How to run the comix algorithm on multi-mujoco tasks? #3
Description
Hi, could you please tell me how to run the comix algorithm on multi-mujoco tasks?
I use the command as follow:
"python3 src/main.py --config=comix_pp --env-config=mujoco_multi with env_args.scenario_name=manyagent_swimmer env_args.agent_conf=10x2 env_args.agent_obsk=1"
and the performance is incongruent with fig 6 in the paper.