README.md

metadata

license: mit
dataset_info:
  - config_name: Derja_tunsi
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 230560
        num_examples: 13037
    download_size: 137033
    dataset_size: 230560
  - config_name: HkayetErwi
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 383744
        num_examples: 946
    download_size: 146010
    dataset_size: 383744
  - config_name: TSAC
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 678557
        num_examples: 8416
    download_size: 383230
    dataset_size: 678557
  - config_name: TunBERT
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 5868844
        num_examples: 67186
    download_size: 3459576
    dataset_size: 5868844
  - config_name: TunSwitchCodeSwitching
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 42240964
        num_examples: 394160
    download_size: 23652042
    dataset_size: 42240964
  - config_name: TunSwitchTunisiaOnly
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 40463064
        num_examples: 380546
    download_size: 22647724
    dataset_size: 40463064
  - config_name: Tunisian_Dialectic_English_Derja
    features:
      - name: text
        dtype: string
    splits:
      - name: Tunisian_Dialectic_English_Derja
        num_bytes: 547732666.7098647
        num_examples: 1220712
    download_size: 273966174
    dataset_size: 547732666.7098647
  - config_name: Tweet_TN
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 3258000
        num_examples: 39637
    download_size: 1431965
    dataset_size: 3258000
configs:
  - config_name: Derja_tunsi
    data_files:
      - split: train
        path: Derja_tunsi/train-*
  - config_name: HkayetErwi
    data_files:
      - split: train
        path: HkayetErwi/train-*
  - config_name: TSAC
    data_files:
      - split: train
        path: TSAC/train-*
  - config_name: TunBERT
    data_files:
      - split: train
        path: TunBERT/train-*
  - config_name: TunSwitchCodeSwitching
    data_files:
      - split: train
        path: TunSwitchCodeSwitching/train-*
  - config_name: TunSwitchTunisiaOnly
    data_files:
      - split: train
        path: TunSwitchTunisiaOnly/train-*
  - config_name: Tunisian_Dialectic_English_Derja
    data_files:
      - split: train
        path: Tunisian_Dialectic_English_Derja/train-*
  - config_name: Tweet_TN
    data_files:
      - split: train
        path: Tweet_TN/train-*

Tunisian Derja Dataset (Tunisian Arabic)

This is a collection of Tunisian dialect textual documents for Language Modeling.

It was used for contunial pretraining of Jais LLM model in Tunisian dialect

Languages

The dataset is available in Tunisian Arabic (Derja).

Data Table

subset	Lines
Tunisian_Dialectic_English_Derja	1220712
HkayetErwi	966
Derja_tunsi	13037
TunBERT	67219
TunSwitchCodeSwitching	394163
TunSwitchTunisiaOnly	380546
Tweet_TN	39637
TRCD	145
TSAC	8479
TuDiCOI	3426

Citations

When using the Tunisian Derja Dataset dataset, please cite:

@dataset{linagora2025LLM-tn,
  author = {Wajdi Ghezaiel and Jean-Pierre Lorré},
  title = {Tunisian Derja Dataset},
  year = {2025},
  month = {January},  
  url = {https://huggingface.co/datasets/Wajdi1976/tunisian_derja_dataset}
}

Citations for included datasets:

• LinTO Texual Dataset in Arabic Tunisian

@misc{linagora2024Linto-tn,
  author = {Hedi Naouara and Jérôme Louradour and Jean-Pierre Lorré},
  title = {LinTO Audio and Textual Datasets to Train and Evaluate Automatic Speech Recognition in Tunisian Arabic Dialect},
  year = {2024},
  month = {October},
  note = {Good Data Workshop, AAAI 2025},
  howpublished = {\url{https://huggingface.co/linagora/linto-asr-ar-tn-0.1}},
}

• Tunisian_English_dialectic_Derja

@dataset{Tunisian_English_dialectic_Derja,
  author = {Khaled Bouzaiene},
  title = {Tunisian-English Dialectic Derja Dataset},
  year = {2024},
  url = {https://huggingface.co/datasets/khaled123/Tunisian_English_dialectic_Derja}
}

• TunSwitch

@misc{abdallah2023leveraging,
      title={Leveraging Data Collection and Unsupervised Learning for Code-switched Tunisian Arabic Automatic Speech Recognition}, 
      author={Ahmed Amine Ben Abdallah and Ata Kabboudi and Amir Kanoun and Salah Zaiem},
      year={2023},
      eprint={2309.11327},
      archivePrefix={arXiv},
      primaryClass={eess.AS}
}