day43 打卡

数据集为眼睛检测数据集 

import kagglehub# Download latest version
path = kagglehub.dataset_download("icebearogo/eye-detection-dataset")print("Path to dataset files:", path)

import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, transforms
from torch.utils.data import DataLoader, random_split
import matplotlib.pyplot as plt
import numpy as np# 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}")# 1. 数据预处理（根据eye-detection数据集调整尺寸和增强策略）
# 假设eye图像为彩色，尺寸可能不统一，先resize到224x224（常见模型输入尺寸）
train_transform = transforms.Compose([transforms.Resize((224, 224)),  # 统一图像尺寸transforms.RandomCrop(224, padding=16),  # 随机裁剪（根据数据集特性调整padding）transforms.RandomHorizontalFlip(),  # 随机水平翻转（适合眼检测任务）transforms.ColorJitter(brightness=0.2, contrast=0.2),  # 颜色抖动（简化，避免干扰特征）transforms.RandomRotation(10),  # 小角度旋转（模拟不同视角）transforms.ToTensor(),# 若没有数据集统计均值标准差，可暂时不标准化或使用ImageNet默认值transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])  # ImageNet默认值
])test_transform = transforms.Compose([transforms.Resize((224, 224)),transforms.ToTensor(),transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
])# 2. 加载本地eye-detection数据集（假设数据集结构为：root/类别名/图像文件）
dataset_path = "C:\\Users\\huawei\\.cache\\kagglehub\\datasets\\icebearogo\\eye-detection-dataset\\versions\\1"# 使用ImageFolder加载自定义图像数据集（要求：每个类别一个子文件夹）
full_dataset = datasets.ImageFolder(root=dataset_path,transform=train_transform  # 训练集使用增强，测试集后续单独调整
)# 手动划分训练集和测试集（原CIFAR-10已内置划分，自定义数据集需手动拆分）
train_size = int(0.8 * len(full_dataset))  # 80%训练集
test_size = len(full_dataset) - train_size
train_dataset, test_dataset = random_split(full_dataset, [train_size, test_size])# 3. 为测试集单独设置transform（避免训练增强影响测试）
test_dataset.dataset.transform = test_transform  # 关键调整：覆盖测试集的transform# 4. 创建数据加载器
batch_size = 32  # 根据GPU内存调整
(train_loader, test_loader) = (DataLoader(train_dataset, batch_size=batch_size, shuffle=True, num_workers=2),DataLoader(test_dataset, batch_size=batch_size, shuffle=False, num_workers=2)
)

class CNN(nn.Module):def __init__(self):super(CNN, self).__init__()# 第一个卷积块（输入224x224，输出112x112）self.conv1 = nn.Conv2d(3, 32, kernel_size=3, padding=1)self.bn1 = nn.BatchNorm2d(32)self.relu1 = nn.ReLU()self.pool1 = nn.MaxPool2d(kernel_size=2, stride=2)  # 224→112# 第二个卷积块（输入112x112，输出56x56）self.conv2 = nn.Conv2d(32, 64, kernel_size=3, padding=1)self.bn2 = nn.BatchNorm2d(64)self.relu2 = nn.ReLU()self.pool2 = nn.MaxPool2d(kernel_size=2, stride=2)  # 112→56# 第三个卷积块（输入56x56，输出28x28）self.conv3 = nn.Conv2d(64, 128, kernel_size=3, padding=1)self.bn3 = nn.BatchNorm2d(128)self.relu3 = nn.ReLU()self.pool3 = nn.MaxPool2d(kernel_size=2, stride=2)  # 56→28# 全连接层（展平后维度：128通道 × 28×28尺寸）self.fc1 = nn.Linear(128 * 28 * 28, 512)self.dropout = nn.Dropout(0.5)self.fc2 = nn.Linear(512, 10)  # 假设目标类别数为10def forward(self, x):# 输入尺寸：[batch_size, 3, 224, 224]x = self.conv1(x)  # [batch_size, 32, 224, 224]x = self.bn1(x)x = self.relu1(x)x = self.pool1(x)  # [batch_size, 32, 112, 112]x = self.conv2(x)  # [batch_size, 64, 112, 112]x = self.bn2(x)x = self.relu2(x)x = self.pool2(x)  # [batch_size, 64, 56, 56]x = self.conv3(x)  # [batch_size, 128, 56, 56]x = self.bn3(x)x = self.relu3(x)x = self.pool3(x)  # [batch_size, 128, 28, 28]x = x.view(x.size(0), -1)  # 展平为 [batch_size, 128*28*28]x = self.fc1(x)x = self.relu3(x)x = self.dropout(x)x = self.fc2(x)return x# 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LRfactor=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 = 0.0correct = 0total = 0for batch_idx, (data, target) in enumerate(train_loader):data, target = data.to(device), target.to(device)  # 移至GPUoptimizer.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 / totaltrain_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 = 0correct_test = 0total_test = 0with 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_testtest_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 = 2  # 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}%")

@浙大疏锦行