Closed turboxin closed 1 month ago
As described in paper, the L2 distance evaluation is aligned with VAD, and we re-implemented the evaluation protocol for collision rate.
The code shows that, the original VAD uses L2 metric averaged overtime, while sparsedrive use the current timestamps L2 metric:
VAD:
Sparsedrive:
Please refer to this issue.
It seems that UniAD and VAD used different eval metric protocal as noted by UAD and PARA-Drive paper
https://openaccess.thecvf.com/content/CVPR2024/papers/Weng_PARA-Drive_Parallelized_Architecture_for_Real-time_Autonomous_Driving_CVPR_2024_paper.pdf
https://arxiv.org/html/2406.17680v1
I'm wondering what did you used for comparision?