# arXiv: Attention Is All You Need v7

Formula-heavy machine-learning paper with tables, vector attention figures, dense paths, and citation links. Categories: complex-columns, formulas, tables, diagrams, vector-heavy, links. Canonical source: https://arxiv.org/abs/1706.03762v7

- Source PDF: [source.pdf](source.pdf)
- Continuous semantic HTML: [semantic-continuous/index.html](semantic-continuous/index.html)
- Side-by-side comparison: [compare.html](compare.html)
- Quality probe: [quality/quality-report.md](quality/quality-report.md)
- Pages: 15
- Text runs: 2610
- Text lines: 656
- Image placements: 3
- Vector paths: 2710
- Exported assets: 13
- Links: 113
- Diagnostics: 23
- Quality status: needs-review
- Quality checks needing review: 2

## Quality Artifacts

- [diff p1](quality/page-1-diff.png)
- [visual report p1](quality/page-1-visual-report.html)
- [diff p2](quality/page-2-diff.png)
- [visual report p2](quality/page-2-visual-report.html)

## Text Preview

Providedproperattributionisprovided,Googleherebygrantspermissionto reproducethetablesandfiguresinthispapersolelyforuseinjournalisticor scholarlyworks. AttentionIsAllYouNeed AshishVaswani∗NoamShazeer∗NikiParmar∗JakobUszkoreit∗ GoogleBrainGoogleBrainGoogleResearchGoogleResearch avaswani@google.comnoam@google.comnikip@google.comusz@google.com LlionJones∗AidanN.Gomez∗†ŁukaszKaiser∗ GoogleResearchUniversityofTorontoGoogleBrain llion@google.comaidan@cs.toronto.edulukaszkaiser@google.com IlliaPolosukhin∗‡ illia.polosukhin@gmail.com Abstract Thedominantsequencetransductionmodelsarebasedoncomplexrecurrentor convolutionalneuralnetworksthatincludeanencoderandadecoder.Thebest performingmodelsalsoconnect...