ProjectFeatured
imitation (HumanCompatibleAI)
Clean implementations of behavioral cloning, DAgger, GAIL, AIRL and more for imitation learning research.
ProjectFeaturedBeginner
Stable-Baselines3
Reliable RL algorithm implementations (PPO, SAC, …) widely used before specialized robot trainers.
Why we recommend it
Default reliable RL baselines for many robot students
Overview
Stable-Baselines3 (SB3) is the default reliable implementation of standard RL algorithms (PPO, SAC, TD3, DQN, …) for many students entering robot learning. It is not robot-specific, but it is the cleanest way to learn on-policy / off-policy loops before specialized stacks like rsl_rl or Isaac Lab trainers.
Pair SB3 with Gymnasium / Gymnasium-Robotics or simple MuJoCo tasks first. Once you need massive parallel locomotion or GPU envs, graduate to robot-native trainers.
On this hub’s Learning Path, SB3 sits in Foundations: get one training curve and one evaluation rollout working before touching VLAs.
Who it is for
Researchers, students, and builders comparing open embodied stacks.
Key highlights
- Reliable RL algorithm implementations (PPO, SAC, …) widely used before specialized robot trainers.
- Type: project; categories: framework
- Marked recommended on the hub
When to use
- You are mapping the open embodied AI landscape
- You need a starting URL and judgment note before deep-diving
When not to use
- You need production SLAs or certified industrial support only
- License or hardware constraints are incompatible with your deployment
Getting started
- 1Open the official URL and read README install pins.
- 2Skim license and citation requirements.
- 3Run the smallest official example before scaling experiments.
- 4Cross-link related hub resources (sim / data / policy) for a full stack.
Install / setup
pip
pip install stable-baselines3[extra]
How it compares
Best generic RL algorithms package; use rsl_rl for high-throughput locomotion.
★ 1.8k
Caveats & pitfalls
- Hub cards are curated summaries — always verify upstream docs.
- Stars and dates drift; treat extra.stars as approximate.
Content reviewed 2026-07-28