Create a Singulary container build of CLMBR

[jason-fries]

Due to the specific system dependencies of CLMBR, can we build a singular container for training models? This will make it much easier to deploy internally as well as share with collaborators.

[EthanSteinberg]

The solution here is to create a compiled pypi package. I'll leave this issue open though to track that though.

[EthanSteinberg]

This is now fixed as the install process has been cleaned up.

[jason-fries]

This is an orthogonal issue to simplifying the installation process. A user might not be able, for security or admin reasons, to install the required packages needed to build clmbr. A containerized distribution is useful in these settings and many others. Let’s leave this open as a todo.