Datasets

Dataset

Granary-v2

by nvidia

Granary-v2: Multilingual Speech Recognition, Translation and Instruction Data Granary-v2 is NVIDIA's second release of generated labels and instruction pairs for multilingual speech.

OpenMDW-1.1automatic-speech-recognition 262 9

published 16 Sept 2026

Preview

The Hugging Face dataset viewer has no preview for this one.

Cite this

OpenMDW-1.1Cite

Tags

task_categories:automatic-speech-recognitiontask_categories:translationtask_categories:audio-text-to-texttask_categories:question-answeringlanguage:arlanguage:aslanguage:bglanguage:bnlanguage:cslanguage:dalanguage:delanguage:ellanguage:enlanguage:eslanguage:etlanguage:falanguage:filanguage:fillanguage:frlanguage:gulanguage:helanguage:hilanguage:hrlanguage:hulanguage:idlanguage:itlanguage:jalanguage:knlanguage:kolanguage:ltlanguage:lvlanguage:mllanguage:mrlanguage:mtlanguage:nllanguage:orlanguage:palanguage:pllanguage:ptlanguage:ro