Hugging Face has released @huggingface/kernels, a JavaScript library for loading and running optimized WebGPU kernels from the Hugging Face Hub, alongside an initial collection of 207 Apache-2.0-licensed kernels. Each kernel is published as a versioned repository with manifests, correctness tests, benchmark cases, and WGSL shader templates. The company also launched Fleet, a browser-based benchmarking tool that crowdsources performance and correctness evidence across real-world GPUs. The kernels serve as a foundational layer for fast browser-based AI inference.
No score is assigned. Sources and their independence are shown in the citation chain below.