org.nd4j.linalg.api.activation

Class Activations

java.lang.Object

org.nd4j.linalg.api.activation.Activations

public class Activations
extends Object

Activation Functions for neural nets

Author:

Adam Gibson

Constructor Summary

Constructors

Constructor and Description
Activations()

Method Summary

Methods

Modifier and Type	Method and Description
static ActivationFunction	exp() The e^x function
static ActivationFunction	hardTanh() Hard Tanh is tanh constraining input to -1 to 1
static ActivationFunction	linear() Linear activation function, just returns the input as is
static ActivationFunction	rectifiedLinear() Rectified linear, the output: rounded
static ActivationFunction	roundedLinear() Rounded linear, the output: rounded
static ActivationFunction	sigmoid() Sigmoid function
static ActivationFunction	softmax() Soft max function used for multinomial classification
static ActivationFunction	softMaxRows() Softmax with row wise features
static ActivationFunction	tanh() Tanh function

Methods inherited from class java.lang.Object

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

Constructor Detail

Activations

public Activations()

Method Detail

softMaxRows

public static ActivationFunction softMaxRows()

Softmax with row wise features

Returns:

the softmax with row wise features

rectifiedLinear

public static ActivationFunction rectifiedLinear()

Rectified linear, the output: rounded

Returns:

the rounded lienar activation function

roundedLinear

public static ActivationFunction roundedLinear()

Rounded linear, the output: rounded

Returns:

the rounded lienar activation function

exp

public static ActivationFunction exp()

The e^x function

Returns:

the e^x activation function

linear

public static ActivationFunction linear()

Linear activation function, just returns the input as is

Returns:

the linear activation function

tanh

public static ActivationFunction tanh()

Tanh function

Returns:

sigmoid

public static ActivationFunction sigmoid()

Sigmoid function

Returns:

hardTanh

public static ActivationFunction hardTanh()

Hard Tanh is tanh constraining input to -1 to 1

Returns:

the hard tanh function

softmax

public static ActivationFunction softmax()

Soft max function used for multinomial classification

Returns: