Hugging Face has unveiled “LeRobot,” a new machine learning model trained for real-world robotics applications. LeRobot functions as a platform, offering a versatile library for data sharing, visualization, and training of advanced models.
LeRobot aims to provide models, datasets, and tools for real-world robotics in PyTorch. The goal is to lower the barrier for entry to robotics so that everyone can contribute and benefit from sharing datasets and pretrained models.
LeRobot simplifies project initiation by offering pretrained models and seamless integration with physics simulators. It was recently evaluated on the AlohaTransferCube environment and compared to a similar model trained with the original ACT repository. Results from 500 episodes demonstrate success rates, providing valuable insights into its performance.
Similarly, LeRobot underwent evaluation on the PushT environment and was compared to a model trained with the original Diffusion Policy code. This evaluation also included success metrics over 500 episodes, offering a comprehensive understanding of LeRobot’s capabilities in real-world scenarios.
Designed to accommodate various robotic hardware, from basic educational arms to sophisticated humanoids in research settings, LeRobot aims to provide an adaptable AI system capable of controlling any robot type, thereby improving versatility and scalability in robotics applications.
LeRobot is a “state-of-the-art machine learning for real-world robotics.”
It aims to provide models, datasets, and tools for real-world robotics in PyTorch. The goal is to lower the barrier for entry to robotics so that everyone can contribute and benefit from sharing datasets and pretrained models.
It also contains state-of-the-art approaches that have been shown to transfer to the real-world with a focus on imitation learning and reinforcement learning: github.com/huggingface/lerobot
LeRobot already provides a set of pretrained models, datasets with human collected demonstrations, and simulated environments so that everyone can get started. In the coming weeks, the plan is to add more and more support for real-world robotics on the most affordable and capable robots out there.
And LeRobot hosts pretrained models and datasets on this HuggingFace community page: huggingface.co/lerobot
Github link: https://github.com/huggingface/lerobot
Press release: https://www.infoq.com/news/2024/05/lerobot-huggingface-robotics/
