Predicting results on unknown data in a k fold validation model python

Question

I am building Regression and classification Models to calculate train and test error and to predict results on an unseen data. To make the model more robust I am using k-fol