Open junhocho opened 1 year ago
acturally conda is ok to do that, surely you use docker maybe based on some reasons like landing projects or your server cuda version which is also used by your friends that can't be changed.
this project is based on centerpoint, you can set env based on that, for spconv v1, I suggest that you:
then maybe you should setup multi_scale_deformable and iou_nms_cuda you should go to det3d/ops/ and setup them
hope this can help you
Hi, thanks for sharing code. I am leaving an issue since I have trouble on running your code. I run a code without ddp
python ./tools/train.py ./configs/nusc/nuscenes_centerformer_separate_detection_head.py
,sh setup.sh
works nicely. but here is follwing error when runningtrain.py
.I have tried hard to run your code on nuscenes dataset. We also have 8gpus of A100 settting as you do. One difference would be that I use docker image. Here is dockerfile.
Through this dockerfile, we build
spconv v1.2.1
oncuda 11.1
andpytorch 1.9.1
environment. This makes exact pytorch, cuda version as your setting. Only difference is python, but I think is not a big difference. (also tried python 3.9.12, but no luck).sh setup.sh
always works nicely.seems following error
might be solved with using other spconv (according to https://github.com/traveller59/spconv/issues/58) , but I have not tried because you specified only
spconv 1.2.1
works.Would there be any idea to sort this issue?
Probably, spconv 1.2.1 does not work in docker accordint to this, but I confirmed spconv 2.2 worked in docker.
If this so, is there any chance this repo be able to support spconv 2.2? (I already tried spconv 2.2 for centerformer and failed a lot)