Hello,there is a logic mistake in your qcnet-decoder, can you correct it and give a new result on argoverse2 val and test?

ZikangZhou / QCNet

[CVPR 2023] Query-Centric Trajectory Prediction

https://openaccess.thecvf.com/content/CVPR2023/papers/Zhou_Query-Centric_Trajectory_Prediction_CVPR_2023_paper.pdf

Apache License 2.0

432 stars 70 forks source link

Hello,there is a logic mistake in your qcnet-decoder, can you correct it and give a new result on argoverse2 val and test? #18

Open USCT-YQJ opened 11 months ago

USCT-YQJ commented 11 months ago

ZikangZhou commented 11 months ago

Thanks for spotting this! This bug was introduced carelessly when I was preparing the open-source code. It causes a noisy feature to the decoder, which may slightly affect the performance. I'll look into this later.