Multilingual LibriSpeech (MLS)

Introduced by Pratap et al. in MLS: A Large-Scale Multilingual Dataset for Speech Research

Multilingual LibriSpeech is a large multilingual corpus suitable for speech research. The dataset is derived from read audiobooks from LibriVox and consists of 8 languages - English, German, Dutch, Spanish, French, Italian, Portuguese, Polish. It includes about 44.5K hours of English and a total of about 6K hours for other languages.

Source: MLS: A Large-Scale Multilingual Dataset for Speech Research

Homepage

Benchmarks

Add a new result Link an existing benchmark

Task	Dataset Variant	Best Model
Speech Recognition	Multilingual LibriSpeech	stt_es_conformer_transducer_large
Automatic Speech Recognition	Multilingual LibriSpeech	openai/whisper-medium
Automatic Speech Recognition	facebook/multilingual_librispeech French	Whisper largeV2 French MLS
Automatic Speech Recognition	facebook/multilingual_librispeech spanish	Whisper largeV2 Spanish MLS
Automatic Speech Recognition	Multilingual LibriSpeech (MLS)	Fine-tuned wav2vec2-FR-7K-large model for ASR in French
Automatic Speech Recognition	facebook/multilingual_librispeech italian	Whisper largeV2 Italian MLS
Automatic Speech Recognition	facebook/multilingual_librispeech german	Whisper largeV2 German MLS
Speech Recognition	facebook/multilingual_librispeech german	TDT 0-4