Chonghao Sima
Chonghao Sima
Thank you for your suggestion, and this is on our TODO list.
It depends. I tend to keep them as in the final detection, those invisible points would be predicted as well.
0. the transformation is largely following [GenLaneNet](https://github.com/yuliangguo/Pytorch_Generalized_3D_Lane_Detection/blob/master/tools/utils.py). You can have a try on R_gv@E method and see if there's any improvement. 1. sorry, that was a typo in OpenLane repo....
Thank you for your code! I struggled with the nan loss for a long time……
Please see [here](https://github.com/OpenDriveLab/DriveLM/tree/main/challenge#results)
looking into it and will get back to you soon.
建议是把微信卸载😈
@CompliceU hello there! I wonder if you could share the training/testing code of CLGo on OpenLane, as we're trying to evaluate CLGo on OpenLane as well. Thanks in advance!
@liuruijin17 Greetings from OpenLane. Do you have a plan to release multi-gpu training code and report a result on OpenLane Dataset?
My experience is that the file is not read correctly and some variables become nonetype. Please check the file loading path and related stuff.