Tweaking Custom Environment Rewards - Reinforcement Learning with Stable Baselines 3 (P.4)