I have a binary classifier, which classifies an input X as class zero if its predicted value is below some threshold (say T), and one otherwise.
I have all
labels are the true labels of the data, scores are the output scores from your classifier (before the threshold) and posclass is the positive class in your labels.