chanb / rl_sandbox_public Goto Github PK

Hi Bryan,
First of all congrats for the repository. Really nice and neat job! 👏
To some extend, I am interested in implementing SOTA RL papers as well and I have gathered all I have managed to implement here.

But unlike you, I have been struggling to put them altogether in a modular manner (the same you have achieved in this repo) rather separate repositories and isolated implementations.
So, I thought to ask you if you could provide me with some tips and tricks and guidelines you followed to build up your repo?
For instance, the way you constructed an inheritance relationship between base_policy, off_policy and on_policy algorithms and why.

Thank you so much. 😄