1/ Introducing HIW-500 (Humanoids-in-the-Wild 500): the largest open-source humanoid teleop dataset collected in real homes Built w/ Unitree Hugging Face across 12 homes in Southeast Asia, it covers: > 500+ hrs > 23K+ episodes > 10+ TB > 10+ household tasks

2/ Humanoid robots have made tremendous progress in locomotion. But whole-body mobile manipulation in real homes remains one of the biggest unsolved challenges in robotics. Most datasets are collected in controlled environments. Real homes are messy, unpredictable, and full of edge cases. To build "robot butlers" that work in the real world, we need data from the real world. HIW-500 is a step in that direction.

3/ To build this dataset, we used the following hardware setup: > Unitree G1 humanoids > Stereo head camera > IR stereo wrist cameras > Grippers This was designed to capture both whole-body motion and manipulation in everyday household environments.

4/ Each episode includes: > Head camera RGB (480p, 30 FPS) > Wrist camera RGB + IR (480p, 30 FPS) > Full robot state & actions (29 DoF, IMU, odometry) > Episode and sub-task annotations

5/ The dataset contains 10+ long-horizon household tasks, with some episodes lasting more than 8 minutes. And 10+ long-horizon tasks spanning navigation, object manipulation, and multi-step activities that require coordination across the entire robot body.

6/ With thousands of demonstrations per task and detailed sub-task annotations, researchers can train and evaluate models across multiple levels of abstraction.

7/ This project represents months of work alongside incredible partners. Anyone who has tried teleoperating a humanoid robot knows how challenging it is. Collecting thousands of high-quality whole-body demonstrations inside real homes is an entirely different level of difficulty. Huge thanks to @UnitreeRobotics for the hardware support. And to @huggingface for partnering with us once again after our previous work on open-sourcing large-scale urban navigation datasets.

8/ Full 10TB dataset in ROSbag + @LeRobotHF ↓ Dataset: LeRobot: For research / commercial use of the expanded HIW-500 dataset → or reach out to us directly!

@UnitreeRobotics @huggingface That's great! Putting web3 x robotics on the map.

@UnitreeRobotics @huggingface Massive step forward for real world humanoid training

@UnitreeRobotics @huggingface 🔥

@UnitreeRobotics @huggingface 🦾

@UnitreeRobotics @huggingface Great

@UnitreeRobotics @huggingface That's huge

@UnitreeRobotics @huggingface 棒棒棒

@UnitreeRobotics @huggingface Keep it up

@UnitreeRobotics @huggingface 🦾

@UnitreeRobotics @huggingface Congratulations team

@UnitreeRobotics @huggingface A real training environment for robots produces great results. Kudos!

@UnitreeRobotics @huggingface Bitrobot power😇

@UnitreeRobotics @huggingface gBot 🦾

@UnitreeRobotics @huggingface Awesome 🧸

@UnitreeRobotics @huggingface 🤖 🦾🦿

@UnitreeRobotics @huggingface 🧡🧡

@UnitreeRobotics @huggingface This is massive, we love to see it 🔥🤖

@UnitreeRobotics @huggingface LFG

@UnitreeRobotics @huggingface great

@UnitreeRobotics @huggingface This is so cool. I’m a huge fan of robotics & humanoids We’re throwing a huge event in LA in a few months partnering up with a big robotics company & Solana fndn. We’d love to have you join and show off your bots :)

@UnitreeRobotics @huggingface huge step for real-world robotics data - this is exactly what’s been missing for humanoid learning 🤖📦

@UnitreeRobotics @huggingface good job!

@UnitreeRobotics @huggingface LFG

@UnitreeRobotics @huggingface ????

@UnitreeRobotics @huggingface Huge infrastructure project

@UnitreeRobotics @huggingface that's the part that's often overlooked collecting and standardizing real-world data is hard, way harder than making a humanoid robot that can follow instructions in a lab

@UnitreeRobotics @huggingface Massive win for open source

@UnitreeRobotics @huggingface Open datasets like this are essential for advancing humanoid robotics.Real-world household data captures the variability, clutter, and unpredictability that robots must learn to handle outside the lab, helping bridge the gap between research and practical deployment. 🤖🏠📊

@UnitreeRobotics @huggingface robots learning from real life is kinda wild ngl