python打卡day43

作业：

kaggle找到一个图像数据集，用cnn网络进行训练并且用grad-cam做可视化

 导入包

import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, transforms
from torch.utils.data import DataLoader
import matplotlib.pyplot as plt
import numpy as np
import os
from torchvision.datasets import ImageFolder
from torch.utils.data import DataLoader, Dataset, random_split

# 设置中文字体支持
plt.rcParams["font.family"] = ["SimHei"]
plt.rcParams['axes.unicode_minus'] = False  # 解决负号显示问题

# 检查GPU是否可用
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print(f"使用设备: {device}")

导入数据集

# 1. 数据预处理
# 训练集：使用多种数据增强方法提高模型泛化能力
IMG_SIZE = 224
data_path = 'c:/Users/83924/杭电/python/python60-days-challenge-master/python60-days-challenge-master/archive/dataset'
train_transform = transforms.Compose([
    transforms.Resize((IMG_SIZE, IMG_SIZE)),
    # 随机水平翻转图像（概率0.5）
    transforms.RandomHorizontalFlip(),
    # 随机颜色抖动：亮度、对比度、饱和度和色调随机变化
    transforms.ColorJitter(brightness=0.2, contrast=0.2, saturation=0.2, hue=0.1),
    # 随机旋转图像（最大角度15度）
    transforms.RandomRotation(15),
    # 将PIL图像或numpy数组转换为张量
    transforms.ToTensor(),
    # 标准化处理：每个通道的均值和标准差，使数据分布更合理
    transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))
])

# 测试集：仅进行必要的标准化，保持数据原始特性，标准化不损失数据信息，可还原
test_transform = transforms.Compose([
    transforms.Resize((IMG_SIZE, IMG_SIZE)),
    transforms.ToTensor(),
    transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))
])
# Full dataset
full_dataset = ImageFolder(data_path, transform=None)
num_classes = len(full_dataset.classes)
print(full_dataset.classes)
train_size=int(0.8*len(full_dataset))
test_size=len(full_dataset)-train_size
train_dataset,test_dataset=random_split(full_dataset,[train_size,test_size])
train_dataset.dataset.transform=train_transform
test_dataset.dataset.transform=test_transform
batch_size = 8
num_worker = os.cpu_count()
train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)
test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False)

模型

class CNN(nn.Module):
    def __init__(self):
        super(CNN, self).__init__()
        # 第一个卷积块
        self.conv1 = nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1)
        self.bn1 = nn.BatchNorm2d(64)
        self.relu1 = nn.ReLU()
        self.pool1 = nn.MaxPool2d(kernel_size=2, stride=2)

        # 第二个卷积块
        self.conv2 = nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1)
        self.bn2 = nn.BatchNorm2d(128)
        self.relu2 = nn.ReLU()
        self.pool2 = nn.MaxPool2d(kernel_size=2, stride=2)

        # 第三个卷积块  
        self.conv3 = nn.Conv2d(128, 256, kernel_size=3, stride=1, padding=1)
        self.bn3 = nn.BatchNorm2d(256)
        self.relu3 = nn.ReLU()
        self.pool3 = nn.MaxPool2d(kernel_size=2, stride=2)
        

        # 全连接层
        self.global_avg_pool = nn.AdaptiveAvgPool2d(1)
        self.fc1 = nn.Linear(256, 512)
        self.relu5 = nn.ReLU()
         # Dropout层：训练时随机丢弃50%神经元，防止过拟合
        self.dropout = nn.Dropout(p=0.5)
        self.fc2 = nn.Linear(512, num_classes)
        

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu1(x)
        x = self.pool1(x)

        x = self.conv2(x)
        x = self.bn2(x)
        x = self.relu2(x)
        x = self.pool2(x)

        x = self.conv3(x)
        x = self.bn3(x)
        x = self.relu3(x)
        x = self.pool3(x)

        x = self.global_avg_pool(x)
        x = x.view(x.size(0), -1)
        x = self.fc1(x)
        x = self.relu5(x)
        x = self.dropout(x)
        x = self.fc2(x)

        return x

训练

# 初始化模型
model = CNN()
model = model.to(device)  # 将模型移至GPU（如果可用）
criterion = nn.CrossEntropyLoss()  # 交叉熵损失函数
optimizer = optim.Adam(model.parameters(), lr=0.001)  # Adam优化器

# 引入学习率调度器，在训练过程中动态调整学习率--训练初期使用较大的 LR 快速降低损失，训练后期使用较小的 LR 更精细地逼近全局最优解。
# 在每个 epoch 结束后，需要手动调用调度器来更新学习率，可以在训练过程中调用 scheduler.step()
scheduler = optim.lr_scheduler.ReduceLROnPlateau(
    optimizer,        # 指定要控制的优化器（这里是Adam）
    mode='min',       # 监测的指标是"最小化"（如损失函数）
    patience=3,       # 如果连续3个epoch指标没有改善，才降低LR
    factor=0.5        # 降低LR的比例（新LR = 旧LR × 0.5）
)
# 5. 训练模型（记录每个 iteration 的损失）
def train(model, train_loader, test_loader, criterion, optimizer, scheduler, device, epochs):
    model.train()  # 设置为训练模式
    
    # 记录每个 iteration 的损失
    all_iter_losses = []  # 存储所有 batch 的损失
    iter_indices = []     # 存储 iteration 序号
    
    # 记录每个 epoch 的准确率和损失
    train_acc_history = []
    test_acc_history = []
    train_loss_history = []
    test_loss_history = []
    
    for epoch in range(epochs):
        running_loss = 0.0
        correct = 0
        total = 0
        
        for batch_idx, (data, target) in enumerate(train_loader):
            data, target = data.to(device), target.to(device)  # 移至GPU
            
            optimizer.zero_grad()  # 梯度清零
            output = model(data)  # 前向传播
            loss = criterion(output, target)  # 计算损失
            loss.backward()  # 反向传播
            optimizer.step()  # 更新参数
            
            # 记录当前 iteration 的损失
            iter_loss = loss.item()
            all_iter_losses.append(iter_loss)
            iter_indices.append(epoch * len(train_loader) + batch_idx + 1)
            
            # 统计准确率和损失
            running_loss += iter_loss
            _, predicted = output.max(1)
            total += target.size(0)
            correct += predicted.eq(target).sum().item()
            
            # 每100个批次打印一次训练信息
            if (batch_idx + 1) % 100 == 0:
                print(f'Epoch: {epoch+1}/{epochs} | Batch: {batch_idx+1}/{len(train_loader)} '
                      f'| 单Batch损失: {iter_loss:.4f} | 累计平均损失: {running_loss/(batch_idx+1):.4f}')
        
        # 计算当前epoch的平均训练损失和准确率
        epoch_train_loss = running_loss / len(train_loader)
        epoch_train_acc = 100. * correct / total
        train_acc_history.append(epoch_train_acc)
        train_loss_history.append(epoch_train_loss)
        
        # 测试阶段
        model.eval()  # 设置为评估模式
        test_loss = 0
        correct_test = 0
        total_test = 0
        
        with torch.no_grad():
            for data, target in test_loader:
                data, target = data.to(device), target.to(device)
                output = model(data)
                test_loss += criterion(output, target).item()
                _, predicted = output.max(1)
                total_test += target.size(0)
                correct_test += predicted.eq(target).sum().item()
        
        epoch_test_loss = test_loss / len(test_loader)
        epoch_test_acc = 100. * correct_test / total_test
        test_acc_history.append(epoch_test_acc)
        test_loss_history.append(epoch_test_loss)
        
        # 更新学习率调度器
        scheduler.step(epoch_test_loss)
        
        print(f'Epoch {epoch+1}/{epochs} 完成 | 训练准确率: {epoch_train_acc:.2f}% | 测试准确率: {epoch_test_acc:.2f}%')
    
    # 绘制所有 iteration 的损失曲线
    plot_iter_losses(all_iter_losses, iter_indices)
    
    # 绘制每个 epoch 的准确率和损失曲线
    plot_epoch_metrics(train_acc_history, test_acc_history, train_loss_history, test_loss_history)
    
    return epoch_test_acc  # 返回最终测试准确率

# 6. 绘制每个 iteration 的损失曲线
def plot_iter_losses(losses, indices):
    plt.figure(figsize=(10, 4))
    plt.plot(indices, losses, 'b-', alpha=0.7, label='Iteration Loss')
    plt.xlabel('Iteration（Batch序号）')
    plt.ylabel('损失值')
    plt.title('每个 Iteration 的训练损失')
    plt.legend()
    plt.grid(True)
    plt.tight_layout()
    plt.show()

# 7. 绘制每个 epoch 的准确率和损失曲线
def plot_epoch_metrics(train_acc, test_acc, train_loss, test_loss):
    epochs = range(1, len(train_acc) + 1)
    
    plt.figure(figsize=(12, 4))
    
    # 绘制准确率曲线
    plt.subplot(1, 2, 1)
    plt.plot(epochs, train_acc, 'b-', label='训练准确率')
    plt.plot(epochs, test_acc, 'r-', label='测试准确率')
    plt.xlabel('Epoch')
    plt.ylabel('准确率 (%)')
    plt.title('训练和测试准确率')
    plt.legend()
    plt.grid(True)
    
    # 绘制损失曲线
    plt.subplot(1, 2, 2)
    plt.plot(epochs, train_loss, 'b-', label='训练损失')
    plt.plot(epochs, test_loss, 'r-', label='测试损失')
    plt.xlabel('Epoch')
    plt.ylabel('损失值')
    plt.title('训练和测试损失')
    plt.legend()
    plt.grid(True)
    
    plt.tight_layout()
    plt.show()

# 8. 执行训练和测试
epochs = 20  # 增加训练轮次以获得更好效果
print("开始使用CNN训练模型...")
final_accuracy = train(model, train_loader, test_loader, criterion, optimizer, scheduler, device, epochs)
print(f"训练完成！最终测试准确率: {final_accuracy:.2f}%")

# # 保存模型
# torch.save(model.state_dict(), 'cifar10_cnn_model.pth')
# print("模型已保存为: cifar10_cnn_model.pth")

训练完成！最终测试准确率: 78.59%

Grad-CAM

import warnings
warnings.filterwarnings("ignore")
import matplotlib.pyplot as plt
# 设置中文字体支持
plt.rcParams["font.family"] = ["SimHei"]
plt.rcParams['axes.unicode_minus'] = False  # 解决负号显示问题
# 选择一个随机图像
classes = ('dew', 'fogsmog', 'frost', 'glaze', 'hail', 'lightning', 'rain', 'rainbow', 'rime', 'sandstorm', 'snow')
# idx = np.random.randint(len(testset))
idx = 44  # 选择测试集中的第101张图片 (索引从0开始)
image, label = test_dataset[idx]
print(f"选择的图像类别: {classes[label]}")
# 定义类别名称，CIFAR-10数据集包含这10个类别

# 转换图像以便可视化
def tensor_to_np(tensor):
    img = tensor.cpu().numpy().transpose(1, 2, 0)
    mean = np.array([0.4914, 0.4822, 0.4465])
    std = np.array([0.2470, 0.2435, 0.2616])
    img = std * img + mean
    img = np.clip(img, 0, 1)
    return img

# 添加批次维度并移动到设备
input_tensor = image.unsqueeze(0).to(device)

# 初始化Grad-CAM（选择最后一个卷积层）
grad_cam = GradCAM(model, model.conv3)

# 生成热力图
heatmap, pred_class = grad_cam.generate_cam(input_tensor)

# 可视化
plt.figure(figsize=(12, 4))

# 原始图像
plt.subplot(1, 3, 1)
plt.imshow(tensor_to_np(image))
plt.title(f"原始图像: {classes[label]}")
plt.axis('off')

# 热力图
plt.subplot(1, 3, 2)
plt.imshow(heatmap, cmap='jet')
plt.title(f"Grad-CAM热力图: {classes[label]}")
plt.axis('off')

# 叠加的图像
plt.subplot(1, 3, 3)
img = tensor_to_np(image)
heatmap_resized = np.uint8(255 * heatmap)
heatmap_colored = plt.cm.jet(heatmap_resized)[:, :, :3]
superimposed_img = heatmap_colored * 0.4 + img * 0.6
plt.imshow(superimposed_img)
plt.title("叠加热力图")
plt.axis('off')

plt.tight_layout()
plt.savefig('grad_cam_result.png')
plt.show()

# print("Grad-CAM可视化完成。已保存为grad_cam_result.png")

@浙大疏锦行