Easy way to evaluate two pre-trained agents

Hello, I was wondering if there is any easy way to directly let **two pre-trained agents** (e.g. agent0 is MEP and agent1 is COLE) **play** with each other just by modifying `eval.py`, `overcooked_runner.py` and the .sh file. I don't think that I would need to change other parts of code, but could you let me know if it's indeed possible to do that? Thank you!