org.deeplearning4j.optimize.api

Interface ConvexOptimizer

All Superinterfaces:

Serializable

All Known Implementing Classes:

BaseOptimizer, ConjugateGradient, GradientAscent, IterationGradientDescent, LBFGS, StochasticHessianFree

public interface ConvexOptimizer
extends Serializable

Convex optimizer.

Author:

Adam Gibson

Method Summary

Modifier and Type	Method and Description
Map<String,org.nd4j.linalg.learning.AdaGrad>	adaGradForVariables() Return the ada grad look up table
int	batchSize() The batch size for the optimizer
org.nd4j.linalg.learning.AdaGrad	getAdaGrad() The adagrad in this model
org.nd4j.linalg.learning.AdaGrad	getAdaGradForVariable(String variable) Get adagrad for a variable
Pair<Gradient,Double>	gradientAndScore() The gradient and score for this optimizer
boolean	optimize() Calls optimize
void	postStep() After the step has been made, do an action
void	preProcessLine(org.nd4j.linalg.api.ndarray.INDArray line) Pre process a line before an iteration
double	score() The score for the optimizer so far
void	setBatchSize(int batchSize) Set the batch size for the optimizer
void	setupSearchState(Pair<Gradient,Double> pair) Based on the gradient and score setup a search state
void	updateGradientAccordingToParams(Gradient gradient, Model params, int batchSize) Update the gradient according to the configuration such as adagrad, momentum, and sparsity
void	updateGradientAccordingToParams(org.nd4j.linalg.api.ndarray.INDArray gradient, org.nd4j.linalg.api.ndarray.INDArray params, int batchSize) Update the gradient according to the configuration such as adagrad, momentum, and sparsity

Method Detail

score

double score()

The score for the optimizer so far

Returns:

the score for this optimizer so far

gradientAndScore

Pair<Gradient,Double> gradientAndScore()

The gradient and score for this optimizer

Returns:

the gradient and score for this optimizer

optimize

boolean optimize()

Calls optimize

Returns:

whether the convex optimizer converted or not

batchSize

int batchSize()

The batch size for the optimizer

Returns:

setBatchSize

void setBatchSize(int batchSize)

Set the batch size for the optimizer

Parameters:

batchSize -

preProcessLine

void preProcessLine(org.nd4j.linalg.api.ndarray.INDArray line)

Pre process a line before an iteration

Parameters:

line -

postStep

void postStep()

After the step has been made, do an action

setupSearchState

void setupSearchState(Pair<Gradient,Double> pair)

Based on the gradient and score setup a search state

Parameters:

pair - the gradient and score

getAdaGrad

org.nd4j.linalg.learning.AdaGrad getAdaGrad()

The adagrad in this model

Returns:

the adagrad in this model

adaGradForVariables

Map<String,org.nd4j.linalg.learning.AdaGrad> adaGradForVariables()

Return the ada grad look up table

Returns:

the ada grad for variables

getAdaGradForVariable

org.nd4j.linalg.learning.AdaGrad getAdaGradForVariable(String variable)

Get adagrad for a variable

Parameters:

variable -

Returns:

updateGradientAccordingToParams

void updateGradientAccordingToParams(Gradient gradient,
                                     Model params,
                                     int batchSize)

Update the gradient according to the configuration such as adagrad, momentum, and sparsity

Parameters:

gradient - the gradient to modify

params - the parameters to update

updateGradientAccordingToParams

void updateGradientAccordingToParams(org.nd4j.linalg.api.ndarray.INDArray gradient,
                                     org.nd4j.linalg.api.ndarray.INDArray params,
                                     int batchSize)

Update the gradient according to the configuration such as adagrad, momentum, and sparsity

Parameters:

gradient - the gradient to modify