a All MoleculeNet datasets are split into training, validation and test subsets following a 80/10/10 ratio. Different splittings are recommended depending on each dataset's contents. For details of splitting methods please refer to the paper.
b Different classification and regress metrics are recommended based on previous works and dataset's contents:
ROC-AUC: Area Under Curve of Receiver Operating Characteristics
PRC-AUC: Area Under Curve of Precision Recall Curve
RMSE: Root-Mean-Square Error
MAE: Mean Absolute Error
For details of metrics please refer to the paper.