scientific-reinforcement-learning

What problem does it solve?

Builds and orchestrates end-to-end reinforcement learning experiments for scientific workflows, reducing setup time and enabling reproducible research with standardized tooling.

Core Features & Use Cases

• Stable-Baselines3 integration for common RL algorithms and training loops.
• Gymnasium environment templates and wrappers for custom scientific tasks.
• PufferLib-powered large-scale multi-agent coordination for collaborative experiments (e.g., molecular design, lab automation).
• Quick-start templates to reproduce typical pipelines and experiments.

Quick Start

Run the included training script to initialize an RL experiment using the provided templates.