Datasets
Dataset
fleurs-indic-speech-service-errors
by burncaps
Speech service errors on FLEURS, 12 Indian languages Speech service errors on 240 FLEURS clips in 12 Indian languages.
CC-BY-4.0automatic-speech-recognition 24 0n < 1K rows
published 16 Oct 2026
Preview
First 1 rows of default / train, from the Hugging Face dataset viewer.
| what | source | scoring | page | clips |
|---|---|---|---|---|
| "Burncaps speech service (Microsoft MAI-Transcribe, North Europe) on 240 Google FLEURS test clips, 2 | "Google FLEURS, test split, licensed CC-BY-4.0 (https://huggingface.co/datasets/google/fleurs). The | "Words and letters wrong = edit distance after lower-casing and removing brackets and punctuation. w | https://burncaps.app/indian-language-speech-to-text-accuracy | [{"language":"Hindi","code":"hi","fleurs_file":"google/fleurs hi_in test id 1766 file 10011266027513 |
Cite this
CC-BY-4.0textCite
Tags
task_categories:automatic-speech-recognitionlanguage:hilanguage:urlanguage:bnlanguage:talanguage:telanguage:gulanguage:mrlanguage:knlanguage:mllanguage:palanguage:nelanguage:orlicense:cc-by-4.0size_categories:n<1Kformat:jsonmodality:textlibrary:datasetslibrary:pandaslibrary:polarslibrary:mlcroissantregion:us