OccWorld

https://github.com/wzzheng/occworld

Admin

OccWorld is a 3D occupancy world model designed for autonomous driving to jointly forecast future 3D...

Tokens:891

Snippets:14

Trust Score:9.1

License:Apache-2.0

Update:2 months ago

Show doc for...

Tokens:

Raw

### Train VQVAE Model

Source: https://github.com/wzzheng/occworld/blob/main/README.md

Command to train the VQVAE component on a 24G GPU.

```bash
python train.py --py-config config/train_vqvae.py --work-dir out/vqvae
```

--------------------------------

### Visualize Results

Source: https://github.com/wzzheng/occworld/blob/main/README.md

Command to visualize model predictions. Specific scene indices can be targeted using the --scene-idx flag.

```bash
python visualize_demo.py --py-config config/train_occworld.py --work-dir out/occworld
```

--------------------------------

### Train OccWorld Model

Source: https://github.com/wzzheng/occworld/blob/main/README.md

Command to train the OccWorld model, ensuring the VQVAE checkpoint path is updated in the config.

```bash
python train.py --py-config config/train_occworld.py --work-dir out/occworld
```

--------------------------------

### Dataset Directory Structure

Source: https://github.com/wzzheng/occworld/blob/main/README.md

The expected directory structure for the OccWorld dataset.

```text
OccWorld/data
    nuscenes                 -    downloaded from www.nuscenes.org
        lidarseg
        maps
        samples
        sweeps
        v1.0-trainval
        gts                  -    download from Occ3d
    nuscenes_infos_train_temporal_v3_scene.pkl
    nuscenes_infos_val_temporal_v3_scene.pkl
```

--------------------------------

### Evaluate OccWorld Model

Source: https://github.com/wzzheng/occworld/blob/main/README.md

Command to evaluate the model, ensuring the OccWorld checkpoint path is updated in the config.

```bash
python eval_metric_stp3.py --py-config config/occworld.py --work-dir out/occworld
```

--------------------------------

### Cite OccWorld in BibTeX

Source: https://github.com/wzzheng/occworld/blob/main/README.md

Use this BibTeX entry to cite the OccWorld paper in academic publications.

```bibtex
@article{zheng2023occworld,
    title={OccWorld: Learning a 3D Occupancy World Model for Autonomous Driving},
    author={Zheng, Wenzhao and Chen, Weiliang and Huang, Yuanhui and Zhang, Borui and Duan, Yueqi and Lu, Jiwen },
    journal={arXiv preprint arXiv: 2311.16038},
    year={2023}
}
```

Tokens:

Raw