Stone Tao

Are there (mobile) manipulation tasks and baselines for those?

3

Hi all, I was wondering if OmniGibson / Behavior-1K has any manipulation tasks yet, and if there are reinforcement learning or learning from demos baselines? Moreover, I saw that there...

### Algorithms / RL **Online** - [ ] APPO? - [ ] RLPD? **Offline** - [ ] CQL - [ ] [Cal-QL](https://arxiv.org/abs/2303.05479) - [ ] AWAC ### Robotics Related -...

Does OpenVLA baseline work in the simulated environments?

1

Was taking a look at the OpenVLA baseline in colab but it seems the final video is a failure. For some context I'm in the middle of porting the simpler...

Citation

1

What would the appropriate citation be for this repo? I can cite the GitHub repo but if there is a arxiv upload or something I can cite that instead. Thanks!

Vectorized/GPU Sim Evaluation / ManiSkill 3 Port

1

- [x] Add GPU sim environments for bridge dataset (in ms3 repo: https://github.com/haosulab/ManiSkill/pull/536) - [x] Modify Octo to perform batched inference - [x] Modify Octo to accept torch tensors/dl packed...

Migrate example environments to ManiSkill 3

2

Intention is to replace the currently git submoduled enviroments with the new GPU accelerated ManiSkill 3 package and some better packaged code designed to be easily extended for user's custom...

Currently we rely on trimesh heavily but it is not neatly parallelized.

A list of tasks planned to be added to ManiSkill and eventually come with demos. - [ ] OpenCabinetDrawer - [ ] Multiple drawing related tasks - [ ] Clean...