Role Play

This repo is the code for paper "Role-Play: Learning Adaptive Role-Specific Strategies in Multi-Agent Interactions" arxiv.

Environment setup

Create an conda virtual environment:

conda create -n rp python==3.10 -y

Install the requirement list:

conda activate rp
pip install -r requirements.txt

Train example

Train 1e7 environment step in two players harverst environment:

python ${workspaceFolder}/experiments/run_meltingpot.py --stop=1e8 --configs harvest

If you want to save the training results in WanDB, add --wandb=True and modify your wandb information in run_meltingpot.py.