CUDA error when training on TuSimple

lucastabelini / LaneATT

Code for the paper entitled "Keep your Eyes on the Lane: Real-time Attention-guided Lane Detection" (CVPR 2021)

https://openaccess.thecvf.com/content/CVPR2021/html/Tabelini_Keep_Your_Eyes_on_the_Lane_Real-Time_Attention-Guided_Lane_Detection_CVPR_2021_paper.html

MIT License

634 stars 167 forks source link

CUDA error when training on TuSimple #111

Closed anir9 closed 2 years ago

anir9 commented 2 years ago

Hello, I am trying to train the model on the TuSimple dataset with the ResNet34 backbone. I keep encountering an unknown CUDA error. I am not sure what the issue is as the program is recognizing that CUDA is available and I cannot seem to find a solution to my problem. I've tried it on the other backbones and have encountered the same issue. I am using Pytorch 1.6.0 and CUDA 10.2.

ResNet34 issue

Here is the issue with ResNet18

I'd really appreciate the help!

lucastabelini commented 2 years ago

Have you tried to isolate the issue by running it inside a docker container? I provided a simple one here that you could try.