In case we want to keep the size of the input unchanged after the convolution layer, we apply same padding where W_{out} = W_{in} ...
確定! 回上一頁