NVIDIA and Hugging Face Expand LeRobot with New Models and Frameworks

NVIDIA and Hugging Face have expanded LeRobot—an open-source library for robotics development. The partners added new pre-trained machine learning models and frameworks to the project, aimed at simplifying the cycle of creating robotic systems.

What Is LeRobot

LeRobot is a Python library with open-source code for training, running, and sharing robotic models, datasets, and control policies. The project allows developers worldwide to exchange ready-made solutions: instead of writing control algorithms from scratch, teams use verified datasets and trained models, significantly accelerating the development of their own robots.

What Updates Were Added

NVIDIA and Hugging Face integrated additional deep learning models and frameworks into LeRobot, facilitating three key developer tasks:

• Training on datasets of real robot manipulations without needing to collect data from scratch
• Rapid prototyping and testing of control policies on different types of robotic systems
• Optimizing computations on NVIDIA GPUs to accelerate model training

Updates allow adaptation of ready-made models to specific tasks: grasping parts, assembling objects, sorting items.

Why This Matters for Developers

Robot development traditionally requires significant R&D investment because each robot has unique kinematics, environments differ, and algorithms are difficult to transfer between projects. LeRobot changes this approach by offering a common platform for experience sharing between teams. Similarly, Hugging Face transformed development in natural language processing, allowing thousands of teams to use a unified infrastructure and ready-made models.

Who Else Drives Robotics Development

LeRobot is one of several initiatives for opening up robotics. In parallel, Figure AI and Boston Dynamics (industrial manipulators), Tesla Optimus (humanoid robot), Google DeepMind (reinforcement learning algorithms), and OpenAI (integrating large language models into robotic systems) are working.

What This Means

Support for LeRobot from NVIDIA and Hugging Face indicates a transition in robotics from closed corporate projects to open ecosystems. This will accelerate the emergence of new robots for automation in logistics, manufacturing, and service professions. It will become easier for developers from small teams and universities to create competitive solutions by leveraging ready-made models and datasets instead of developing everything from scratch.