org.nd4j.linalg.learning

Interface GradientUpdater<T extends IUpdater>

All Known Implementing Classes:

AdaDeltaUpdater, AdaGradUpdater, AdaMaxUpdater, AdamUpdater, AMSGradUpdater, NadamUpdater, NesterovsUpdater, NoOpUpdater, RmsPropUpdater, SgdUpdater

public interface GradientUpdater<T extends IUpdater>

Gradient modifications: Calculates an update and tracks related information for gradient changes over time for handling updates.

Author:

Alex Black

Method Summary

Modifier and Type	Method and Description
void	applyUpdater(INDArray gradient, int iteration, int epoch) Modify the gradient to be an update.
T	getConfig()
Map<String,INDArray>	getState()
void	setState(Map<String,INDArray> stateMap, boolean initialize)
void	setStateViewArray(INDArray viewArray, long[] gradientShape, char gradientOrder, boolean initialize) For the internal updater state (if any): set this to use the provided array.

Method Detail

getConfig

T getConfig()

setState

void setState(Map<String,INDArray> stateMap,
              boolean initialize)

getState

Map<String,INDArray> getState()

setStateViewArray

void setStateViewArray(INDArray viewArray,
                       long[] gradientShape,
                       char gradientOrder,
                       boolean initialize)

For the internal updater state (if any): set this to use the provided array. Used during initialization, and when restoring the updater state (after serialization, for example)

Parameters:

viewArray - Array (that is a view of a larger array) to use for the state.

gradientShape -

gradientOrder -

initialize - If true: the updater must initialize the view array. If false: no change to view array contents

applyUpdater

void applyUpdater(INDArray gradient,
                  int iteration,
                  int epoch)

Modify the gradient to be an update. Note that this is be done in-place

Parameters:

gradient - the gradient to modify

iteration -