DriveLM: Driving with Graph Visual Question Answering.

We facilitate Perception, Prediction, Planning, Behavior, Motion tasks with human-written reasoning logic as a connection. We propose the task of GVQA to connect the QA pairs in a graph-style structure. To support this novel task, we provide the DriveLM-Data.

DriveLM-Data comprises two distinct components: DriveLM-nuScenes and DriveLM-CARLA. In the case of DriveLM-nuScenes, we construct our dataset based on the prevailing nuScenes dataset. As for DriveLM-CARLA, we collect data from the CARLA simulator. For now, only the training set of DriveLM-nuScenes is publicly available.

Prepare DriveLM-nuScenes Dataset

Our DriveLM-nuScenes contains a collection of questions and answers. The dataset is named v1_0_train_nus.json. We offer a subset of image data that includes all the images used in our DriveLM. You can also download the full nuScenes dataset HERE.

Usage

Download nuScenes subset image data (or full nuScenes dataset) and v1_0_train_nus.json.
Organize the data structure as follows:

DriveLM
├── data/
│   ├── QA_dataset_nus/
│   │   ├── v1_0_train_nus.json
│   ├── nuscenes/
│   │   ├── samples/

License and Citation

This language dataset is licensed under CC-BY-NC-SA 4.0. If you use this dataset, please cite our work:

@article{drivelm_paper2023,
  title={DriveLM: Driving with Graph Visual Question Answering},
  author={Sima, Chonghao and Renz, Katrin and Chitta, Kashyap and Chen, Li and Zhang, Hanxue and Xie, Chengen and Luo, Ping and Geiger, Andreas and Li, Hongyang},
  journal={arXiv preprint arXiv:2312.14150},
  year={2023}
}

@misc{drivelm_repo2023,
  title={DriveLM: Driving with Graph Visual Question Answering},
  author={DriveLM contributors},
  howpublished={\url{https://github.com/OpenDriveLab/DriveLM}},
  year={2023}
}

For more information and updates, please visit our GitHub repository.