英伟达推出新的开源模型产品线，用于在企业和边缘计算场景中运行AI代理1。其中，面向企业的Nemotron 3.5 Super VL拥有1200亿参数，性能与Nemotron 3 Ultra相当1；针对个人电脑的Bolt系列包括740亿参数的Bolt 74B-A9B和轻量级会话代理模型Bolt E4B1。Super VL和Bolt 74B-A9B已正式上市，Bolt E4B将于本月晚些时候发布1。
这些模型设计可在本地基础设施上部署，无需依赖云端服务1。英伟达高管Kari Briski指出："与闭源前沿模型不同，组织可以在自己的数据上进行后训练，在自己的基础设施上部署"1。Skydance、EY、Crowdstrike等企业已在使用Super VL1。新发布的AI个人电脑搭载RTX Spark芯片，通过Microsoft和Dell等合作伙伴推向市场1。
Nvidia has introduced a new line of open-source models designed to enable organizations and individual users to run AI agents locally without cloud dependency.1 The enterprise-focused Nemotron 3.5 Super VL features 120 billion parameters and delivers performance comparable to Nemotron 3 Ultra, while the Bolt series offers lighter-weight alternatives for personal computers, including the 74-billion-parameter Bolt 74B-A9B and the conversational Bolt E4B model.1 The Nemotron 3.5 Super VL and Bolt 74B-A9B are now available, with Bolt E4B set to launch later this month.1
According to Kari Briski, "unlike closed-source cutting-edge models, organizations can perform additional training on their own data and deploy on their own infrastructure."1 These models address enterprise use cases requiring edge computing and on-premises deployment, and have already been adopted by companies including Skydance, EY, Ramp, Crowdstrike, Perplexity, and Salesforce.1 Nvidia is also promoting new AI PCs powered by RTX Spark chips, which will be distributed through partners such as Microsoft and Dell.1
评论
还没有评论，欢迎留下第一条。