从0写一个ResNet网络，并在最后讲述训练过程

在这篇文章中，我将从头开始实现一个ResNet（残差网络），并展示如何训练这个网络。ResNet是由何凯明等人在2015年提出的，它通过引入残差模块解决了深层神经网络训练困难的问题。

1. 什么是ResNet？

ResNet的核心思想是通过引入“残差块”（Residual Block）来构建深层网络。残差块通过跳跃连接（skip connection）使得梯度可以直接传递，从而缓解了梯度消失问题。

2. 实现ResNet

我们将使用PyTorch来实现ResNet。首先，我们需要定义一个残差块。

import torch
import torch.nn as nn

class ResidualBlock(nn.Module):
    def __init__(self, in_channels, out_channels, stride=1, downsample=None):
        super(ResidualBlock, self).__init__()
        self.conv1 = nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=stride, padding=1, bias=False)
        self.bn1 = nn.BatchNorm2d(out_channels)
        self.relu = nn.ReLU(inplace=True)
        self.conv2 = nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=1, padding=1, bias=False)
        self.bn2 = nn.BatchNorm2d(out_channels)
        self.downsample = downsample

    def forward(self, x):
        identity = x
        if self.downsample is not None:
            identity = self.downsample(x)
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)
        out = self.conv2(out)
        out = self.bn2(out)
        out += identity
        out = self.relu(out)
        return out

接下来，我们定义ResNet网络结构。

class ResNet(nn.Module):
    def __init__(self, block, layers, num_classes=1000):
        super(ResNet, self).__init__()
        self.in_channels = 64
        self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3, bias=False)
        self.bn1 = nn.BatchNorm2d(64)
        self.relu = nn.ReLU(inplace=True)
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
        self.layer1 = self._make_layer(block, 64, layers[0])
        self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
        self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
        self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
        self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
        self.fc = nn.Linear(512 * block.expansion, num_classes)

    def _make_layer(self, block, out_channels, blocks, stride=1):
        downsample = None
        if stride != 1 or self.in_channels != out_channels * block.expansion:
            downsample = nn.Sequential(
                nn.Conv2d(self.in_channels, out_channels * block.expansion, kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(out_channels * block.expansion),
            )
        layers = []
        layers.append(block(self.in_channels, out_channels, stride, downsample))
        self.in_channels = out_channels * block.expansion
        for _ in range(1, blocks):
            layers.append(block(self.in_channels, out_channels))
        return nn.Sequential(*layers)

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)
        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)
        x = self.avgpool(x)
        x = torch.flatten(x, 1)
        x = self.fc(x)
        return x

我们可以通过以下代码来实例化一个ResNet-18模型：

def resnet18():
    return ResNet(ResidualBlock, [2, 2, 2, 2])

3. 训练ResNet

接下来，我们将展示如何训练这个ResNet模型。我们将使用CIFAR-10数据集进行训练。

import torch.optim as optim
import torchvision
import torchvision.transforms as transforms

# 数据预处理
transform = transforms.Compose([
    transforms.RandomResizedCrop(224),
    transforms.RandomHorizontalFlip(),
    transforms.ToTensor(),
    transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),
])

trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=128, shuffle=True, num_workers=2)

testset = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=100, shuffle=False, num_workers=2)

# 实例化模型、定义损失函数和优化器
net = resnet18()
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(), lr=0.1, momentum=0.9, weight_decay=5e-4)

# 训练模型
for epoch in range(10):  # 训练10个epoch
    net.train()
    running_loss = 0.0
    for i, data in enumerate(trainloader, 0):
        inputs, labels = data
        optimizer.zero_grad()
        outputs = net(inputs)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()
        running_loss += loss.item()
        if i % 100 == 99:  # 每100个batch打印一次损失
            print(f'[Epoch {epoch + 1}, Batch {i + 1}] loss: {running_loss / 100:.3f}')
            running_loss = 0.0

print('Finished Training')

4. 总结

在这篇文章中，我们从头实现了一个ResNet网络，并展示了如何在CIFAR-10数据集上进行训练。通过引入残差块，ResNet能够有效地训练深层神经网络。如果你有任何问题或建议，欢迎在评论区留言。