Usage of sum of pytorch

Usage of () and ()

The function of the flatten() function is to flatten the tensor into one dimension

(input, start_dim=0, end_dim=- 1) → Tensor

input (Tensor) – the input tensor.
start_dim (int) – the first dim to flatten
end_dim (int) – the last dim to flatten

start_dim and end_dim form the entire range of dimensions you want to choose to pave.

The following examples are given

x = ([[1,2], [3,4], [5,6]])
x = (0)
x
------------------------
tensor([1, 2, 3, 4, 5, 6])

For image data, we often expect the dimensions entering the fc layer to be (channels, N)

x = ([[[1,2],[3,4]], [[5,6],[7,8]]])
x = (1)
x
-------------------------
tensor([[1, 2],
        [3, 4],
        [5, 6]])

Note:

(start_dim=1, end_dim=- 1)

start_dim defaults to 1

So when building a network, the following two are equivalent

class Classifier():
    def __init__(self):
        super(Classifier, self).__init__()
        # The arguments for commonly used modules:
        # .Conv2d(in_channels, out_channels, kernel_size, stride=1, padding=0)
        # .MaxPool2d(kernel_size, stride=None, padding=0)

        # input image size: [3, 128, 128]
        self.cnn_layers = (
            nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(64),
            (),
            nn.MaxPool2d(kernel_size=2, stride=2, padding=0),

            nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(128),
            (),
            nn.MaxPool2d(kernel_size=2, stride=2, padding=0),

            nn.Conv2d(128, 256, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(256),
            (),
            nn.MaxPool2d(kernel_size=4, stride=4, padding=0),
        )
        self.fc_layers = (
            (256 * 8 * 8, 256),
            (),
            (256, 256),
            (),
            (256, 11)
        )

    def forward(self, x):
        # input (x): [batch_size, 3, 128, 128]
        # output: [batch_size, 11]

        # Extract features by convolutional layers.
        x = self.cnn_layers(x)

        # The extracted feature map must be flatten before going to fully-connected layers.
        x = (1)

        # The features are transformed by fully-connected layers to obtain the final logits.
        x = self.fc_layers(x)
        return x

class Classifier():
    def __init__(self):
        super(Classifier, self).__init__()

         = (
            nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(64),
            (),
            nn.MaxPool2d(kernel_size=2, stride=2, padding=0),

            nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(128),
            (),
            nn.MaxPool2d(kernel_size=2, stride=2, padding=0),

            nn.Conv2d(128, 256, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(256),
            (),
            nn.MaxPool2d(kernel_size=4, stride=4, padding=0),

            (),

            (256 * 8 * 8, 256),
            (),
            (256, 256),
            (),
            (256, 11)
        )

    def forward(self, x):
       
        x = (x)

        return x

Summarize

The above is personal experience. I hope you can give you a reference and I hope you can support me more.