CornerNet-Lite icon indicating copy to clipboard operation
CornerNet-Lite copied to clipboard

Published 20 hours ago • verified publisher icon princeton-vl

Our Nvidia Driver is CUDA9.0

Open momo1986 opened this issue 5 years ago • 5 comments

Hello, Thanks a lot for the excellent job from Princeton-VL, our system is based on the Nvidia-GPU driver of CUDA-9.0. Can I just directly try the open source code on my system with the configuration file change. Or should I upgrade my driver? Thanks a lot!

momo1986 avatar Apr 29 '19 09:04 momo1986

I only tested it with CUDA10 so I am not sure if it would run on CUDA9. I would recommend upgrading your driver if possible.

heilaw avatar Apr 29 '19 15:04 heilaw

Hello. Did you successfully reproduce it？I also used cuda 9.0 ，but it made a mistake.

yinyuhangi avatar May 08 '19 01:05 yinyuhangi

@yinyuhangi @momo1986 I have run the project on CUDA9 correctly with the following steps:

  1. conda create --name CornerNet_Lite --file conda_packagelist.txt --channel pytorch
  2. source activate CornerNet_Lite
  3. conda uninstall pytorch cuda100
  4. conda install pytorch=1.0.0 torchvision -c pytorch perfom the 4th step will install the following modules. pytorch pytorch/linux-64::pytorch-1.0.0-py3.7_cuda9.0.176_cudnn7.4.1_1 torchvision pytorch/noarch::torchvision-0.2.2-py_3 and you can run python demo.py to check ! hope to help you!

brooks0519 avatar May 09 '19 06:05 brooks0519

@chiminghui Thank you very much. I have run the project on cuda 10.1 correctly.

yinyuhangi avatar May 09 '19 07:05 yinyuhangi

Thanks a lot!!!

SeeeeShiwei avatar May 13 '19 10:05 SeeeeShiwei