instadeepai / Mava

🦁 A research-friendly codebase for fast experimentation of multi-agent reinforcement learning in JAX
Apache License 2.0
709 stars 83 forks source link

feat(lbf): generalise wrapper and logger #952

Closed WiemKhlifi closed 10 months ago

WiemKhlifi commented 10 months ago

NB: This PR should be reviewed after this one #948 :crossed_fingers:

What?

Based on the changes made in this first PR #948 for the RwareWrapper, we added a separate LbfWrapper as well as some changes to support the LBF environment. -> Merged them into one single wrapper 🤞

Why?

LbfWrapper is useful in this case to choose what reward will be used during training whether it is the shared reward or not.

How?