pytorch CNN model 中的 BatchNorm2d 錯誤

Question

我的數據庫有大小為 128 * 128* 1 的灰度圖像，每個批次大小 =10 我正在使用 cnn model 但我在 BatchNorm2d 中遇到了這個錯誤
預期 4D 輸入（獲得 2D 輸入）

我發布了我用來轉換圖像的方式（灰度 - 張量 - 歸一化）並將其分成批次

data_transforms = {
    'train': transforms.Compose([
        transforms.Grayscale(num_output_channels=1),
        transforms.Resize(128),
        transforms.CenterCrop(128),
        transforms.ToTensor(),
        transforms.Normalize([0.5], [0.5])
    ]),
    'val': transforms.Compose([
        transforms.Grayscale(num_output_channels=1),
        transforms.Resize(128),
        transforms.CenterCrop(128),
        transforms.ToTensor(),
        transforms.Normalize([0.5], [0.5])
    ]),
}


data_dir = '/content/drive/My Drive/Colab Notebooks/pytorch'
dsets = {x: datasets.ImageFolder(os.path.join(data_dir, x), data_transforms[x])
         for x in ['train', 'val']}
dset_loaders = {x: torch.utils.data.DataLoader(dsets[x], batch_size=10,
                                               shuffle=True, num_workers=25)
                for x in ['train', 'val']}
dset_sizes = {x: len(dsets[x]) for x in ['train', 'val']}
dset_classes = dsets['train'].classes

我用這個 model

class HeartNet(nn.Module):
    def __init__(self, num_classes=7):
        
        super(HeartNet, self).__init__()

        self.features = nn.Sequential(
            nn.Conv2d(1, 64, kernel_size=3, stride=1, padding=1),
            nn.ELU(inplace=True),
            nn.BatchNorm2d(64),
            nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1),
            nn.ELU(inplace=True),
            nn.BatchNorm2d(64),
            nn.MaxPool2d(kernel_size=2, stride=2),
            nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1),
            nn.ELU(inplace=True),
            nn.BatchNorm2d(128),
            nn.Conv2d(128, 128, kernel_size=3, stride=1, padding=1),
            nn.ELU(inplace=True),
            nn.BatchNorm2d(128),
            nn.MaxPool2d(kernel_size=2, stride=2),
            nn.Conv2d(128, 256, kernel_size=3, stride=1, padding=1),
            nn.ELU(inplace=True),
            nn.BatchNorm2d(256),
            nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1),
            nn.ELU(inplace=True),
            nn.BatchNorm2d(256),
            nn.MaxPool2d(kernel_size=2, stride=2)
            )

        self.classifier = nn.Sequential(
            nn.Dropout(0.5),
            nn.Linear(16*16*256, 2048),
            nn.ELU(inplace=True),
            nn.BatchNorm2d(2048),
            nn.Linear(2048, num_classes)
            )

        nn.init.xavier_uniform_(self.classifier[1].weight)
        nn.init.xavier_uniform_(self.classifier[4].weight)

    def forward(self, x):
        x = self.features(x)
        x = x.view(x.size(0), 16 * 16 * 256)
        x = self.classifier(x)
        return x

我怎么解決這個問題？

Answer 1

您的self.classifier子網絡中的批處理規范層存在問題：雖然您的self.features子網絡是完全卷積的並且需要BatchNorm2d ，但self.classifier子網絡是一個完全連接的多層感知器 (MLP) 網絡並且本質上是一維的。 注意forward function 在將其饋送到分類器之前如何從特征 map x中刪除空間維度。

嘗試用BatchNorm2d替換self.classifier中的BatchNorm1d 。