the size of training set?

[knimet]

hi, I recheck the size of training set, there are 1227 questions and it's different with the statistic you reported in the paper. Should it be 1227 or I made a mistake? If you don't mind, would you please check the training set ? :)

[knimet]

here is the script I use :

t = pd.read_csv('train-all.txt',sep='\t',header=None) print(t[0].unique().len()) #1227