uctnlp/mzansi-text
MzansiText MzansiText is a curated multilingual pretraining corpus for all eleven official South African languages. Dataset details Splits: 3,943,584 train rows, 19,379 validation rows, and 19,341 test rows lang values: afr, eng, nbl, nso, sot, ssw, tsn, tso, ven, xho, zul Schema: { "text": "string", "lang": "string" } Validation and test sets are capped at approximately 2M tokens per language to prevent high-resource languages from dominating early… See the full description on the dataset page: https://huggingface.co/datasets/uctnlp/mzansi-text.
| Author / organization | uctnlp |
|---|---|
| License (as declared) | apache-2.0 |
| Downloads | 583 (as reported by Hugging Face) |
| Likes | 12 |
| Created | 11 January 2026 |
| Last updated at source | 3 October 2026 |
Popularity, last 3 days
583 now (-22 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
Related papers
More from uctnlp
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.