Pytorch implementation for the paper: TrafficPredict: Trajectory Prediction for Heterogeneous Traffic-Agents (AAAI), Oral, 2019
The repo has been forked initially from Anirudh Vemula's repository for his paper Social Attention: Modeling Attention in Human Crowds (ICRA 2018). If you find this code useful in your research then please also cite Anirudh Vemula's paper.
| Methods | Paper ADE | This repo ADE | Paper FDE | This repo FDE |
|---|---|---|---|---|
| pedestrian | 0.091 | 0.088 | 0.150 | 0.132 |
| bicycle | 0.083 | 0.075 | 0.139 | 0.115 |
| vehicle | 0.080 | 0.090 | 0.131 | 0.153 |
| total | 0.085 | 0.084 | 0.141 | 0.133 |