Genre and Speaker Label Supoort for Coref task

[acalejos]

I see that the pairwise encoder used in the coref model support annotating the genre and Speaker labels for the input text. I didn't see anywhere in the documentation that mentioned this capability, nor any way to input them with the top level API. For now, I've added a processor to put this info into the doc objects, but I was wondering if there's any suggested or standard way to input these features.

Thanks!