File size: 2,296 Bytes
ee22b18
 
a5b4ff6
 
 
 
 
ee22b18
 
8485284
a5b4ff6
8485284
ee22b18
35521e3
9417189
dbfc31e
4d19a27
 
 
dbfc31e
9417189
b2e8ebc
ee22b18
e32a581
 
 
a5b4ff6
b7c050e
ee22b18
1adb8a4
 
ee22b18
b7c050e
ee22b18
1adb8a4
1b9c02b
 
b9a0bb7
 
 
ee22b18
b7c050e
5d8f5d2
 
b7c050e
 
ee22b18
b7c050e
 
58cf78f
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
---
license: apache-2.0
tags:
- CARLA
- NAVSIM
- Imitation-Learning
- Closed-Loop-Driving
---

# LEAD: Minimizing Learner–Expert Asymmetry in End-to-End Driving

[**Project Page**](https://ln2697.github.io/lead) | [**Paper**](https://huggingface.co/papers/2512.20563) | [**Code**](https://github.com/autonomousvision/lead)

Official model weights for **Latent TransFuser v6 (LTFv6)**, a NAVSIM checkpoint accompanies our paper LEAD: Minimizing Learner–Expert Asymmetry in End-to-End Driving. 

> We release the complete pipeline required to achieve state-of-the-art closed-loop performance on the Bench2Drive benchmark. Built around the CARLA simulator, the stack features a data-centric design with:
> 
> - Extensive visualization suite and runtime type validation for easier debugging.
> - Optimized storage format, packs 72 hours of driving in ~200GB.
> - Native support for NAVSIM and Waymo Vision-based E2E and extending those benchmarks through closed-loop simulation and synthetic data for additional supervision during training.

Find more information on [https://github.com/autonomousvision/lead](https://github.com/autonomousvision/lead).

<p align="center">
  <img src="https://ln2697.github.io/lead/static/images/tfv6.png" alt="TFv6 Architecture" width="80%" >
</p>

## Usage

Install dependencies

```bash
pip install torch timm numpy opencv-python jaxtyping beartype omegaconf huggingface_hub
```

See [example.ipynb](https://huggingface.co/ln2697/tfv6_navsim/blob/main/example.ipynb) to inspect data format and example inference.

## Data Format

We also provide example NAVSIM cache [here](https://huggingface.co/ln2697/tfv6_navsim/tree/main/data).

**Input:**
- RGB: (256, 1920, 3), range [0, 255]
- Command: [left, straight, right, unknown], e.g. [0, 1, 0, 0] for straight
- Speed: m/s
- Acceleration: m/s²

**Output:**
- `waypoints`: (N, 2) predicted positions
- `headings`: (N,) predicted angles

## Citation

If you find this work useful, please cite:

```bibtex
@article{Nguyen2025ARXIV,
  title={LEAD: Minimizing Learner-Expert Asymmetry in End-to-End Driving},
  author={Nguyen, Long and Fauth, Micha and Jaeger, Bernhard and Dauner, Daniel and Igl, Maximilian and Geiger, Andreas and Chitta, Kashyap},
  journal={arXiv preprint arXiv:2512.20563},
  year={2025}
}
```