UniNum is a database of number names for 186 languages, locales, and scripts made available by Google.

In those languages where numbers take on different pronunciations depending on their morphosyntactic context, a citation form is given.

See codes.tsv for the names and scripts involved.

If you find errors in this database, please feel free to submit corrections.

For more information about this database, please see:

Gorman, K., and Sproat, R. 2016. Minimally supervised number normalization. Transactions of the Association for Computational Linguistics 4: 507-519.

Ritchie, S., Sproat, R., Gorman, K., van Esch, D., Schallhart, C., Bampounis, N., Brard, B., Mortensen, J. F., Holt, M., and Mahon, E. 2019. Unified verbalization for speech recognition & synthesis across languages. In Proc. INTERSPEECH, pages 3530-3534.

If you use UniNum in your research, the authors would appreciate if you refer your readers to the Ritchie et al. paper, and to this repository.

This is not an official Google product.