Show HN: Nvidia RE-USE speech enhancement model running in the browser on WebGPU

Category: ai-ml

Tags: webgpu, speech-enhancement, browser-inference, custom-kernels, onnx

Score: 5.5/10 (Innovation: 6, Technical: 7, Documentation: 4, Utility: 5)

A browser-based demo of NVIDIA's RE-USE speech enhancement model running on hand-written WebGPU kernels without ONNX runtime or server-side processing. It's technically interesting for demonstrating custom GPU kernel inference in JavaScript, but the minimal README and narrow demo scope limit its broader impact.

Target audience: ML engineers and frontend developers interested in in-browser GPU inference

Repository: https://re-use-demo.geronimo-labs.com · JavaScript · MIT · 1 stars

View on Hacker News