org.deeplearning4j.nn.conf.constraint

Class MaxNormConstraint

java.lang.Object

org.deeplearning4j.nn.conf.constraint.BaseConstraint

org.deeplearning4j.nn.conf.constraint.MaxNormConstraint

All Implemented Interfaces:

Serializable, Cloneable, LayerConstraint

public class MaxNormConstraint
extends BaseConstraint

Constrain the maximum L2 norm of the incoming weights for each unit to be less than or equal to the specified value. If the L2 norm exceeds the specified value, the weights will be scaled down to satisfy the constraint.

Author:

Alex Black

See Also:

Serialized Form

Field Summary

Fields inherited from class org.deeplearning4j.nn.conf.constraint.BaseConstraint

DEFAULT_EPSILON, dimensions, epsilon, params

Constructor Summary

Constructors

Constructor and Description
MaxNormConstraint(double maxNorm, int... dimensions) Apply to weights but not biases by default
MaxNormConstraint(double maxNorm, Set<String> paramNames, int... dimensions)

Method Summary

Modifier and Type	Method and Description
void	apply(INDArray param)
MaxNormConstraint	clone()

Methods inherited from class org.deeplearning4j.nn.conf.constraint.BaseConstraint

applyConstraint, getBroadcastDims

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.deeplearning4j.nn.api.layers.LayerConstraint

getParams, setParams

Constructor Detail

MaxNormConstraint

public MaxNormConstraint(double maxNorm,
                         Set<String> paramNames,
                         int... dimensions)

Parameters:

maxNorm - Maximum L2 value

paramNames - Which parameter names to apply constraint to

dimensions - Dimensions to apply to. For DenseLayer, OutputLayer, RnnOutputLayer, LSTM, etc: this should be dimension 1. For CNNs, this should be dimensions [1,2,3] corresponding to last 3 of parameters which have order [depthOut, depthIn, kH, kW]

MaxNormConstraint

public MaxNormConstraint(double maxNorm,
                         int... dimensions)

Apply to weights but not biases by default

Parameters:

maxNorm - Maximum L2 value

dimensions - Dimensions to apply to. For DenseLayer, OutputLayer, RnnOutputLayer, LSTM, etc: this should be dimension 1. For CNNs, this should be dimensions [1,2,3] corresponding to last 3 of parameters which have order [depthOut, depthIn, kH, kW]

Method Detail

apply

public void apply(INDArray param)

Specified by:

apply in class BaseConstraint

clone

public MaxNormConstraint clone()

Specified by:

clone in interface LayerConstraint

Specified by:

clone in class BaseConstraint