PyTorch实现CIFAR100图像分类:从基础到高级模型

pytorch-cifar100

PyTorch实现CIFAR100图像分类:从基础到高级模型

CIFAR100是计算机视觉领域广泛使用的图像分类数据集之一。本文将详细介绍如何使用PyTorch在CIFAR100数据集上实现图像分类,涵盖从简单到复杂的多种深度学习模型,并分析它们的性能表现。

CIFAR100数据集介绍

CIFAR100数据集包含60,000张32x32的彩色图像,分为100个类别,每个类别600张图像。其中50,000张用于训练,10,000张用于测试。这个数据集的特点是:

图像尺寸小(32x32像素)
类别数量多(100个类别)
每个类别的样本数量相对较少(600张/类)

这些特点使得CIFAR100成为一个具有挑战性的图像分类任务,非常适合用来评估和比较不同的深度学习模型。

环境配置

要在CIFAR100上训练模型,我们需要以下环境:

Python 3.6+
PyTorch 1.6.0+
torchvision
tensorboard (可选,用于可视化)

可以使用以下命令安装所需的包:

pip install torch torchvision tensorboard

数据加载与预处理

PyTorch提供了方便的API来加载CIFAR100数据集:

from torchvision import datasets, transforms

# 定义数据预处理
transform_train = transforms.Compose([
    transforms.RandomCrop(32, padding=4),
    transforms.RandomHorizontalFlip(),
    transforms.ToTensor(),
    transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])

transform_test = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])

# 加载数据集
trainset = datasets.CIFAR100(root='./data', train=True, download=True, transform=transform_train)
testset = datasets.CIFAR100(root='./data', train=False, download=True, transform=transform_test)

# 创建数据加载器
trainloader = torch.utils.data.DataLoader(trainset, batch_size=128, shuffle=True, num_workers=2)
testloader = torch.utils.data.DataLoader(testset, batch_size=100, shuffle=False, num_workers=2)

这里我们对训练集应用了数据增强技术(随机裁剪和水平翻转),以提高模型的泛化能力。

模型架构

接下来,我们将介绍几种不同复杂度的模型架构,从简单的逻辑回归到复杂的ResNet。

1. 逻辑回归

逻辑回归是最简单的分类模型,可以作为基线模型:

class LogisticRegression(nn.Module):
    def __init__(self):
        super(LogisticRegression, self).__init__()
        self.linear = nn.Linear(32*32*3, 100)
    
    def forward(self, x):
        x = x.view(x.size(0), -1)
        return self.linear(x)

2. 简单卷积神经网络(CNN)

一个基础的CNN模型:

class SimpleCNN(nn.Module):
    def __init__(self):
        super(SimpleCNN, self).__init__()
        self.conv1 = nn.Conv2d(3, 32, 3, padding=1)
        self.conv2 = nn.Conv2d(32, 64, 3, padding=1)
        self.pool = nn.MaxPool2d(2, 2)
        self.fc1 = nn.Linear(64 * 8 * 8, 512)
        self.fc2 = nn.Linear(512, 100)

    def forward(self, x):
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = x.view(-1, 64 * 8 * 8)
        x = F.relu(self.fc1(x))
        x = self.fc2(x)
        return x

3. ResNet18

ResNet是一种更深层的网络架构,通过残差连接解决了深度网络的梯度消失问题:

class ResidualBlock(nn.Module):
    def __init__(self, in_channels, out_channels, stride=1):
        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.conv2 = nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=1, padding=1, bias=False)
        self.bn2 = nn.BatchNorm2d(out_channels)

        self.shortcut = nn.Sequential()
        if stride != 1 or in_channels != out_channels:
            self.shortcut = nn.Sequential(
                nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(out_channels)
            )

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

class ResNet18(nn.Module):
    def __init__(self, num_classes=100):
        super(ResNet18, self).__init__()
        self.in_channels = 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(ResidualBlock, 64, 2, stride=1)
        self.layer2 = self._make_layer(ResidualBlock, 128, 2, stride=2)
        self.layer3 = self._make_layer(ResidualBlock, 256, 2, stride=2)
        self.layer4 = self._make_layer(ResidualBlock, 512, 2, stride=2)
        self.fc = nn.Linear(512, num_classes)

    def _make_layer(self, block, out_channels, num_blocks, stride):
        strides = [stride] + [1] * (num_blocks - 1)
        layers = []
        for stride in strides:
            layers.append(block(self.in_channels, out_channels, stride))
            self.in_channels = out_channels
        return nn.Sequential(*layers)

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

模型训练

以下是一个通用的训练函数,可以用于训练上述所有模型:

def train(model, trainloader, criterion, optimizer, device):
    model.train()
    running_loss = 0.0
    correct = 0
    total = 0
    for batch_idx, (inputs, targets) in enumerate(trainloader):
        inputs, targets = inputs.to(device), targets.to(device)
        optimizer.zero_grad()
        outputs = model(inputs)
        loss = criterion(outputs, targets)
        loss.backward()
        optimizer.step()

        running_loss += loss.item()
        _, predicted = outputs.max(1)
        total += targets.size(0)
        correct += predicted.eq(targets).sum().item()

    accuracy = 100. * correct / total
    return running_loss / len(trainloader), accuracy

def test(model, testloader, criterion, device):
    model.eval()
    test_loss = 0
    correct = 0
    total = 0
    with torch.no_grad():
        for batch_idx, (inputs, targets) in enumerate(testloader):
            inputs, targets = inputs.to(device), targets.to(device)
            outputs = model(inputs)
            loss = criterion(outputs, targets)

            test_loss += loss.item()
            _, predicted = outputs.max(1)
            total += targets.size(0)
            correct += predicted.eq(targets).sum().item()

    accuracy = 100. * correct / total
    return test_loss / len(testloader), accuracy