💬 Chat with Qwen_Qwen2.5-7B-Instruct in your browser
Runs entirely on your machine via WebGPU (requires a recent Chrome/Edge or other WebGPU-capable browser) — nothing is uploaded, and no torrent or download is required. This mirror ships no ONNX weights, so the demo loads WebLLM's prebuilt MLC build of the same model (~5.0 GBin VRAM), not this site's files.
Checking your browser's capabilities…