OpenAI Gym (https://github.com/openai/gym) is an open source library that was written to allow access to a standardized set of reinforcement learning tasks. It provides a toolkit that can be used to compare and develop reinforcement learning algorithms.
OpenAI Gym consists of a collection of environments, all presenting a common interface called env. This interface decouples the various environments from the agents, which can be implemented in any way we like—the only requirement from the agent is that it can interact with the environment(s) via the env interface. This will be described in the next subsection.
The basic package, gym, provides access to several environments and can be installed as follows:
pip install gym
Several other packages are available, such as 'Atari', 'Box2D', and 'MuJoCo', that provide access to numerous...