Foresight News reported that NVIDIA (NVDA.O) CEO Jensen Huang announced the launch of a full series of models based on Llama, using the NVIDIA Llama Nemo Tron language base model, named Nemotron. Additionally, NVIDIA has released NVIDIA Cosmos, a base model aimed at understanding the real world, also using the NVIDIA Llama Nemo Tron language base model.
The Cosmos model can accept text, image, or video prompts to generate virtual world states as unique video outputs for autonomous driving and robotics applications. Developers can leverage Cosmos to generate AI feedback for reinforcement learning, improving policy models and testing performance in different scenarios. NVIDIA is collaborating with the global developer ecosystem to drive the next wave of physical AI advancement. Cosmos is the world's first base model trained on 20 million hours of video, focusing on physical dynamics and human interaction.
