k-fold r2
I am performing a stepwise procedure and evaluating the performance using a k-fold r-squared. Currently, the only k-fold r-square that is provided is averaged across all 5 folds. Is it possible to secure the r-squared for each individual k-fold in a model?