What is the best way to restructure this sklearn class wrapper to avoid redundancy?

Question

I have trained multiple models in the form of Pipelines through sklearn, and I am creating a module of multiple classes that load these trained models.

For example, I tra