import math
import torch
from mqbench.fake_quantize.quantize_base import QuantizeBase
from mqbench.utils import is_tracing_state
from mqbench.utils.hook import PerChannelLoadHook
[docs]def dsq_function_per_tensor(x, scale, zero_point, quant_min, quant_max, alpha):
tanh_scale = 1 / (1 - alpha)
tanh_k = math.log((tanh_scale + 1) / (tanh_scale - 1))
x = x / scale + zero_point
x = torch.clamp(x, quant_min, quant_max)
x = x.floor() + (tanh_scale * torch.tanh(tanh_k * (x - x.floor() - 0.5))) * 0.5 + 0.5
x = (x.round() - x).detach() + x
x = (x - zero_point) * scale
return x
[docs]def dsq_function_per_channel(x, scale, zero_point, quant_min, quant_max, ch_axis, alpha):
new_shape = [1] * len(x.shape)
new_shape[ch_axis] = x.shape[ch_axis]
scale = scale.reshape(new_shape)
zero_point = zero_point.reshape(new_shape)
tanh_scale = 1 / (1 - alpha)
tanh_k = math.log((tanh_scale + 1) / (tanh_scale - 1))
x = x / scale + zero_point
x = torch.clamp(x, quant_min, quant_max)
x = x.floor() + (tanh_scale * torch.tanh(tanh_k * (x - x.floor() - 0.5))) * 0.5 + 0.5
x = (x.round() - x).detach() + x
x = (x - zero_point) * scale
return x
[docs]class DSQFakeQuantize(QuantizeBase):
def __init__(self, observer, alpha=0.4, **observer_kwargs):
super(DSQFakeQuantize, self).__init__(observer, **observer_kwargs)
self.register_buffer('scale', torch.tensor([1.0], dtype=torch.float))
self.register_buffer('zero_point', torch.tensor([0], dtype=torch.int))
self.alpha = alpha
self.load_state_dict_hook = PerChannelLoadHook(self)
[docs] def forward(self, X):
if self.training:
self.activation_post_process(X.detach())
_scale, _zero_point = self.activation_post_process.calculate_qparams()
_scale, _zero_point = _scale.to(self.scale.device), _zero_point.to(self.zero_point.device)
if self.scale.shape != _scale.shape:
self.scale.resize_(_scale.shape)
self.zero_point.resize_(_zero_point.shape)
self.scale.copy_(_scale)
self.zero_point.copy_(_zero_point.float())
if self.fake_quant_enabled[0] == 1:
if self.is_per_channel:
if is_tracing_state():
X = FakeQuantizeDSQPerchannel.apply(
X, self.scale, self.zero_point, self.quant_min, self.quant_max, self.ch_axis, self.alpha)
else:
X = dsq_function_per_channel(
X, self.scale, self.zero_point, self.quant_min, self.quant_max, self.ch_axis, self.alpha)
else:
if is_tracing_state():
X = FakeQuantizeDSQPertensor.apply(
X, self.scale, self.zero_point, self.quant_min, self.quant_max, self.alpha)
else:
X = dsq_function_per_tensor(
X, self.scale, self.zero_point, self.quant_min, self.quant_max, self.alpha)
return X
[docs]class FakeQuantizeDSQPerchannel(torch.autograd.Function):
[docs] @staticmethod
def forward(ctx, x, scale, zero_point, quant_min, quant_max, ch_axis, alpha):
return dsq_function_per_channel(x, scale, zero_point, quant_min, quant_max, ch_axis, alpha)
[docs] @staticmethod
def symbolic(g, x, scale, zero_point, quant_min, quant_max, ch_axis, alpha):
return g.op("::FakeQuantizeDSQPerchannel", x, scale, zero_point, quant_min_i=quant_min, quant_max_i=quant_max, alpha_f=alpha)
[docs]class FakeQuantizeDSQPertensor(torch.autograd.Function):
[docs] @staticmethod
def forward(ctx, x, scale, zero_point, quant_min, quant_max, alpha):
return dsq_function_per_tensor(x, scale, zero_point, quant_min, quant_max, alpha)
[docs] @staticmethod
def symbolic(g, x, scale, zero_point, quant_min, quant_max, alpha):
return g.op("::FakeQuantizeDSQPertensor", x, scale, zero_point, quant_min_i=quant_min, quant_max_i=quant_max, alpha_f=alpha)