add training_dataset.nf pipeline