deepseek-ai / FlashMLA
FlashMLA: Efficient Multi-head Latent Attention Kernels
- Repository stars
- 13,241All-time #3,843
- Fork
- 1,258
- Last push
- updated 10 days ago
- Repository created
- since 2025
About
Summarised by AI from the README; the original prevails
Author description: FlashMLA: Efficient Multi-head Latent Attention Kernels
Basic listing based on official public information; functionality has not been independently evaluated.
- Platform
- Not verified
- Pricing
- Pricing not verifiedCheck current official software, hosting, model and infrastructure pricing.
- Maintenance notes
- Not verified