Nvidia Corp. (NVDA), a leading AI chip manufacturer, has unveiled NVIDIA Cosmos. This groundbreaking platform includes cutting-edge generative world foundation models (WFMs), sophisticated tokenizers, guardrails, and an expedited video processing pipeline. It is specifically designed to enhance the development of physical AI systems, such as autonomous vehicles (AVs) and robotics.
The company highlighted the substantial costs and extensive real-world data and testing needed to create physical AI models. With the introduction of Cosmos WFMs, developers can easily produce vast amounts of photorealistic, physics-based synthetic data for training and evaluating their current models.
Additionally, developers have the option to create customized models by refining Cosmos WFMs, as stated by the company.
To further support the robotics and AV community, Nvidia is releasing Cosmos models under an open model license. Developers can access previews of the initial models through the NVIDIA API catalog or download the full suite of models and fine-tuning framework from both the NVIDIA NGC catalog and Hugging Face.
Nvidia also introduced the Nvidia Llama Nemotron large language models and Nvidia Cosmos Nemotron vision language models. These resources are intended for enterprise AI applications, including industries like healthcare, financial services, and manufacturing.
The material has been provided by InstaForex Company -
www.instaforex.com
