What percentage of the original dataset was classified as spam before balancing the classes?

What percentage of the original dataset was classified as spam before balancing the classes?: a task in SmolDataEnvs-harbor-train-sandbox: 📊 SmolDataEnvs: Harbor (train) (Harbor dataset). Files (in /home/user/input, no subfolders): - spam.csv