Recurrent Discrete Soft Actor-Critic (SAC)

This repository contains work-in-progress code for research into offline reinforcement learning with POMDPs, currently utilizing SAC as the backbone algorithm.