I don't think it is possible - but even if it is, the memory gained by erasing a loaded module is usually minimal.
Running del module in the code of some file will not remove it from memory: all loaded modules are available from the global sys.modules dictionary. If you delete the module from sys.modules, you can gain some space, but I think Python will still hold references to at least the classes that were defined in the module, (even if you have no instances of, or other references to those classes in your code.)
all in all, most modules would get you a few kilobytes back. If you plan to "unload" modules that have a multi-megabyte impact in memory (think of Pandas/matplotlib/sqlalchemy), you will be even less likely to be successful as the complexity of these code bases mean they have a lot of internal cross references, and you'd need to delete everything they import internally from sys.modules.
If the setup is based on a multi-megabyte framework like these, and you are running on a very limited system (raspberry-class mini PC, with a few MB, for example) you could try to perform the initialisation in a subprocess, and kill that process when you are done - that would give you some guarantees the memory is freed up. But limited hardware apart you should simply not bother about it.