Congo-digital-service/audios-lingala-annotatees

Annotated Lingala Dataset – Full Version

Deprecated. Do not use for new work. This first release is kept online, unchanged, only as the provenance record of the models trained from it. It published relative audio paths without the audio itself, and pre-computed spectrograms. The statistics below are those of the original card and are kept for the record. Use audios-lingala-annotatees-v3.2.

Datasheet (short)

Item	Description
Core purpose	First release of the MALOBA Lingala ASR data; provenance record of the first Whisper fine-tuning.
Source	Radio Rurale of the Republic of Congo, sole owner of the recordings, which consented to their use in the MALOBA project.
Annotators	A team of Lingala-speaking annotators, mostly graduates of the Faculté des Lettres et des Langues Vivantes, Université Marien Ngouabi (UMNG), trained in AI and IT and paid for their work by CDS. Identities not published.
Status	Deprecated; superseded by v2, then v3.2. Kept unchanged.
Licence	NOODL-1.0, full text in LICENSE.md.
Full datasheet	See audios-lingala-annotatees-v3.2.
Contact	contact@congo-digital.com

Description

This dataset gathers annotated Lingala audio data, intended for open-source automatic speech recognition (ASR) research and for fine-tuning Whisper-type models.

It includes:

• the original audio files (viewable directly in the Hugging Face viewer)
• text transcriptions
• Mel spectrograms
• tokenized labels

Overall statistics

Metric	Value
Total volume	5 h 0 min 18 s
Number of audio segments	2,940
Speakers	31 speaker IDs (estimated 25–31 unique speakers)
Transcribed words	26,238
Distinct vocabulary	4,064 words

Audio segmentation

Metric	Value
Average segment duration	6.13 seconds
Shortest segment	0.95 seconds
Longest segment	24.09 seconds

Training split

Split	Segments	Duration	Speakers
Train	2,557	4 h 20 min 20 s	22
Validation	383	39 min 59 s	9

Note: speaker IDs are disjoint between the train and validation sets, allowing evaluation on voices unseen during training.

Content and context

• Language: Lingala (code lin)
• Main topics: daily life, family, relationships, economy (work, trade), religion, travel, health, and education
• Linguistic variety: urban Kinshasa Lingala, oral, everyday, conversational register, with borrowings from French
• Geography of mentioned locations: mainly communes of Kinshasa (Barumbu, N'Djili, Kintambo, Masina, Gombe, Kimbanseke, Limete, Lemba)

Normalization status

The corpus requires some technical normalization before definitive use, notably:

• harmonization of Unicode characters (e.g. ε vs ɛ)
• standardization of lexical spelling (e.g. po na vs mpo na)
• handling of digits (numbers intended for ASR should be spelled out in full)
• clarification of certain text values (e.g. fkm)

Dataset link

https://huggingface.co/datasets/Congo-digital-service/audios-lingala-annotatees

License

This dataset is published under the Nwulite Obodo Open Data License (NOODL-1.0), a license designed for equitable sharing of African datasets.

This dataset, audios-lingala-annotatees (Congo-digital-service/audios-lingala-annotatees), was created by Congo Digital Services (CDS SARL) (https://www.congo-digital.com/) in August 2026 in collaboration with the MALOBA community (https://maloba.congo-digital.com/). It is licensed under the Nwulite Obodo Open Data License (https://licensingafricandatasets.com/nwulite-obodo-license).

In exchange for use of this dataset, users from developing countries are required to provide: Use of the dataset (no additional benefit required). Users from high-income countries or commercial entities are additionally required to publicly acknowledge and credit the Maloba Project (UNDP Republic of Congo — language digitalisation initiative) in any publication, product, model, or output derived from this dataset. To fulfil this requirement, contact contact@congo-digital.com or visit https://www.congo-digital.com/contact.

Creators

• Congo Digital Services (CDS SARL) — https://www.congo-digital.com/
• In collaboration with:
  • the MALOBA community — https://maloba.congo-digital.com/
  • Radio Rurale
  • the Faculté des Lettres et des Langues Vivantes, Université Marien Ngouabi (UMNG)
  • UNDP Congo
  • the Ministry of Posts, Telecommunications and Digital Economy of the Republic of Congo
• Created in August 2026