텐서플로우:: 작전:: FakeQuantWithMinMaxVarsGradient

#include <array_ops.h>

FakeQuantWithMinMaxVars 작업에 대한 기울기를 계산합니다.

요약

인수:

선택적 속성( Attrs 참조):

  • num_bits: 양자화의 비트폭. 2부터 8까지.
  • arrow_range: 2^num_bits - 1개의 고유 값으로 양자화할지 여부입니다.

보고:

  • Output backprops_wrt_input: 역전파된 경사 wrt 입력: gradients * (inputs >= min && inputs <= max) .
  • Output backprop_wrt_min: 역전파된 기울기 wrt 최소 매개변수: sum(gradients * (inputs < min)) .
  • Output backprop_wrt_max: 역전파된 기울기 wrt max 매개변수: sum(gradients * (inputs > max)) .

생성자와 소멸자

FakeQuantWithMinMaxVarsGradient (const :: tensorflow::Scope & scope, :: tensorflow::Input gradients, :: tensorflow::Input inputs, :: tensorflow::Input min, :: tensorflow::Input max)
FakeQuantWithMinMaxVarsGradient (const :: tensorflow::Scope & scope, :: tensorflow::Input gradients, :: tensorflow::Input inputs, :: tensorflow::Input min, :: tensorflow::Input max, const FakeQuantWithMinMaxVarsGradient::Attrs & attrs)

공개 속성

backprop_wrt_max
backprop_wrt_min
backprops_wrt_input
operation

공개 정적 함수

NarrowRange (bool x)
NumBits (int64 x)

구조체

텐서플로우:: ops:: FakeQuantWithMinMaxVarsGradient:: Attrs

FakeQuantWithMinMaxVarsGradient 에 대한 선택적 속성 설정자입니다.

공개 속성

backprop_wrt_max

::tensorflow::Output backprop_wrt_max

backprop_wrt_min

::tensorflow::Output backprop_wrt_min

backprops_wrt_input

::tensorflow::Output backprops_wrt_input

작업

Operation operation

공공 기능

FakeQuantWithMinMaxVarsGradient

 FakeQuantWithMinMaxVarsGradient(
  const ::tensorflow::Scope & scope,
  ::tensorflow::Input gradients,
  ::tensorflow::Input inputs,
  ::tensorflow::Input min,
  ::tensorflow::Input max
)

FakeQuantWithMinMaxVarsGradient

 FakeQuantWithMinMaxVarsGradient(
  const ::tensorflow::Scope & scope,
  ::tensorflow::Input gradients,
  ::tensorflow::Input inputs,
  ::tensorflow::Input min,
  ::tensorflow::Input max,
  const FakeQuantWithMinMaxVarsGradient::Attrs & attrs
)

공개 정적 함수

좁은 범위

Attrs NarrowRange(
  bool x
)

NumBits

Attrs NumBits(
  int64 x
)