search

hirotomusiker / CLRerNet

The official implementation of "CLRerNet: Improving Confidence of Lane Detection with LaneIoU"
Apache License 2.0
181 stars 19 forks source link

Is GPU and cuda essential for implementation? #46

Open mansooreh1 opened 3 months ago

mansooreh1 commented 3 months ago

Hello Thanks for your great code. Do I need GPU and cuda to implement your code?

hirotomusiker commented 3 months ago

You can run our code on your cpu environment but we recommend to use gpu to train the model. Thanks!

mansooreh1 commented 3 months ago

Hi, If I want to run your code on CPU, do I need to change the code settings?

hirotomusiker commented 3 months ago

Thank you. I have checked again with my cpu-only environment and found that nms module needs GPU and cuda. Unfortunately we need to implement (lane) nms for cpu. Sorry for inconvenience.

MehmetKaTR commented 1 month ago

I would like to ask If I can run this module on an Apple Silicon Chip I Have a M2 chip MacBook Pro I couldn't run the module on it could you think you can help me with that ? thanks in advanced !

hirotomusiker commented 3 weeks ago

Thank you, unfortunately you can run this repo only in the cuda environment.