图片预处理：读取训练集和测试集的个数

dataset_path = './data3_dir/'

dataset = datasets.ImageFolder(root=dataset_path, transform=transform)

# 划分数据集为训练集、测试集和验证集

train_size = int(0.8 * len(dataset))

#val_size =  int(0.1 * len(dataset))

test_size=len(dataset) - train_size

#训练集、验证集、测试集比例分为8.1.1

train_dataset,test_dataset = random_split(dataset, [train_size,test_size])

# 创建数据加载器

batch_size = 32

train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)

#val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False)

test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False)

# 输出训练集和测试集的数量

print(f"训练集数量: {len(train_dataset)}")

print(f"测试集数量: {len(test_dataset)}")

# 在训练循环之前检查测试数据

for images, labels in train_loader:

    print("Train data shape:", images.shape)

    print("Train labels shape:", labels.shape)

    break

# 在测试循环之前检查测试数据

for images, labels in test_loader:

    print("Test data shape:", images.shape)

    print("Test labels shape:", labels.shape)

    break