proroklab / rllib_differentiable_comms

This is a minimal example to demonstrate how multi-agent reinforcement learning with differentiable communication channels and centralized critics can be realized in RLLib. This example serves as a reference implementation and starting point for making RLLib more compatible with such architectures.

39 stars 3 forks source link

Make it run on ray 2.7 #3

Open cmarschner opened 1 year ago

cmarschner commented 1 year ago

Description

Update ray to 2.7, which involves several API changes and an update from OpenAI Gym to Gymnasium (with its own API changes)
Fix a typo
Expose num_gpus (since I run on a laptop without GPU...)

cmarschner commented 1 year ago

cc @matteobettini @janblumenkamp

janblumenkamp commented 1 year ago

Looks good to me! Have you tested whether the test train script still works with the expected results?