DLR-RM/stable-baselines3
PyTorch version of Stable Baselines, reliable implementations of reinforcement learning algorithms.
Implements canonical on-policy (PPO, A2C) and off-policy (DQN, SAC, TD3) algorithms with a unified sklearn-like API, supporting Dict observation spaces and custom policies via modular network architectures. Integrates with Gymnasium for environment interaction, TensorBoard for experiment tracking, and Weights & Biases/Hugging Face for model management and sharing. Includes companion tools like RL Zoo for hyperparameter tuning and SB3-Contrib for experimental features (masked action support, recurrent policies).
12,878 stars. Used by 12 other packages. Actively maintained with 3 commits in the last 30 days. Available on PyPI.
Stars
12,878
Forks
2,081
Language
Python
License
MIT
Category
Last pushed
Feb 21, 2026
Commits (30d)
3
Dependencies
6
Reverse dependents
12
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/DLR-RM/stable-baselines3"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
google-deepmind/dm_control
Google DeepMind's software stack for physics-based simulation and Reinforcement Learning...
Denys88/rl_games
RL implementations
flatland-association/flatland-rl
The Flatland Framework is a multi-purpose environment to tackle problems around resilient...
takuseno/d3rlpy
An offline deep reinforcement learning library
pytorch/rl
A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.