Minor fix for RMSRescaler to avoid NaN

facebookresearch / rlmeta

RLMeta is a light-weight flexible framework for Distributed Reinforcement Learning Research.

MIT License

284 stars 28 forks source link

Closed xiaomengy closed 2 years ago