Adequately partition a labeled data set for training, hyper-parameter tuning, and accuracy assessment

Adequately partition a labeled data set for training, hyper-parameter tuning, and accuracy assessment