Padding with Conv2DTranspose. I have an input of 8x8x32 tensor, which I am compressing to 4x4x16 using Conv2D: torch.nn.
確定! 回上一頁