All Tools
R
OtherFreeOpen Source
RLAX
JAX library of RL building blocks by DeepMind
Apache-2.0
ABOUT
Implementing reinforcement learning algorithms from scratch is error-prone and hard to compose, and most RL libraries are designed for single frameworks. RLax provides modular, composable building blocks that work with JAX's functional programming model, enabling researchers and engineers to prototype and experiment with RL algorithms without reimplementing core components.
INSTALL
pip install rlaxINTEGRATION GUIDE
1. Build custom reinforcement learning agents with composable value and policy components
2. Implement and benchmark distributional RL algorithms for game-playing and robotics
3. Create RL training pipelines that leverage JAX's automatic differentiation and hardware acceleration
TAGS
reinforcement-learningjaxdeepmindmachine-learningpythonresearch