wolof-sft
61,971 fine-tuning lines
Official site consultedAI: datasets
Last verified: 1 October 2026
Dataset of 61,971 instruction-response pairs, mostly in Wolof, for supervised fine-tuning. It includes translations between Wolof, English and French and content on science, history, literature and general knowledge.
Services
- instruction, input and output columns
- Single train split, Parquet format (16.1 MB)
Access requirements
Free download on Hugging Face; no license is specified.
Sources
- wolof-sft (new tab), verified on 1 October 2026Official site consulted