metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: transcript
dtype: string
- name: transliteration
dtype: string
- name: translation
dtype: string
- name: language
dtype: string
- name: duration_ms
dtype: int64
splits:
- name: train
num_bytes: 5031890088.468
num_examples: 3727
- name: test
num_bytes: 3798649241.41
num_examples: 1585
download_size: 8055397450
dataset_size: 8830539329.878
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
license: cc-by-nc-sa-4.0
Mixat
Mixat is a dataset of Emirati speech code-mixed with English. The dataset consists of 15 hours of speech derived from two public podcasts featuring native Emirati speakers. The data collection process, annotation, and dataset statistics are described in detail in the Mixat paper. The original dataset only contains the transcriptions with the English code-switching in latin characters. We augment these transcriptions with two additional dimensions: transliterations and translations. If you use this dataset, please cite the following papers:
@inproceedings{,
}
@inproceedings{al-ali-aldarmaki-2024-mixat,
title = "Mixat: A Data Set of Bilingual Emirati-{E}nglish Speech",
author = "Al Ali, Maryam Khalifa and
Aldarmaki, Hanan",
booktitle = "Proceedings of the 3rd Annual Meeting of the Special Interest Group on Under-resourced Languages @ LREC-COLING 2024",
month = may,
year = "2024",
address = "Torino, Italia",
publisher = "ELRA and ICCL",
url = "https://aclanthology.org/2024.sigul-1.26",
pages = "222--226"
}
