MetricsHelper

public class MetricsHelper

These are helper methods for Metrics and will be module private when Java modularity is applied to TensorFlow Java. These methods should not be used outside of the metrics packages.

Constants

float NEG_INF

Public Constructors

MetricsHelper()

Public Methods

static <T extends TNumber> Op	assertBroadcastable(Ops tf, Operand<T> sampleWeights, Operand<T> values) Asserts that the `sampleWeights` can be broadcast to the same shape as `values` In losses and metrics, limited weight broadcasting is supported.
static Operand<TFloat64>	booleanMean(Ops tf, Operand<TBool> x) Calculate the mean of the operand, along all axes and `keepDims` is `false`
static Operand<TFloat64>	booleanMean(Ops tf, Operand<TBool> x, boolean keepDims) Calculates the mean of the boolean operand, alongside all axes.
static Operand<TFloat64>	booleanMean(Ops tf, Operand<TBool> x, Operand<? extends TIntegral> axes) Calculate the mean of the operand, alongside the specified axis with `keepDims` is `false`
static Operand<TFloat64>	booleanMean(Ops tf, Operand<TBool> x, Operand<? extends TIntegral> axes, boolean keepDims) Calculates the mean of the boolean operand, alongside the specified axes.
static <T extends TNumber> Operand<T>	broadcastWeights(Ops tf, Operand<T> weights, Operand<T> values) Broadcast `weights` to the same shape as `values`.
static <T extends TNumber> Operand<T>	mean(Ops tf, Operand<T> x) Calculate the mean of the operand, along all axes and `keepDims` is `false`
static <T extends TNumber> Operand<T>	mean(Ops tf, Operand<T> x, Operand<? extends TIntegral> axes) Calculate the mean of the operand, alongside the specified axis with `keepDims` is `false`
static <T extends TNumber> Operand<T>	mean(Ops tf, Operand<T> x, Operand<? extends TIntegral> axes, boolean keepDims) Calculates the mean of the operand, alongside the specified axes.
static <T extends TNumber> Operand<T>	mean(Ops tf, Operand<T> x, boolean keepDims) Calculates the mean of the operand, along all axes.

Inherited Methods

From class java.lang.Object

boolean	equals(Object arg0)
final Class<?>	getClass()
int	hashCode()
final void	notify()
final void	notifyAll()
String	toString()
final void	wait(long arg0, int arg1)
final void	wait(long arg0)
final void	wait()

Constants

public static final float NEG_INF

Constant Value: -1.0E10

Public Constructors

public MetricsHelper ()

Public Methods

public static Op assertBroadcastable (Ops tf, Operand<T> sampleWeights, Operand<T> values)

Asserts that the sampleWeights can be broadcast to the same shape as values

In losses and metrics, limited weight broadcasting is supported. Weights must be either scalar, or the same rank as the target values, with each dimension either 1, or the same as the corresponding values dimension.

Parameters

tf	the TensorFlow Ops
sampleWeights	the sample weights.
values	the values to which weights are applied.

Returns

Operation with control dependencies to ensure sampleWeight can be broadcast to values

Throws

NotBroadcastableException	If static checks determine `sampleWeights` has an incorrect shape that prohibit broadcasting to `values`

public static Operand<TFloat64> booleanMean (Ops tf, Operand<TBool> x)

Calculate the mean of the operand, along all axes and keepDims is false

Parameters

tf	the TensorFlow Ops
x	the Operand used to calculate the mean

Returns

the mean of the operand containing floating point numbers

public static Operand<TFloat64> booleanMean (Ops tf, Operand<TBool> x, boolean keepDims)

Calculates the mean of the boolean operand, alongside all axes.

Parameters

tf	the TensorFlow Ops
x	the boolean Operand used to calculate the mean
keepDims	Indicates whether to keep the dimensions or not. If `keepdims` is `false`, the rank of the tensor is reduced by 1 for each entry in `axes`. If `keepdims` is `true`, the reduced dimensions are retained with length 1.

Returns

the mean of elements of x containing floating point numbers

public static Operand<TFloat64> booleanMean (Ops tf, Operand<TBool> x, Operand<? extends TIntegral> axes)

Calculate the mean of the operand, alongside the specified axis with keepDims is false

Parameters

tf	the TensorFlow Ops
x	the Operand used to calculate the mean
axes	Axes to compute the mean.

Returns

the mean of the operand, along the specified axes, containing floating point numbers

public static Operand<TFloat64> booleanMean (Ops tf, Operand<TBool> x, Operand<? extends TIntegral> axes, boolean keepDims)

Calculates the mean of the boolean operand, alongside the specified axes.

Parameters

tf	the TensorFlow Ops
x	the boolean Operand used to calculate the mean
axes	Axes to compute the mean.
keepDims	Indicates whether to keep the dimensions or not. If `keepdims` is `false`, the rank of the tensor is reduced by 1 for each entry in `axes`. If `keepdims` is `true`, the reduced dimensions are retained with length 1.

Returns

the mean of elements of x containing floating point numbers

public static Operand<T> broadcastWeights (Ops tf, Operand<T> weights, Operand<T> values)

Broadcast weights to the same shape as values.

Parameters

tf	the TensorFlow ops
weights	Operand whose shape is broadcastable to `values`.
values	Operand of any shape

Returns

weights broadcast to values shape

public static Operand<T> mean (Ops tf, Operand<T> x)

Calculate the mean of the operand, along all axes and keepDims is false

Parameters

tf	the TensorFlow Ops
x	the Operand used to calculate the mean

Returns

the mean of the operand

public static Operand<T> mean (Ops tf, Operand<T> x, Operand<? extends TIntegral> axes)

Calculate the mean of the operand, alongside the specified axis with keepDims is false

Parameters

tf	the TensorFlow Ops
x	the Operand used to calculate the mean
axes	Axes to compute the mean.

Returns

the mean of the operand, along the specified axes.

public static Operand<T> mean (Ops tf, Operand<T> x, Operand<? extends TIntegral> axes, boolean keepDims)

Calculates the mean of the operand, alongside the specified axes.

Parameters

tf	the TensorFlow Ops
x	the Operand used to calculate the mean
axes	Axes to compute the mean.
keepDims	Indicates whether to keep the dimensions or not. If `keepdims` is `false`, the rank of the tensor is reduced by 1 for each entry in `axes`. If `keepdims` is `true`, the reduced dimensions are retained with length 1.

Returns

the mean of elements of x.

public static Operand<T> mean (Ops tf, Operand<T> x, boolean keepDims)

Calculates the mean of the operand, along all axes.

Parameters

tf	the TensorFlow Ops
x	the Operand used to calculate the mean
keepDims	Indicates whether to keep the dimensions or not. If `keepdims` is `false`, the rank of the tensor is reduced by 1 for each entry in `axes`. If `keepdims` is `true`, the reduced dimensions are retained with length 1.

Returns

the mean of elements of x.