purejaxrl
High-performance, end-to-end reinforcement learning implementations fully written in JAX for massive parallelization on GPUs.
What is purejaxrl?
PureJaxRL is a high-performance reinforcement learning library that implements end-to-end training pipelines entirely in JAX. It solves the problem of slow RL experimentation by enabling massive parallelization of agents on GPUs, achieving speedups of over 1000x compared to traditional PyTorch implementations. The library allows researchers to JIT compile complete training loops including environments for optimal performance.
Reinforcement learning researchers and practitioners who need to run large-scale experiments, hyperparameter tuning, or meta-RL algorithms efficiently on GPU hardware.
Developers choose PureJaxRL for its unmatched performance gains through full JAX implementation, enabling parallel training of thousands of agents simultaneously. Its single-file, research-friendly design makes it ideal for algorithm discovery and rapid prototyping compared to modular but slower alternatives.
Overview
Really Fast End-to-End Jax RL Implementations
Use Cases
Best For
- Running thousands of parallel RL agent seeds on a single GPU
Related Projects
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
