Ours_Waymo_ADGS

AD-GS checkpoints trained on ten Waymo Open Dataset scenes for non-commercial research.

Training setup

• Dataset: Waymo Open Dataset
• Scenes: 10
• Cameras: FRONT, FRONT_LEFT, and FRONT_RIGHT
• Split: LiNSpace, 50% training frames
• Final checkpoint: iteration 60,000
• Source code: https://github.com/Dororo99/AD-GS

Each scene directory contains:

<scene_id>/
├── cfg_args
├── cameras.json
├── results.json
└── point_cloud/iteration_60000/
    ├── deform.pth
    ├── env.pth
    └── point_cloud.ply

W&B runs, TensorBoard event files, evaluation renders, ground-truth images, and local logs are intentionally excluded.

Waymo license notice

This checkpoint collection was made using the Waymo Open Dataset and is released under the Waymo Dataset License Agreement for Non-Commercial Use. Any downstream use or modification is governed by that agreement, including its non-commercial restrictions. Users must review and comply with the complete agreement before downloading or using these files.

The underlying Waymo sensor data and preprocessed dataset files are not included in this model repository.

@misc{waymo_open_dataset,
  title   = {Waymo Open Dataset: An autonomous driving dataset},
  website = {https://www.waymo.com/open},
  year    = {2019--2025}
}