Add AGO nuScenes checkpoints
Browse files- README.md +50 -0
- SHA256SUMS +3 -0
- ago_nuscenes_open_world_finetuning.pth +3 -0
- ago_nuscenes_open_world_pretraining.pth +3 -0
- ago_nuscenes_self_supervised.pth +3 -0
README.md
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
license_name: nuscenes-terms-of-use
|
| 4 |
+
license_link: https://www.nuscenes.org/terms-of-use
|
| 5 |
+
tags:
|
| 6 |
+
- occupancy-prediction
|
| 7 |
+
- open-vocabulary
|
| 8 |
+
- autonomous-driving
|
| 9 |
+
- nuscenes
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# AGO: Adaptive Grounding for Open World 3D Occupancy Prediction
|
| 13 |
+
|
| 14 |
+
Pretrained weights for [AGO](https://github.com/EdwardLeeLPZ/AGO) (ICCV 2025) on Occ3D-nuScenes. Paper: [arXiv:2504.10117](https://arxiv.org/abs/2504.10117).
|
| 15 |
+
|
| 16 |
+
## Weights
|
| 17 |
+
|
| 18 |
+
Each file contains the model `state_dict`. The CLIP weights are not included and are loaded from FeatUp at runtime.
|
| 19 |
+
|
| 20 |
+
| File | Config | Known mIoU / Unknown mIoU / mIoU |
|
| 21 |
+
| :--- | :--- | :-: |
|
| 22 |
+
| `ago_nuscenes_self_supervised.pth` | `ago_nuscenes_self_supervised.py` | - / - / 19.16 |
|
| 23 |
+
| `ago_nuscenes_open_world_pretraining.pth` | `ago_nuscenes_open_world_pretraining.py` | 23.23 / 0.00 / 11.61 |
|
| 24 |
+
| `ago_nuscenes_open_world_pretraining.pth` (zero-shot) | `ago_nuscenes_open_world_pretraining.py` | 33.95 / 0.42 / 7.13 |
|
| 25 |
+
| `ago_nuscenes_open_world_finetuning.pth` | `ago_nuscenes_open_world_finetuning.py` | 36.88 / 8.78 / 14.40 |
|
| 26 |
+
|
| 27 |
+
## Usage
|
| 28 |
+
|
| 29 |
+
```shell
|
| 30 |
+
hf download EdwardLeeLPZ/AGO --include "*.pth" --local-dir checkpoints
|
| 31 |
+
bash tools/dist_test.sh projects/configs/ago_nuscenes_self_supervised.py checkpoints/ago_nuscenes_self_supervised.pth 8 --tag self_supervised
|
| 32 |
+
```
|
| 33 |
+
|
| 34 |
+
See the [code repository](https://github.com/EdwardLeeLPZ/AGO) for installation and data preparation.
|
| 35 |
+
|
| 36 |
+
## License
|
| 37 |
+
|
| 38 |
+
The weights are trained on nuScenes and subject to the [nuScenes terms of use](https://www.nuscenes.org/terms-of-use).
|
| 39 |
+
|
| 40 |
+
## Citation
|
| 41 |
+
|
| 42 |
+
```
|
| 43 |
+
@inproceedings{li2025ago,
|
| 44 |
+
title={AGO: Adaptive Grounding for Open World 3D Occupancy Prediction},
|
| 45 |
+
author={Li, Peizheng and Ding, Shuxiao and Zhou, You and Zhang, Qingwen and Inak, Onat and Triess, Larissa and Hanselmann, Niklas and Cordts, Marius and Zell, Andreas},
|
| 46 |
+
booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
|
| 47 |
+
pages={8645--8655},
|
| 48 |
+
year={2025}
|
| 49 |
+
}
|
| 50 |
+
```
|
SHA256SUMS
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
110173bd60be38c2c192842d7faa5997cea8e51b78223e1bd55f02a24e1d9675 ago_nuscenes_open_world_finetuning.pth
|
| 2 |
+
493906ad5690ce87792a4ecf26d492a384bc321e7b784a7633336ebb4e4b389c ago_nuscenes_open_world_pretraining.pth
|
| 3 |
+
968b8f6e8eeae8716a31f08b7ebb7a06c128fb41d1cf774175da11b599f4c1a0 ago_nuscenes_self_supervised.pth
|
ago_nuscenes_open_world_finetuning.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:110173bd60be38c2c192842d7faa5997cea8e51b78223e1bd55f02a24e1d9675
|
| 3 |
+
size 272927269
|
ago_nuscenes_open_world_pretraining.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:493906ad5690ce87792a4ecf26d492a384bc321e7b784a7633336ebb4e4b389c
|
| 3 |
+
size 272928200
|
ago_nuscenes_self_supervised.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:968b8f6e8eeae8716a31f08b7ebb7a06c128fb41d1cf774175da11b599f4c1a0
|
| 3 |
+
size 272921683
|