RayLink is a RL framework used to build and train RL algorithms. RayLink was used to build a RL framework, and tested in a large-scale multi-agent training environment. It can help you easily build your own workload. RayLink will connect every working node, and let you efficiently arrange your own dataflow. Because you can define a node type based on its functionality, all dataflow are more clear to researchers and developers.

It's based on ray for now, and it has implemented several new features to improve user experience.

RayLink was used to build a reinforcement learning framework, and tested in a large-scale multi-agent training environment. It also can be used as a common actor framework.

Clone this project and use pip install . to install raylink.

Check out demo_script.py in script folder for quick start.