deepnetts.net.layers.activation

Class Relu

java.lang.Object

deepnetts.net.layers.activation.Relu

All Implemented Interfaces:

ActivationFunction, Serializable

public final class Relu
extends Object
implements ActivationFunction, Serializable

Rectified Linear Activation and its Derivative. y = max(0, x) - | 1, x > 0 y' = < | 0, 0x<=0 -

Author:

Zoran Sevarac

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
Relu()

Method Summary

Modifier and Type	Method and Description
void	apply(Tensor tensor, int channel)
void	apply(Tensor tensor, int from, int to)
float	getPrime(float y) Returns the first derivative of activation function for specified output y
float	getValue(float x) Returns the value of activation function for specified input x

Methods inherited from class java.lang.Object

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

Methods inherited from interface deepnetts.net.layers.activation.ActivationFunction

create

Constructor Detail

Relu

public Relu()

Method Detail

getValue

public float getValue(float x)

Description copied from interface: ActivationFunction

Returns the value of activation function for specified input x

Specified by:

getValue in interface ActivationFunction

Parameters:

x - input for activation

Returns:

value of activation function

getPrime

public float getPrime(float y)

Description copied from interface: ActivationFunction

Returns the first derivative of activation function for specified output y

Specified by:

getPrime in interface ActivationFunction

Parameters:

y - output of activation function

Returns:

first derivative of activation function

apply

public void apply(Tensor tensor,
                  int channel)

Specified by:

apply in interface ActivationFunction

apply

public void apply(Tensor tensor,
                  int from,
                  int to)

Specified by:

apply in interface ActivationFunction