[Question] How to test recurrent + maskable + dependent multidiscrete actions?

[maxmax1992]

❓ Question

Hey I plan to contribute to the repo, but not sure what env to choose as POC approval of the solution.
The env needs to:
test a dependent actions (each action is selected in a multidiscrete fashion, but it depends what action was chosen before. Additionally e.g some action may result into an action mask for future actions e.g choosing navigation action instead of some other high-level action). Also Ideally the policy will be recurrent (e2e). Problem is following: I had look at some repo like microRTS for python but it seems it won’t be trivial to also test recurrent policy feature as well (even though it’s good for action mask testing).

Checklist

• I have checked that there is no similar issue in the repo
• I have read the documentation
• If code there is, it is minimal and working
• If code there is, it is formatted using the markdown code blocks for both code and stack traces.

[araffin]

Hello,

Unfortunately, I don't have any specific task in mind, you might have a look at #101 (and ask there).