@浙大疏锦行
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 # 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 # 训练集：使用多种数据增强方法提高模型泛化能力 train_transform = transforms.Compose([ # 随机裁剪图像，从原图中随机截取32x32大小的区域 transforms.RandomCrop(32, padding=4), # 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.ToTensor(), transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)) ]) # 2. 加载CIFAR-10数据集 train_dataset = datasets.CIFAR10( root='./data', train=True, download=True, transform=train_transform # 使用增强后的预处理 ) test_dataset = datasets.CIFAR10( root='./data', train=False, transform=test_transform # 测试集不使用增强 ) # 3. 创建数据加载器 batch_size = 64 train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True) test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False)import torch import torch.nn as nn import torch.optim as optim import torch.nn.functional as F # 设置设备 device = torch.device("cuda" if torch.cuda.is_available() else "cpu") print(f"使用设备: {device}") # 1. 定义CNN模型 class CNN(nn.Module): def __init__(self): super(CNN, self).__init__() # ---------------------- 卷积特征提取部分 ---------------------- # 第一个卷积块 self.conv_block1 = nn.Sequential( nn.Conv2d(3, 32, kernel_size=3, padding=1), # [batch, 3, 32, 32] -> [batch, 32, 32, 32] nn.BatchNorm2d(32), nn.ReLU(inplace=True), nn.MaxPool2d(2) # [batch, 32, 32, 32] -> [batch, 32, 16, 16] ) # 第二个卷积块 self.conv_block2 = nn.Sequential( nn.Conv2d(32, 64, kernel_size=3, padding=1), # [batch, 32, 16, 16] -> [batch, 64, 16, 16] nn.BatchNorm2d(64), nn.ReLU(inplace=True), nn.MaxPool2d(2) # [batch, 64, 16, 16] -> [batch, 64, 8, 8] ) # 第三个卷积块 self.conv_block3 = nn.Sequential( nn.Conv2d(64, 128, kernel_size=3, padding=1), # [batch, 64, 8, 8] -> [batch, 128, 8, 8] nn.BatchNorm2d(128), nn.ReLU(inplace=True), nn.MaxPool2d(2) # [batch, 128, 8, 8] -> [batch, 128, 4, 4] ) # ---------------------- 全连接分类部分 ---------------------- self.classifier = nn.Sequential( nn.Linear(128 * 4 * 4, 512), nn.ReLU(inplace=True), nn.Dropout(0.5), nn.Linear(512, 256), nn.ReLU(inplace=True), nn.Dropout(0.3), nn.Linear(256, 10) ) def forward(self, x): # 卷积特征提取 x = self.conv_block1(x) x = self.conv_block2(x) x = self.conv_block3(x) # 展平 x = x.view(x.size(0), -1) # [batch, 128, 4, 4] -> [batch, 2048] # 分类 x = self.classifier(x) return x # 2. 初始化模型 model = CNN().to(device) print(f"模型参数量: {sum(p.numel() for p in model.parameters()):,}") print(f"可训练参数量: {sum(p.numel() for p in model.parameters() if p.requires_grad):,}") # 3. 定义损失函数和优化器 criterion = nn.CrossEntropyLoss() optimizer = optim.Adam(model.parameters(), lr=0.001, weight_decay=1e-4) # 添加L2正则化 scheduler = optim.lr_scheduler.ReduceLROnPlateau( optimizer, # 第一个参数是optimizer，不要用关键字参数 mode='min', factor=0.5, patience=5, threshold=0.01, min_lr=1e-5 )训练完成！最终测试准确率: 79.96%
