Questions about dataset

[Neakail]

Hello. We tried to process the dataset according to the description on GitHub: "For the training, validation and test sets, we discard invalid molecules and molecules with only one atom." There are certain differences in the results. I wonder if it's convenient for you to share the data and the inference scripts? Thank you!