org.tensorflow.op.nn

Class FusedBatchNormGrad<T extends TNumber,U extends TNumber>

java.lang.Object

org.tensorflow.op.RawOp

org.tensorflow.op.nn.FusedBatchNormGrad<T,U>

Type Parameters:

T - data type for x_backprop output

U - data type for scale_backprop output

All Implemented Interfaces:

Op

@Operator(group="nn")
public final class FusedBatchNormGrad<T extends TNumber,U extends TNumber>
extends RawOp

Gradient for batch normalization. Note that the size of 4D Tensors are defined by either "NHWC" or "NCHW". The size of 1D Tensors matches the dimension C of the 4D Tensors.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	FusedBatchNormGrad.Inputs<T extends TNumber,U extends TNumber>
static class	FusedBatchNormGrad.Options Optional attributes for FusedBatchNormGrad

Field Summary

Fields

Modifier and Type	Field and Description
static String	OP_NAME The name of this op, as known by TensorFlow core engine

Fields inherited from class org.tensorflow.op.RawOp

operation

Constructor Summary

Constructors

Constructor and Description
FusedBatchNormGrad(Operation operation)

Method Summary

Modifier and Type	Method and Description
static <T extends TNumber,U extends TNumber> FusedBatchNormGrad<T,U>	create(Scope scope, Operand<T> yBackprop, Operand<T> x, Operand<TFloat32> scale, Operand<U> reserveSpace1, Operand<U> reserveSpace2, Operand<U> reserveSpace3, FusedBatchNormGrad.Options... options) Factory method to create a class wrapping a new FusedBatchNormGradV3 operation.
static FusedBatchNormGrad.Options	dataFormat(String dataFormat) Sets the dataFormat option.
static FusedBatchNormGrad.Options	epsilon(Float epsilon) Sets the epsilon option.
static FusedBatchNormGrad.Options	isTraining(Boolean isTraining) Sets the isTraining option.
Output<U>	offsetBackprop() Gets offsetBackprop.
Output<U>	reserveSpace4() Gets reserveSpace4.
Output<U>	reserveSpace5() Gets reserveSpace5.
Output<U>	scaleBackprop() Gets scaleBackprop.
Output<T>	xBackprop() Gets xBackprop.

Methods inherited from class org.tensorflow.op.RawOp

equals, hashCode, op, toString

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface org.tensorflow.op.Op

env

Field Detail

OP_NAME

public static final String OP_NAME

The name of this op, as known by TensorFlow core engine

See Also:

Constant Field Values

Constructor Detail

FusedBatchNormGrad

public FusedBatchNormGrad(Operation operation)

Method Detail

create

@Endpoint(describeByClass=true)
public static <T extends TNumber,U extends TNumber> FusedBatchNormGrad<T,U> create(Scope scope,
                                                                                                                   Operand<T> yBackprop,
                                                                                                                   Operand<T> x,
                                                                                                                   Operand<TFloat32> scale,
                                                                                                                   Operand<U> reserveSpace1,
                                                                                                                   Operand<U> reserveSpace2,
                                                                                                                   Operand<U> reserveSpace3,
                                                                                                                   FusedBatchNormGrad.Options... options)

Factory method to create a class wrapping a new FusedBatchNormGradV3 operation.

Type Parameters:

T - data type for FusedBatchNormGradV3 output and operands

U - data type for FusedBatchNormGradV3 output and operands

Parameters:

scope - current scope

yBackprop - A 4D Tensor for the gradient with respect to y.

x - A 4D Tensor for input data.

scale - A 1D Tensor for scaling factor, to scale the normalized x.

reserveSpace1 - When is_training is True, a 1D Tensor for the computed batch mean to be reused in gradient computation. When is_training is False, a 1D Tensor for the population mean to be reused in both 1st and 2nd order gradient computation.

reserveSpace2 - When is_training is True, a 1D Tensor for the computed batch variance (inverted variance in the cuDNN case) to be reused in gradient computation. When is_training is False, a 1D Tensor for the population variance to be reused in both 1st and 2nd order gradient computation.

reserveSpace3 - When is_training is True, a 1D Tensor for some intermediate results to be reused in gradient computation. When is_training is False, a dummy empty Tensor will be created.

options - carries optional attribute values

Returns:

a new instance of FusedBatchNormGrad

epsilon

public static FusedBatchNormGrad.Options epsilon(Float epsilon)

Sets the epsilon option.

Parameters:

epsilon - A small float number added to the variance of x.

Returns:

this Options instance.

dataFormat

public static FusedBatchNormGrad.Options dataFormat(String dataFormat)

Sets the dataFormat option.

Parameters:

dataFormat - The data format for y_backprop, x, x_backprop. Either "NHWC" (default) or "NCHW".

Returns:

this Options instance.

isTraining

public static FusedBatchNormGrad.Options isTraining(Boolean isTraining)

Sets the isTraining option.

Parameters:

isTraining - A bool value to indicate the operation is for training (default) or inference.

Returns:

this Options instance.

xBackprop

public Output<T> xBackprop()

Gets xBackprop. A 4D Tensor for the gradient with respect to x.

Returns:

xBackprop.

scaleBackprop

public Output<U> scaleBackprop()

Gets scaleBackprop. A 1D Tensor for the gradient with respect to scale.

Returns:

scaleBackprop.

offsetBackprop

public Output<U> offsetBackprop()

Gets offsetBackprop. A 1D Tensor for the gradient with respect to offset.

Returns:

offsetBackprop.

reserveSpace4

public Output<U> reserveSpace4()

Gets reserveSpace4. Unused placeholder to match the mean input in FusedBatchNorm.

Returns:

reserveSpace4.

reserveSpace5

public Output<U> reserveSpace5()

Gets reserveSpace5. Unused placeholder to match the variance input in FusedBatchNorm.

Returns:

reserveSpace5.