AI AGENTS
rl-baselines3-zoo
by DLR-RM
Overview
Training framework for Stable Baselines3 reinforcement-learning agents with scripts for training, evaluation, hyperparameter tuning, plotting, videos, and pretrained agents.
Details
RL Baselines3 Zoo is a DLR-RM project for working with Stable Baselines3 reinforcement-learning agents. Its documentation describes it as a training framework that provides scripts for training and evaluating agents, tuning hyperparameters, plotting results, and recording videos. The related DLR-RM rl-trained-agents repository hosts a collection of pretrained reinforcement-learning agents using Stable Baselines3.
When to Use
Train or evaluate Stable Baselines3 reinforcement-learning agents using an established project workflow. Tune hyperparameters plot experiment results or record videos for RL experiments. Start from or compare against pretrained Stable Baselines3 agents hosted by DLR-RM.
Getting Started
- Open the GitHub repository at https://github.com/DLR-RM/rl-baselines3-zoo to review the project source.
- Read the documentation at https://rl-baselines3-zoo.readthedocs.io/en/master/index.html for the available training
- evaluation
- tuning
- plotting
- and video-recording scripts.
- Review pretrained agents in https://github.com/DLR-RM/rl-trained-agents if you want existing Stable Baselines3 agent checkpoints or baselines.
Key Features
- •Training framework for Stable Baselines3 reinforcement-learning agents.
- •Scripts for training and evaluating agents.
- •Hyperparameter tuning support.
- •Plotting results and recording videos for experiments.
- •Access to a related collection of pretrained Stable Baselines3 agents.
Capabilities
- •reinforcement-learning-agent-training
- •agent-evaluation
- •hyperparameter-optimization
- •experiment-plotting
- •video-recording
- •pretrained-agent-reuse
Last updated Jun 5, 2026