Derox News
NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot

NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot

By editorial News

NVIDIA and Hugging Face have joined forces to integrate the NVIDIA Isaac GR00T 1.7 model and the NVIDIA Isaac Teleop framework into LeRobot, Hugging Face’s open-source robotics library. The partnership also plans to add NVIDIA Cosmos 3, a frontier model for physical AI, to the platform in the near future. These additions aim to provide developers with a more accessible, standardized path for end-to-end robot development, fostering innovation and collaboration across the open robotics community.

The companies emphasized that open-source AI has demonstrated rapid innovation when models, data, and tools are shared, and they see a similar opportunity in robotics. However, progress in physical AI has often been hindered by costly and fragmented resources, including large datasets, robot foundation models, simulation tools, and validation capabilities.

“Open source is how a field turns advanced research into something people can study, adapt and build on,” said Thomas Wolf, co-founder and chief science officer at Hugging Face. “With NVIDIA Isaac GR00T 1.7 and Isaac TeleOp in LeRobot today, robotics developers can use shared models, data, and workflows to train and evaluate robots in the open. And with NVIDIA Cosmos 3 planned next, the community will have a path to bring frontier world models into that same collaborative loop.”

NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot

Bridging Robotics and AI Communities

Hugging Face LeRobot is an open-source library designed for training, running, and sharing robot datasets, models, policies, and workflows. Through this ongoing partnership, NVIDIA connects its 3 million robotics developers with Hugging Face’s 16 million AI builders, expanding access to advanced physical AI tools via open workflows.

By integrating NVIDIA’s physical AI capabilities into LeRobot, developers gain a unified approach to collecting and standardizing data, training and fine-tuning robot foundation models, evaluating performance, and deploying models. The key integrations include:

  • NVIDIA Isaac Teleop: An open-source framework for robot data collection that allows developers to capture high-quality human demonstrations from external devices using standardized, interoperable formats. Data can then be expanded and shared directly within the LeRobot community.

  • NVIDIA Isaac GR00T 1.7: Described as the first open and commercially viable robot foundation model, this VLA (vision-language-action) model for humanoid robots makes it easier to post-train and deploy models through LeRobot workflows. Developers can adapt GR00T to new robot embodiments and tasks with benchmarked performance.

  • NVIDIA Cosmos 3: A frontier world foundation model for physical AI that is coming soon to LeRobot. It will help developers generate and augment robotics data, simulate scenarios, and support policy development when real-world data is limited or too expensive to collect.

Existing NVIDIA Technology Already Available in LeRobot

These new integrations build on a broader set of NVIDIA resources already connected to LeRobot. The library currently offers a large open-source physical AI dataset, which has been downloaded more than 15 million times and includes over 350,000 real and simulated trajectories and 57 million grasps to jump-start development.

Additionally, LeRobot features simulation frameworks based on NVIDIA Isaac Sim and Isaac Lab. These tools enable developers to set up environments, generate robot data, test policies, and validate behaviors before moving to physical hardware. The NVIDIA Isaac Lab-Arena is also part of the LeRobot Environment Hub, allowing for rapid prototyping of complex simulation environments that can be registered and used to train and evaluate generalist robot policies such as GR00T, Pi, and SmolVLA.

Furthermore, the integration of NVIDIA Jetson Thor with LeRobot’s Reachy 2 supports the deployment of VLA models on open-source humanoid robots, extending the ecosystem’s reach into real-world robotics applications.

The source for this article is https://www.therobotreport.com/nvidia-hugging-face-bring-new-models-frameworks-lerobot/.