Topic: "flash-mla"
xlite-dev/Awesome-LLM-Inference
📚A curated list of Awesome LLM/VLM Inference Papers with codes: WINT8/4, FlashAttention, PagedAttention, Parallelism, MLA, etc.
Language: Python - Size: 115 MB - Last synced at: about 1 hour ago - Pushed at: about 2 hours ago - Stars: 4,100 - Forks: 283

xlite-dev/ffpa-attn
📚FFPA(Split-D): Extend FlashAttention with Split-D for large headdim, O(1) GPU SRAM complexity, 1.8x~3x↑🎉 faster than SDPA EA.
Language: Cuda - Size: 4.21 MB - Last synced at: 7 days ago - Pushed at: 30 days ago - Stars: 183 - Forks: 8
