pytorch_quantization.calib.max.MaxCalibrator Class Reference

Inheritance diagram for pytorch_quantization.calib.max.MaxCalibrator:

Collaboration diagram for pytorch_quantization.calib.max.MaxCalibrator:

Public Member Functions
def	__init__ (self, num_bits, axis, unsigned, track_amax=False)
def	amaxs (self)
def	collect (self, x)
def	reset (self)
def	compute_amax (self)
def	__str__ (self)
def	__repr__ (self)
def	compute_amax (self, *args, **kwargs)

Private Attributes
	_track_amax
	_amaxs
	_calib_amax
	_num_bits
	_axis
	_unsigned

Detailed Description

Max calibrator, tracks the maximum value globally

Args:
    calib_desc: A MaxCalibDescriptor.
    num_bits: An integer. Number of bits of quantization.
    axis: A tuple. see QuantDescriptor.
    unsigned: A boolean. using unsigned quantization.

Readonly Properties:
    amaxs: A list of amax. Numpy array is saved as it is likely to be used for some plot.

Constructor & Destructor Documentation

__init__()

def pytorch_quantization.calib.max.MaxCalibrator.__init__	(		self,
			num_bits,
			axis,
			unsigned,
			track_amax = False
	)

Member Function Documentation

amaxs()

def pytorch_quantization.calib.max.MaxCalibrator.amaxs

(

self

)

collect()

def pytorch_quantization.calib.max.MaxCalibrator.collect	(		self,
			x
	)

Tracks the absolute max of all tensors

Args:
    x: A tensor

Raises:
    RuntimeError: If amax shape changes

Reimplemented from pytorch_quantization.calib.calibrator._Calibrator.

reset()

def pytorch_quantization.calib.max.MaxCalibrator.reset

(

self

)

Reset the collected absolute max

Reimplemented from pytorch_quantization.calib.calibrator._Calibrator.

compute_amax() [1/2]

def pytorch_quantization.calib.max.MaxCalibrator.compute_amax

(

self

)

Return the absolute max of all tensors collected

__str__()

def pytorch_quantization.calib.max.MaxCalibrator.__str__

(

self

)

__repr__()

def pytorch_quantization.calib.max.MaxCalibrator.__repr__

(

self

)

Reimplemented from pytorch_quantization.calib.calibrator._Calibrator.

compute_amax() [2/2]

def pytorch_quantization.calib.calibrator._Calibrator.compute_amax (   self, *  args, **  kwargs  )

inherited

Abstract method: compute the amax from the collected data

Returns:
    amax: a tensor

Member Data Documentation

_track_amax

pytorch_quantization.calib.max.MaxCalibrator._track_amax

private

_amaxs

pytorch_quantization.calib.max.MaxCalibrator._amaxs

private

_calib_amax

pytorch_quantization.calib.max.MaxCalibrator._calib_amax

private

_num_bits

pytorch_quantization.calib.calibrator._Calibrator._num_bits

privateinherited

_axis

pytorch_quantization.calib.calibrator._Calibrator._axis

privateinherited

_unsigned

pytorch_quantization.calib.calibrator._Calibrator._unsigned

privateinherited

---

The documentation for this class was generated from the following file:

• max.py