jsat.classifiers.neuralnetwork.initializers

Class ConstantInit

java.lang.Object

jsat.classifiers.neuralnetwork.initializers.ConstantInit

All Implemented Interfaces:

Serializable, BiastInitializer

public class ConstantInit
extends Object
implements BiastInitializer

This initializes all bias values to a single constant value

Author:

Edward Raff

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ConstantInit(double c)

Method Summary

Modifier and Type	Method and Description
ConstantInit	clone()
double	getConstant()
void	init(Vec b, int fanIn, Random rand) Performs the initialization of the given vector of bias values
void	setConstant(double c)

Methods inherited from class java.lang.Object

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

Constructor Detail

ConstantInit

public ConstantInit(double c)

Parameters:

c - the constant to set all biases to

Method Detail

setConstant

public void setConstant(double c)

Parameters:

c - the constant value to use

getConstant

public double getConstant()

Returns:

the constant value that will be used for initialization

init

public void init(Vec b,
                 int fanIn,
                 Random rand)

Description copied from interface: BiastInitializer

Performs the initialization of the given vector of bias values

Specified by:

init in interface BiastInitializer

Parameters:

b - the vector to store the biases in

fanIn - the number of connections coming into the layer that these biases are for.

rand - the source of randomness for initialization

clone

public ConstantInit clone()

Specified by:

clone in interface BiastInitializer

Overrides:

clone in class Object