STATIC QUANTIZATION WITH EAGER MODE IN PYTORCH

最新推荐文章于 2024-09-22 00:06:49 发布

JachinMa

最新推荐文章于 2024-09-22 00:06:49 发布

阅读量679

点赞数

本文链接：https://blog.csdn.net/JachinMa/article/details/104933635

版权

本文介绍了如何在PyTorch中使用Eager Mode进行静态量化，这是一种优化深度学习模型的方法，旨在减少内存消耗和提高推理速度。通过链接提供的详细教程，读者可以学习到量化的基本概念和实施步骤。

摘要由CSDN通过智能技术生成

关于pytorch的量化，可以看https://blog.csdn.net/zlgahu/article/details/104662203/

import numpy as np
import torch
import torch.nn as nn
import torchvision
from torch.utils.data import DataLoader
import torchvision.transforms as transforms
import os
import time
import sys
import torch.quantization
import requests
import warnings

#用于定义需要量化的参数
from torch.quantization import QuantStub,DeQuantStub

#确保通道数能被divisor整除
def _make_divisible(v,divisor,min_value=None):
	if min_value is None:
		min_value = divisor
	new_v = max(min_value,int(v + divisor / 2) // divisor * divisor)
	#防止损失过多通道
	if new_v < 0.9 * v:
		new_v += divisor
	return new_v
#定义几个网络的组合层
class ConvBNReLU(nn.Sequential):
	def __init__(self,in_planes,out_planes,kernel_size=3,stride=1,groups=1):
		padding = (kernel_size - 1) // 2
		super().__init__(
		nn.Conv2d(in_planes,out_planes,kernel_size,stride,padding,groups=groups,bias=False),
		nn.BatchNorm2d(out_planes,momentum=0.1),
		nn.ReLU(inplace=False)
		)
		
#定义残差块
class InvertedResidual(nn.Module):
	def __init__(self,inp,oup,stride,expand_ratio):
		super().__init__()
		self.stride = stride
		assert stride in [1,2]

		hidden_dim = int(round(inp * expand_ratio))
		self.use_res_connect = self.stride == 1 and inp == oup

		layers = []
		if expand_ratio != 1:
			layers.append(ConvBNReLU(inp,hidden_dim,kernel_size=1))

		layers.extend([
		ConvBNReLU(hidden_dim,hidden_dim,stride=stride,groups=hidden_dim),
		nn.Conv2d(hidden_dim,oup,1,1,0,bias),
		nn.BatchNorm2d(oup,momentum=0.1),
		])

		self.conv = nn.Sequential(*layers)
		#将无状态的相加操作转化为FloatFunctional，猜测是因为量化后加法操作不能正常运作
		self.skip_add = nn.quantized.FloadFunctional()
		
	def forward(self,x):
		if self.use_res_connect:
			return self.skip_add.add(x,self.conv(x))
		else:
			return self.conv(x)

#定义网络结构
class MobileNetV2(nn.Module):
	def __init__(self,num_classes=1000,width_mult=1.0,inverted_residual_setting=None,round_nearest=8):
		super().__init__()
		block = InvertedResidual
		input_channel = 32
		last_channel = 1280
		
		#定义网络结构参数
		if inverted_residual_setting is None:
			inverted_residual_setting = [
					[1, 16, 1, 1],
                	[6, 24, 2, 2],
                	[6, 32, 3, 2],
                	[6, 64, 4, 2],
                	[6, 96, 3, 1],
                	[6, 160, 3, 2],
                	[6, 320, 1, 1],
			]

		if len(inverted_residual_setting) == 0 or len(inverted_residual_setting[0]