How to contribute a proposal for an adapted advantage computation to RLlib

Hey guys,

I implemented a proposal for an adapted advantage compution (which could be used by PPO). My implemented example shows how this adapted advantage compution can be done via the custom callback function on_postprocess_trajectory. Maybe it is also helpful and interesting for other people, therefore how can I contribute my proposal to RLlib? Open an issue on Github?

1 Like

Ho @klausk55 ,

great initiative! If you want to contribute to rllib read first the notes on Getting Involved.

In your case I suppose the right place for your proposal would be in rllib/examples.

Hope this helps to get it on

1 Like

See PR

1 Like

@klausk55 Nice work! I wrote a comment in the PR regarding the CI tests that did not pass.

1 Like