前言

        延续之前所讲，基本上项目代码都是数据集的读入和处理，模型定义、超参数定义以及训练流程，那么接下来也是按照这个顺序进行。

数据集读入和处理

class food_Dataset(Dataset):
    def __init__(self, path, mode="train"):
        self.mode = mode
        if mode == "semi":
            self.X = self.read_file(path)
        else:
            self.X, self.Y = self.read_file(path)
            self.Y = torch.LongTensor(self.Y)  #标签转为长整形\

        if mode == "train":
            self.transform = train_transform
        else:
            self.transform = val_transform

    def read_file(self, path):
        if self.mode == "semi":
            file_list = os.listdir(path)
            xi = np.zeros((len(file_list), HW, HW, 3), dtype=np.uint8)
            # 列出文件夹下所有文件名字
            for j, img_name in enumerate(file_list):
                img_path = os.path.join(path, img_name)
                img = Image.open(img_path)
                img = img.resize((HW, HW))
                xi[j, ...] = img
            print("读到了%d个数据" % len(xi))
            return xi
        else:
            for i in tqdm(range(11)):
                file_dir = path + "/%02d" % i
                file_list = os.listdir(file_dir)

                xi = np.zeros((len(file_list), HW, HW, 3), dtype=np.uint8)
                yi = np.zeros(len(file_list), dtype=np.uint8)

                # 列出文件夹下所有文件名字
                for j, img_name in enumerate(file_list):
                    img_path = os.path.join(file_dir, img_name)
                    img = Image.open(img_path)
                    img = img.resize((HW, HW))
                    xi[j, ...] = img
                    yi[j] = i

                if i == 0:
                    X = xi
                    Y = yi
                else:
                    X = np.concatenate((X, xi), axis=0)
                    Y = np.concatenate((Y, yi), axis=0)
            print("读到了%d个数据" % len(Y))
            return X, Y

        不同项目的实现方式各有差异：文件读取功能会根据数据集在本地的存储路径进行配置。Dataset类负责数据读取，将原始数据转换为三维数值表示；而Dataloader则用于批量加载数据，支持数据打乱和多批次处理功能。

        建议在数据读入阶段引入数据增强技术。通过旋转、裁剪等图像变换操作，不仅能扩充训练数据规模，还能有效提升模型对目标物体的识别准确率。

模型定义    

class myModel(nn.Module):
    def __init__(self, num_class):
        super(myModel, self).__init__()
        #3 *224 *224  -> 512*7*7 -> 拉直 -》全连接分类
        self.conv1 = nn.Conv2d(3, 64, 3, 1, 1)    # 64*224*224
        self.bn1 = nn.BatchNorm2d(64)
        self.relu = nn.ReLU()
        self.pool1 = nn.MaxPool2d(2)   #64*112*112


        self.layer1 = nn.Sequential(
            nn.Conv2d(64, 128, 3, 1, 1),    # 128*112*112
            nn.BatchNorm2d(128),
            nn.ReLU(),
            nn.MaxPool2d(2)   #128*56*56
        )
        self.layer2 = nn.Sequential(
            nn.Conv2d(128, 256, 3, 1, 1),
            nn.BatchNorm2d(256),
            nn.ReLU(),
            nn.MaxPool2d(2)   #256*28*28
        )
        self.layer3 = nn.Sequential(
            nn.Conv2d(256, 512, 3, 1, 1),
            nn.BatchNorm2d(512),
            nn.ReLU(),
            nn.MaxPool2d(2)   #512*14*14
        )

        self.pool2 = nn.MaxPool2d(2)    #512*7*7
        self.fc1 = nn.Linear(25088, 1000)   #25088->1000
        self.relu2 = nn.ReLU()
        self.fc2 = nn.Linear(1000, num_class)  #1000-11

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.pool1(x)
        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.pool2(x)
        x = x.view(x.size()[0], -1)
        x = self.fc1(x)
        x = self.relu2(x)
        x = self.fc2(x)
        return x

     模型定义相对简单且相似。值得一提的是，数据作为新时代的石油资源，我们训练的模型通常难以与投入数百万美元训练的大模型相媲美，因此可以采用迁移学习策略。迁移学习主要分为微调和线性探测两种方式，二者的核心区别在于是否冻结主干网络的参数。

训练流程

def train_val(model, train_loader, val_loader, no_label_loader, device, epochs, optimizer, loss, thres, save_path):
    model = model.to(device)
    plt_train_loss = []
    plt_val_loss = []

    plt_train_acc = []
    plt_val_acc = []

    max_acc = 0.0

    for epoch in range(epochs):
        train_loss = 0.0
        val_loss = 0.0
        train_acc = 0.0
        val_acc = 0.0
        model.train()
        for batch_x, batch_y in train_loader:
            x, target = batch_x.to(device), batch_y.to(device)
            pred = model(x)
            train_bat_loss = loss(pred, target)
            train_bat_loss.backward()
            optimizer.step()  # 更新参数 之后要梯度清零否则会累积梯度
            optimizer.zero_grad()
            train_loss += train_bat_loss.cpu().item()
            train_acc += np.sum(np.argmax(pred.detach().cpu().numpy(), axis=1) == target.cpu().numpy())
  
        model.eval()
        with torch.no_grad():
            for batch_x, batch_y in val_loader:
                x, target = batch_x.to(device), batch_y.to(device)
                pred = model(x)
                val_bat_loss = loss(pred, target)
                val_loss += val_bat_loss.cpu().item()
                val_acc += np.sum(np.argmax(pred.detach().cpu().numpy(), axis=1) == target.cpu().numpy())
        plt_val_loss.append(val_loss / val_loader.__len__())
        plt_val_acc.append(val_acc / val_loader.dataset.__len__())
        if val_acc > max_acc:
            torch.save(model, save_path)
            max_acc = val_acc

        主干训练流程也是大体相似，这里和回归模型在于多了计算精确率，因为分类任务输出label可以直接知道整体预测正确率。