Skip to content

Latest commit

 

History

History
12 lines (11 loc) · 634 Bytes

CONTRIBUTING.md

File metadata and controls

12 lines (11 loc) · 634 Bytes

Can I open an issue with a PR?

I'd prefer not since there isn't much more to add, other than quality-of-life features. Feel free to fork and make it your own: I'm using a MIT license!

What needs to be done?

In order of decreasing priority:

  • The ability to just pass the whole training sequence into the RNN, i.e. handle the passing of the most recent hidden states within the class.
  • Maybe the ability to handle some arbitrary, parameterized distribution for the likelihood? Not sure how to do this nicely without becoming less PyTorch-like. Would be complicated depending on whether or not you're using the expected value.