crowdAI / marLo

Multi Agent Reinforcement Learning using MalmÖ
MIT License
245 stars 46 forks source link

specified number of rollouts in experiment #14

Closed AndKram closed 6 years ago

AndKram commented 6 years ago

run a specified number of rollouts and then join up with the background agents before exiting. I've also removed the AsyncEvaluator class (can be put back if needed).