google-deepmind / dqn_zoo

DQN Zoo is a collection of reference implementations of reinforcement learning agents developed at DeepMind based on the Deep Q-Network (DQN) agent.
Apache License 2.0
451 stars 78 forks source link

Is the AtariEnvironmentWrapper not being used? #19

Closed raymondchua closed 1 year ago

raymondchua commented 1 year ago

Hi, Is this function being used anywhere?

jqdm commented 1 year ago

Thanks for your interest, indeed AtariEnvironmentWrapper is not used anywhere. It was originally used during development of these agents to make it easier to compare having the Atari preprocessing in the environment versus in the agent itself. We eventually went with having the preprocessing in the agent for a number of reasons, but since it is more common to have the preprocessing in the environment, AtariEnvironmentWrapper was kept for the convenience of users.

raymondchua commented 1 year ago

I see. Thanks for your reply, John :)