🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

I believe the world model is the most promising way to absorb diverse data sources.

Wow🤯

🤖Introducing DEMO3: our new model-based RL framework for multi-stage robotic manipulation from visual inputs and sparse rewards. 🧵🔽 📜 Paper: [arxiv.org/abs/2503.01837] 🌍 Project Page: [adrialopezescoriza.github.io/demo3/] 💻 Code: [github.com/adrialopezesco…]

✈️Heading to Singapore 🇸🇬 for #ICLR2025! Super excited to catch up with old friends and meet new folks. 💬DMs open if you wanna talk robot learning, scaling robot data, or video foundation models!

Drop by our poster at Hall 3 #35 tomorrow morning! (Apr 25, 10:30 AM) #ICLR2025

XLeRobot 0.2.0 Trailer: Some simple keyboard joint controls in Maniskill. Pick a bottle, open the fridge, place the bottle, close the fridge. Done in one run, just clipped it to save some time. Not perfect, but real enough.

Robot learning is replicating the success of scaling models and data in LLMs — but what else can we scale? In this work, we study the scaling law with respect to the *number of robot bodies*.

Congrats to my PhD advisor Hao Su🎉🎉 Hao Su

Compete for a $560,000 Prize Pool at IROS 2025 AgiBot World Challenge! 💰 The AgiBot World Challenge – Manipulation Track is LIVE! Hosted by @AgiBot and OpenDriveLab at #IROS2025. 🚀 Challenge: Tackle 10 complex Sim2Real manipulation tasks. 🛠️ Resources: Access a unique

Probably the creepiest robot I have ever seen🥴

The new MeanFlow paper (arxiv.org/abs/2505.13447) claims strong 1-step generation results. I tested it on a simple 2D toy example, and the 1-step performance is indeed impressive! See the visualization below Grab my minimal one-file implementation here: github.com/tongzhoumu/mea…