Skip to content

Latest commit

 

History

History
12 lines (8 loc) · 364 Bytes

README.md

File metadata and controls

12 lines (8 loc) · 364 Bytes

BetaNN

This project implements computation kernels using WebGPU, with C++ APIs. It is in very early stage and not ready for any use.

Most of the code are highly inspired by MLX's Metal kernels.

Targets:

  • Power a WebGPU backend for MLX.
  • Provide a library for porting machine learning frameworks to WebGPU.
  • Support WebAssembly target and run in browsers.