Conv2d of torch

for instance .

I have a color picture , On the computer is R G B Channels are stored in three matrices , in other words , My pictures are stored in the computer size yes （1,3,64,64）

1 representative 1 A picture ,3 representative Three color channels ,64×64 representative , Every channel is （64,64） Matrix .

torch.nn.Conv2d()

Parameters ：

in_channels=3 This parameter is equal to size The second value of , That is, the number of channels of your photos .

out_channels=2 Represents the number of convolution kernels

kernel_size = 2              The size of the convolution kernel is (2,2) Matrix

stride=2              Represents the step length moving to the right , After the end of the line , Continue from below and move to the right