Seeed Embodied AI Hackathon
Build Transformers and Diffusers Powered Robots with LeRobot
Hugging Face's LeRobot open-source framework, combined with NVIDIA AI and robotics technology, enables researchers and developers to drive advancements in AI robotics. Join our weekend hackathon at Circuit Launch in Mountain View!
Event Overview
Explore imitation learning with the open-source SO-ARM100 robotics arm kit
Utilize Hugging Face's LeRobot framework for advanced AI architectures like ACT and Diffusion Policy
Calling developers, engineers, and robotics enthusiasts eager to explore advanced AI and robotics.
Check out Embodied AI Hacks-01 highlights：
Check out our wiki documentation
Check out SO-ARM100 Dev Kit to start building your own robot arm
Hands-on Experience Includes:
Working with the SO-ARM100, a powerful 6DoF robotics arm designed for quick assembly, synchronized teaching tasks, and seamless integration with Hugging Face's LeRobot platform.
Collecting data through teleoperation and simulation in NVIDIA Issac Lab.
Training robot policies with imitation learning and developing validated policies on real-world robots powered by NVIDIA Jetson for real-time inference.
🎉🎉 Winner Prize
1st team: 2x bundles, each bundle includes 1x SO-ARM100 Kit and 1x NVIDIA Orin Nano Super Developer Kit
2nd team: 5x $100 shopping coupon for Seeed Studio bazaar
3rd team: 5x reCamera AI cameras
Hackathon Workflow
The planned collaborative workflow involves collecting data through teleoperation and simulation in Issac Lab, and storing them in the standard LeRobot Dataset format. Then they'll be used to train a robot policy with imitation learning, which is subsequently evaluated in simulation. Finally, the validated policy is deployed on real-world robots with NVIDIA Jetson for real-time inference.
We'd like to see you complete the LeRobot getting-started part by following the next steps, and also create the coolest application demo video using LeRobot to perform creative tasks :
Assembled two arms using SO-ARM100 robotics arm kit and 3D printed parts
Setting up imitation learning environments
Data acquisition and training
Deploying tasks on dual robotic arm
Deploy the validated model using NVIDIA Jetson Orin
Then more tasks built upon the getting-started part would be even better:
Build with reCamera, such as start performing a task(pour coffee/tidy up desk, etc) by recognizing a dedicated situation through reCamera
Voice command controls robot arm movements with reSpeaker
Propose your task ideas, let's see how we can help and discuss further!
⏰ Timeline
Saturday
10:00 AM: Check-in & Kickoff
12:00 PM: Lunch
6:00 PM: Dinner
8:00 PM: Quiet Hours
Sunday
8:00 AM: Breakfast
12:00 PM: Lunch
4:00 PM: Demos & Snacks & Networking
We’ll conclude the hackathon with a demo session showcasing the technical projects participants have built—no company pitches.
If you’re attending only the demo session, please indicate this during signup so we can plan for food.
What We Provide
SO-ARM100 Low-Cost AI Arm Kit:
Includes servo motors, control board, power supply, and cables
6DoF robotic arm with leading and following arms
3D-printed components for participants
reComputer J4012 Powered by Jetson Orin NX 16GB:
compact and powerful computing platform for model deployment
reCamera:
Tiny AI camera in modular design, with customizable base-board, Linux-based OS, and Node-RED integration for no-code development
Integrates voice processing and audio playback to be a smart voice assistant
Online Resources and tech support in-person:
Assembly, calibration, teleoperation, and more steps are available on our wiki
If you have any questions in mind, feel free to contact us anytime via email: edgeai@seeed.cc
What You'll Learn:
Assembly and Calibration: Get hands-on with assembling and fine-tuning the robotic arm
Collecting real-world robot data through teleoperation
Access NVIDIA Issac Lab to accelerate LeRobot's data collection, training, and verification workflow
Generated motion data is then used to train a policy with imitation learning
The Robotics Community Thrives When We Build Together
Join forces with LeRobot, NVIDIA, and Seeed Studio to explore building advanced robotic arms and accelerate research and innovation in AI robotics.
LeRobot, developed by Hugging Face, extends the successful paradigms from its Transformers and Diffusers libraries into the robotics domain. LeRobot offers a comprehensive suite of tools for sharing data collection, model training, and simulation environments along with designs for low-cost manipulator kits.
NVIDIA’s AI technology, simulation, and open-source robot learning modular framework such as NVIDIA Isaac Lab can accelerate LeRobot’s data collection, training, and verification workflow. Researchers and developers can share their models and datasets built with LeRobot and Isaac Lab, creating a data flywheel for the robotics community.
Hackster.io is a dynamic community for makers and developers, where you can explore and collaborate on exciting hardware and software projects. Known for its focus on DIY electronics, IoT, and robotics, Hackster is the perfect platform to dive into hands-on experiences, enabling collaborations and knowledge-sharing with cutting-edge technologies such as edge computing, robotics, and physical AI.
Seeed is a global leader in open hardware innovation and an NVIDIA Elite Partner, We specialize in cutting-edge open hardware platforms. We aim to democratize technology through IoT and AI, empowering developers worldwide to break barriers in hardware innovation.