Geometry-Aware Spatio-Temporal Context Modeling for 4D Occupancy Forecasting (ECCV 2026)
June 24, 2026 ยท View on GitHub
Environment
- Create the conda environment
conda create -n occ python=3.7
conda activate occ
- Install the following dependencies
pip install torch==1.12.1+cu116 torchvision==0.13.1+cu116 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu116
pip install torch-scatter -f https://data.pyg.org/whl/torch-1.12.0+cu116.html
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu116/torch1.12.1/index.html
pip install spconv-cu116 tensorboard numba einops pyquaternion nuscenes-devkit
Prepare Dataset
-
Download Occ3D-nuScenes from Occ3D-nuScenes.
-
Download the generated info file from Google Drive.
Model Zoo
| Method | Input | Ego traj. | mIoU (%) | IoU (%) | Weights |
|---|---|---|---|---|---|
| GAST | 3D-Occ | Pred. | 27.38 | 35.90 | Google-drive |
| GAST | 3D-Occ | GT | 47.40 | 56.24 | Google-drive |
Training and Evaluation
-
Modify the configuration file
config/gast_nuscenes.yaml. -
Enter the
tasks/occ3d_nuscenesfolder, modifyrun.shand run the command./run.sh.