framework-internalsdevelopment
flash-attention
Optimize transformer attention with Flash Attention — 2-4x speedup, 10-20x memory reduction for long sequences on CUDA GPUs.
maintainer
AlexAI-MCP
更新於 4/7/2026
星標
62
分支
14
quick start
Installation and usage
Optimize transformer attention with Flash Attention — 2-4x speedup, 10-20x memory reduction for long sequences on CUDA GPUs.
安裝
$ install --globalskills.sh
使用
安裝後,您可以通過在終端運行以下命令來使用此技能:
skills use flash-attention