Hugging Face releases 207 open-source WebGPU kernels for browser inference

nicodotdev · x · 2026-09-01

Hugging Face ships @huggingface/kernels on npm, providing 207 open-source WebGPU kernels loaded straight from the Hub. This addresses the bottom layer of the fast browser inference stack. They also launched Fleet, a tool to benchmark your GPU in the browser and get a card for your device.

Related event: Hugging Face open-sources 207 WebGPU kernels for faster browser inference(2 posts)→

Original post →

More from Infra

Infra channel →