---
title: GiTBench_sim_train_lerobotv3
canonical_url: "https://www.modelscope.cn/datasets/xgq12345/GiTBench_sim_train_lerobotv3"
md_url: "https://www.modelscope.cn/datasets/xgq12345/GiTBench_sim_train_lerobotv3.md"
repository: xgq12345/GiTBench_sim_train_lerobotv3
last_updated: 2026-10-16
license: apache-2.0
storage_size: "9.0 GB"
downloads: 129
stars: 0
---

# GiTBench_sim_train_lerobotv3

> GiTBench_sim_train_lerobotv3 - xgq12345 在 ModelScope 开源的数据集。in prose gets mangled into a tag -->

xgq12345/GiTBench_sim_train_lerobotv3 是 ModelScope 魔搭社区上的数据集，存储大小 9.0 GB，采用 apache-2.0 许可。

- **Repository**: xgq12345/GiTBench_sim_train_lerobotv3
- **License**: apache-2.0
- **Storage size**: 9.0 GB
- **Downloads**: 129
- **Stars**: 0
- **Last updated**: 2026-10-16

Source: https://www.modelscope.cn/datasets/xgq12345/GiTBench_sim_train_lerobotv3

---

<!-- ============================================================= -->

<!--  DATASET CARD — works on both HuggingFace and ModelScope.      -->

<!--  Still to fill in by hand (search the ALL-CAPS keywords):      -->

<!--    TEASER  ->  asset/GIT.png                                  -->

<!--    GITHUB  ->  every <GITHUB_URL>                              -->

<!--    ARXIV   ->  every <ARXIV_URL> and <ARXIV_ID>                -->

<!--    <Authors> / <YEAR> / <EMAIL>                                  -->

<!--  ------------------------------------------------------------  -->

<!--  WARNING: this file is formatted by Prettier on save (VSCode). -->

<!--  Prettier rewrites angle-bracket placeholders, so:             -->

<!--    * keep placeholders inside HTML attributes or code fences — -->

<!--      a bare `<Word Word>` in prose gets mangled into a tag     -->

<!--    * write links as  <a href="<GITHUB_URL>">GitHub</a>         -->

<!--      NOT as the markdown form [GitHub](<GITHUB_URL>), whose    -->

<!--      brackets Prettier strips (silently breaking the link)     -->

<!-- ============================================================= -->

# GiTBench training data

<!-- TEASER: asset/GIT.png (1920x1080). Keep this repo path in sync with
     wherever the file actually lands, or the image will 404 on the card. -->

<p align="center">
  <img src="asset/gitbench.png" alt="Grounded in Time — teaser" width="90%">
</p>

<p align="center">
  <a href="https://kk-stephen.github.io/grounded-in-time/"><img src="https://img.shields.io/badge/%F0%9F%8C%90%20Website-GiT-blue" alt="Website"></a>
  <a href="https://github.com/xgq2005/GiTBench"><img src="https://img.shields.io/badge/GitHub-Code-black?logo=github" alt="GitHub"></a>
  <a href="https://arxiv.org/abs/2610.04255"><img src="https://img.shields.io/badge/arXiv-Paper-b31b1b?logo=arxiv" alt="arXiv"></a>
  <a href="https://huggingface.co/datasets/XGQ12345/GiTBench_sim_train_lerobotv3"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20HuggingFace-Dataset-yellow" alt="HuggingFace"></a>
  <a href="https://www.modelscope.cn/datasets/xgq12345/GiTBench_sim_train_lerobotv3"><img src="https://img.shields.io/badge/ModelScope-Dataset-blue" alt="ModelScope"></a>
</p>

<!-- ===================== LINKS ===================== -->

<p align="center">
  <b>[ <a href="https://kk-stephen.github.io/grounded-in-time/">Website</a> ]
     [ <a href="https://github.com/xgq2005/GiTBench">GitHub</a> ]
     [ <a href="https://arxiv.org/abs/2610.04255">arXiv</a> ]
     [ <a href="https://huggingface.co/datasets/XGQ12345/GiTBench_sim_train_lerobotv3">HuggingFace</a> ]
     [ <a href="https://www.modelscope.cn/datasets/xgq12345/GiTBench_sim_train_lerobotv3">ModelScope</a> ]</b>
</p>

<!-- =================================================== -->

Lerobot v3 format of GiTBench training data. There are a total of 9 tasks, each with 160 training episodes.

## 📜 **Citation**

If you find this dataset and benchmark useful, please cite:

```bibtex
@article{wang2026grounded,
  title={Grounded in Time: A Multi-Source Dataset and Benchmark for Temporal Grounding in Robotic Manipulation},
  author={Wang, Yi and Yang, Yang and Xu, Guangqi and Lin, Sumin and Kang, Ning and Lu, Pengxiang and Chen, Xiaotong and Xue, Zeyu and Deng, Ping and Liu, Xing and others},
  journal={arXiv preprint arXiv:2610.04255},
  year={2026}
}
```

<!-- TODO: fill in the BibTeX above, and make sure <ARXIV_ID> matches the
     arXiv URL used at the top of this card. -->

---

## 📄 License

Released under the Apache-2.0 license. See `LICENSE` for details.

## 🙏 Acknowledgments

We thank the ManiSkill-3 team for their open-source simulation platform and the work that makes this benchmark's manipulation environments possible.

<!-- TODO: funding, lab, robot platform, simulator, and anyone who helped
     with data collection or annotation. -->

## 📮 Contact

For questions, issues, or collaboration, please open an issue on
<a href="https://github.com/xgq2005/GiTBench">GitHub</a>.
