org.tensorflow.op.core

Class FakeQuantWithMinMaxVarsPerChannelGradient

java.lang.Object

org.tensorflow.op.PrimitiveOp

org.tensorflow.op.core.FakeQuantWithMinMaxVarsPerChannelGradient

All Implemented Interfaces:

Op

@Operator
public final class FakeQuantWithMinMaxVarsPerChannelGradient
extends PrimitiveOp

Compute gradients for a FakeQuantWithMinMaxVarsPerChannel operation.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	FakeQuantWithMinMaxVarsPerChannelGradient.Options Optional attributes for FakeQuantWithMinMaxVarsPerChannelGradient

Field Summary

Fields inherited from class org.tensorflow.op.PrimitiveOp

operation

Method Summary

Modifier and Type	Method and Description
Output<Float>	backpropsWrtInput() Backpropagated gradients w.r.t.
Output<Float>	backpropWrtMax() Backpropagated gradients w.r.t.
Output<Float>	backpropWrtMin() Backpropagated gradients w.r.t.
static FakeQuantWithMinMaxVarsPerChannelGradient	create(Scope scope, Operand<Float> gradients, Operand<Float> inputs, Operand<Float> min, Operand<Float> max, FakeQuantWithMinMaxVarsPerChannelGradient.Options... options) Factory method to create a class to wrap a new FakeQuantWithMinMaxVarsPerChannelGradient operation to the graph.
static FakeQuantWithMinMaxVarsPerChannelGradient.Options	narrowRange(Boolean narrowRange)
static FakeQuantWithMinMaxVarsPerChannelGradient.Options	numBits(Long numBits)

Methods inherited from class org.tensorflow.op.PrimitiveOp

equals, hashCode, toString

Methods inherited from class java.lang.Object

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

Method Detail

create

public static FakeQuantWithMinMaxVarsPerChannelGradient create(Scope scope,
                                                               Operand<Float> gradients,
                                                               Operand<Float> inputs,
                                                               Operand<Float> min,
                                                               Operand<Float> max,
                                                               FakeQuantWithMinMaxVarsPerChannelGradient.Options... options)

Factory method to create a class to wrap a new FakeQuantWithMinMaxVarsPerChannelGradient operation to the graph.

Parameters:

scope - current graph scope

gradients - Backpropagated gradients above the FakeQuantWithMinMaxVars operation, shape one of: `[d]`, `[b, d]`, `[b, h, w, d]`.

inputs - Values passed as inputs to the FakeQuantWithMinMaxVars operation, shape same as `gradients`. min, max: Quantization interval, floats of shape `[d]`.

min -

max -

options - carries optional attributes values

Returns:

a new instance of FakeQuantWithMinMaxVarsPerChannelGradient

numBits

public static FakeQuantWithMinMaxVarsPerChannelGradient.Options numBits(Long numBits)

Parameters:

numBits - The bitwidth of the quantization; between 2 and 16, inclusive.

narrowRange

public static FakeQuantWithMinMaxVarsPerChannelGradient.Options narrowRange(Boolean narrowRange)

Parameters:

narrowRange - Whether to quantize into 2^num_bits - 1 distinct values.

backpropsWrtInput

public Output<Float> backpropsWrtInput()

Backpropagated gradients w.r.t. inputs, shape same as `inputs`: `gradients * (inputs >= min && inputs <= max)`.

backpropWrtMin

public Output<Float> backpropWrtMin()

Backpropagated gradients w.r.t. min parameter, shape `[d]`: `sum_per_d(gradients * (inputs < min))`.

backpropWrtMax

public Output<Float> backpropWrtMax()

Backpropagated gradients w.r.t. max parameter, shape `[d]`: `sum_per_d(gradients * (inputs > max))`.