Fetching the paper…

SEPT: Towards Efficient Scene Representation Learning for Motion Prediction · Around