How can I implement or obtain diminishing return in sklearn-linear regression?

Question

I have a campaign data for which I\'d like to obtain the diminishing return.

I\'m trying to implement linear regression analysis for the Cost of Ad Vs App downloads in Py