---
dataset_info:
  features:
  - name: audio
    dtype: audio
  - name: transcription
    dtype: string
  splits:
  - name: train
    num_bytes: 5042141880.712
    num_examples: 60677
  - name: test
    num_bytes: 453476003.856
    num_examples: 5164
  download_size: 3269805000
  dataset_size: 5495617884.568
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---

This repo contains the dataset presented by Almeman et al., 2013. If you use this dataset, please cite the authors.

```
@INPROCEEDINGS{6487288,
  author={Almeman, Khalid and Lee, Mark and Almiman, Ali Abdulrahman},
  booktitle={2013 1st International Conference on Communications, Signal Processing, and their Applications (ICCSPA)}, 
  title={Multi dialect Arabic speech parallel corpora}, 
  year={2013},
  volume={},
  number={},
  pages={1-6},
  keywords={Speech;Training;Microphones;Cities and towns;Speech recognition;Receivers;NIST;Multi-Dialect;Speech Corpora;Parallel;Arabic Dialects},
  doi={10.1109/ICCSPA.2013.6487288}}
```