Reinforcement Learning of Quadruped Locomotion Gaits in OpenAI Gym