CoolFace
Datasetpublic

Harcuracy/google_waxal_asr_challenge

WaxalNLP ASR — Cleaned Subset (Lingala, Shona, Luganda) This dataset is a cleaned, corrected subset of google/WaxalNLP, covering the train and validation splits for three languages: lin_asr — Lingala sna_asr — Shona lug_asr — Luganda The test split from the original dataset is intentionally excluded. What was changed The original transcriptions for these three languages contained a number of errors. A corrected transcription file was applied on top of the… See the full description on the dataset page: https://huggingface.co/datasets/Harcuracy/google_waxal_asr_challenge.

sourceHugging Facecc-by-4.0updated 3mo agoView on Hugging Face
0likes195downloads
5 commits on main
1b53b0e3mo ago

Update README.md

Harcuracy
0d459803mo ago

Upload dataset

Harcuracy
0cbe5273mo ago

Upload dataset

Harcuracy
21885aa3mo ago

Upload dataset

Harcuracy
b1e96f83mo ago

initial commit

Harcuracy