Abstract
The dataset consists of the 723 early imperial (1 BC to AD 401) funerary and honorific inscriptions in Greek, Latin, and Hebrew from the I.Sicily database. The dataset is stored on Zenodo in the form of two .conllu and two .vert files, one for the (primarily) Latin and one for the (primarily) Greek inscriptions, containing the tokenized, lemmatized, part-of-speech tagged, and verticalized inscription texts with their ISicXXXXXX identifiers. The Python scripts used to prepare the dataset and the modified corpus configuration file to implement the .vert files into Sketch Engine are provided for reuse and development.