I'm still new to docker. If I use a Python script to build a model, which uses GPUs and some python libraries, I then make a docker image, and submit the model. How can you run my model? I guess you need to at least have GPU and those libraries, right? Thanks, Yichao

Created by yichao li unfashionable
Dear Yichao, The python libraries will be "wrapped up" in your docker container so they are not a problem. But there are no GPUs to run your container on in our MM Challenge instances so they will not be supported. You could use the GPU on your training side for feature selection and then retrain your model without GPUs but that is up to you. Kind Regards, Mike

a docker question page is loading…