Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Attention is all you need, Vaswani+, arXiv'17 #245

Open
AkihikoWatanabe opened this issue Jan 19, 2018 · 4 comments
Open

Attention is all you need, Vaswani+, arXiv'17 #245

AkihikoWatanabe opened this issue Jan 19, 2018 · 4 comments

Comments

@AkihikoWatanabe
Copy link
Owner

https://arxiv.org/abs/1706.03762

@AkihikoWatanabe
Copy link
Owner Author

AkihikoWatanabe commented Jan 19, 2018

Transformer (self-attentionを利用) 論文
解説スライド:https://www.slideshare.net/DeepLearningJP2016/dlattention-is-all-you-need
解説記事:https://qiita.com/nishiba/items/1c99bc7ddcb2d62667c6

  • 新しい翻訳モデル(Transformer)を提案。既存のモデルよりも並列化に対応しており、短時間の訓練で(既存モデルの1/4以下のコスト)高いBLEUスコアを達成した。
  • TransformerはRNNやCNNを使わず、attentionメカニズムに基づいている。

(解説より)

@AkihikoWatanabe
Copy link
Owner Author

分かりやすい:
https://qiita.com/halhorn/items/c91497522be27bde17ce

@AkihikoWatanabe
Copy link
Owner Author

Transformerの各コンポーネントでのoutputのshapeや、attention_maskの形状、実装について記述されており有用:
https://qiita.com/FuwaraMiyasaki/items/239f3528053889847825

@AkihikoWatanabe
Copy link
Owner Author

集合知

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant