This dataset is a collection of single-turn and multi-turn conversational data designed for Supervised Fine-Tuning (SFT) of large language models, specifically focusing on supporting multiple African languages alongside English. It is intended to train models to be helpful and friendly assistants capable of understanding and generating responses across a diverse linguistic landscape.
Kiya-SFT combines several existing instruction-following and… See the full description on the dataset page:
https://huggingface.co/datasets/NaolBM/Kiya-SFT.
