instadeepai / Mava

🦁 A research-friendly codebase for fast experimentation of multi-agent reinforcement learning in JAX
Apache License 2.0
737 stars 90 forks source link

Feat:Sebulba [3] ff-ippo #978

Closed OmaymaMahjoub closed 4 months ago

OmaymaMahjoub commented 10 months ago

What?

Implement Sebulba architecture with feedforward IPPO on Rware.

Why?

Integrate Sebulba's architecture due to its effectiveness in scenarios involving non-jitted/non-jax environments.

How?

Enhance the existing Cleanba code to support marl algorithm and ensure compatibility with Mava's key components, including logger, evaluator, etc.

Extra

Before reviewing this PR, we need to merge #976 and #977 The early draft of sebulba ff-ippo can be found in #959 Action Item:

OmaymaMahjoub commented 4 months ago

@Louay-Ben-nessir is working on maintaining this PR in a new one that will be added soon :)