[D] ML dev in containers

So like many others out there I’m in a predicament where I have a Linux development environment that I access through SSH (pretty awesome machines) but there are relatively bare metal and come with docker, Nvidia drivers, and some python.
The catch: it’s all offline.

Instead of trying to guess and bring up eight versions etc, I’m working to pursue using containers (which only need to confirm compatibility with the nvidia driver).

I have two main questions:

1. Is there …

containers dev development docker environment linux machinelearning machines metal nvidia offline python ssh through versions