datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
fongbe-speech-zenodo
Fongbe Speech Dataset (Complete & Tone-Preserved)
Dataset Summary
This dataset is a unified, high-quality collection of Fongbe speech data, specifically curated to preserve the linguistic integrity of this tonal language. It acts as a complete, unsegmented, and tone-accurate assembly of the Fongbe Continuous Speech Recognition corpora, merging:
The foundational ALFFA Project data (Train/Test splits, 2016).
The expanded Zenodo release (Validation split, 2022).… See the full description on the dataset page: https://huggingface.co/datasets/Professor/fongbe-speech-zenodo.fongbe-hausa-asr-dataset
Fongbe-Hausa ASR Dataset (Semi-Supervised)
This dataset provides ~6,770 audio-transcription pairs for Fongbe (fon) and Hausa (hau). It was created using a semi-supervised pipeline to convert long-form video content into a training-ready format for Automatic Speech Recognition (ASR).
Dataset Details
Total Examples: 6,770
Audio Format: WAV (16kHz, Mono)
Languages: Fongbe (Benin), Hausa (Nigeria/West Africa)
Annotation: Semi-supervised (Machine-generated labels)
License:… See the full description on the dataset page: https://huggingface.co/datasets/Professor/fongbe-hausa-asr-dataset.fongbe-speechmece-fongbe-corpus
MƐCE — Corpus d'instruction Fongbe (Tune_Pigier)
Corpus d'instruction / conversation centré sur le Fongbe (fon), utilisé pour
fine-tuner l'assistant vocal MƐCE (mémoire de fin d'études, École PIGIER Bénin).
Contenu
Format : ChatML — chaque ligne JSON est un objet {"messages": [...]} avec des
rôles system / user / assistant.
Taille : ~140 000 exemples — train.jsonl (126 725) + eval.jsonl (14 089).
Langues : fon (principal, avec tons/diacritiques), fr, en.… See the full description on the dataset page: https://huggingface.co/datasets/CapitainVigs/mece-fongbe-corpus.fongbe-whisperfongbe-speech-dataset-female
Dataset Card for Dataset Name
This dataset card aims to be a base template for new datasets. It has been generated using this raw template.
Dataset Details
Dataset Description
Curated by: [More Information Needed]
Funded by [optional]: Fréjus LALEYE
Shared by [optional]: Fréjus LALEYE
Language(s) (NLP): Fongbe
License: [More Information Needed]
Dataset Sources [optional]
Repository: https://github.com/laleye/pyFongbe
Paper [optional]:… See the full description on the dataset page: https://huggingface.co/datasets/beethogedeon/fongbe-speech-dataset-female.fongbefongbe-guide-lexique
Lexique Fongbe - Guide du Stagiaire (Peace Corps Benin)
Description
Ce dataset contient 1784 entrees lexicales et phrases en fongbe (Fon),
avec traduction francaise, extraites des lecons 5 a 14 du "Guide du Stagiaire
en Fon", un manuel d'apprentissage linguistique developpe pour les volontaires
du Peace Corps au Benin.
Chaque entree est categorisee par lecon et theme (marche, sante, orientation,
activites quotidiennes, saisons, invitations, sentiments, urgences).… See the full description on the dataset page: https://huggingface.co/datasets/wachinou/fongbe-guide-lexique.dataset_fongbefrench-fongbe-corpus
French-Fongbe Parallel Corpus (Fon-Dataset-Generator)
Description
Ce dataset est un corpus parallèle Français-Fongbe (Bénin) généré par IA et structuré pour l'entraînement de modèles de langue (LLM Fine-tuning). Il couvre divers aspects de la vie quotidienne : santé, commerce, émotions, proverbes, et interactions sociales.
Structure des données
Le dataset est au format JSONL. Chaque entrée contient :
messages : Un format compatible avec le fine-tuning… See the full description on the dataset page: https://huggingface.co/datasets/Shads229/french-fongbe-corpus.fongbe_datasetfongbe-speech-dataset-malefongbe-asrOriginal dataset at this repo
We transformed the original repo to take into account the waveform values directly in the csv.
Using IPython.diplay module, you can load an audio by doing:
import pandas as pd
from IPython.display import Audio, display
train = pd.read_csv("train.csv")
sample = train.sample(1).values[0]
print(f"Text: {sample[2]}")
display(Audio(sample[3], rate=16000, autoplay=True))
Text: alin ɔ ɖo xwe tεntin
Audio :