Implementation of multiple RL algorithms in pytorch (DQN, AC, ACER) and trained agents to play digital games