pipette
ESEspañol

Wolof-Adapted Industry-Ready Open-Weight Language Models

Gw\'enol\'e Lecorv\'e, Delphine Charlet, Pierre Adam, Floriane Behanzin, Lionel Delphin-Poulat, Johannes Heinecke, G\'eraldine Damnati, Pelin Kaytmaz, Nourredine Khaous, Charbel Kindji, Meriam Maadi, Abdoulaye Mbaye, Claire Perroux, Isma\"el Rousseau, Papa S\'ega Wad\'e, Gilles Xiberras, Olivier Zhang, Malek Zitouni

PreprintReal-world useCode available

In the authors' words

Wolof is spoken by 23M people, mostly in Senegal, but remains weakly supported by open-weight language models, while API models raise cost, governance, and sovereignty constraints for local deployments. In this paper, we present 0.6B-14B open-weight models adapted to Wolof through instruction tuning, preference alignment, and model merging. After presenting the pipeline and data, extensive experiments on 63~public and in-house tasks show good performance on Wolof understanding and generation while remaining effective on other languages and cross-lingual skills, with the 8B and 14B models being particularly competitive with substantially larger commercial systems. The models are currently being evaluated in internal customer-care pilots and are publicly available at https://huggingface.co/collections/Orange/wolof-lm .

Main resultLimitation the authors admit

Appeared: Friday, October 16. arXiv. Preprint, not yet peer-reviewed.

Authors' comment: 6 + 47 pages (main + appendices), 27 tables, 1 figure