datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
ISSAI-KSC2-Structured
KSC2 Structured: Punctuation & Capitalization Restored
Developed by: Inflexion Lab
Based on: ISSAI KSC2 (Kazakh Speech Corpus 2)
License: CC BY 4.0
Dataset Description
This dataset is an enhanced version of the Kazakh Speech Corpus 2 (KSC2). While the original KSC2 dataset provides high-quality audio, the transcripts are provided in plain lowercase without punctuation. This limits the ability of ASR models to learn proper sentence structure.
KSC2 Structured solves this by… See the full description on the dataset page: https://huggingface.co/datasets/InflexionLab/ISSAI-KSC2-Structured.global-french-speech_structured
Zeldeo/global-french-speech_structured
Dataset ASR restructuré depuis SilencioNetwork/global-french-speech
(config=french_canada, split=train).
Nombre d'exemples : 25.
Métadonnées ajoutées : source_dataset, type, langue_accent.
Normalisation texte : aucune.
Colonnes conservées
audio
gender
dialect
emotions
language
location
noise_sources
transcript
age_band
Usage
from datasets import load_dataset
ds =… See the full description on the dataset page: https://huggingface.co/datasets/Zeldeo/global-french-speech_structured.french-conversation_structured
Zeldeo/french-conversation_structured
Dataset ASR restructuré depuis Snit/french-conversation
(config=default, split=train).
Nombre d'exemples : 98.
Métadonnées ajoutées : source_dataset, type, langue_accent.
Normalisation texte : aucune.
Colonnes conservées
audio
transcription
id
part
audio_path
Usage
from datasets import load_dataset
ds = load_dataset("Zeldeo/french-conversation_structured", split="train")
print(ds[0])