deepseek-ai

deepseek-ai / FlashMLA

#5

12,228955+30 todayC++

FlashMLA: Efficient Multi-head Latent Attention Kernels

📊 Project Info

Language: C++
Stars: ⭐ 12,228
Forks: 955
Today: +30
Ranking: #5
Collection: Overall
Trending Date: January 22, 2026