Possible Hierarchical RL PR

Open peasant98 opened this issue 4 years ago • 1 comments

Hello, I am a RL researcher, and my team and I have recently implemented HIRO (Data Efficient Hierarchical Reinforcement Learning with Off-Policy Correction) with PFRL. I'm wondering if a PR of an HRL algorithm (which required some large changes) would be encouraged on this platform.

Thanks!

Mar 24 '21 18:03 peasant98

Hi, the developer team thinks it is possible to merge such a new algorithm PR, and we would really appreciate such a contribution! To see how easy a specific PR could be merged beforehand, can you let us know what your PR would look like, especially in the following aspects?

what kind of changes the PR would make e.g.
- how "large" is it?
- could it affect other algorithms?
- could it break backward compatibility of some API?
how the implementation has been verified e.g.
- is there any significant performance gap between the official HIRO implementation and yours?

Mar 25 '21 04:03 muupan