awesome-transformer, should classified into Deep Learning。
Transformer is a powerful model applied in sequence to sequence learning. However, when I was using transformer as my baseline in NMT research I found no good & reliable guide to reproduce approximate result as reported in original paper(even official tensor2tensor implementation), which means my research would be unauthentic. I searched on the Internet, found some implementations, obtained some performance-reproducable approaches and other learning materials, which eventually formed this project.
I hope more researchers can benefit from it.
awesome-transformer, should classified into
Deep Learning
。 Transformer is a powerful model applied in sequence to sequence learning. However, when I was using transformer as my baseline in NMT research I found no good & reliable guide to reproduce approximate result as reported in original paper(even official tensor2tensor implementation), which means my research would be unauthentic. I searched on the Internet, found some implementations, obtained some performance-reproducable approaches and other learning materials, which eventually formed this project. I hope more researchers can benefit from it.