1. Download nuScenes
Download the nuScenes dataset to ./data/nuscenes
.
We modify data preparation in MMDetection3D
, which addtionally creates 2D annotations and temporal information for training/evaluation.
python tools/create_data_nusc.py --root-path ./data/nuscenes --out-dir ./data/nuscenes --extra-tag nuscenes2d --version v1.0
Using the above code will generate nuscenes2d_temporal_infos_{train,val}.pkl
.
We also privided the processed train, val and test pkl.
cd /path/to/StreamPETR
mkdir ckpts
Please download the pretrained weights to ./ckpts. To verify the performance on the val set, we provide the pretrained V2-99 weights. The V2-99 is pretrained on DDAD15M (weights) and further trained on nuScenes train set with FCOS3D. For the results on test set in the paper, we use the DD3D pretrained weights. The nuImage pretrained weights of R50 model can be found here.
- After preparation, you will be able to see the following directory structure:
Folder structure
StreamPETR
├── projects/
├── mmdetection3d/
├── tools/
├── configs/
├── ckpts/
├── data/
│ ├── nuscenes/
│ │ ├── maps/
│ │ ├── samples/
│ │ ├── sweeps/
│ │ ├── v1.0-test/
| | ├── v1.0-trainval/
| | ├── nuscenes2d_temporal_infos_train.pkl
| | ├── nuscenes2d_temporal_infos_val.pkl