tensorflow::ops::QuantizedInstanceNorm::Attrs

#include <array_ops.h>

Optional attribute setters for QuantizedInstanceNorm.

Summary

Public attributes

given_y_max_ = 0.0f
float
given_y_min_ = 0.0f
float
min_separation_ = 0.001f
float
output_range_given_ = false
bool
variance_epsilon_ = 1e-05f
float

Public functions

GivenYMax(float x)
TF_MUST_USE_RESULT Attrs
Output in y_max if output_range_given is True.
GivenYMin(float x)
TF_MUST_USE_RESULT Attrs
Output in y_min if output_range_given is True.
MinSeparation(float x)
TF_MUST_USE_RESULT Attrs
Minimum value of y_max - y_min
OutputRangeGiven(bool x)
TF_MUST_USE_RESULT Attrs
If True, given_y_min and given_y_min and given_y_max are used as the output range.
VarianceEpsilon(float x)
TF_MUST_USE_RESULT Attrs
A small float number to avoid dividing by 0.

Public attributes

given_y_max_

float tensorflow::ops::QuantizedInstanceNorm::Attrs::given_y_max_ = 0.0f

given_y_min_

float tensorflow::ops::QuantizedInstanceNorm::Attrs::given_y_min_ = 0.0f

min_separation_

float tensorflow::ops::QuantizedInstanceNorm::Attrs::min_separation_ = 0.001f

output_range_given_

bool tensorflow::ops::QuantizedInstanceNorm::Attrs::output_range_given_ = false

variance_epsilon_

float tensorflow::ops::QuantizedInstanceNorm::Attrs::variance_epsilon_ = 1e-05f

Public functions

GivenYMax

TF_MUST_USE_RESULT Attrs tensorflow::ops::QuantizedInstanceNorm::Attrs::GivenYMax(
  float x
)

Output in y_max if output_range_given is True.

Defaults to 0

GivenYMin

TF_MUST_USE_RESULT Attrs tensorflow::ops::QuantizedInstanceNorm::Attrs::GivenYMin(
  float x
)

Output in y_min if output_range_given is True.

Defaults to 0

MinSeparation

TF_MUST_USE_RESULT Attrs tensorflow::ops::QuantizedInstanceNorm::Attrs::MinSeparation(
  float x
)

Minimum value of y_max - y_min

Defaults to 0.001

OutputRangeGiven

TF_MUST_USE_RESULT Attrs tensorflow::ops::QuantizedInstanceNorm::Attrs::OutputRangeGiven(
  bool x
)

If True, given_y_min and given_y_min and given_y_max are used as the output range.

Otherwise, the implementation computes the output range.

Defaults to false

VarianceEpsilon

TF_MUST_USE_RESULT Attrs tensorflow::ops::QuantizedInstanceNorm::Attrs::VarianceEpsilon(
  float x
)

A small float number to avoid dividing by 0.

Defaults to 1e-05