Title: Possible label errors in BBB_Martins test set — 2 molecules appear mislabeled
Description: While auditing the BBB_Martins dataset from the tdc.single_pred.ADME module, we identified two molecules in the test split that appear to have incorrect labels. These discrepancies affect evaluation metrics and introduce a ceiling on achievable accuracy (e.g., 99.51% max if 2/406 are mislabelled).
from tdc.single_pred import ADME
data = ADME(name = 'BBB_Martins')
split = data.get_split()
Thank you for your time.
Title: Possible label errors in BBB_Martins test set — 2 molecules appear mislabeled
Description: While auditing the BBB_Martins dataset from the tdc.single_pred.ADME module, we identified two molecules in the test split that appear to have incorrect labels. These discrepancies affect evaluation metrics and introduce a ceiling on achievable accuracy (e.g., 99.51% max if 2/406 are mislabelled).
from tdc.single_pred import ADME
data = ADME(name = 'BBB_Martins')
split = data.get_split()
Thank you for your time.