Invalid GPU Request
193 views
Skip to first unread message
William Rudman
Mar 31, 2022, 10:42:11 AM3/31/22
to physionet-challenges
Hi,
I was wondering if anyone else has encountered the error "invalid GPU request" when submitting code? Would you all be able to provide an example of a successful GPU request? I can only test my code on Docker with a CPU.
Thank you in advance!
Best,
William
PhysioNet Challenge
Mar 31, 2022, 10:43:42 AM3/31/22
to physionet-challenges
Dear William,
Unfortunately, GPUs are one of the more difficult things to support with containerization.
One of the easiest solutions is to use a base image (the first line of your Dockerfile) that provides "out-of-the-box" GPU support. Here are several base images that have worked recently with a GPU:
nvidia/cuda:10.0-cudnn7-devel
pytorch/pytorch:1.7.1-cuda11.0-cudnn8-devel
pytorch/pytorch:1.8.0-cuda11.1-cudnn8-runtime
tensorflow/tensorflow:2.4.1-gpu-jupyter
Others may have other suggestions as well, but this should be a good place to start.
We'll also provide more detailed error messages to make any errors easier to debug.
Best,
Matt
(On behalf of the Challenge team.)
https://PhysioNetChallenges.org/
https://PhysioNet.org/
Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email challenge at physionet.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.
Unfortunately, GPUs are one of the more difficult things to support with containerization.
One of the easiest solutions is to use a base image (the first line of your Dockerfile) that provides "out-of-the-box" GPU support. Here are several base images that have worked recently with a GPU:
nvidia/cuda:10.0-cudnn7-devel
pytorch/pytorch:1.7.1-cuda11.0-cudnn8-devel
pytorch/pytorch:1.8.0-cuda11.1-cudnn8-runtime
tensorflow/tensorflow:2.4.1-gpu-jupyter
Others may have other suggestions as well, but this should be a good place to start.
We'll also provide more detailed error messages to make any errors easier to debug.
Best,
Matt
(On behalf of the Challenge team.)
https://PhysioNetChallenges.org/
https://PhysioNet.org/
Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email challenge at physionet.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.
Reply all
Reply to author
Forward
0 new messages