SAME = use zero-padding so that the output has the same size with the input. The subtlety is how to compute given the padding. As I tested on Tensorflow 1.8 ...
確定! 回上一頁