# Swahili-Safi Dataset A relatively clean dataset for Swahili language modeling, built by combining and cleannig several existing datasets. Sources include: ``` mc4-sw oscar-sw swahili_news IWSLT XNLI flores 101 gamayun-swahili-minikit broadcastnews-sw ```