Micro tutorial on how to run and scale hyperparameter optimization with LightGBM and Tune

It is for you if:

you model structured data using LightGBM (classification or regression tasks).
want to run or scale hyperparameter optimization in your project.
Looking around for quick ways to try hyperparameter optimization in your project.

What will you do?

Run hyperparameter tuning with LightGBM on the structured data.
Configure scheduler for more efficient tuning.

What will you learn?

Few bits about Ray and Tune fundamentals.
How to use Tune to run hyperparameter optimization - quick start.
Few more bits about scheduler - to better define how tuning should progress.

Where to start?

First, make sure that you have an environment ready. Please follow the instructions on environment setup page (5 minutes read).
Once your env is ready, go ahead and start ray_tune_micro_tutorial.ipynb.

What to do next?

Check the user guides for more in-depth introduction to Tune.
Learn about Distributed LightGBM on Ray that enables multi-node and multi-GPU training.
Have a closer look at Tune docs to learn more about other search algorithms and schedulers.
Go to the Ray tutorials page to

How to connect with community, learn more, join other trainings?

Feel free to reach out on Ray-distributed Slack. Join #tutorials channel, say hello and ask questions.