正在加载视频...

视频加载失败

Introducing LDA, a latent world action foundation model that, for the first time, unifies the utilization of heterogeneous embodied data across simulation and reality, humans and robots, and varying levels of action quality and annotation. By breaking long-standing data silos in embodied intelligence, LDA enables the field, much like...

40,883 次观看 • 6 个月前 •via X (Twitter)

5 条评论

Jiangran Lyu 的头像
Jiangran Lyu6 个月前

We have open-sourced the model and code! Project Page: arXiv:

Sharpa 的头像
Sharpa6 个月前

Nothing better than Sharpa hands to cook a burger just right! Joke aside, it's great to see the ecosystem coming up with new recipes to scale robot learning👍

abdel 的头像
abdel6 个月前

oh no why do you release it now.. i just finished my essay on world models haha and I am not sure yet in which category your world model would fit in. seems exciting, great job.

Tradeye 的头像
Tradeye6 个月前

LDA just dropped the mic @GalbotRobotics 🔥 Unifying sim + real, human + robot data? Absolute game changer! Tradeye is out here supplying the messy real-world chaos you need: narrated POV video from licensed HVAC/plumbing/electrical crews in actual tight attics, crawlspaces & lived-in homes (C20 licensed of course). DM me if you want some authentic trades data to supercharge the model? Let’s make robots unstoppable together 🚀

Akito 的头像
Akito6 个月前

Can consumers-captured video serve as training data for the model??

相关视频