Class Softmax

java.lang.Object

org.tensorflow.framework.activations.AbstractActivation

org.tensorflow.framework.activations.Softmax

All Implemented Interfaces:

Activation

public class Softmax
extends AbstractActivation

Softmax converts a real vector to a vector of categorical probabilities.

The elements of the output vector are in range (0, 1) and sum to 1.

Each vector is handled independently. The axis argument sets which axis of the input the function is applied along.

Softmax is often used as the activation for the last layer of a classification network because the result could be interpreted as a probability distribution.

The softmax of each vector x is computed as: exp(x) / tf.sum(exp(x)).

The input values in are the log-odds of the resulting probability.

Field Summary

Fields

Modifier and Type	Field	Description
static final String	NAME	The activation name as known by TensorFlow

Fields inherited from class AbstractActivation

NAME_KEY, tf

Constructor Summary

Constructors

Constructor	Description
Softmax()	Creates a softmax activation where the default axis is AXIS_DEFAULT which indicates the last dimension.
Softmax(int axis)	Creates a Softmax activation
Softmax(Map<String,Object> config)	Creates a Softmax activation from a config map.

Method Summary

Modifier and Type	Method	Description
<T extends TNumber> Operand<T>	call(Ops tf, Operand<T> input)	Gets the calculation operation for the activation.
int	getAxis()	Gets the axis along which the softmax normalization is applied.
Map<String,Object>	getConfig()	Gets a configuration map with entries axis and value set with axis.
String	getName()	Get the name of the activation as known by the TensorFlow Engine
static <T extends TNumber> Operand<T>	softmax(Ops tf, Operand<T> input)	Converts a vector of values to a probability distribution along the last axis.
static <T extends TNumber> Operand<T>	softmax(Ops tf, Operand<T> input, Operand<TInt32> axis)	Converts a vector of values to a probability distribution.

Methods inherited from class AbstractActivation

checkClassName, checkConfigKeys, getDefaultConfig, getTF, setTF

Methods inherited from class Object

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

Field Details

NAME

public static final String NAME

The activation name as known by TensorFlow

See Also:

• Constant Field Values

Constructor Details

Softmax

public Softmax()

Creates a softmax activation where the default axis is AXIS_DEFAULT which indicates the last dimension.

Softmax

public Softmax(int axis)

Creates a Softmax activation

Parameters:

axis - The dimension softmax would be performed on.

Softmax

public Softmax(Map<String,Object> config)

Creates a Softmax activation from a config map.

Parameters:

config - the configuration map, if the map contains an entry for axis that value is used, otherwise AXIS_DEFAULT is used.

Throws:

IllegalArgumentException - if the configuration contains unsupported keys for this class or if the value for the name key does not match the name for the Activation

Method Details

softmax

public static <T extends TNumber> Operand<T> softmax(Ops tf,
 Operand<T> input)

Converts a vector of values to a probability distribution along the last axis.

The elements of the output vector are in range (0, 1) and sum to 1.

Each vector is handled independently. The axis argument sets which axis of the input the function is applied along.

Softmax is often used as the activation for the last layer of a classification network because the result could be interpreted as a probability distribution.

The softmax of each vector x is computed as exp(x) / tf.reduce_sum(exp(x)). The input values in are the log-odds of the resulting probability.

Type Parameters:

T - teh data type of the input

Parameters:

tf - the TensorFlow Ops

input - the input

Returns:

the output of softmax transformation (all values are non-negative and sum to 1).

softmax

public static <T extends TNumber> Operand<T> softmax(Ops tf,
 Operand<T> input,
 Operand<TInt32> axis)

Converts a vector of values to a probability distribution.

The elements of the output vector are in range (0, 1) and sum to 1.

Each vector is handled independently. The axis argument sets which axis of the input the function is applied along.

Softmax is often used as the activation for the last layer of a classification network because the result could be interpreted as a probability distribution.

The softmax of each vector x is computed as exp(x) / tf.reduce_sum(exp(x)). The input values in are the log-odds of the resulting probability.

Type Parameters:

T - teh data type of the input

Parameters:

tf - the TensorFlow Ops

input - the input

axis - Integer, axis along which the softmax normalization is applied.

Returns:

the output of softmax transformation (all values are non-negative and sum to 1).

getConfig

public Map<String,Object> getConfig()

Gets a configuration map with entries

• axis and value set with axis.

Specified by:

getConfig in class AbstractActivation

Returns:

config the configuration map

call

public <T extends TNumber> Operand<T> call(Ops tf,
 Operand<T> input)

Gets the calculation operation for the activation.

Type Parameters:

T - the data type of the input and the result

Parameters:

tf - the TensorFlow Ops

input - the input tensor

Returns:

The operand for the activation

getName

public String getName()

Get the name of the activation as known by the TensorFlow Engine

Specified by:

getName in class AbstractActivation

Returns:

the name of the activation as known by the TensorFlow Engine

getAxis

public int getAxis()

Gets the axis along which the softmax normalization is applied.

Returns:

the axis along which the softmax normalization is applied.