tensorflow::ops::QuantizeAndDequantize::Attrs

#include <array_ops.h>

Optional attribute setters for QuantizeAndDequantize.

Summary

Public attributes

input_max_ = 0.0f
float
input_min_ = 0.0f
float
num_bits_ = 8
int64
range_given_ = false
bool
signed_input_ = true
bool

Public functions

InputMax(float x)
If range is given, this is the max of the range.
InputMin(float x)
If range is given, this is the min of the range.
NumBits(int64 x)
The bitwidth of the quantization.
RangeGiven(bool x)
If the range is given or should be computed from the tensor.
SignedInput(bool x)
If the quantization is signed or unsigned.

Public attributes

input_max_

float tensorflow::ops::QuantizeAndDequantize::Attrs::input_max_ = 0.0f

input_min_

float tensorflow::ops::QuantizeAndDequantize::Attrs::input_min_ = 0.0f

num_bits_

int64 tensorflow::ops::QuantizeAndDequantize::Attrs::num_bits_ = 8

range_given_

bool tensorflow::ops::QuantizeAndDequantize::Attrs::range_given_ = false

signed_input_

bool tensorflow::ops::QuantizeAndDequantize::Attrs::signed_input_ = true

Public functions

InputMax

Attrs tensorflow::ops::QuantizeAndDequantize::Attrs::InputMax(
  float x
)

If range is given, this is the max of the range.

Defaults to 0

InputMin

Attrs tensorflow::ops::QuantizeAndDequantize::Attrs::InputMin(
  float x
)

If range is given, this is the min of the range.

Defaults to 0

NumBits

Attrs tensorflow::ops::QuantizeAndDequantize::Attrs::NumBits(
  int64 x
)

The bitwidth of the quantization.

Defaults to 8

RangeGiven

Attrs tensorflow::ops::QuantizeAndDequantize::Attrs::RangeGiven(
  bool x
)

If the range is given or should be computed from the tensor.

Defaults to false

SignedInput

Attrs tensorflow::ops::QuantizeAndDequantize::Attrs::SignedInput(
  bool x
)

If the quantization is signed or unsigned.

Defaults to true