Saving state for later

I apologize in advance for my ignorance.

I am totally enjoying my learning experience with Jeremy and don’t really have any issues understanding the material. I do however appear to have a conceptual problem with something that is probably very simple.

In NLP, once I have the classifier (learn.predict) giving excellent results, what exactly do I need to save (and how) in order to reload and do predictions at a later date without having to repeat the whole training process? Is this done through some conventional Pytorch utilities.

I have seen a number of discussion threads about deployment but they don’t seem to address this simple scenario.