Implement "Curious Replay" into Dreamer?

Hi,
I recently saw the paper [Curious Replay for Model-based Adaptation](https://arxiv.org/pdf/2306.15934), which proposes a fairly straight forward curiosity-based sampling from the Replay Buffer.

Are there any plans to potentially integrate this into the SheepRL Dreamer implementations?

Thanks

No files, tests, or entry points are named. Start by reading the linked Curious Replay paper and locating the Dreamer implementations and their replay-buffer sampling path. A useful next step would require defining the integration scope and acceptance criteria with maintainers before implementation.

Written by the indexing model from the issue text.