Add example of saving

[hyanwong]

For running on a cluster, it could be helpful to show in the example notebook how to save and reload results from the fit function (i.e. a list of DemographicModel objects). E.g. can these simply be pickled, or would that take up too much space?

More generally, what's the recommended way of saving output from long runs?

[terhorst]

Yes, they can simply be pickled, and that is the intended usage. DemographicModel is just a NamedTuple (so a tuple) and occupies minimal memory (about 3 * # time points * sizeof(float)) so there should be no issue with saving a large number of them.

Will add to docs!