Maloba

Documents corpus · Earlier version

qwen-vl-lingala-dataset-augmented

Document set augmented by rotation, noise and contrast variation.

View on Hugging Face (opens in a new window)All corpora

How it was built

  • Set augmented from the annotated regions of the SIL Congo collection: rotation, noise and contrast variation applied to the images.

The corpus, as charts

Corpus sizes comparedThe project's four reference corpora, on the same scale. The speech corpus dominates in number of items; translation, the smallest, is the one that moved its model the most.
05 06710 13415 20120 268Speech — segments18 767Text — annotated examples4 818Documents — examples467Translation — human pairs399

Source : Livrable 4, chiffres arrêtés au 2026-09-30

Show the chart data
CorpusItems
Speech — segments18 767
Text — annotated examples4 818
Documents — examples467
Translation — human pairs399

Limits

  • Augmentation multiplies examples without adding information: it did not improve the model that was kept.
  • Historical version; the reference set is dataset-qwen-vl-lingala-qlora-vf.

Citation

Congo Digital Services (2026). qwen-vl-lingala-dataset-augmented. Hugging Face. NOODL-1.0 licence.