We are excited to announce Ray 2.3.0. Here are the highlights:
Observability enhancements to the Ray Dashboard
Ray Dataset Streaming (developer preview)
Boost in Ray core scheduling performance
Additions of Gym/Gymnasium library to RLlib
Support for ARM and Python 3.11
Support for multiple applications for Ray Serve (developer preview)
You can read the blog and release notes for details.
We want to thank you all contributors for your valuable contributions to this new release of Ray 2.3. Your enduring support continues to foster the wider use of Ray adoption.