{"id":13442300,"url":"https://github.com/hustvl/VAD","last_synced_at":"2025-03-20T13:33:22.609Z","repository":{"id":153827671,"uuid":"602453355","full_name":"hustvl/VAD","owner":"hustvl","description":"[ICCV 2023] VAD: Vectorized Scene Representation for Efficient Autonomous Driving","archived":false,"fork":false,"pushed_at":"2025-03-03T03:06:03.000Z","size":4663,"stargazers_count":884,"open_issues_count":67,"forks_count":99,"subscribers_count":30,"default_branch":"main","last_synced_at":"2025-03-16T22:07:27.309Z","etag":null,"topics":["autonomous-driving","end-to-end"],"latest_commit_sha":null,"homepage":"https://arxiv.org/abs/2303.12077","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/hustvl.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2023-02-16T08:41:37.000Z","updated_at":"2025-03-16T09:53:56.000Z","dependencies_parsed_at":"2024-01-16T02:46:36.171Z","dependency_job_id":"ac19bba4-c49e-4bbf-9a8f-65b3e67873ae","html_url":"https://github.com/hustvl/VAD","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hustvl%2FVAD","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hustvl%2FVAD/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hustvl%2FVAD/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hustvl%2FVAD/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/hustvl","download_url":"https://codeload.github.com/hustvl/VAD/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":244619276,"owners_count":20482390,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["autonomous-driving","end-to-end"],"created_at":"2024-07-31T03:01:44.073Z","updated_at":"2025-03-20T13:33:22.603Z","avatar_url":"https://github.com/hustvl.png","language":"Python","funding_links":[],"categories":["Python","💻 Open-Source Projects"],"sub_categories":["Papers"],"readme":"## VAD v1 \u0026 v2\n\n[project page](https://hgao-cv.github.io/VADv2/)\n\nhttps://user-images.githubusercontent.com/45144254/229673708-648e8da5-4c70-4346-9da2-423447d1ecde.mp4\n\nhttps://github.com/hustvl/VAD/assets/45144254/153b9bf0-5159-46b5-9fab-573baf5c6159\n\n\n\u003e [**VAD: Vectorized Scene Representation for Efficient Autonomous Driving**](https://arxiv.org/abs/2303.12077)\n\u003e\n\u003e [Bo Jiang](https://github.com/rb93dett)\u003csup\u003e1\u003c/sup\u003e\\*, [Shaoyu Chen](https://scholar.google.com/citations?user=PIeNN2gAAAAJ\u0026hl=en\u0026oi=sra)\u003csup\u003e1\u003c/sup\u003e\\*, Qing Xu\u003csup\u003e2\u003c/sup\u003e, [Bencheng Liao](https://github.com/LegendBC)\u003csup\u003e1\u003c/sup\u003e, Jiajie Chen\u003csup\u003e2\u003c/sup\u003e, [Helong Zhou](https://scholar.google.com/citations?user=wkhOMMwAAAAJ\u0026hl=en\u0026oi=ao)\u003csup\u003e2\u003c/sup\u003e, [Qian Zhang](https://scholar.google.com/citations?user=pCY-bikAAAAJ\u0026hl=zh-CN)\u003csup\u003e2\u003c/sup\u003e, [Wenyu Liu](http://eic.hust.edu.cn/professor/liuwenyu/)\u003csup\u003e1\u003c/sup\u003e, [Chang Huang](https://scholar.google.com/citations?user=IyyEKyIAAAAJ\u0026hl=zh-CN)\u003csup\u003e2\u003c/sup\u003e, [Xinggang Wang](https://xwcv.github.io/)\u003csup\u003e1,\u0026#8224;\u003c/sup\u003e\n\u003e \n\u003e \u003csup\u003e1\u003c/sup\u003e Huazhong University of Science and Technology, \u003csup\u003e2\u003c/sup\u003e Horizon Robotics\n\u003e\n\u003e \\*: equal contribution, \u003csup\u003e\u0026#8224;\u003c/sup\u003e: corresponding author.\n\u003e\n\u003e[arXiv Paper](https://arxiv.org/abs/2303.12077), ICCV 2023\n\n## News\n* **`27 Feb, 2025`:** Check out our latest work, [DiffusionDrive](https://github.com/hustvl/DiffusionDrive), accepted to CVPR 2025! This study explores multi-modal end-to-end driving using diffusion models for real-time and real-world applications.\n* **`19 Feb, 2025`:** Checkout our new work [RAD](https://hgao-cv.github.io/RAD) 🥰, end-to-end autonomous driving with large-scale 3DGS-based Reinforcement Learning post-training.\n* **`30 Oct, 2024`:** Checkout our new work [Senna](https://github.com/hustvl/Senna) 🥰, which combines VAD/VADv2 with large vision-language models to achieve more accurate, robust, and generalizable autonomous driving planning.\n* **`20 Sep, 2024`:** Core code of VADv2 (config and model) is available in the `VADv2` folder. Easy to integrade it into the VADv1 framework for training and inference.\n* **`17 June, 2024`:** CARLA implementation of VADv1 is available on [Bench2Drive](https://github.com/Thinklab-SJTU/Bench2Drive?tab=readme-ov-file).\n* **`20 Feb, 2024`:** VADv2 is available on arXiv    [paper](https://arxiv.org/pdf/2402.13243)    [project page](https://hgao-cv.github.io/VADv2/).\n* **`1 Aug, 2023`:** Code \u0026 models are released!\n* **`14 July, 2023`:** VAD is accepted by ICCV 2023🎉! Code and models will be open source soon!\n* **`21 Mar, 2023`:** We release the VAD paper on [arXiv](https://arxiv.org/abs/2303.12077). Code/Models are coming soon. Please stay tuned! ☕️\n\n## Introduction\n\u003e VAD is a vectorized paradigm for end-to-end autonomous driving.\n\n\u003cdiv align=\"center\"\u003e\n\u003cimg src=\"./assets/arch.png\" /\u003e\n\u003c/div\u003e\n\n- We propose VAD, an end-to-end unified vectorized paradigm for autonomous driving. VAD models the driving scene as a fully vectorized representation, getting rid of computationally intensive dense rasterized representation and hand-designed post-processing steps.\n- VAD implicitly and explicitly utilizes the vectorized scene information to improve planning safety, via query interaction and vectorized planning constraints.\n- VAD achieves SOTA end-to-end planning performance, outperforming previous methods by a large margin. Not only that, because of the vectorized scene representation and our concise model design, VAD greatly improves the inference speed, which is critical for the real-world deployment of an autonomous driving system.\n\n## Models\n\n| Method | Backbone | avg. L2 | avg. Col. | FPS | Config | Download |\n| :---: | :---: | :---: | :---: |  :---: | :---: | :---: |\n| VAD-Tiny | R50 | 0.78 | 0.38 | 16.8 | [config](projects/configs/VAD/VAD_tiny_stage_2.py) | [model](https://drive.google.com/file/d/1KgCC_wFqPH0CQqdr6Pp2smBX5ARPaqne/view?usp=sharing) |\n| VAD-Base | R50 | 0.72 | 0.22 | 4.5 | [config](projects/configs/VAD/VAD_base_stage_2.py) | [model](https://drive.google.com/file/d/1FLX-4LVm4z-RskghFbxGuYlcYOQmV5bS/view?usp=sharing) |\n\n## Results\n- Open-loop planning results on [nuScenes](https://github.com/nutonomy/nuscenes-devkit). See the [paper](https://arxiv.org/abs/2303.12077) for more details.\n\n| Method | L2 (m) 1s | L2 (m) 2s | L2 (m) 3s | Col. (%) 1s | Col. (%) 2s | Col. (%) 3s | FPS |\n| :---: | :---: | :---: | :---: | :---:| :---: | :---: | :---: |\n| ST-P3 | 1.33 | 2.11 | 2.90 | 0.23 | 0.62 | 1.27 | 1.6 |\n| UniAD | 0.48 | 0.96 | 1.65 | **0.05** | 0.17 | 0.71 | 1.8 |\n| VAD-Tiny | 0.46 | 0.76 | 1.12 | 0.21 | 0.35 | 0.58 | **16.8** |\n| VAD-Base | **0.41** | **0.70** | **1.05** | **0.07** | **0.17** | **0.41** | 4.5 |\n\n- Closed-loop simulation results on [CARLA](https://github.com/carla-simulator/carla).\n\n| Method | Town05 Short DS | Town05 Short RC | Town05 Long DS | Town05 Long RC |\n| :---: | :---: | :---: | :---: | :---:|\n| CILRS | 7.47 | 13.40 | 3.68 | 7.19 |\n| LBC | 30.97 | 55.01 | 7.05 | 32.09 |\n| Transfuser\\* | 54.52 | 78.41 | 33.15 | 56.36 |\n| ST-P3 | 55.14 | 86.74 | 11.45 | 83.15 |\n| VAD-Base | **64.29** | **87.26** | **30.31** | 75.20 |\n\n\u003e \\*: LiDAR-based method.\n\n## Getting Started\n- [Installation](docs/install.md)\n- [Prepare Dataset](docs/prepare_dataset.md)\n- [Train and Eval](docs/train_eval.md)\n- [Visualization](docs/visualization.md)\n\n## Catalog\n- [x] Code \u0026 Checkpoints Release\n- [x] Initialization\n\n## Contact\nIf you have any questions or suggestions about this repo, please feel free to contact us (bjiang@hust.edu.cn, outsidercsy@gmail.com).\n\n## Citation\nIf you find VAD useful in your research or applications, please consider giving us a star \u0026#127775; and citing it by the following BibTeX entry.\n\n```BibTeX\n@article{jiang2023vad,\n  title={VAD: Vectorized Scene Representation for Efficient Autonomous Driving},\n  author={Jiang, Bo and Chen, Shaoyu and Xu, Qing and Liao, Bencheng and Chen, Jiajie and Zhou, Helong and Zhang, Qian and Liu, Wenyu and Huang, Chang and Wang, Xinggang},\n  journal={ICCV},\n  year={2023}\n}\n\n@article{chen2024vadv2,\n  title={Vadv2: End-to-end vectorized autonomous driving via probabilistic planning},\n  author={Chen, Shaoyu and Jiang, Bo and Gao, Hao and Liao, Bencheng and Xu, Qing and Zhang, Qian and Huang, Chang and Liu, Wenyu and Wang, Xinggang},\n  journal={arXiv preprint arXiv:2402.13243},\n  year={2024}\n}\n```\n\n## License\nAll code in this repository is under the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0).\n\n## Acknowledgement\nVAD is based on the following projects: [mmdet3d](https://github.com/open-mmlab/mmdetection3d), [detr3d](https://github.com/WangYueFt/detr3d), [BEVFormer](https://github.com/fundamentalvision/BEVFormer) and [MapTR](https://github.com/hustvl/MapTR). Many thanks for their excellent contributions to the community.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhustvl%2FVAD","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fhustvl%2FVAD","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhustvl%2FVAD/lists"}