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.