---
task_categories:
- robotics
license: mit
tags:
- robotics
- multimodal
---
The FuSe dataset contains 26,866 trajectories collected on a WidowX robot at the RAIL lab @ UC Berkeley, USA. It contains visual, tactile, sound and action data collected across several environments, annotated with natural language.

Paper: https://huggingface.co/papers/2501.04693
Code: https://github.com/oier-mees/FuSe

If you use the dataset please cite:
```bibtex
@inproceedings{jones25fuse,
  title={Beyond Sight: Finetuning Generalist Robot Policies with Heterogeneous Sensors via Language Grounding},
  author={Joshua Jones and Oier Mees and Carmelo Sferrazza and Kyle Stachowicz and Pieter Abbeel and Sergey Levine},
  booktitle = {Proceedings of the IEEE International Conference on Robotics and Automation (ICRA)},
  year={2025},
  address = {Atlanta, USA}
}
```