Hands-On Intelligent Agents with OpenAI Gym

Creating a template for custom Gym environment implementations

Based on the anatomy of the Gym environment we have already discussed, we will now lay out a basic version of a custom environment class implementation named CustomEnv, which will be a subclass of gym.Env and implement the essential methods and arguments required to make it a Gym-compatible environment. A template for a minimal implementation is as follows:

import gymclass CustomEnv(gym.Env):    """    A template to implement custom OpenAI Gym environments    """    metadata = {'render.modes': ['human']}    def __init__(self):        self.__version__ = "0.0.1"        # Modify the observation space, low, high and shape values according to your custom environment's needs self.observation_space = gym.spaces.Box(low=0.0, ...