TWB Parallel Sentence kits - Congo Swahili (25k)

The Congo Swahili portion of CLEAR Global's Gamayun Language Data Kits — 25,305 parallel French–Congo Swahili sentences in total, distributed across three independent kit sizes: 5,000 (`kit5k`), 10,000 (`kit10k`), and 10,305 (`kit15k`, partial delivery of the medium-kit target). French source sentences were drawn from the Tatoeba repository using a selection algorithm that ensures representation of the most frequently used words in French; Congo Swahili translations were produced by professionals and volunteers of the Translators without Borders (now CLEAR Global) translator community. Each kit TSV includes a `swc_clean` column with parenthetical content removed, in addition to the raw translation. Each kit TSV is accompanied by a core kit TSV containing the original French source sentences with their Tatoeba IDs. Note that kit sizes are independent selections — `kit10k` is not a superset of `kit5k`. Gamayun is part of CLEAR Global's initiative to develop open-source language technology for under-resourced languages used in humanitarian contexts.