I have an input of 8x8x32 tensor, which I am compressing to 4x4x16 using Conv2D: torch.nn.Conv2d(32, 16, kernel_size = 3, stride = 2, ...
確定! 回上一頁