Classroom is an open source Python library for teaching reinforcement learning agents about human preferences. It comes packaged with a browser-based GUI for providing human feedback to RL agents. Classroom builds upon OpenAI's popular Gym library for reinforcement learning environments, and aims to be compatible with both PyTorch and JAX. The goal of Classroom is to make preference-based reinforcement learning easy, requiring only a few extra lines of code on top of the usual RL pipeline.
This repo is currently a work in progress and is not yet production-ready. In order to use it, you'll need to have Python 3.10 or later, as well as a Node.js installation for building the GUI, which is written in Svelte and TypeScript.