NVIDIA announced that it will bring the recently emerging generative artificial intelligence (AI) to the Jetson robotics platform, accelerating applications such as robotic computer vision and deep learning to meet requirements such as defect detection, real-time asset tracking, autonomous planning and navigation, and human-machine interaction, thereby strengthening edge computing and IoT applications.
By leveraging generative AI technology, NVIDIA says it will be able to integrate large natural language models, including Transformer, into platforms like Jetson Orin, enabling faster-response and more accurate AI applications on robotics platforms.
Automatically generated artificial intelligence significantly allows users to more easily operate machines through natural language. It also makes it easier for machines to understand users' actual operational needs, allowing them to more quickly engage in workflows such as detection, segmentation, tracking, and searching, and even reprogram work content at any time.
The introduction of generative AI technology in the Jetson Isaac framework and Metropolis platform, combined with the Transformer model and AI computing power, will make robotic operations more intelligent and efficient.
For example, users can currently use the NVIDIA TAO toolkit to build low-code artificial intelligence models for visual applications, and can easily fine-tune and optimize them, thereby quickly allowing robots to help detect product defects using computer vision. They can also use the NVIDIA Metropolis developer tool to create solutions using artificial intelligence computers.
NVIDIA also announced a series of selected artificial intelligence reference workflows based on the Metropolis and Isaac frameworks, allowing developers to quickly adopt and integrate the entire workflow or selectively integrate individual components of these workflows, thereby reducing development time and costs.
