Transformer之前是用RNN(recurrent structure). 而 T用自我attention

來源: 2023-02-10 20:37:43 [博客] [舊帖] [給我悄悄話] 本文已被閱讀: