Tuning, training and test sets used to train the first and second BERT model used by SnoBIRD

Tuning, training and test sets used to train the first and second BERT model used by SnoBIRD, respectively for the identification (C/D box snoRNA prediction in the general sense) and refinement steps (prediction of expressed C/D box snoRNA vs C/D box snoRNA pseudogene) of SnoBIRD. Both these datasets were data augmented using different strategies described in Fafard-Couture et al. 2025 and were split into tuning (used for hyperparameter tuning of the BERT model), training (used to train/fine-tune the BERT model) and test (used to assess SnoBIRD's performance on an independent dataset) sets.