Chester-Xi

> > Yes, correct. > > I forgot one important issue which is your work/code now is just suit for "no negative samples" (all data in your case have event),...