完整教程：【无标题】

Day45

# 在data_loader.py中增强数据预处理def create_data_loaders(data_dir, batch_size, img_size):    # 训练集增强：随机裁剪、翻转、旋转、亮度调整等    train_transform = transforms.Compose([        transforms.RandomResizedCrop(img_size),        transforms.RandomHorizontalFlip(),        transforms.RandomRotation(10),        transforms.ColorJitter(brightness=0.2, contrast=0.2),        transforms.ToTensor(),        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])  # ImageNet标准    ])        # 测试集只需调整大小和标准化    test_transform = transforms.Compose([        transforms.Resize((img_size, img_size)),        transforms.ToTensor(),        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])    ])        # 加载数据集    train_dataset = datasets.ImageFolder(os.path.join(data_dir, 'train'), transform=train_transform)    test_dataset = datasets.ImageFolder(os.path.join(data_dir, 'test'), transform=test_transform)        # 创建数据加载器    train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True, num_workers=4)    test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False, num_workers=4)        return train_loader, test_loader, train_dataset.classes # 在train.py中添加早停和学习率调度def train_model(model, train_loader, test_loader, criterion, optimizer, num_epochs, device):    best_acc = 0.0    best_model_weights = copy.deepcopy(model.state_dict())    early_stop_counter = 0    early_stop_patience = 5  # 如果5个epoch没有提升，则停止训练        # 学习率调度器：当验证集准确率停滞时，降低学习率    scheduler = optim.lr_scheduler.ReduceLROnPlateau(        optimizer, mode='max', factor=0.5, patience=2, verbose=True    )        for epoch in range(num_epochs):        # 训练阶段        model.train()        running_loss = 0.0        correct = 0        total = 0                for batch_idx, (inputs, targets) in enumerate(train_loader):            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()                        if (batch_idx + 1) % 100 == 0:                print(f'Epoch: {epoch+1}/{num_epochs} | Batch: {batch_idx+1}/{len(train_loader)} '                      f'| 单Batch损失: {loss.item():.4f} | 累计平均损失: {running_loss/(batch_idx+1):.4f}')                train_loss = running_loss / len(train_loader)        train_acc = 100. * correct / total        print(f'Epoch {epoch+1}/{num_epochs} | 训练准确率: {train_acc:.2f}%')                # 验证阶段        model.eval()        test_loss = 0        correct = 0        total = 0                with torch.no_grad():            for inputs, targets in test_loader:                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()                test_acc = 100. * correct / total        print(f'Epoch {epoch+1}/{num_epochs} | 测试准确率: {test_acc:.2f}%')                # 更新学习率        scheduler.step(test_acc)                # 保存最佳模型        if test_acc > best_acc:            best_acc = test_acc            best_model_weights = copy.deepcopy(model.state_dict())            torch.save(best_model_weights, 'best_model.pth')            early_stop_counter = 0            print("已保存最佳模型")        else:            early_stop_counter += 1            print(f"早停计数器: {early_stop_counter}/{early_stop_patience}")                # 早停检查        if early_stop_counter >= early_stop_patience:            print(f"早停: 在{early_stop_patience}个epoch内验证集准确率没有提升")            break        # 加载最佳模型    model.load_state_dict(best_model_weights)    return model  # 在main.py中调整优化器参数# 定义损失函数和优化器，添加L2正则化criterion = nn.CrossEntropyLoss()optimizer = optim.Adam(model.parameters(), lr=learning_rate, weight_decay=1e-5)  # 添加L2正则化

@浙大疏锦行