pip install torch torchvision matplotlib
在这个项目中，我们将使用一个包含多标签图像的数据集。您可以选择适用于您项目的多标签图像数据集，确保每张图像都标有一个或多个标签。在这里，我们以简单的示例使用PyTorch内置的CIFAR-100数据集。
import torchfrom torchvision import transforms, datasets# 定义数据变换transform = transforms.Compose([ transforms.ToTensor(),])# 下载并准备CIFAR-100数据集train_dataset = datasets.CIFAR100(root='./data', train=True, download=True, transform=transform)test_dataset = datasets.CIFAR100(root='./data', train=False, download=True, transform=transform)步骤3：定义数据加载器
# 定义数据加载器batch_size = 64train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=True)test_loader = torch.utils.data.DataLoader(test_dataset, batch_size=batch_size, shuffle=False)步骤4：构建多标签图像分类模型
import torch.nn as nnimport torch.optim as optim# 定义多标签图像分类模型class MultiLabelClassifier(nn.Module): def __init__(self, num_classes): super(MultiLabelClassifier, self).__init__() self.conv1 = nn.Conv2d(3, 64, kernel_size=3, padding=1) self.relu = nn.ReLU() self.pool = nn.MaxPool2d(kernel_size=2, stride=2) self.fc1 = nn.Linear(64 16 16, num_classes) def forward(self, x): x = self.conv1(x) x = self.relu(x) x = self.pool(x) x = x.view(-1, 64 16 16) x = self.fc1(x) return x# 初始化模型num_classes = 100 # CIFAR-100有100个类别model = MultiLabelClassifier(num_classes)# 定义损失函数和优化器criterion = nn.CrossEntropyLoss()optimizer = optim.Adam(model.parameters(), lr=0.001)步骤5：训练模型
# 训练模型num_epochs = 10device = torch.device("cuda" if torch.cuda.is_available() else "cpu")model.to(device)for epoch in range(num_epochs): model.train() running_loss = 0.0 for inputs, labels in train_loader: inputs, labels = inputs.to(device), labels.to(device) optimizer.zero_grad() outputs = model(inputs) loss = criterion(outputs, labels) loss.backward() optimizer.step() running_loss += loss.item() print(f"Epoch {epoch + 1}/{num_epochs}, Loss: {running_loss / len(train_loader)}")步骤6：评估模型
# 评估模型model.eval()correct = 0total = 0with torch.no_grad(): for inputs, labels in test_loader: inputs, labels = inputs.to(device), labels.to(device) outputs = model(inputs) _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum().item()accuracy = correct / totalprint(f"Test Accuracy: {accuracy 100:.2f}%")
通过这个实际项目，您学到了如何构建一个多标签图像分类器，包括准备数据、定义模型、训练模型和评估模型。希望这个实战教程对您在PyTorch项目开发中的实际应用有所帮助！
