Xenova open-sources 200+ WebGPU kernels for in-browser AI
Xenova open-sourced what it calls the world's fastest WebGPU kernel collection on Hugging Face, covering 200+ ML operators optimized for various hardware and enabling fully local in-browser AI inference.
2026-10-01 ~ 2026-10-01 · 2 related posts
- Hugging Face open-sources 200+ WebGPU kernels for in-browser local AI — xenovatech · 2026-10-01
- Xenova open-sources WebGPU kernels for 200+ ML ops, running fully in-browser — CatAstro_Piyush · 2026-10-01