Hi James, I'm currently buidling something very similar to what you're talking about.
I've currently got it set up so that I can access multiple Dockers, containing isolated machine learning models, through a Jupyter notebook (located in a third Docker), via SSH. It wasn't super difficult to do, although I'm not claiming it was done very elegantly. If you're interested, you can take a look here https://github.com/stojan211287/DockerSSH. I've uploaded a minimal example, consisting of one "drone" and one "overlord" container. The overlord issues commands via SSH, the drone complies and delivers. As it stands now, I've based the images on Alpine 3.6 and am currently using them as base images for further development.- the overlord get Jupyter installed on top of it, and the drone, for example, can host scikit-learn. On Wednesday, 24 May 2017 23:10:34 UTC+2, James wrote: > > Hey, sorry to revive this thread again, but having docker container > kernels (and not whole jupyter server systems) would be very useful for me. > My use case is having certain hard to build scientific software installed > within the container. That way you could call out to them using python's > subprocess calls from within the notebook. My goal would be to make > several kernels, accessible from the same notebook server, to act as a > toolkit of sorts for my lab. Ideally having the kernels in containers > would make them easy to share and install in sister labs at other > institutions for use in their Jupyter ecosystem. Thank you for any > guidance! > -- You received this message because you are subscribed to the Google Groups "Project Jupyter" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/jupyter/b22a6d05-25c6-4e24-ae2b-df9ea723ed6c%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
