ResNet18 结构分析

一、概述

ResNet（残差神经网络）于 2015 年提出，是 ImageNet 图像分类任务上的冠军。其核心思想是通过引入跨层连接（shortcut connection）解决了深层神经网络训练过程中梯度消失和梯度爆炸的问题。ResNet18 是 ResNet 的一个较小版本，共有 18 层，包括 16 层卷积神经网络层和 2 层全连接层。

二、ResNet18 细节

1.卷积层

ResNet18 的卷积层共包含16层，在 3×3 的卷积核后，采用 ReLU 激活函数，步长为1，不使用池化层。其中，前7层卷积没有跨层连接，8-16 层使用了跨层连接。具体来说，8-16 层中每隔一个残差块就有一条跨层连接，用于将上一层特征图与下一层残差块的输出相加。

2.残差块

ResNet18 的每个残差块包含 2-3 个卷积层，每个卷积层都跟随着 Batch Normalization 层和 ReLU 激活函数。具体来说，第一个卷积层的卷积核大小为 3 × 3，第二个卷积层的卷积核大小也为 3 × 3。如果跨层连接存在，还需要添加一层 1 × 1 的卷积层用于调整维度。

3.全连接层

ResNet18 的最后两层全连接层都含有 512 个神经元，倒数第二层使用 ReLU 激活函数，而最后一层使用 Softmax 函数产生对类的预测输出。

三、代码示例

1.定义 ResNet18 结构

import torch.nn as nn

class BasicBlock(nn.Module):
    expansion = 1

    def __init__(self, in_planes, planes, stride=1):
        super(BasicBlock, self).__init__()
        self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, stride=stride, padding=1, bias=False)
        self.bn1 = nn.BatchNorm2d(planes)
        self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=1, padding=1, bias=False)
        self.bn2 = nn.BatchNorm2d(planes)

        self.shortcut = nn.Sequential()
        if stride != 1 or in_planes != self.expansion*planes:
            self.shortcut = nn.Sequential(
                nn.Conv2d(in_planes, self.expansion*planes, kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(self.expansion*planes)
            )

    def forward(self, x):
        out = nn.functional.relu(self.bn1(self.conv1(x)))
        out = self.bn2(self.conv2(out))
        out += self.shortcut(x)
        out = nn.functional.relu(out)
        return out

class ResNet18(nn.Module):
    def __init__(self, num_classes=10):
        super(ResNet18, self).__init__()
        self.in_planes = 64

        self.conv1 = nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1, bias=False)
        self.bn1 = nn.BatchNorm2d(64)
        self.layer1 = self._make_layer(64, 2, stride=1)
        self.layer2 = self._make_layer(128, 2, stride=2)
        self.layer3 = self._make_layer(256, 2, stride=2)
        self.layer4 = self._make_layer(512, 2, stride=2)
        self.linear = nn.Linear(512*BasicBlock.expansion, num_classes)

    def _make_layer(self, planes, num_blocks, stride):
        strides = [stride] + [1]*(num_blocks-1)
        layers = []
        for stride in strides:
            layers.append(BasicBlock(self.in_planes, planes, stride))
            self.in_planes = planes * BasicBlock.expansion
        return nn.Sequential(*layers)

    def forward(self, x):
        out = nn.functional.relu(self.bn1(self.conv1(x)))
        out = self.layer1(out)
        out = self.layer2(out)
        out = self.layer3(out)
        out = self.layer4(out)
        out = nn.functional.avg_pool2d(out, 4)
        out = out.view(out.size(0), -1)
        out = self.linear(out)
        return out

2.实例化 ResNet18 模型

resnet18 = ResNet18()

3.模型训练

criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(resnet18.parameters())

for epoch in range(num_epochs):
    running_loss = 0.0
    for i, data in enumerate(trainloader, 0):
        inputs, labels = data
        optimizer.zero_grad()
        outputs = resnet18(inputs)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()
        
        running_loss += loss.item()
        if i % 2000 == 1999:    
            print('[%d, %5d] loss: %.3f' %
                  (epoch + 1, i + 1, running_loss / 2000))
            running_loss = 0.0