Layer which contains the neurons of one layer. More...

#include <layer.h>

Public Types
enum	WeightNormalisation { WEIGHT_NORM_NONE = 0, WEIGHT_NORM_LAYER_EUCLEDIAN = 1, WEIGHT_NORM_NEURON_EUCLEDIAN = 2, WEIGHT_NORM_LAYER_MANHATTAN = 3, WEIGHT_NORM_NEURON_MANHATTAN = 4, WEIGHT_NORM_LAYER_INFINITY = 5, WEIGHT_NORM_NEURON_INFINITY = 6 }
	Weight normalisation constants Defines if weights are normalised layer-wide or for every neuron separately.

Public Member Functions
	Layer (int _nNeurons, int _nInputs)
	Constructor. More...

	~Layer ()
	Destructor Frees all memory. More...

void	calcOutputs ()
	Calculates the output values in all neurons. More...

void	doLearning ()
	Adjusts the weights. More...

void	setError (double _error)
	Sets the global error for all neurons. More...

void	setError (int i, double _error)
	sets the error individually More...

void	setErrors (double *_errors)
	Sets all errors from an input array. More...

double	getError (int i)
	Retrieves the error. More...

void	setBias (double _bias)
	Sets the global bias for all neurons. More...

void	setInput (int inputIndex, double input)
	Set the input value of one input. More...

void	setInputs (double *_inputs)
	Sets all inputs from an input array. More...

void	setLearningRate (double _learningRate)
	Sets the learning rate of all neurons. More...

void	setActivationFunction (Neuron::ActivationFunction _activationFunction)
	Set the activation function. More...

void	setMomentum (double _momentum)
	Set the momentum of all neurons in this layer. More...

void	setDecay (double _decay)
	Sets the weight decay scaled by the learning rate. More...

void	initWeights (double _max=1, int initBiasWeight=1, Neuron::WeightInitMethod weightInitMethod=Neuron::MAX_OUTPUT_RANDOM)
	Inits the weights. More...

double	getOutput (int index)
	Gets the outpuut of one neuron. More...

Neuron *	getNeuron (int index)
	Gets a pointer to one neuron. More...

int	getNneurons ()
	Gets the number of neurons. More...

int	getNinputs ()
	Number of inputs. More...

void	setConvolution (int width, int height)
	Defines a 2D geometry for the input layer of widthxheight. More...

void	setMaxDetLayer (int _m)
	Maxium detection layer. More...

void	setNormaliseWeights (WeightNormalisation _normaliseWeights)
	Normalise the weights. More...

void	setDebugInfo (int layerIndex)
	Sets the layer index within the whole network. More...

void	setStep (long int step)
	Sets the simulation step in the layer for debug purposes. More...

double	getWeightDistanceFromInitialWeights ()
	Get weight distance from the start of the simulation. More...

void	doNormaliseWeights ()
	Performs the weight normalisation. More...

void	setUseThreads (int _useThreads)
	Sets if threads should be used. More...

int	saveWeightMatrix (char *filename)
	Save weight matrix for documentation and debugging. More...

Detailed Description

Layer which contains the neurons of one layer.

It performs all computations possible in a layer. In particular it calls all neurons in separate threads and triggers the compuations there. These functions are all called from the parent class.

Constructor & Destructor Documentation

◆ Layer()

Layer::Layer	(	int	_nNeurons,
		int	_nInputs
	)

Constructor.

GNU GENERAL PUBLIC LICENSE Version 3, 29 June 2007.

Parameters

_nNeurons	Number of neurons in the layer.
_nInputs	Number of inputs to the Layer.
_nFilters	Number of lowpass filters at each input.
_minT	Minimum time of the lowpass filter.
_maxT	Maximum time of the lowpass filter.

◆ ~Layer()

Layer::~Layer ( )

Destructor Frees all memory.

Member Function Documentation

◆ calcOutputs()

void Layer::calcOutputs ( )

Calculates the output values in all neurons.

◆ doLearning()

void Layer::doLearning ( )

Adjusts the weights.

◆ doNormaliseWeights()

void Layer::doNormaliseWeights ( )

Performs the weight normalisation.

◆ getError()

double Layer::getError ( int i )

Retrieves the error.

Parameters

i	Index of the neuron

◆ getNeuron()

Neuron* Layer::getNeuron ( int index )

inline

Gets a pointer to one neuron.

Parameters

index The index number of the neuron.

Returns: A pointer to a Layer class.

◆ getNinputs()

int Layer::getNinputs ( )

inline

Number of inputs.

Returns: The number of inputs

◆ getNneurons()

int Layer::getNneurons ( )

inline

Gets the number of neurons.

Returns: The number of neurons.

◆ getOutput()

double Layer::getOutput ( int index )

inline

Gets the outpuut of one neuron.

Parameters

index The index number of the neuron.

Returns: Retuns the double valye of the output.

◆ getWeightDistanceFromInitialWeights()

double Layer::getWeightDistanceFromInitialWeights ( )

Get weight distance from the start of the simulation.

Returns: The distance from the initial (random) weight setup.

◆ initWeights()

void Layer::initWeights	(	double	_max = `1`,
		int	initBiasWeight = `1`,
		Neuron::WeightInitMethod	weightInitMethod = `Neuron::MAX_OUTPUT_RANDOM`
	)

Inits the weights.

Parameters

_max	Maximum value if using random init.
initBiasWeight	if one also the bias weight is initialised.
weightInitMethod	The methid employed to init the weights.

◆ saveWeightMatrix()

int Layer::saveWeightMatrix ( char * filename )

Save weight matrix for documentation and debugging.

Parameters

filename The filename it should be saved to.

◆ setActivationFunction()

void Layer::setActivationFunction ( Neuron::ActivationFunction _activationFunction )

Set the activation function.

Parameters

_activationFunction The activation function. See: Neuron::ActivationFunction

◆ setBias()

void Layer::setBias ( double _bias )

Sets the global bias for all neurons.

Parameters

_bias The bias for all neurons

◆ setConvolution()

void Layer::setConvolution	(	int	width,
		int	height
	)

Defines a 2D geometry for the input layer of widthxheight.

Parameters

width	The width of the convolutional window.
height	The height of the convolution window.

◆ setDebugInfo()

void Layer::setDebugInfo ( int layerIndex )

Sets the layer index within the whole network.

Parameters

layerIndex The layer index in the whole network.

◆ setDecay()

void Layer::setDecay ( double _decay )

Sets the weight decay scaled by the learning rate.

Parameters

_decay The decay rate of the weights

◆ setError() [1/2]

void Layer::setError ( double _error )

Sets the global error for all neurons.

Parameters

_error Sets the error in the whole layer

◆ setError() [2/2]

void Layer::setError	(	int	i,
		double	_error
	)

sets the error individually

Parameters

i	Index of the neuron
_error	The error to be set

◆ setErrors()

void Layer::setErrors ( double * _errors )

Sets all errors from an input array.

Parameters

_errors is an array of errors

◆ setInput()

void Layer::setInput	(	int	inputIndex,
		double	input
	)

Set the input value of one input.

Parameters

inputIndex	The index number of the input.
input	The value of the input

◆ setInputs()

void Layer::setInputs ( double * _inputs )

Sets all inputs from an input array.

Parameters

_inputs array of all inputs

◆ setLearningRate()

void Layer::setLearningRate ( double _learningRate )

Sets the learning rate of all neurons.

Parameters

_learningRate The learning rate

◆ setMaxDetLayer()

void Layer::setMaxDetLayer ( int _m )

inline

Maxium detection layer.

Experimental. This hasn't been implemented.

◆ setMomentum()

void Layer::setMomentum ( double _momentum )

Set the momentum of all neurons in this layer.

Parameters

_momentum The momentum for all neurons in this layer.

◆ setNormaliseWeights()

void Layer::setNormaliseWeights ( WeightNormalisation _normaliseWeights )

Normalise the weights.

Parameters

_normaliseWeights Metod of normalisation.

◆ setStep()

void Layer::setStep ( long int step )

Sets the simulation step in the layer for debug purposes.

Parameters

step	Step number.

◆ setUseThreads()

void Layer::setUseThreads ( int _useThreads )

inline

Sets if threads should be used.

Parameters

_useThreads 0 = no Threads, 1 = Threads

The documentation for this class was generated from the following files:

layer.h
layer.cpp

Public Types

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ Layer()

◆ ~Layer()

Member Function Documentation

◆ calcOutputs()

◆ doLearning()

◆ doNormaliseWeights()

◆ getError()

◆ getNeuron()

◆ getNinputs()

◆ getNneurons()

◆ getOutput()

◆ getWeightDistanceFromInitialWeights()

◆ initWeights()

◆ saveWeightMatrix()

◆ setActivationFunction()

◆ setBias()

◆ setConvolution()

◆ setDebugInfo()

◆ setDecay()

◆ setError() [1/2]

◆ setError() [2/2]

◆ setErrors()

◆ setInput()

◆ setInputs()

◆ setLearningRate()

◆ setMaxDetLayer()

◆ setMomentum()

◆ setNormaliseWeights()

◆ setStep()

◆ setUseThreads()