Nvidia的新Llama 尼莫坦 AI模型 提供20%的精确度 和五倍的更快的推断速度。 Nvidia's new Llama Nemotron AI models offer 20% more accuracy and five times faster inference speeds.

Nvidia在GTC 2025年会议上介绍了新的Llama 尼莫坦 AI模型,其精确度提高了20%,比以前的模型的推论速度提高了五倍。 Nvidia introduced its new Llama Nemotron AI models at the GTC 2025 conference, featuring a 20% increase in accuracy and a five-fold boost in inference speed over previous models. 这些模型是针对不同的硬件设计的,从个人计算机到多GPU服务器,并被微软和SAP等合作伙伴用来加强AI服务。 The models are tailored for different hardware, from personal computers to multi-GPU servers, and are being used by partners like Microsoft and SAP to enhance AI services. Nvidia还启动了更多的AI结构构件,以支持先进的AI模型的开发。 Nvidia also launched additional AI building blocks to support the development of advanced AI models.