A collection of token classification models designed for diacritic tagging and restoration.