kfoldpredict do I understand it correctly?

Okay, I have built a crossvalidated regression tree. now i want to predict its accuracy for that i need the predicted value of y in order to compare it to the actual y. Imagining I did a 10 kfold crossvalidation. so i have tested now ten times the training data on my test data. which gives me 10 times predicted y values. does kfold predict gives me an average of all of them?

Any help welcome, I am going a bit mad on this...