Goal:
Implement a way to a cancel Logistic Regression training during runtime without impacting the performance too much.

Solution proposal:
Identify code paths that are CPU intensive such as tight loops and place a check in them for cancellation signal.

Below is the performance graphs of Logistic Regression run from this example. The graph will have a function selected and red stripes indicate the position in the graph it is called.

The checkpoint should be at cursor.MoveNext() in TrainCore method. As we can see in the graph this function is called periodically in the training process. Everything before this is CPU cycles consumed by the transforms and it is not related to Logistic Regression training cycles.