sb3-soft Documentation

Reinforcement learning algorithms with soft Q-targets, built for Stable-Baselines3.

sb3-soft currently provides:

• SQL: Discrete Soft Q-Learning

• SDSAC: Stable Discrete Soft Actor-Critic

User Guide

• Quickstart
  • Installation
  • Basic SQL usage
  • Basic SDSAC usage
  • Notes

API Reference

• API Reference
  • SQL API
  • SDSAC API