GTK+ IOStream  Beta
<< GTK+ >> add C++ IOStream operators to GTK+. Now with extra abilities ... like network serialisation
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Pages
Public Member Functions | Public Attributes | Protected Attributes | List of all members
NeuralLayer< TYPE > Class Template Reference

#include <NeuralNetwork.H>

Inheritance diagram for NeuralLayer< TYPE >:
Inheritance graph
[legend]
Collaboration diagram for NeuralLayer< TYPE >:
Collaboration graph
[legend]

Public Member Functions

 NeuralLayer (int inputSize, int outputSize)
 
template<typename Derived >
 NeuralLayer (const Eigen::MatrixBase< Derived > &weightsIn, const Eigen::MatrixBase< Derived > &biasIn)
 
virtual ~NeuralLayer (void)
 Destructor. More...
 
virtual Eigen::Matrix< TYPE,
Eigen::Dynamic, 1 > & 
activate (const Eigen::Matrix< TYPE, Eigen::Dynamic, 1 > &input)
 

Public Attributes

Eigen::Matrix< TYPE,
Eigen::Dynamic, 1 > 
output
 The output from this layer. More...
 

Protected Attributes

Eigen::Matrix< TYPE,
Eigen::Dynamic, Eigen::Dynamic > 
weights
 The neural weights for this layer. More...
 
Eigen::Matrix< TYPE,
Eigen::Dynamic, 1 > 
bias
 The biases for this layer. More...
 

Detailed Description

template<typename TYPE>
class NeuralLayer< TYPE >

Implements a single neural layer
Template Parameters
TYPEthe precision of the data to use, e.g. float, double

Definition at line 28 of file NeuralNetwork.H.

Constructor & Destructor Documentation

template<typename TYPE >
NeuralLayer< TYPE >::NeuralLayer ( int  inputSize,
int  outputSize 
)
inline
Generate a neural layer of particular size
Parameters
inputSizeThe number of the inputs
outputSizeThe number of outputs

Definition at line 40 of file NeuralNetwork.H.

template<typename TYPE >
template<typename Derived >
NeuralLayer< TYPE >::NeuralLayer ( const Eigen::MatrixBase< Derived > &  weightsIn,
const Eigen::MatrixBase< Derived > &  biasIn 
)
inline
Generate a neural layer of particular size providing the weights
Parameters
weightsInThe weights to set
biasInThe biases to set
Template Parameters
Derivedis used by Eigen's Curiously recurring template pattern (CRTP)

Definition at line 52 of file NeuralNetwork.H.

template<typename TYPE >
virtual NeuralLayer< TYPE >::~NeuralLayer ( void  )
inlinevirtual

Destructor.

Definition at line 59 of file NeuralNetwork.H.

Member Function Documentation

template<typename TYPE >
virtual Eigen::Matrix<TYPE, Eigen::Dynamic, 1>& NeuralLayer< TYPE >::activate ( const Eigen::Matrix< TYPE, Eigen::Dynamic, 1 > &  input)
inlinevirtual
The activation function

This evaluates the neural network

Parameters
inputThe input to this layer
Returns
The result of the layer after processing the input

Reimplemented in TanhLayer< TYPE >, and SigmoidLayer< TYPE >.

Definition at line 66 of file NeuralNetwork.H.

Here is the caller graph for this function:

Member Data Documentation

template<typename TYPE >
Eigen::Matrix<TYPE, Eigen::Dynamic, 1> NeuralLayer< TYPE >::bias
protected

The biases for this layer.

Definition at line 31 of file NeuralNetwork.H.

template<typename TYPE >
Eigen::Matrix<TYPE, Eigen::Dynamic, 1> NeuralLayer< TYPE >::output

The output from this layer.

Definition at line 34 of file NeuralNetwork.H.

template<typename TYPE >
Eigen::Matrix<TYPE, Eigen::Dynamic, Eigen::Dynamic> NeuralLayer< TYPE >::weights
protected

The neural weights for this layer.

Definition at line 30 of file NeuralNetwork.H.


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