Oriel — AI & technology worth noticing, every day.

Data-Science-Nigeria/dsn-2026-lid-train

Dataset · https://huggingface.co/datasets/Data-Science-Nigeria/dsn-2026-lid-train

DSN 2026 spoken language identification: training data Training and validation data for the DSN 2026 AI Bootcamp competition Find the Language I Am Speaking. The task: given a short audio clip, predict which of 12 Nigerian languages is spoken, or other. The hidden test set is not part of this dataset. Contents data/train-*.parquet (24774 clips) and data/validation-*.parquet (4352 clips). Columns: id, audio (16 kHz mono FLAC clips of 2 to 30 s, stored inside the… See the full description on the dataset page: https://huggingface.co/datasets/Data-Science-Nigeria/dsn-2026-lid-train.

First seen by Oriel on .

AI summary

The DSN 2026 training dataset provides 29,126 audio clips for identifying 12 Nigerian languages or other speech, with substantial class imbalance and speaker-separated validation.

  • It contains 24,774 training clips and 4,352 validation clips, stored in Parquet as 16 kHz mono FLAC audio lasting 2 to 30 seconds.
  • Labels use ISO 639-3 codes; some languages have under one hour of training audio, while “other” has the most at 12.98 hours.
  • Speakers are kept within one split where speaker information exists, and clips are deduplicated and segmented at quiet points when recordings exceed 30 seconds.
  • The dataset combines sources with different license terms; users must consult LICENSE.md and preserve source attributions.
  • The data is intended for language identification, has no transcripts, and includes read recordings that may not reflect everyday speech. The hidden test set is excluded.

#speech recognition #language identification #nigerian languages #audio datasets

Generated by gpt-5.6-sol. Check the original for details.

Metric trends

Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained

Likes

Total likes reported by the source.

  • Hugging Face Datasets: 24, latest observation ; +1 over observed period since

Downloads (30d)

Rolling 30-day downloads. A change in this value is not new downloads during the observed period.

  • Hugging Face Datasets: 615, latest observation ; +45 over observed period since

Aliases

SourceExternal IDURL
Hugging Face DatasetsData-Science-Nigeria/dsn-2026-lid-trainhttps://huggingface.co/datasets/Data-Science-Nigeria/dsn-2026-lid-train Read discussion