Which wine quality class had the fewest samples in the original dataset before binning into binary categories?

Which wine quality class had the fewest samples in the original dataset before binning into binary categories?: a task in SmolDataEnvs-harbor-train (Harbor dataset). Files (in /home/user/input, no subfolders): - winequality-red.csv