Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,35 @@
|
|
1 |
-
---
|
2 |
-
license: apache-2.0
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
tags:
|
6 |
+
- autonomous-driving
|
7 |
+
- end-to-end
|
8 |
+
- personalization
|
9 |
+
pretty_name: StyleDrive
|
10 |
+
size_categories:
|
11 |
+
- 1B<n<10B
|
12 |
+
---
|
13 |
+
We release the checkpoints and StyleDrive's preference annotations in this huggingface repo.
|
14 |
+
|
15 |
+
<B>StyleDrive</B> is the first large-scale real-world dataset designed to advance personalized end-to-end autonomous driving (E2EAD).
|
16 |
+
|
17 |
+
This repository includes:
|
18 |
+
|
19 |
+
🧠 Model checkpoints.
|
20 |
+
|
21 |
+
📊 StyleDrive datasets' annotations.
|
22 |
+
|
23 |
+
These resources support training and evaluating models under our proposed benchmark for personalized E2EAD, enabling more human-aligned and context-aware driving behaviors.
|
24 |
+
|
25 |
+
For env configs, code, training pipelines, and benchmark details, please visit our [GitHub repository](https://github.com/AIR-THU/StyleDrive) and [Paper](https://arxiv.org/abs/2506.23982).
|
26 |
+
|
27 |
+
If you find StyleDrive is useful in your research or applications, please consider giving us a star 🌟 on GitHub and citing it by the following BibTeX entry:
|
28 |
+
```bibtex
|
29 |
+
@article{hao2025styledrive,
|
30 |
+
title={StyleDrive: Towards Driving-Style Aware Benchmarking of End-To-End Autonomous Driving},
|
31 |
+
author={Hao, Ruiyang and Jing, Bowen and Yu, Haibao and Nie, Zaiqing},
|
32 |
+
journal={arXiv preprint arXiv:2506.23982},
|
33 |
+
year={2025}
|
34 |
+
}
|
35 |
+
```
|