Source code for aimet_torch.v2.quantization.float.quantizer

# -*- mode: python -*-
# =============================================================================
#  @@-COPYRIGHT-START-@@
#
#  Copyright (c) 2024, Qualcomm Innovation Center, Inc. All rights reserved.
#
#  Redistribution and use in source and binary forms, with or without
#  modification, are permitted provided that the following conditions are met:
#
#  1. Redistributions of source code must retain the above copyright notice,
#     this list of conditions and the following disclaimer.
#
#  2. Redistributions in binary form must reproduce the above copyright notice,
#     this list of conditions and the following disclaimer in the documentation
#     and/or other materials provided with the distribution.
#
#  3. Neither the name of the copyright holder nor the names of its contributors
#     may be used to endorse or promote products derived from this software
#     without specific prior written permission.
#
#  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
#  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
#  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
#  ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
#  LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
#  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
#  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
#  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
#  CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
#  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
#  POSSIBILITY OF SUCH DAMAGE.
#
#  SPDX-License-Identifier: BSD-3-Clause
#
#  @@-COPYRIGHT-END-@@
# =============================================================================
# pylint: disable=redefined-builtin
""" Float quantizers """

import contextlib
import functools
from typing import Optional, List, Dict
import math

import torch
from aimet_torch.v2.quantization.encoding_analyzer import EncodingAnalyzer
from aimet_torch.v2.quantization.base import QuantizerBase
from aimet_torch.v2.quantization.float import FloatEncoding
from aimet_torch.v2.utils import StatisticsNotFoundError, patch_attr
from aimet_torch.fp_quantization import fake_cast_to_ieee_float


__all__ = ['QuantizeDequantize', 'FloatQuantizeDequantize']


def _ieee_float_max_representable_value(exponent_bits, mantissa_bits):
    exponent_max = 2 ** exponent_bits - 1
    exponent_bias = exponent_max // 2
    return (2 - 2**-mantissa_bits) * 2 ** (exponent_max - exponent_bias - 1)


_IEEE_FLOAT16_EXPONENT_BITS = 5
_IEEE_FLOAT16_MANTISSA_BITS = 10
assert _ieee_float_max_representable_value(_IEEE_FLOAT16_EXPONENT_BITS, _IEEE_FLOAT16_MANTISSA_BITS) == \
        torch.finfo(torch.float16).max

_BFLOAT16_EXPONENT_BITS = 8
_BFLOAT16_MANTISSA_BITS = 7
assert _ieee_float_max_representable_value(_BFLOAT16_EXPONENT_BITS, _BFLOAT16_MANTISSA_BITS) == \
        torch.finfo(torch.bfloat16).max


[docs]class FloatQuantizeDequantize(QuantizerBase): # pylint: disable=abstract-method r""" Simulates quantization by fake-casting the input If dtype is provided, this is equivalent to .. math:: out = x.to(dtype).to(x.dtype) \\ If the exponent and mantissa bits are provided, this is equivalent to .. math:: out = \left\lceil\frac{x_c}{scale}\right\rfloor * scale where .. math:: x_c &= clamp(x, -max, max) \\ bias &= 2^{exponent} - \log_2(max) + \log_2(2 - 2^{-mantissa}) - 1 \\ scale &= 2 ^ {\left\lfloor \log_2 |x_c| + bias \right\rfloor - mantissa - bias} \\ The IEEE standard computes the maximum representable value by .. math:: max = (2 - 2^{-mantissa}) * 2^{(\left\lfloor 0.5 * exponent\_max \right\rfloor)} \\ where .. math:: exponent\_max = 2^{exponent} - 1 \\ Args: exponent_bits (int): Number of exponent bits to simulate mantissa_bits (int): Number of mantissa bits to simulate dtype (torch.dtype): torch.dtype to simulate. This argument is mutually exclusive with exponent_bits and mantissa_bits. encoding_analyzer (EncodingAnalyzer): If specified, the maximum value to represent will be determined dynamically based on the input statistics for finer precision. Examples: >>> import aimet_torch.v2.quantization as Q >>> input = torch.tensor([[ 1.8998, -0.0947],[-1.0891, -0.1727]]) >>> qdq = Q.float.FloatQuantizeDequantize(mantissa_bits=7, exponent_bits=8) >>> # Unlike AffineQuantizer, FloatQuantizer is initialized without calling compute_encodings() >>> qdq.is_initialized() True >>> qdq.is_bfloat16() True >>> qdq.bitwidth 16 >>> qdq(input) tensor([[ 1.8984, -0.0947], [-1.0859, -0.1729]]) >>> from aimet_torch.v2.quantization.encoding_analyzer import MinMaxEncodingAnalyzer >>> encoding_analyzer = MinMaxEncodingAnalyzer(shape=[]) >>> qdq = Q.float.FloatQuantizeDequantize(dtype=torch.float16, encoding_analyzer=encoding_analyzer) >>> qdq.is_float16() True >>> qdq.bitwidth 16 >>> qdq(input) tensor([[ 1.8994, -0.0947], [-1.0889, -0.1727]]) """ maxval: torch.Tensor def __init__(self, exponent_bits: int = None, mantissa_bits: int = None, dtype: torch.dtype = None, encoding_analyzer: EncodingAnalyzer = None): super().__init__() if dtype is None: if exponent_bits is None or mantissa_bits is None: raise ValueError('Neither "dtype" nor "exponent/mantissa_bits" was specified.') if dtype is not None: if exponent_bits is not None or mantissa_bits is not None: raise ValueError( 'Argument "dtype" is mutually exclusive with "exponent/mantissa_bits".') if dtype not in (torch.half, torch.float16, torch.bfloat16): raise ValueError( f"Float quantizer only supports torch.float16 and torch.bfloat16. Got {dtype}.") if dtype in (torch.half, torch.float16): exponent_bits = _IEEE_FLOAT16_EXPONENT_BITS mantissa_bits = _IEEE_FLOAT16_MANTISSA_BITS else: exponent_bits = _BFLOAT16_EXPONENT_BITS mantissa_bits = _BFLOAT16_MANTISSA_BITS self.exponent_bits = exponent_bits self.mantissa_bits = mantissa_bits self.encoding_analyzer = encoding_analyzer if self.encoding_analyzer: shape = self.encoding_analyzer.observer.shape maxval = _ieee_float_max_representable_value(exponent_bits, mantissa_bits) self.register_buffer('maxval', torch.full(shape, maxval)) else: self.register_buffer('maxval', None) def get_extra_state(self): extra_state_dict = {} extra_state_dict['exponent_bits'] = torch.tensor(self.exponent_bits) extra_state_dict['mantissa_bits'] = torch.tensor(self.mantissa_bits) super_extra_state = super().get_extra_state() extra_state_dict.update(super_extra_state) return extra_state_dict def set_extra_state(self, state): self.exponent_bits = state['exponent_bits'].item() self.mantissa_bits = state['mantissa_bits'].item() super().set_extra_state(state) def load_state_dict(self, state_dict, strict: bool = True): if 'maxval' in state_dict: if self.maxval is None: del self.maxval self.register_buffer('maxval', state_dict['maxval']) elif self.maxval is not None: del self.maxval self.register_buffer('maxval', None) ret = super().load_state_dict(state_dict, strict) return ret @property def bitwidth(self): """ Returns bitwidth of the quantizer """ return self.exponent_bits + self.mantissa_bits + 1 def is_float16(self): """ Returns true if current configuration simulates IEEE float16 """ return self.exponent_bits == _IEEE_FLOAT16_EXPONENT_BITS and \ self.mantissa_bits == _IEEE_FLOAT16_MANTISSA_BITS def is_bfloat16(self): """ Returns true if current configuration simulates bfloat16 """ return self.exponent_bits == _BFLOAT16_EXPONENT_BITS and \ self.mantissa_bits == _BFLOAT16_MANTISSA_BITS def get_legacy_encodings(self) -> Optional[List[Dict]]: """ :meta private: """ return [{'bitwidth': self.bitwidth, 'dtype': 'float'}] def set_legacy_encodings(self, encodings: List[Dict]): """ :meta private: Set encodings represented in the same format as the output of get_legacy_encodings as below: [ {'bitwidth': int, 'dtype': str}, ... ] """ if encodings[0]['bitwidth'] != 16: raise RuntimeError(f"{self.__class__} can only import 16-bit legay encodings.") self.exponent_bits = 5 self.mantissa_bits = 10 def get_encoding(self) -> Optional[FloatEncoding]: if self.is_initialized(): return FloatEncoding(self.mantissa_bits, self.exponent_bits, self.maxval) return None @contextlib.contextmanager def compute_encodings(self): """ Observe inputs and update quantization parameters based on the input statistics. During ``compute_encodings`` is enabled, the quantizer forward pass performs dynamic quantization using the batch statistics. """ if not self.encoding_analyzer or not self._allow_overwrite: yield return original_forward = self.forward @functools.wraps(original_forward) def forward_wrapper(input): input = input.as_subclass(torch.Tensor) batch_statistics = self.encoding_analyzer.update_stats(input) num_steps = math.pow(2, self.bitwidth) - 1 dynamic_min, dynamic_max =\ self.encoding_analyzer.compute_encodings_from_stats(batch_statistics, num_steps, is_symmetric=False) dynamic_absmax = torch.maximum(dynamic_min.abs(), dynamic_max.abs()) dynamic_absmax = dynamic_absmax.to(dtype=self.maxval.dtype, device=self.maxval.device).expand_as(self.maxval) with patch_attr(self, 'maxval', dynamic_absmax): return original_forward(input) self.encoding_analyzer.reset_stats() try: with patch_attr(self, 'forward', forward_wrapper): yield except: # pylint: disable=try-except-raise raise else: try: num_steps = math.pow(2, self.bitwidth) - 1 min, max = self.encoding_analyzer.compute_encodings(num_steps, is_symmetric=False) except StatisticsNotFoundError: return if min is None or max is None: return absmax = torch.maximum(min.abs(), max.abs()).expand_as(self.maxval) with torch.no_grad(): self.maxval.copy_(absmax) def forward(self, input: torch.Tensor): """ :param input: Input to quantize and dequantize :return: Quantize-dequantized output """ maxval = self.maxval exponent_bits = self.exponent_bits mantissa_bits = self.mantissa_bits if maxval is None: if self.is_float16() or self.is_bfloat16(): # Fast forward using type casting orig_dtype = input.dtype dtype = torch.float16 if self.is_float16() else torch.bfloat16 return input.to(dtype).to(orig_dtype) maxval = _ieee_float_max_representable_value(exponent_bits, mantissa_bits) # Subclasses of torch.Tensor with custom __torch_function__ (in our case, QuantizedTensorBase) # is known to introduce substantial CPU overhead. # Cast types of the inputs to plain torch.Tensor for faster execution. return fake_cast_to_ieee_float(input.as_subclass(torch.Tensor), maxval, exponent_bits, mantissa_bits) def extra_repr(self): """ :meta private: """ return f'exponent_bits={self.exponent_bits}, mantissa_bits={self.mantissa_bits}'
[docs]class QuantizeDequantize(FloatQuantizeDequantize): r""" Alias of FloatQuantizeDequantize """