metadata
license: bsd-2-clause
pipeline_tag: robotics
Action Chunking with Transformers (ACT) - MujocoUR5eCable
This model is a part of the RoboManipBaselines framework. It implements the Action Chunking with Transformers (ACT) architecture and was trained on the MujocoUR5eCable dataset.
- Paper: RoboManipBaselines: A Unified Framework for Imitation Learning in Robotic Manipulation across Real and Simulation Environments
- Project Page: RoboManipBaselines Project Page
- Repository: GitHub - isri-aist/RoboManipBaselines
Installation
See GitHub for installation instructions.
Policy rollout
To run the trained policy, use the following command from the top directory of the repository:
$ cd robo_manip_baselines
$ python ./bin/Rollout.py Act MujocoUR5eCable --checkpoint ./checkpoint/Act/<checkpoint_name>/policy_last.ckpt
Technical Details
For more information on the technical details of the ACT architecture, please see the following paper:
@INPROCEEDINGS{ACT_RSS23,
author = {Tony Z. Zhao and Vikash Kumar and Sergey Levine and Chelsea Finn},
title = {Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware},
booktitle = {Proceedings of Robotics: Science and Systems},
year = {2023},
month = {July},
doi = {10.15607/RSS.2023.XIX.016}
}
Citation
If you use RoboManipBaselines in your work, please cite the framework paper:
@article{RoboManipBaselines_Murooka_2025,
title={RoboManipBaselines: A Unified Framework for Imitation Learning in Robotic Manipulation across Real and Simulation Environments},
author={Murooka, Masaki and Motoda, Tomohiro and Nakajo, Ryoichi and Oh, Hanbit and Makihara, Koshi and Shirai, Keisuke and Ogata, Tetsuya and Domae, Yukiyasu},
journal={arXiv preprint arXiv:2509.17057},
year={2025}
}