Skip to content

Dataset · FBK-MT

EU multilingual speech corpus

FBK-MT/mosel

Up to 950K hours of open speech across 24 official EU languages with automatic transcripts.

What's inside

MOSEL is a collection of open-source speech recordings. It includes 441K hours of unlabeled speech from VoxPopuli and LibriLight with Whisper transcripts. It covers 24 official EU languages, with en, bg, hr, cs, da, nl, and others listed in the metadata.

What you can do with it

  • Train a multilingual ASR model like Whisper
  • Fine-tune TTS on speech across various EU languages
  • Evaluate transcription quality on low-resource languages
Tasks
Automatic speech recognition, Text → speech
Languages
English, Bulgarian, Croatian, Czech, Danish, Dutch, Estonian, Finnish and 4 more
Size
1M–10M rows
Downloads this month
8K
Likes
94

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("FBK-MT/mosel")