~ Lowering the barrier to entry for robotics ~ Crafted with care by @HuggingFace 🤗 Join our discord: https://xitter.cf/t.co/Sx2jdT0jeF
Worldwide
Joined July 2024
- Tweets662
- Following28
- Followers21.3K
- Likes950
Pinned Tweet
Training a robot to grasp an object means learning from many attempts, viewpoints and situations. One robot with three cameras at 20 fps produces 216,000 images per hour. Now that robotics is scaling, how can we efficiently train on all this data?
Read how we help @microagi scale their infra on @huggingface!
As our customers scale to training robot models using millions of hours of our recordings, we needed to build a new kind of data infrastructure.
Terabytes of data delivered in minutes, and the capacity to deliver over a Petabyte per day to each of our customers, to multiple customers at the same time. That’s more than 10 Gigabytes a second, delivered into any cloud, wherever our customers need it.
Using our customer portal, customers select precisely the data they want and it starts arriving in minutes at their destination, including AWS, GCP, Cloudflare, Hugging Face and any cloud provider who can handle the ingest speed.
We’ve written about how we worked with @huggingface and @Cloudflare to build this secure, scalable infrastructure. Link in comments.
New in LeRobot: DM0.5 by @Dexmal_AI 🤖
A 4B VLM with a 680M flow-matching action expert, pretrained on 7 robot embodiments. Right now it's the top open-source model on RoboColiseum.
Fine-tune it on your robot:
--policy.path=lerobot/dm05_base
Docs 👇
huggingface.co/docs/lerobot/…
We checked the port on LIBERO: lerobot/dm05_libero solves 197/200 episodes across all four suites.
DM0.5 is on main now and ships in the next LeRobot release.
Thanks @Dexmal_AI for the model and the video 🎥 How it works 👇
dexmal.com/blog/dm0.5
Another dataset now on Hugging Face, 600 hours of manipulation data!
We release Gen-HumanWholeBody, an egocentric whole-body manipulation dataset with industry's highest precision. The dataset is now open-source! 🚀 huggingface.co/datasets/genr…
🚶600 hours of whole-body manipulation data
🎯 ~3 cm whole-body MPJPE
🌍 16,976 real-world scenarios
🦾 1,281 human skills
We’re opening this dataset to help researchers and developers advance whole-body motion learning, human-to-robot transfer and embodied foundation models.
Contact Us:[email protected]
Discord: discord.gg/rSSb5thgu
So many SO-ARMs! 💪
Here is the robotics workshop organized at @NVIDIAGTC Berlin that uses SO-ARM101 and LeRobot to control real and simulated robotic arms. The light box makes it easier to get policies working in any location and light conditions.
We trained π0.5 for 8 hours on 4× H100, streaming a 10.7 TB dataset straight from the Hub.
Time spent waiting for data: 0.52% of step time.
No full download first.
Just add --dataset.streaming=true
How it works 👇
huggingface.co/spaces/lerobo…
Tell your robot what you want, in plain words.
New in LeRobot: a vision-language model breaks your goal into skills your robot's brain already knows, steers it step by step, and answers questions as it works.
Demo: huggingface.co/spaces/lerobo…
💥 LeRobot × Zenoh
Remote inference just landed in LeRobot: the policy runs on a GPU server or cloud, the robot keeps moving. Cameras, motors and recording stay on the robot computer, actions stream back.
What this unblocks:
🤖 Robot side: a laptop or a small board is enough for cameras, motors, teleop and recording.
🖥️ Server side: big model checkpoints run on a GPU server or in the cloud.
🏃 Smooth motion, with or without RTC: the next action chunk is predicted while the current one executes, and new chunks are aligned and blended into the ongoing motion.
🔌 Works on the same machine, over the lab network, or across the internet through a Zenoh router with TLS.
Join us to talk with LeRobot engineers, ask questions and see what's coming soon to LeRobot.
LeRobot Live - Thursday 8th of October at 4:00 PM CET
Join on our Discord: discord.gg/GZaVGvBcZ?event=1…
LeRobot retweeted
Royal visit from @LeRobotHF at @QualiaRobotics today - Steven doesn’t have X because it’s too much noise - fair.
The robotics data ecosystem keeps getting stronger 🦾
@Voxel51 now reads LeRobot datasets natively, with episode visualization, embedding exploration and semantic search across embodiments 🤗
you balanced your robot dataset by embodiment. the embeddings say you balanced it by recording session
i pulled 10 episodes from each of the 50 robot types in lerobot's community dataset and embedded every clip with qwen3-vl
1. 495 of 497 clips have their nearest neighbor in the same session
2. 292 score below 0.1 uniqueness
3. the two "most unique" clips are single-frame recordings, 0.02 seconds long
filter the session out and the cross-embodiment signal shows up: a cloth fold on one rig lands next to so100 arms folding cloth
i indexed all 497 episodes in fiftyone so you can search by text, cut the near-duplicates, and export a curated lerobot v3 subset
one prompt to an agent with fiftyone skills built the whole thing
start here, read the full blog: huggingface.co/blog/harpreet…
read the dataset card: huggingface.co/datasets/Voxe…
@LeRobotHF @Alibaba_Qwen
LeRobot retweeted
you balanced your robot dataset by embodiment. the embeddings say you balanced it by recording session
i pulled 10 episodes from each of the 50 robot types in lerobot's community dataset and embedded every clip with qwen3-vl
1. 495 of 497 clips have their nearest neighbor in the same session
2. 292 score below 0.1 uniqueness
3. the two "most unique" clips are single-frame recordings, 0.02 seconds long
filter the session out and the cross-embodiment signal shows up: a cloth fold on one rig lands next to so100 arms folding cloth
i indexed all 497 episodes in fiftyone so you can search by text, cut the near-duplicates, and export a curated lerobot v3 subset
one prompt to an agent with fiftyone skills built the whole thing
start here, read the full blog: huggingface.co/blog/harpreet…
read the dataset card: huggingface.co/datasets/Voxe…
@LeRobotHF @Alibaba_Qwen
LeRobot retweeted
Introducing MC-EgoHands, the first all-in-one harness for turning raw ego video into training-ready data.
→ Frontier hand pose (<7mm) across 4 public benchmarks
→ Human grasps retargeted to Panda/LIBERO and Yam in one pass
→ Occluded frames flagged as interpolated or missing
→ Structured episodes in @huggingface @LeRobotHF v3
LeRobot retweeted
The Era of Evals is coming to AI robotics.
Today we’re launching Reality Check: our leaderboard & first public robot manipulation benchmark.
14,400 real-world rollouts. Four models. Multiple tasks & data regimes. Object placements. Confidence intervals. FR3 Duo stations.
1/10
New robot added to LeRobot! 💪
Both DM and RS versions of the reBot B601 from @seeedstudio are now officially supported. The arm features a 2.5kg payload capacity and 6 degrees of freedom + gripper.
In the example below, the robot performs towel folding fully autonomously. The ACT model was trained on just 200 examples.
