Have you finetured CLRNet and CondlaneNet?

[ppbangKGT]

A nice work!
When I read your paper, I found the performances of CLRNet and CondlaneNet are quite bad on both datasets. So have you finetuned them on these datasets? Or you used the pretrained model to evaluate to see the generalization. If so, which model did you choose? CULane or Tusimple?
Thanks for your time.