facebookresearch/Pearl
A production-ready reinforcement learning library for building AI agents, developed by Meta's Applied Reinforcement Learning team.
Not currently ranked — collecting fresh signals.
star history
Pearl is a production-ready Reinforcement Learning AI agent library developed by Meta’s Applied Reinforcement Learning team. It enables researchers and practitioners to develop AI agents that learn from cumulative long-term feedback rather than immediate rewards. The library supports environments with limited observability, sparse feedback, and high stochasticity, providing tools to build state-of-the-art RL agents.
Frequently asked
- What is facebookresearch/Pearl?
- A production-ready reinforcement learning library for building AI agents, developed by Meta's Applied Reinforcement Learning team.
- Is Pearl open source?
- Yes — facebookresearch/Pearl is open source, released under the MIT license.
- What language is Pearl written in?
- facebookresearch/Pearl is primarily written in Jupyter Notebook.
- How popular is Pearl?
- facebookresearch/Pearl has 3k stars on GitHub.
- Where can I find Pearl?
- facebookresearch/Pearl is on GitHub at https://github.com/facebookresearch/Pearl.