Tutorials on how to package various kinds of models

[bfirsh]

Currently users have to figure out the best way to package their model.

We could give people a tutorial and some best practice boilerplate for PyTorch, Tensorflow, sklearn, Hugging Face, etc models.