I am trying to troubleshoot using an external service (AWS: Amazon Web Service) for NetTrain with a GPU.

My RemoteBatchSubmit has been stuck at "Runable" for about 10 hours as described in the SE discussion here (Intuitive Explanations for AWS).

I haven't read all the AWS documentation, but can someone give a workflow for troubleshooting? Is 10 hours too long and should I wait? Is there a way to determine whether the mistake is in my RemoteBatchSubmissionEnvironment?

At this point, I am just trying random things.

There are several similar threads in SE:

SE: training neural networks on AWS

Tutorial Examples fail

Anyone had success with NetTrain on AWS

  • $\begingroup$ Be careful not to run out of wolfram credits! Using that framework requires money per hour on top of AWS costs. Also you need to have access to more than just ec2, hopefully you can work with an admin to provision the AWS services you need to your IAM... $\endgroup$
    – user5601
    Jan 19, 2022 at 17:41
  • $\begingroup$ If it's stuck in runnable for so many hours, most likely, your quota doesn't permit you to launch EC2 instances with a GPU. Check your limits and request an increase. $\endgroup$
    – stathisk
    Feb 6, 2022 at 20:56
  • $\begingroup$ Yes, that's it. I'm working with AWS and some help from WRI to get this sorted. Will report back. $\endgroup$ Feb 7, 2022 at 21:05


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.