Deep Learning for Natural Language Processing

  Transformer是首个完全抛弃RNN的recurrence，CNN的convolution，仅用attention来做特征抽取的模型。Deep Learning for Natural Language Processing：The Transformer model_up.pdf: https://url39.ctfile.com/f/2501739-805100030-845fca?p=2096 (访问密码: 2096)

参考文献：
[1]Deep Learning for Natural Language Processing：The Transformer model_up.pdf: https://url39.ctfile.com/f/2501739-805100030-845fca?p=2096 (访问密码: 2096)