How many unique classes are present in the dataset based on the label distribution?

How many unique classes are present in the dataset based on the label distribution?: a task in SmolDataEnvs-harbor-train-sandbox: 📊 SmolDataEnvs: Harbor (train) (Harbor dataset). Files (in /home/user/input, no subfolders): - sign mnist train.csv - sign mnist test.csv