search

tanliyon / gym-xiangqi

This repo sets up the environment to play Xiang Qi (chinese chess) following the OpenAI Gym framework.
GNU Lesser General Public License v3.0
32 stars 6 forks source link

[Feature Request] PettingZoo Support #94

Open jkterry1 opened 3 years ago

jkterry1 commented 3 years ago

Hey, have you considered using the PettingZoo API (https://github.com/PettingZoo-Team/PettingZoo)? It's like a multi-agent version of Gym and most of the major MARL libraries support it.

hojoungjang commented 3 years ago

Definitely sounds interesting. Will you be able to share just a short description or examples on how this project can benefit from using PettingZoo?

jkterry1 commented 3 years ago

Right now, you're using the Gym API. This only is for one player, which is inconvenient to work with for multi-player games like yours, and does not allow for using any of the multi-agent RL libraries. If you switch your API to PettingZoo, you can just use all the major multi-agent RL libraries for learning.

hojoungjang commented 3 years ago

We will look into this. Thanks for sharing!