deepseek-ai

deepseek-ai / FlashMLA

#9

12,313960+184 todayC++

FlashMLA: Efficient Multi-head Latent Attention Kernels

📊 Project Info

Language: C++
Stars: ⭐ 12,313
Forks: 960
Today: +184
Ranking: #9
Collection: Overall
Trending Date: January 23, 2026