I would like to ask whether these indicators are obtained from the test set. Does graph clustering need to divide test set and training set

[bz154]