eric-yyjau/pytorch-superpoint

about export.py

Closed this issue · 2 comments

Hi!
Thank you for your great work!!
I have a question about procedure (2).

I tried exporting detection on MS-COCO by GPU, but it is failed.
Do you know why it doesn't work?
No error, but tqdm bar doesn't progress.

スクリーンショット 2020-08-14 1 12 28
スクリーンショット 2020-08-14 1 13 24

Now, I can exporting by CPU.

Hi @coper16 ,
Thank you for raising this issue.
I'm not sure why. I tested with KITTI dataset just now and it worked ok with GPU.
Is it because of too many dataloader workers?


You can try to set it to 0.

Thank you for your quick reply.
I tried changing workers_test parameter(2 to 0) and it worked well!

Thank you very much.