Qwen3.5-2B at 620 tok/s inside Chrome: model-specific WebGPU kernels, 6.6× WebLLM on the same model

🌐 链接: https://www.reddit.com/r/LocalLLaMA/comments/1xmv3lc/qwen352b_at_620_toks_inside_chrome_modelspecific/
🏷️ 分组: Reddit｜AI·自托管·VPS
🕒 时间: 2026-11-03 22:56:02
Reddit
From the LocalLLaMA community on Reddit: Qwen3.5-2B at 620 tok/s inside Chrome: model-specific WebGPU kernels, 6.6× WebLLM on the…
From the LocalLLaMA community on Reddit: Qwen3.5-2B at 620 tok/s inside Chrome: model-specific WebGPU kernels, 6.6× WebLLM on the…
Explore this post and more from the LocalLLaMA community