yuanty/LIBERO-fastwam

---
pretty_name: LIBERO MuJoCo 3.3.2 (Preprocessed LeRobot v2.1 Release)
license: cc-by-4.0
tags:
- robotics
- robot-learning
- imitation-learning
- lerobot
- libero
- embodied-ai
---

# LIBERO MuJoCo 3.3.2 (Preprocessed LeRobot v2.1 Release)

This repository releases **our preprocessed LIBERO dataset** in LeRobot `v2.1` format for the open-source release of **Fast-WAM: Do World Action Models Need Test-time Future Imagination?**

This is not an official upstream LIBERO data dump. It is our paper-specific processed release prepared to support training, evaluation, and reproducibility for our project.

This Hugging Face repository distributes the dataset as four `.tar.gz` archives, one per LIBERO subset:

```text
README.md
libero_10_no_noops_lerobot.tar.gz
libero_goal_no_noops_lerobot.tar.gz
libero_object_no_noops_lerobot.tar.gz
libero_spatial_no_noops_lerobot.tar.gz
```

## Summary

- Provenance: preprocessed from LIBERO for the Fast-WAM open-source release
- Format: LeRobot `v2.1`
- Environment backend: `MuJoCo 3.3.2`
- Robot type: `franka`
- Number of subsets: `4`

## Project

- Project page: https://yuantianyuan01.github.io/FastWAM/
- Paper: https://arxiv.org/abs/2603.16666
- Code repository: https://github.com/yuantianyuan01/FastWAM

## Important

This dataset is **MuJoCo-version sensitive**. It was generated with **MuJoCo `3.3.2`**, and this version matters for reproducibility. If you use a different MuJoCo version, you should expect dataset or environment mismatch.

## Download and Extract

Download **all** archive files from this repository, then extract all of them locally:

```bash
for f in *.tar.gz; do
tar -xzf "$f"
done
```

After extraction, you will get:

```text
libero_10_no_noops_lerobot/
libero_goal_no_noops_lerobot/
libero_object_no_noops_lerobot/
libero_spatial_no_noops_lerobot/
```

Each subset follows the standard LeRobot `v2.1` layout:

```text
libero_10_no_noops_lerobot/
data/
meta/
videos/
```

## Loading

After extraction, each subset can be loaded as a standard LeRobot `v2.1` dataset.

## Notes

- The release is distributed as per-subset archives. Users need to extract a subset locally before loading it with LeRobot-compatible tooling.
- Please review the upstream LIBERO redistribution terms before making this repository public.

## License

This release is a preprocessed derivative of the LIBERO datasets for Fast-WAM.

It is released under CC BY 4.0, consistent with the upstream LIBERO dataset license. Please also refer to the official LIBERO project for upstream attribution and terms.

## Citation

If you use this release, please cite the Fast-WAM paper.

```bibtex
@misc{yuan2026fastwam,
title={Fast-WAM: Do World Action Models Need Test-time Future Imagination?},
author={Tianyuan Yuan and Zibin Dong and Yicheng Liu and Hang Zhao},
year={2026},
note={arXiv preprint arXiv:2603.16666}
}
```

If you use the underlying LIBERO benchmark or data source, please also cite the original LIBERO work and project.