Layer Class Reference

A single layer within the neural net. More...

Inherited by ConvolutionalLayer, ForceBackpropLayer, FullyConnectedLayer, InputLayer, LossLayer, NormalizationLayer, PoolingLayer, RandomPatches, and RandomTranslations.

Public Member Functions
virtual publicapi float *	getResults ()=0
virtual publicapi int	getPersistSize () const =0
	Get the size of array needed for persisting to/from an array. More...
virtual publicapi int	getResultsSize () const =0
	Get the size of the activated output from this layer. More...
publicapi	Layer (Layer *previousLayer, LayerMaker2 *maker)
publicapi VIRTUAL void	setTraining (bool training)
	Are we training or predicting? Only affects the Random translations and patches layers currently. More...
publicapi VIRTUAL void	setBatchSize (int batchSize)
	used to set up internal buffers and stuff More...
publicapi VIRTUAL bool	getBiased () const
publicapi VIRTUAL bool	hasResultsWrapper () const
publicapi VIRTUAL CLWrapper *	getResultsWrapper ()
publicapi VIRTUAL ActivationFunction const *	getActivationFunction ()
publicapi VIRTUAL int	getOutputCubeSize () const
publicapi VIRTUAL int	getOutputPlanes () const
publicapi VIRTUAL int	getOutputImageSize () const
publicapi VIRTUAL void	backProp (float learningRate)
publicapi VIRTUAL int	getWeightsSize () const
publicapi VIRTUAL int	getBiasWeightsSize () const
publicapi VIRTUAL void	persistToArray (float *array)
	store the current weights and biases to array Note that you need to allocate array first More...
publicapi VIRTUAL void	unpersistFromArray (float const *array)
	initialize the current weights and biases from array More...
publicapi VIRTUAL std::string	asString () const
	Get a string representation of the layer. More...

Detailed Description

A single layer within the neural net.

Public API

Constructor & Destructor Documentation

publicapi Layer::Layer	(	Layer *	previousLayer,
		LayerMaker2 *	maker
	)

Public API

Member Function Documentation

virtual publicapi float* Layer::getResults ( )

pure virtual

Public API

virtual publicapi int Layer::getPersistSize ( ) const

pure virtual

Get the size of array needed for persisting to/from an array.

Public API

virtual publicapi int Layer::getResultsSize ( ) const

pure virtual

Get the size of the activated output from this layer.

Public API

publicapi VIRTUAL void Layer::setTraining

(

bool

training

)

Are we training or predicting? Only affects the Random translations and patches layers currently.

Public API

publicapi VIRTUAL void Layer::setBatchSize

(

int

batchSize

)

used to set up internal buffers and stuff

Public API

publicapi VIRTUAL bool Layer::getBiased

(

)

const

Public API

publicapi VIRTUAL bool Layer::hasResultsWrapper

(

)

const

Public API

publicapi VIRTUAL CLWrapper * Layer::getResultsWrapper

(

)

Public API

publicapi VIRTUAL ActivationFunction const * Layer::getActivationFunction

(

)

Public API

publicapi VIRTUAL int Layer::getOutputCubeSize

(

)

const

Public API

publicapi VIRTUAL int Layer::getOutputPlanes

(

)

const

Public API

publicapi VIRTUAL int Layer::getOutputImageSize

(

)

const

Public API

publicapi VIRTUAL void Layer::backProp

(

float

learningRate

)

Public API

publicapi VIRTUAL int Layer::getWeightsSize

(

)

const

Public API

publicapi VIRTUAL int Layer::getBiasWeightsSize

(

)

const

Public API

publicapi VIRTUAL void Layer::persistToArray

(

float *

array

)

store the current weights and biases to array Note that you need to allocate array first

Public API

publicapi VIRTUAL void Layer::unpersistFromArray

(

float const *

array

)

initialize the current weights and biases from array

Public API

publicapi VIRTUAL std::string Layer::asString

(

)

const

Get a string representation of the layer.

Public API

---

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

• Layer.h
• Layer.cpp