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
更新于 4/7/2026
星标
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.
安装
$ install --globalskills.sh
使用
安装后,您可以通过在终端运行以下命令来使用此技能:
skills use flash-attention