Releases: anyscale/academy
Prep for ODSC West
Prep for Ray Summit tutorials
- updated yaml config (resolve
pipvs.condaconflict for PyTorch)
Prep for Ray Summit tutorials
- simplified the
recsysdependencies - sync'ed the conda and pip dependencies
Robustness Improvements for Anyscale Platform
Minor updates that only affect deployment to the Anyscale platform.
Ray Summit Pre-Release
This release is targeted for Ray Summit 2020 tutorials, although it's possible that minor updates will be made before then.
This release adds full support for running the Academy in a Docker image on the Anyscale platform. (Docker support on a local machine is still TODO.)
There are also updates to the RL & RLlib tutorial, namely a new "keystone" example of using RL for recommender systems. See the ray-rllib/recsys directory.
This release also supports Ray 0.8.7 and improves the support for running on Windows.
Colorized "Game of Life", Refinements and Bug Fixes
This release adds a fun update to Game of Life, adding colored graphs, vs. monochrome, to indicate the age of each cell.
Also includes numerous bug fixes and text refinements.
New Tutorials for Ray Tune and Ray Serve - Bug Fixes
This patch release fixes two issues in the notebook ray-tune/03-Search-Algos-and-Schedulers.ipynb:
- When running on Python 3.6, you have to force an update to a library. The beginning of the notebook provides instructions.
- There was a typo in the link to the exercise solution.
New Tutorials for Ray Tune and Ray Serve
This release completes the new tutorials for Tune and Serve. It is the content used for the July 22 live Anyscale Academy event.
Draft of tutorials for Ray Tune and Ray Serve
This is a draft release with Ray Tune and Ray Serve tutorials.
Support for Ray 0.8.6, experimental Windows support
This release upgrades to Ray 0.8.6, which has experimental support for Windows. See the README for details.