MaxStrange / retrowrapper

Wrapper for OpenAI Retro envs for parallel execution

retrowrapper

Wrapper for OpenAI Retro envs for parallel execution

OpenAI's Retro exposes an OpenAI gym interface for Deep Reinforcement Learning, but unfortunately, their back-end only allows one emulator instance per process. To get around this, I wrote this class.

To Use

To use it, just instantiate it like you would a normal retro environment, and then treat it exactly the same, but now you can have multiples in a single python process. Magic!

import retrowrapper

if __name__ == "__main__":
    game = "SonicTheHedgehog-Genesis"
    state = "GreenHillZone.Act1"
    env1 = retrowrapper.RetroWrapper(game, state=state)
    env2 = retrowrapper.RetroWrapper(game, state=state)
    _obs = env1.reset()
    _obs = env2.reset()

    done = False
    while not done:
        action = env1.action_space.sample()
        _obs, _rew, done, _info = env1.step(action)
        env1.render()

        action = env2.action_space.sample()
        _obs, _rew, done, _info = env2.step(action)
        env2.render()

Using a custom make function

Sometimes you will need a custom make function, for example the retro_contest repository requires you to use their make function rather than retro.make.

In these cases you can use the retrowrapper.set_retro_make() to set a new make function.

Example usage:

import retrowrapper
from retro_contest.local import make

retrowrapper.set_retro_make( make )

env1 = retrowrapper.RetroWrapper(
    game='SonicTheHedgehog2-Genesis', 
    state='MetropolisZone.Act1'
)
env2 = retrowrapper.RetroWrapper(
    game='SonicTheHedgehog2-Genesis', 
    state='MetropolisZone.Act2'
)