2  Model Serving

Under Development

Model serving on Databricks is well documented for Python. It is a common misconception that you cannot serve R models on Databricks, you can!

For most model flavours the amount of effort is not worth it - but that decision is not mine to make, I will show you how and you can decide for yourself.

2.1 Serving Overview

Caution

Work In Progress

2.2 XGBoost

Caution

Work In Progress

XGBoost is relatively straightforward since it the artifacts generated when saving the model are universally understood across each language the library is available in.

2.3 ONNX

Caution

Work In Progress

2.4 Crate

Caution

Work In Progress