ASR_pulaar
17,253 lines of Pulaar speech, CC-BY
Official site consultedAI: datasets
Last verified: 1 October 2026
Fula (Pulaar dialect) speech recognition dataset with transcriptions: 17,253 rows (15,500 train, 1,730 test), about 7.81 GB. It is under CC-BY-4.0 in Parquet format.
Services
- Audio with speaker ID, duration and timestamps
Access requirements
Open access on Hugging Face, no gating, CC-BY-4.0 license.
Sources
- ASR_pulaar (new tab), verified on 1 October 2026Official site consulted