org.deeplearning4j.optimize

Class VectorizedBackTrackLineSearch

java.lang.Object

org.deeplearning4j.optimize.VectorizedBackTrackLineSearch

All Implemented Interfaces:

LineOptimizerMatrix

public class VectorizedBackTrackLineSearch
extends Object
implements LineOptimizerMatrix

Nested Class Summary

Nested classes/interfaces inherited from interface org.deeplearning4j.optimize.LineOptimizerMatrix

LineOptimizerMatrix.ByGradient

Constructor Summary

Constructors

Constructor and Description
VectorizedBackTrackLineSearch(OptimizableByGradientValueMatrix optimizable)

Method Summary

Methods

Modifier and Type	Method and Description
double	optimize(org.jblas.DoubleMatrix line, double initialStep) Returns the last step size used.
void	setAbsTolx(double tolx) Sets the tolerance of absolute diff in function value.
void	setRelTolx(double tolx) Sets the tolerance of relative diff in function value.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

VectorizedBackTrackLineSearch

public VectorizedBackTrackLineSearch(OptimizableByGradientValueMatrix optimizable)

Method Detail

setRelTolx

public void setRelTolx(double tolx)

Sets the tolerance of relative diff in function value. Line search converges if abs(delta x / x) < tolx for all coordinates.

setAbsTolx

public void setAbsTolx(double tolx)

Sets the tolerance of absolute diff in function value. Line search converges if abs(delta x) < tolx for all coordinates.

optimize

public double optimize(org.jblas.DoubleMatrix line,
              double initialStep)

Description copied from interface: LineOptimizerMatrix

Returns the last step size used.

Specified by:

optimize in interface LineOptimizerMatrix