To train your own model, you need to download the Lyft Motion Prediction Dataset. These files contain Training Dataset(8.4GB), validation Dataset (8.2GB), Aerial Map and Semantic Map. You also need to download Lyft's Python software kit. The detailed tutorial can be found at https://woven-planet.github.io/l5kit/introduction.html. All data files should be stored in a single folder to match this structure: https://woven-planet.github.io/l5kit/dataset.html.
Run train.py to use single GPU to train the model.
python train.py Run multi_train.py to use multi GPU to train the model.
python -m torch.distributed.launch --nproc_per_node=2 multi_train.py