Pinned
Excited to introduce PyRoki ("Python Robot Kinematics"): easier IK, trajectory optimization, motion retargeting... with an open-source toolkit on both CPU and GPU
00:00
Chung Min Kim
102 posts
Chung Min Kim
@ChungMinKim
PhD @ UC Berkeley 🧸🌅
Joined October 2021
- 🪴Should the leaves of a plant be considered separate or part of the whole? Answer: it really depends! Points can, and should, belong to multiple groups. With GARField, points can belong to multiple groups, with physical scale 📏 as an extra dimension. garfield.studio
00:00 - choosing viser has simply been one of the best research decisions I've made -- interacting with your 3D scene does wonders for your debugging speed!
00:52July has been a big month for Viser! - Released v1.0.0😊 - We did some writing Some demos👇 - Replying to @ChungMinKimThis work wouldn't have been possible without my collaborators @brenthyi @redstone_hong @YiMaTweets @Ken_Goldberg @akanazawa ! The toolkit is available here: Website: pyroki-toolkit.github.io Code: github.com/chungmin99/pyr…
- Introducing our #CoRL2023 (oral) project: 🥡 LERF-TOGO uses LERF for zero-shot semantic grasping. Check out our website at lerftogo.github.io! LERF (lerf.io) can localize the "handle" of the mug or the "neck" of a bottle... so we tell robots where to grasp!Language Embedded Radiance Fields for Zero-Shot Task-Oriented Grasping paper page: huggingface.co/papers/2309.07… Grasping objects by a specific part is often crucial for safety and for executing downstream tasks. Yet, learning-based grasp planners lack this behavior unless they are
00:00 - It’s notoriously difficult to model the mechanics of compliant robot jaw tips during grasping! We found that a new tool from computer graphics can help. IPC-GraspSim, from @AUTOLab_Cal @UCBerkeley. Paper, data, video: sites.google.com/berkeley.edu/i… (1/9)
00:00 - Replying to @arthurallshire @redstone_hong and 2 otherswalking with videomimic 😀🤝🤖
00:00 - BAIR is filled with so much fun, come join us! 🧸This week in the @berkeley_ai CompVision meeting we did 5 min presentations: short powerpoint presentations about anything you want, as long as it isn't related to work. Here is a sampling of what transpired:
- Replying to @ChungMinKimPyRoki's focus is flexibility, but it's also fast: up to 1.4-1.7x faster than cuRobo for GPU-accelerated IK! (with higher success and lower errors, see paper ✨)
- Robot see 👀 (recover part motions) Robot do 🦾 (optimize robot traj for motion)Robot See, Robot Do allows you to teach a robot articulated manipulation with just your hands and a phone! RSRD imitates from 1) an object scan and 2) a human demonstration video, reconstructing 3D motion to plan a robot trajectory. robot-see-robot-do.github.io #CoRL2024 (Oral)
00:00 - Excited to present 📎LERF🚜 at #ICCV2023! Come see our poster this Friday afternoon 02:30 PM-04:30 PM, at poster 163 in room "Foyer Sud"! Come chat with @justkerrding* and me* :) Our talk is at Room "Paris Sud" 4:30-6:30Replying to @akanazawaLERF: Language Embedded Radiance Fields, w/@justkerrding*, @ChungMinKim*, @Ken_Goldberg, Matt Tancik Embed CLIP📎 in NeRF🚜 for semantically queryable 3D!! 🔎 🌐lerf.io Friday afternoon Oral🥳& Poster!
00:00 - One of the videos: 🧸👋 (bear waves hello)
00:00I’m not at CVPR this year🥲, but @justkerrding and @ChungMinKim are doing a talk *Tuesday 2:15pm* on GARField, LERF, and the latest updates🔥 It's gonna be an awesome talk with many cool videos✨ Please check it out! At opensun3d workshop 2:15pm 6/18 Arch 211 #CVPR2024 - Replying to @ChungMinKimWhat's GARField? Group Anything with Radiance Fields! The groupings in 3D are optimized using multi-level masks from Segment-Anything (SAM). Once trained, these groups can be *extracted* and moved around...
00:00
