deepseek-ai/FlashMLA

Star 趋势上榜 4 天 · 累计 +377 · 日均 +125

项目介绍与用途

「deepseek-ai/FlashMLA」是由 deepseek-ai 维护、主要使用 C++ 开发的开源项目。FlashMLA: Efficient Multi-head Latent Attention Kernels。目前该项目在 GitHub 上累计获得 13.2k 个 Star、1.2k 次 Fork（今日新增 35 Star）。如果你关注 C++ 方向的开源生态，它值得收藏、学习与参考。

项目简介

FlashMLA: Efficient Multi-head Latent Attention Kernels

相关项目

php/php-srcThe PHP Interpreter

vllm-project/vllmA high-throughput and memory-efficient inference and serving engine for LLMs

supabase/supabaseThe Postgres development platform. Supabase gives you a dedicated Postgres database to build your web, mobile, and AI applications.

trycua/cuaScale computer-use 2.0 with open-source drivers, cross-OS sandboxes, and benchmarks for training, evaluation, and data generation.

ModernRelay/omnigraphObject-storage native graph database with branching and typed ontology

tursodatabase/tursoA SQL database in Rust: SQLite-compatible, now also speaking Postgres (experimental). The LLVM of databases.

项目体检

分享