rlworkgroup/garage
A reinforcement learning research toolkit providing modular components for implementing and evaluating RL algorithms.
Not currently ranked — collecting fresh signals.
star history
Garage provides a comprehensive framework for developing and evaluating reinforcement learning algorithms. It includes composable neural network models, replay buffers, high-performance samplers, and an expressive experiment definition interface. The toolkit supports reproducibility through global random seeds, logging to TensorBoard, and reliable experiment checkpointing and resuming. It ships with implementations of state-of-the-art RL algorithms and interfaces for popular benchmark suites.
Frequently asked
- What is rlworkgroup/garage?
- A reinforcement learning research toolkit providing modular components for implementing and evaluating RL algorithms.
- Is garage open source?
- Yes — rlworkgroup/garage is open source, released under the MIT license.
- What language is garage written in?
- rlworkgroup/garage is primarily written in Python.
- How popular is garage?
- rlworkgroup/garage has 2.1k stars on GitHub.
- Where can I find garage?
- rlworkgroup/garage is on GitHub at https://github.com/rlworkgroup/garage.