USPTO_500_MT miss '334' in the 'labels' column?

HelloJocelynLu / t5chem

Transformer-based model for chemical reactions

MIT License

58 stars 14 forks source link

USPTO_500_MT miss '334' in the 'labels' column? #5

Closed DaShenZi721 closed 2 years ago

DaShenZi721 commented 2 years ago

I found there are 499 labels without '334'(i.e. [0, 1, ... 333, 335, ... 498, 499]) in the 'labels' column of the dataset USPTO_500_MT. Is it correct?

HelloJocelynLu commented 2 years ago

Hi, It is correct. It is only a subset of original USPTO_1k_TPL after removing some sparse classes.