Step 1: Download Waymo Open Motion Dataset (we use the scenario protocol form dataset), and organize the data as follows:
MTR
├── data
│ ├── waymo
│ │ ├── scenario
│ │ │ ├──training
│ │ │ ├──validation
│ │ │ ├──testing
├── mtr
├── tools
Step 2: Install the Waymo Open Dataset API as follows:
pip install waymo-open-dataset-tf-2-6-0
Step 3: Preprocess the dataset:
cd mtr/datasets/waymo
python data_preprocess.py ../../../data/waymo/scenario/ ../../../data/waymo
These two paths indicate the raw data path and the output data path.
Then, the processed data will be saved to data/waymo/ directory as follows:
MTR
├── data
│ ├── waymo
│ │ ├── processed_scenarios_training
│ │ ├── processed_scenarios_validation
│ │ ├── processed_scenarios_training_infos.pkl
│ │ ├── processed_scenarios_val_infos.pkl
├── mtr
├── tools