search

takuseno / d4rl-atari

Datasets for data-driven deep reinforcement learning with Atari (wrapper for datasets released by Google)
MIT License
101 stars 14 forks source link

why using NoFrameskip environment? #12

Closed GilgameshD closed 2 years ago

GilgameshD commented 2 years ago

In the document of the DQN Replay Dataset, it says:

"Note that the dataset consists of approximately 50 million tuples due to frame skipping (i.e., repeating a selected action for k consecutive frames) of 4"

but I notics that the online environment is built with NoFrameskip. Is there any reason for this?

takuseno commented 2 years ago

You do not need to worry about this. It's handled inside AtariPreprocessing. https://github.com/takuseno/d4rl-atari/blob/799428bbc570a224c2df58a78c878f3410b0fa59/d4rl_atari/envs.py#L27

GilgameshD commented 2 years ago

cool, thanks for your answer.