Rainbow Hyperparameters

[rfali]

Hi. The paper mentions

Rainbow (Hessel et al., 2018) is based on Q-Learning and combines several advances, including for exploration. The defaults for Atari did not work well, so we tuned the hyper parameters for Crafter and found a compromise between Atari defaults and the data-efficient version of the method (van Hasselt et al., 2019) to be ideal.

Are these hyperparameters available anywhere, and if not, can you please share these? The crafter baselines for Rainbow has the same hyperparameters as Kaixhin's Rainbow (which are designed for ALE), and there are no hyper parameter values mentioned in the paper. Thanks for your time!

[danijar]

Yes, those are the hyper parameters. I tried changing a handful of them but nothing I tried ended up working as well as the defaults.