org.tensorflow.op.nn

Class BatchNormWithGlobalNormalizationGrad<T>

java.lang.Object

org.tensorflow.op.PrimitiveOp

org.tensorflow.op.nn.BatchNormWithGlobalNormalizationGrad<T>

Type Parameters:

T - data type for dx() output

All Implemented Interfaces:

Op

@Operator(group="nn")
public final class BatchNormWithGlobalNormalizationGrad<T>
extends PrimitiveOp

Gradients for batch normalization.

This op is deprecated. See `tf.nn.batch_normalization`.

Field Summary

Fields inherited from class org.tensorflow.op.PrimitiveOp

operation

Method Summary

Modifier and Type	Method and Description
static <T> BatchNormWithGlobalNormalizationGrad<T>	create(Scope scope, Operand<T> t, Operand<T> m, Operand<T> v, Operand<T> gamma, Operand<T> backprop, Float varianceEpsilon, Boolean scaleAfterNormalization) Factory method to create a class wrapping a new BatchNormWithGlobalNormalizationGrad operation.
Output<T>	db() 1D backprop tensor for beta.
Output<T>	dg() 1D backprop tensor for gamma.
Output<T>	dm() 1D backprop tensor for mean.
Output<T>	dv() 1D backprop tensor for variance.
Output<T>	dx() 4D backprop tensor for input.

Methods inherited from class org.tensorflow.op.PrimitiveOp

equals, hashCode, op, toString

Methods inherited from class java.lang.Object

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

Method Detail

create

public static <T> BatchNormWithGlobalNormalizationGrad<T> create(Scope scope,
                                                                 Operand<T> t,
                                                                 Operand<T> m,
                                                                 Operand<T> v,
                                                                 Operand<T> gamma,
                                                                 Operand<T> backprop,
                                                                 Float varianceEpsilon,
                                                                 Boolean scaleAfterNormalization)

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

Parameters:

scope - current scope

t - A 4D input Tensor.

m - A 1D mean Tensor with size matching the last dimension of t. This is the first output from tf.nn.moments, or a saved moving average thereof.

v - A 1D variance Tensor with size matching the last dimension of t. This is the second output from tf.nn.moments, or a saved moving average thereof.

gamma - A 1D gamma Tensor with size matching the last dimension of t. If "scale_after_normalization" is true, this Tensor will be multiplied with the normalized Tensor.

backprop - 4D backprop Tensor.

varianceEpsilon - A small float number to avoid dividing by 0.

scaleAfterNormalization - A bool indicating whether the resulted tensor needs to be multiplied with gamma.

Returns:

a new instance of BatchNormWithGlobalNormalizationGrad

dx

public Output<T> dx()

4D backprop tensor for input.

dm

public Output<T> dm()

1D backprop tensor for mean.

dv

public Output<T> dv()

1D backprop tensor for variance.

db

public Output<T> db()

1D backprop tensor for beta.

dg

public Output<T> dg()

1D backprop tensor for gamma.