Description
Arcade Learning Environment experiments can be run from Python for reinforcement-learning research. This library is for researchers and developers benchmarking agents on Atari-style environments. It is a research platform, and results depend on ROM availability, seeds, and experiment configuration.