shark::SoftmaxNeuron< VectorType > Struct Template Reference

Computes the softmax activation function. More...

#include <shark/Models/NeuronLayers.h>

Public Types
typedef EmptyState	State

Public Member Functions
template<class Arg , class Device >
void	evalInPlace (blas::vector_expression< Arg, Device > &arg) const
template<class Arg , class Device >
void	evalInPlace (blas::matrix_expression< Arg, Device > &arg) const
template<class Arg , class Device >
void	evalInPlace (blas::matrix_expression< Arg, Device > &arg, State &) const
template<class Output , class Derivative >
void	multiplyDerivative (Output const &output, Derivative &der, State const &s) const

Detailed Description

template<class VectorType = RealVector>
struct shark::SoftmaxNeuron< VectorType >

Computes the softmax activation function.

\[ f_i(x)= \exp(x_i) \ \sum_j \exp(x_j) \]

computes the exponential function of the inputs and normalizes the outputs to sum to one. This is the NormalizerNeuron just without the constraint of values being positive

Definition at line 218 of file NeuronLayers.h.

Member Typedef Documentation

State

template<class VectorType = RealVector>

typedef EmptyState shark::SoftmaxNeuron< VectorType >::State

Definition at line 219 of file NeuronLayers.h.

Member Function Documentation

evalInPlace() [1/3]

template<class VectorType = RealVector>

template<class Arg , class Device >

void shark::SoftmaxNeuron< VectorType >::evalInPlace ( blas::matrix_expression< Arg, Device > &  arg ) const

inline

Definition at line 228 of file NeuronLayers.h.

evalInPlace() [2/3]

template<class VectorType = RealVector>

template<class Arg , class Device >

void shark::SoftmaxNeuron< VectorType >::evalInPlace ( blas::matrix_expression< Arg, Device > &  arg, State &    ) const

inline

Definition at line 234 of file NeuronLayers.h.

References shark::SoftmaxNeuron< VectorType >::evalInPlace().

evalInPlace() [3/3]

template<class VectorType = RealVector>

template<class Arg , class Device >

void shark::SoftmaxNeuron< VectorType >::evalInPlace ( blas::vector_expression< Arg, Device > &  arg ) const

inline

Definition at line 222 of file NeuronLayers.h.

Referenced by shark::SoftmaxNeuron< VectorType >::evalInPlace().

multiplyDerivative()

template<class VectorType = RealVector>

template<class Output , class Derivative >

void shark::SoftmaxNeuron< VectorType >::multiplyDerivative ( Output const &  output, Derivative &  der, State const &  s  ) const

inline

Definition at line 239 of file NeuronLayers.h.

---

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

• include/shark/Models/NeuronLayers.h