This repository illustrates some ML challenges of training a custom ML (multiclass text classification) model for each user.

This repository complements my presentation at Data Science fwdays'21 conference: Custom ML Models for Each User.

When training a separate model for each user, there are multiple training data sets each of its own size.

Here you can find: how different model metrics change with training data set size.

The data set used is based on GitHub Issue Classification tutorial that illistrates ML.NET usage.