NVIDIA推出Nemotron 3 Nano Omni模型,多模态推理效率提升9倍

Odaily星球日报讯 NVIDIA 在 X 平台发文表示,今日推出开源多模态模型 Nemotron 3 Nano Omni。该模型采用 30B-A3B 混合专家架构(MoE),支持 256K 上下文,可统一处理视频、音频、图像及文本输入。与同类交互水平的开源全能模型相比,该模型吞吐量提升达 9 倍,大幅降低推理成本并提高扩展性。目前 Nemotron 3 Nano Omni 已在 Hugging Face、OpenRouter 及 NVIDIA NIM 上线,并获得 Aible、Applied Scientific Intelligence、H Company 等企业采用。
Disclaimer: OKX Orbit content is provided for informational purposes only. Learn more
Replies
Related Flash News
Analyst: HYPE and AI tokens may lead the next altcoin rally, and market risk appetite is returning
Iran proposed opening Hormuz in exchange for U.S. compensation and called for discussions on sanctions and funding freezes
The U.S. CLARITY Act may spur a new "revenue as a service" track, promoting AI-driven compliance revenue infrastructure development
US Media: Trump Willing to Give US-Iran Negotiations 'More Time'
Analysis: In the previous two bear markets, the BTC loss at the bottom was 10.6 million coins, corresponding to a price of $60,000 in this round
An Ethereum OG address bought 3,942 ETH on dips today, at an average price of $2,049
Ethereum OG, which once received a 376x return, has bottom-fished again, having already bought over $8 million worth of ETH
U.S. ETF inflows hit a record high for the year, with a daily net inflow of $8.5 billion
Analysis: The continued rise in U.S. Treasury yields has weakened market willingness to allocate to Bitcoin
Catcher Predict: "LoL: Team WE vs. LNG Esports (BO5) - LPL Play-In" "Total kills exceed/under 27.5 in game 1?" Winning rate soars to 49.45%


