Home

Welcome to the RL-Framework wiki!

Some quick notes.

The Environment wrapper is primarily for control frequency and determining when an episode has ended. The Actor is used to perform any kind of special operation on the state space, in this case creating the hierarchy and computing LLP actions. The [implementtps://github.com/Neo-X/RL-Framework/blob/master/algorithm/AlgorithmInterface.py) class is really an abstract interface the learning algorithm should impliment to that in can "plugin" to the learning framework that collection trajectories and performs training updates.