Meta 推出 Llama 3.2：支援視覺多模態與邊緣裝置運行的輕量級模型，Hugging Face 全面支援
Original: Llama can now see and run on your device - welcome Llama 3.2
Meta has officially introduced the Llama 3.2 family of open-source models, marking a significant architectural upgrade with two major…
Meta 正式發布 Llama 3.2，最大亮點為首次推出 11B 和 90B 的視覺多模態模型，能理解圖像與圖表；同時推出專為手機與邊緣裝置設計的 1B 和 3B 超輕量文字模型，支援 128k 脈絡長度。Hugging Face 已在 Transformers、TGI 及微調工具中全面支援，方便開發者即刻部署與應用。
Meta has officially introduced the Llama 3.2 family of open-source models, marking a significant architectural upgrade with two major breakthroughs: multimodal vision capabilities, and the ability to run natively on edge devices (on-device).
Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.
See Pro plans →Want the original English / full article?
Read on Hugging Face Blog →Summaries are AI-generated; the original article is authoritative.