Hugging Face and Pollen Robotics unveil 'Le Robot,' an open-source humanoid robot capable of performing home tasks. Discover how AI training using VR teleoperation sessions led to Reachy2's impressive capabilities. Explore the open-source data and model on GitHub.
Hugging Face's Open-Source Humanoid Robot 'Le Robot' Shows Promising Results
Hugging Face, a Brooklyn-based AI company, created an open-source robotics application named “Le Robot” earlier this year and has made its first results public.
Remi Cadene, an engineer at the company, shared a video on X, which appears to be the first demonstration of a functional humanoid robot from Hugging Face’s Le Robot program.
Also Read: AI Virtual Rat by Harvard and DeepMind Sheds Light on Brain-Body Connection
Known as “Reachy2,” it is a humanoid robot created by Bordeaux, France-based Pollen Robotics, another open-source robotics startup. Hugging Face collaborated with Pollen Robotics to teach the robot various home tasks and how to securely interact with people and dogs.
Cadene elaborated on the unique training methodology of Reachy2 in a direct message sent over X to VentureBeat. He clarified that the robot was first “teleoperated” by a person donning a virtual reality headset, who guided it through a variety of tasks, such as setting cups on a dishrack and skillfully and safely handing an apple or other object to a person.
After watching 50 movies of the VR teleoperation sessions, each lasting around 15 seconds, a machine learning system learned how to perform each task independently and directed Reachy2 to complete it.
Every video is connected to a separate Reachy2 robot sensor. After training for 40,000 and 60,000 steps, Reachy2 was able to turn an apple, hand, and back to its starting position.
Also Read: China Creates World’s First 3D e-Skin, Helping Robots Feel Human-like Sensing
The company has made the Hugging Face dataset and the model it used for the demonstration available to the robotics company as open-source projects.
The information is there for everyone to use, as Cadene emphasizes. “On X, he said, you can do the same at home on smaller robots.”
With its all-new 7-DoF bio-inspired arms (7-pound payload each) and AI-ready electronics that enable enhanced perception, interaction, and manipulation, Pollen Robotics Reachy2 is perfect for such applications, supporting the research.
In March 2024, Hugging Face—which keeps one of the biggest collections of open-source AI models and codes—started working on the Le Robot project.
Its goal is to produce an open-source toolset that will empower AI robotics and encourage the next wave of roboticists. You may get it right now on GitHub.
Also Read: AI Steve: A businessman proposes the first AI MP Candidate for Brighton in the UK General Election
Hugging Face hopes that by making it open-source, a worldwide community of academics, developers, and enthusiasts will be able to contribute to and gain from the overall advancement of AI robots.
This post was last modified on June 13, 2024 2:57 am
Rish Gupta is an Indian entrepreneur who serves as the chief executive officer (CEO) of…
Are you looking to advance your engineering career in the field of robotics? Check out…
Artificial intelligence is a topic that has recently made internet users all over the world…
Boost your learning journey with the power of AI communities. The article below highlights the…
Demystify the world of Artificial Intelligence with our comprehensive AI Glossary and Terminologies Cheat Sheet.…
Scott Wu is the co-founder and Chief Executive Officer of Cognition Labs, an artificial intelligence…