the metrics that are calculated for the training set measure of the goodness of fit of the fitted model and those calculated for the validation set are used to measure predictive performance. A large difference between the training and validation errors is the likelihood of overfitting the training data