Exciting things at LeRobot! 🤖 We’ve just integrated Meta-World: a benchmark for testing multi-task and generalization abilities in robotic manipulation: huggingface.co/docs/lerobot/e… We’ve also cleaned up our environments and standardized on Gymnasium ≥ 1.x.x and MuJoCo ≥ 3.0.0

Finally see some decent research using SO101 as experiment platform!

This is what it looks like preparing for the Embodied AI hackathon by Seeed Studio NVIDIA Hugging Face in Shenzhen

Why Open-source With Isaac sim, GR00T, SmolVLA, 660$ XLeRobot can organize stuff into drawer and zipping up bag in 2 days Made by MakerMods Isaac Sin QI LIU Ryan Chan on Seeed Studio Hugging Face NVIDIA Robotics Home Robot Hackathon in Shenzhen (Bay Area next)

Our hackathon submission was posted by the author of XLeRobot himself Vector Wang

🥉 3rd place at the Embodied AI Hackathon by NVIDIA Robotics Seeed Studio Hugging Face - I made flipped dual-arm layout with elevating mechanism XLeRobot - Isaac Sin got it working with MakerMods AI - QI LIU , Bubbles and Kaho tuned SmolVLA for a smooth demo

It's absolutely incredible for us. We never imagined we could design the hardware for a liftable platform and debug the VLA model for the workflow all within a single day. This has been an incredibly satisfying solution for us.

Running a GR00t N1.5 model for matcha pouring. Getting ready for hackathon Seeed Studio NVIDIA Robotics Vector Wang

Try out this awesome work: LoGoPlanner on your LeKiwi now!

🌁NeurIPS Conference World Model Workshop Chelsea Finn (Physical Intelligence): Evaluate VLA Agent 🚀π0.6 scalable RL 1️⃣Mixed data: policy rollout+human correct 2️⃣Progress-to-go: pretrain VLM cond on image 3️⃣Adv-cond SFT learn→tell model 📊2×throughput >90%☕️success 👉🏻linkedin.com/pulse/92-chels…

We just upgraded XLeRobot 🚀 Built by the MakerMods team Isaac Sin, Mr Thompson and QI LIU. • Easier to build • Improved chassis • Reduced 3D-print time and material • Designed in collaboration with the original author Vector Wang Fully open-source Full build

NVIDIA Robotics Hugging Face Get started here: huggingface.co/docs/lerobot/e… Blog: huggingface.co/blog/nvidia/ge…

Building upon SimpleVLA-RL, we have implemented real-world RL on long-horizon dexterous tasks and witnessed a non-trivial (~relatively 300%) performance improvement over the SFT model, along with surprising capabilities on auto-recovery. Blog coming soon. The entire process uses

chat with papers for any arXiv link to HF paper you can now chat using Hugging Chat All Hugging Face Papers now include a built-in assistant, powered by HuggingChat and the Hugging Face MCP server. It helps you quickly understand papers by answering questions, summarizing

This is the most dexterous task I’ve seen a humanoid do so far. Fully autonomous powered by Sharpa’s CraftNet (VTLA) — using tactile feedback to continuously fine-tune the last-millimeter interaction.

Meta × TBD Lab × CMU × UChicago × UMaryland In our latest work, we introduce Token-Level LLM Collaboration via FusionRoute 📝: arxiv.org/pdf/2601.05106 LLMs have come a long way, but we continue to face the same trade-off: – one huge model that kind of does everything, but is

roberto Good step there. AGI still loading…

Rid Yep! Our NEO data is 98.5% pick and place, and most tasks we show are zero-shot with no similar robot data at all

@WesRothMoney I this is a game changer, too bad this will be limited by the amount of available compute. $IREN