Granary: Speech Recognition and Translation Dataset in 25 European Languages

TL;DR

Granary is a large-scale, open-source speech dataset for 25 European languages, enabling improved recognition and translation, especially for low-resource languages, through innovative data processing techniques.

Contribution

This work introduces the first extensive open-source speech dataset for recognition and translation in 25 European languages, with novel data augmentation and filtering methods.

Findings

Models trained on Granary data perform comparably with 50% less data.

Enhanced data quality improves speech recognition and translation accuracy.

Pipeline efficiently processes large datasets within hours.

Abstract

Multi-task and multilingual approaches benefit large models, yet speech processing for low-resource languages remains underexplored due to data scarcity. To address this, we present Granary, a large-scale collection of speech datasets for recognition and translation across 25 European languages. This is the first open-source effort at this scale for both transcription and translation. We enhance data quality using a pseudo-labeling pipeline with segmentation, two-pass inference, hallucination filtering, and punctuation restoration. We further generate translation pairs from pseudo-labeled transcriptions using EuroLLM, followed by a data filtration pipeline. Designed for efficiency, our pipeline processes vast amount of data within hours. We assess models trained on processed data by comparing their performance on previously curated datasets for both high- and low-resource languages. Our findings show that these models achieve similar performance using approx. 50% less data. Dataset will be made available at https://hf.co/datasets/nvidia/Granary