Using phonological feature input

[ecooper7]

Hello, thank you for this very nice and easy to use codebase! I was able to get it up and running very easily, and train a small model and run synthesis from it using the Wizard.

Now, I would like to try using phonological feature input. This appears to be implemented, and I can see the options for that in the preprocessing and training code, but I haven't been able to figure out how to activate it....is there some option I can give on the command line, or something I can drop in to the config files?