Datasets:

antoineedy
/

multilingual_librispeech_fr_punctuated

license: apache-2.0
task_categories:
  - automatic-speech-recognition
language:
  - fr
tags:
  - audio
  - speech
  - multilingual-librispeech
size_categories:
  - 10K<n<100K

Multilingual LibriSpeech French (Punctuated)

This dataset is a converted version of BrunoHays/multilingual_librispeech_fr_punctuated in the new Hugging Face datasets format (Parquet-based, without loading scripts).

Original Dataset

The original dataset contains French speech data from Multilingual LibriSpeech with punctuated transcriptions.

Changes

Converted from old loading script format to new Parquet-based format
Maintains all original features and data
Compatible with the latest Hugging Face datasets library

Usage

from datasets import load_dataset

dataset = load_dataset("antoineedy/multilingual_librispeech_fr_punctuated")

Citation

Please cite the original Multilingual LibriSpeech dataset:

@article{pratap2020mls,
  title={MLS: A Large-Scale Multilingual Dataset for Speech Research},
  author={Pratap, Vineel and Xu, Qiantong and Sriram, Anuroop and Synnaeve, Gabriel and Collobert, Ronan},
  journal={arXiv preprint arXiv:2012.03411},
  year={2020}
}

License

Same license as the original dataset.