A Tour of Driverless AI's Expert Settings
- Predictive AI
- H2O Driverless AI
This lesson gives a high-level tour of the Expert Settings window in Driverless AI, starting with the keyword filter that lets you jump straight to a setting like duplicate row handling instead of scrolling through every option. It covers the System tab, which handles failure behavior and core or GPU allocation, and the Experiment documentation tab, which controls what appears in the generated AutoDoc. Most of the lesson focuses on the Training tab's seven sub-tabs, covering data, feature engineering, models, the genetic algorithm, validation, and deployment, and flags that none of these settings can change once an experiment has launched.
Key takeaways
- How the keyword filter in Expert Settings helps you jump straight to a specific option like duplicate rows
- What the System tab controls, including failure handling and the number of cores or GPUs used
- Why the Training tab's seven sub-tabs cover everything from feature engineering to deployment settings
- Why Expert Settings can no longer be changed once you click the launch button on an experiment
Start owning your AI