Downlowd V2X-Seq-SPD dataset, organize the files as official instructions and then follow the steps below to prepare the dataset.
a. Create V2X-Seq-SPD-Example (as an example, you can give it a name that you like) according to given sequence ids.
python ./tools/spd_data_converter/gen_example_data.py --input YOUR_V2X-Seq-SPD_ROOT --output ./datasets/V2X-Seq-SPD-Example --sequences 0000 0001 0002 --update-label
Note that if you want to use all sequences, you can directly use --sequences 'all'. We report results on all sequences.
b. Generate data info for training and nuscenes style dataset for evaluation.
You should modify ./tools/spd_example_converter.sh according to the path used in the previous step, and then execute the following commands in the terminal.
bash ./tools/spd_example_converter.sh vehicle-side
bash ./tools/spd_example_converter.sh infrastructure-side
bash ./tools/spd_example_converter.sh cooperative
Download checkpoints of R101DCN and BEVFormer from this project, and create a new folder named ckpt and place the downloaded models in it.
As a result, you will see the overall structure as follows:
V2XTrack
├── projects/
├── tools/
├── ckpts/
│ ├── r101_dcn_fcos3d_pretrain.pth
│ ├── bevformer_r101_dcn_24ep.pth
├── data/
│ ├── infos/
| | ├── V2X-Seq-SPD-Example
| | | ├── cooperative
| | │ │ ├── spd_infos_temporal_train.pkl
| | │ │ ├── spd_infos_temporal_val.pkl
| | | ├── vehicle-side
| | | ├── infrastructure-side
│ ├── split_datas/
│ │ ├── cooperative-split-data-spd.json
├── datasets/
│ ├── V2X-Seq-SPD-Example/
│ │ ├── cooperative/
│ │ │ ├── calib /
│ │ │ ├── image /
│ │ │ ├── label /
│ │ │ ├── velodyne /
│ │ │ ├── v1.0-trainval /
│ │ │ ├── data_info.json
│ │ ├── vehicle-side/
│ │ ├── infrastructure-side/
<- Last Page: Installation
-> Next Page: Train/Eval UniAD