Will RLlib consider implementing more distributed RL algorithms?

Hi there,

I’m a student working on RL research. Ray and RLlib are amazing projects, where I could browse and use various RL algorithms.

I learned a lot from RLlib’s implementation of APE-X and IMPALA. But RLlib lacks some recently proposed distributed RL baselines. Will RLlib consider implementing some recent algorithms such as SEED RL or even newer ones?


@chitandaeru , rllib algorithms You can find here:

We have no plans on the roadmap to implement new distributed RL baselines.

But we’ll gladly accept contributions!