Turkish NLP Suite is a nonprofit organization that aims to deliver resources for Turkish NLP, including corpus, pretrained models, benchmarking resources and much more. We aim to bring excellence to the Turlish language and speech research by publishing and open-sourcig for all areas of Turkish processing including morphology, language modelling, subword modelling, corpora building and more. The organization is run by Duygu, your research scientist (originally )from Istanbul, (currently) living in Berlin and San Diego, CA.
know more
We proudly present the first part of our Turkish subword manifesto: how Turkish language modeling benefits from morphology. In this article, we compare character-, word-, and morphology-aware subword tokenization, where...
Read More
For a long time, Turkish models climbed leaderboards written somewhere else—translated datasets, English‑first assumptions, noisy web scrapes. Useful, yes. Representative of real Turkish? Not quite. TrGLUE is our answer: a...
Read More