news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Fast Technology reported on July 30 that at the top computer graphics conference SIGGRAPH 2024, NVIDIA demonstrated its latest progress in rendering, simulation and generative AI, the most eye-catching of which was the new upgrade of NVIDIA NIM.

NIM is NVIDIA's new trump card in the era of generative AI. Through this platform, users can build AI applications in minutes, greatly simplifying the development and deployment process of AI models.

Huang Renxun described NIM as "AI-in-a-Box", which means artificial intelligence in a box. This upgrade further optimizes performance through the complex deployment of standardized AI models, making Nvidia's leadership in the field of AI more solid.

The NIM ecosystem provides a range of pre-trained AI models covering multiple fields, including text processing, speech synthesis, 3D scene creation, robotics, and digital biology.

These models can be quickly deployed as optimized "containers" in the cloud, data centers, or workstations, greatly accelerating the development and deployment of AI applications.

Nvidia also announced that the Hugging Face inference-as-a-service platform is now powered by Nvidia NIM and runs in the cloud, further expanding the application scope of NIM.

In addition, NVIDIA NIM microservices enable industries to build physics-based virtual worlds and digital twins by enhancing the understanding of the OpenUSD (Universal Scenario Description) language.

At the SIGGRAPH conference, NVIDIA also demonstrated the application of generative AI on the open USD and Omniverse platforms, including a new USD connector that connects robotics and industrial simulation data formats and developer tools, promoting the development of 3D modeling, robotics simulation, and digital twin technologies.