Question: Where did you get the history-length concept from?

Kaixhin / Rainbow

Rainbow: Combining Improvements in Deep Reinforcement Learning

MIT License

1.56k stars 282 forks source link

Closed albertwujj closed 5 years ago

albertwujj commented 5 years ago

Is it in a paper or did you decide to also train on adjacent previous transitions yourself?

Kaixhin commented 5 years ago

That's maybe not named well but it's the 4-stack of frames that has been used since the original DQN paper.