From NVIDIA: 2024-11-06 11:00:32

Hugging Face and NVIDIA collaborate at CoRL in Munich to accelerate robotics research and development. LeRobot open AI platform combined with NVIDIA AI technology aims to drive advances in manufacturing, healthcare, and logistics industries.

Open-source robotics frameworks are crucial for the era of physical AI. Hugging Face’s platform serves over 5 million ML researchers, offering tools to streamline AI development. LeRobot by Hugging Face extends successful paradigms into robotics, providing tools for data collection, model training, and simulation environments.

NVIDIA’s AI technology and open-source robot learning framework like NVIDIA Isaac Lab can accelerate LeRobot’s workflow. By sharing models and datasets built with LeRobot and Isaac Lab, a data flywheel is created for the robotics community.

Isaac Lab enables robot training in simulation using high-fidelity rendering and physics simulation to generate vast amounts of training data. This iterative process leverages real-world data accuracy and scalability of simulated synthetic data to ensure robust and reliable robotic systems.

Collaborative workflow involves collecting data through teleoperation and simulation, training a robot policy with imitation learning, and deploying it on real-world robots with NVIDIA Jetson for real-time inference. Initial steps have shown LeRobot software running on NVIDIA Jetson Orin Nano for deployment.

Hugging Face and NVIDIA aim to accelerate innovation in AI for robotics by combining open-source communities and hardware. Together, they support open models and libraries optimizing inference for large language models, small language models, and multimodal vision-language models, driving progress in industries from transportation to manufacturing and logistics.



Read more at NVIDIA: Hugging Face and NVIDIA to Accelerate Open-Source AI Robotics Research and Development