Lyft Motion prediction with Sampling and Fitting Framework

Kaggle Competition: Lyft Motion Prediction for Autonomous Vehicles Build motion prediction models for self-driving vehicles

Reproduce this paper for Pytorch version: Overcoming Limitations of Mixture Density Networks: A Sampling and Fitting Framework for Multimodal Future Prediction

Requirements

• Pytorch 1.6.0
• tqdm, typing, prettytable
• l5kit

Data

To check our result, Please download data from [Link]

Refernece

[1] Overcoming Limitations of Mixture Density Networks: A Sampling and Fitting Framework for Multimodal Future Prediction in CVPR 2019