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
ChainCatcher•4h agoSenator: The OCC should not grant a national trust license to crypto businesses like Ripple and Coinbase
ChainCatcher•4h agoDeloitte acquires crypto infrastructure company Blocknative
币界网•4h agoWalsh is about to take office and accelerate the completion of financial asset divestiture commitments
ChainCatcher•4h agoCoinbase, Kraken, and Gemini urge the Senate to remove restrictions on crypto token listings
Odaily•5h agoOpenAI联合创始人Andrej Karpathy宣布加入Anthropic
ChainCatcher•6h agoData: A whale withdrew 85,226 HYPE from Bybit, and the address has purchased a total of 2,220,000 HYPE in the past month
币界网•6h agoAnchorage-linked wallets withdrew 85,226 HYPE and purchased 2.22 million in the past month
ChainCatcher•6h agoCatcher Predict：“Dota 2：Team Falcons 对阵 Tundra Esports（BO3）- DreamLeague 季后赛” “总击杀数在第 2 局超过 72.5？” 胜率飙升 49.95%
Odaily•6h agoCrypto companies pause IPO plans as trading volume declines and the AI boom reshapes the tech market
ChainCatcher•6h agoThe probability of the Fed raising interest rates by the end of 2026 is over 80%