How to use containers in gradient job runner?


#1

I want to run a recurrent neural network on gradient job runner. But reading documentation, the command that is used has --container Test-Container. Also documentation has base container list, but I have no idea how to use those container. Tell me on how to specify containers in command line.

I am running the project using PyTorch.


#2

Hi there,

You can learn more about how to use different containers with Gradient here: https://paperspace.zendesk.com/hc/en-us/articles/360003415434-Private-containers-and-workspaces

So if you are using PyTorch you would likely want to use a container like ufoym/deepo:pytorch-py27 which you can learn more about here: https://hub.docker.com/r/ufoym/deepo/

You could also build your own container which is a bit more advanced.

Does that help?


#3

yes, thank you!.
I ran the job, but it is not giving any output. I mean no logs are coming, even though I am writing to stdout. I am using floydhub pytorch container.

I have put this in the other question. Can you please help me on that?