• Follow the environment setup instructions on https://github.com/StanfordASL/Trajectron-plus-plus

• Add the current directorty to PYTHONPATH

Modify the "main" function in process_data.py such that data_path points to the nuscenes dataset, output_path points to the directory where the processed data pickle files are stored. Then run python process_data.py

Training and evaluation code are in the trajectron folder. Instructions are the same as https://github.com/StanfordASL/Trajectron-plus-plus