) |
Sets globally the bias. More... | |
int | getNumLayers () |
Gets the total number of layers. More... | |
Layer * | getLayer (int i) |
Gets a pointer to a layer. More... | |
Layer * | getOutputLayer () |
Gets the output layer. More... | |
int | getNumInputs () |
Gets the number of inputs. More... | |
Layer ** | getLayers () |
Returns all Layers. More... | |
bool | saveModel (const char *name) |
Saves the whole network. More... | |
bool | loadModel (const char *name) |
Loads the while network. More... | |
Derived classes of the FeedforwardClosedloopLearning class for special functionality.
FeedforwardClosedloopLearningWithFilterbank::FeedforwardClosedloopLearningWithFilterbank | ( | int | num_of_inputs, |
int * | num_of_neurons_per_layer_array, | ||
int | num_layers, | ||
int | num_filtersInput, | ||
double | minT, | ||
double | maxT | ||
) |
FeedforwardClosedloopLearning with Filterbank at each input.
Constructor: FCL with a filter bank at the input Every input feeds internally into a has a filter bank of num_filtersInput filters. This allows for a temporal distribution of the inputs.
num_of_inputs | Number of inputs in the input layer |
num_of_neurons_per_layer_array | Number of neurons in each layer |
_num_layers | Number of layer (needs to match with array above) |
num_filtersInput | Number of filters at the input layer, 0 = no filterbank |
num_filters | Number of filters in the hiddel layers (usually zero) |
_minT | Minimum/first temporal duration of the 1st filter |
_maxT | Maximum/last temporal duration of the last filter |
void FeedforwardClosedloopLearningWithFilterbank::doStep | ( | double * | input, |
double * | error | ||
) |
Performs the simulation step.
input | Array with the input values |
error | Array of the error signals |
void FeedforwardClosedloopLearningWithFilterbank::doStep | ( | double * | input, |
int | n1, | ||
double * | error, | ||
int | n2 | ||
) |
Python wrapper function.
Not public.