ResNet网络搭建

最新推荐文章于 2024-05-30 16:08:30 发布

YDoMore

最新推荐文章于 2024-05-30 16:08:30 发布

阅读量166

点赞数

文章标签：网络深度学习 python

本文链接：https://blog.csdn.net/YZBFOREVERNB/article/details/127330298

版权

import torch
import torch.nn as nn
import torch.nn.functional as F
#基本单元  完成跳连结构前向推理
class ResBlock(nn.Module):
    def __init__(self,in_channel,out_channel,stride = 1):
        super(ResBlock, self).__init__()
        #主干分支
        self.layer = nn.Sequential(
            nn.Conv2d(in_channel, out_channel, kernel_size=3, stride=stride, padding=1),
            nn.BatchNorm2d(out_channel),
            nn.ReLU(),
            nn.Conv2d(out_channel, out_channel, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(out_channel),
        )
        self.shortcut = nn.Sequential()
        #判断in_channel与out_channel是否相同 以及stride是否为1，再去决定是否卷积操作
        if in_channel != out_channel or stride >1:
            #跳连分支 将输入结果直连layer层
            self.shortcut = nn.Sequential(
                nn.Conv2d(in_channel, out_channel, kernel_size=3, stride=stride, padding=1),
                nn.BatchNorm2d(out_channel),
            )

    def forward(self,x):
        out1 = self.layer(x)
        out2 = self.shortcut(x)
        out = out1 + out2 #out1 out2 下采样倍数需要相同 不然无法相加
        out = F.relu(out)
        return out

class ResNet(nn.Module):
    #定义多个重复的层
    def make_layer(self,block,out_channel,stride,num_block):
        layers_list = []
        for i in range(num_block):
            if i == 0:
                in_stride = stride
            else:
                in_stride = 1
            layers_list.append(block(self.in_channel,out_channel,in_stride))
            self.in_channel = out_channel
        return nn.Sequential(*layers_list)


    def __init__(self,ResBlock):
        super(ResNet, self).__init__()
        self.in_channel = 32
        #第一次采用标准卷积
        self.conv1 = nn.Sequential(
            nn.Conv2d(3, 32, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(32),
            nn.ReLU()
        )

        self.layer1 =  \
            self.make_layer(ResBlock,64,2,2)
        self.layer2 =  \
            self.make_layer(ResBlock,128,2,2)
        self.layer3 =  \
            self.make_layer(ResBlock,256,2,2)
        self.layer4 =  \
            self.make_layer(ResBlock,512,2,2)
        self.fc = nn.Linear(512,10)

    def forward(self,x):
        out = self.conv1(x)
        out = self.layer1(out)
        out = self.layer2(out)
        out = self.layer3(out)
        out = self.layer4(out)
        out = F.avg_pool2d(out,2)
        out = out.view(out.size(0),-1)
        out = self.fc(out)
        return out

def resnet():
    return ResNet(ResBlock)