The key is a while loop on the mean_reward variable when evaluating your agent🥰📚

Training more and learning the routine

Trained 1mil steps once - LunarLander-v2 trained agent