takara-ai / go-attention
A full attention mechanism and transformer in pure go.
Why this job: works in the areas we cover, but does not say which job it does
What it does
- Loop Unrolling. 8x unrolled dot product for maximum throughput
- Memory Pooling. Object pools reduce allocation overhead in hot paths
- Cache-Friendly Projections. projectVector uses row-major loop order so weight rows are read sequentially (fixes column-major stride that caused cache misses on Q/K/V and feed-forward matmuls)
- Zero Dependencies. Pure Go implementation with no external requirements
Get started
No single install command. Read the steps in the README (opens in new tab)
- Maintenance
- Active
- Last commit 5 months ago
- Stars
- 480
- +1 this week
- Licence
- MIT
- Allows business use
- Downloads for
- None listed
- No release files
- Latest version
- None yet
- No releases
- Built with
- Go
- 100% of the code
Is it looked after?
- Contributors
- Not stated
- Last 90 days
- Issues
- 88% closed
- 1 still open
- Pull requests
- 3 merged
- 0 waiting
- Releases
- None
- Never published
- Contributing guide (opens in new tab)
- No security policy
- Code of conduct (opens in new tab)
Similar repos
| Repo | Stars | Last commit | Licence | Downloads for |
|---|---|---|---|---|
| modal-labs / modal-examples | 1,292 | 2 days ago | MIT | No downloads |
| fynnfluegge / rocketnotes | 1,496 | 9 months ago | Apache-2.0 | No downloads |
| bodaay / HuggingFaceModelDownloader | 1,212 | 39 days ago | Apache-2.0 | Mac, Windows, Linux |
More: languages, topics and project record
Languages
- Go 100%
Topics
- artificial-intelligence
- golang
- machine-learning
- serverless
Project record
- JobOther AI tools
- StatusListed
- Commits24
- Forks20
- Watching6
- Stars in 28 days+1
- CreatedFeb 10, 2025
- First seen by usNov 7, 2026
- CheckedNov 8, 2026