Research paper · AI and research
Attention Is All You Need, the Transformer paper (2017)
The paper by Ashish Vaswani and seven co-authors proposed the Transformer, a neural network architecture based solely on attention mechanisms, without recurrence or convolutions. In English-to-German translation it beat the best results of the time by more than 2 BLEU points; in English-to-French it reached 41.8 BLEU after 3.5 days of training on eight GPUs.
June 12, 2017
Attention Is All You Need
Research paper · AI and research · Active
Dates
- June 12, 2017 Publication · exact date [1]
What the sources say
Names
- Attention Is All You Need — official name
- Transformer — period nickname
Historical captures
No archived capture has been located for this record yet.
Sources
- Attention Is All You Need — Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, Illia Polosukhin — arXiv, June 12, 2017. Source primary · accessed November 2, 2026 · look for a copy on the Internet Archive
What still needs documenting
- Presentation at NeurIPS 2017
- Source for the Transformer's influence on later models
On GD
GD has followed Google since 2006. Search GD for “Attention Is All You Need” (in Portuguese)
How to cite
"Attention Is All You Need, the Transformer paper (2017)", Google Museum (GD), https://gd.eurisko.com.br/museu/en/collection/attention-is-all-you-need-2017/, CC BY 4.0. Reviewed on November 2, 2026.
ID: attention-is-all-you-need-2017 · Reviewed on November 2, 2026