Skip to content

DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

License

Notifications You must be signed in to change notification settings

drewjin/DeepGEMM

 
 

Repository files navigation

DeepGEMM 学习仓库

我简单重构了一下仓库的一些组织,用以适配我自己的开发习惯。

因为目前项目本身的快速开始并没做好,构建的时候一堆问题。

主要是用 uv 的虚拟环境,用来方便我自己进行学习与开发。

流程

About

DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Cuda 41.0%
  • C++ 39.1%
  • Python 19.1%
  • Other 0.8%