README.md

metadata

dataset_info:
  - config_name: Algeria
    features:
      - name: ID
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: transcription
        dtype: string
    splits:
      - name: validation
        num_bytes: 212803455
        num_examples: 1600
      - name: train
        num_bytes: 220829707
        num_examples: 1600
    download_size: 433384710
    dataset_size: 433633162
  - config_name: Egypt
    features:
      - name: ID
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: transcription
        dtype: string
    splits:
      - name: validation
        num_bytes: 213453522
        num_examples: 1600
      - name: train
        num_bytes: 232133794
        num_examples: 1600
    download_size: 445027125
    dataset_size: 445587316
  - config_name: Jordan
    features:
      - name: ID
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: transcription
        dtype: string
    splits:
      - name: validation
        num_bytes: 218000219
        num_examples: 1600
      - name: train
        num_bytes: 222851937
        num_examples: 1600
    download_size: 439912124
    dataset_size: 440852156
  - config_name: Mauritania
    features:
      - name: ID
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: transcription
        dtype: string
    splits:
      - name: validation
        num_bytes: 188544149
        num_examples: 1600
      - name: train
        num_bytes: 191847587
        num_examples: 1600
    download_size: 379825993
    dataset_size: 380391736
  - config_name: Morocco
    features:
      - name: ID
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: transcription
        dtype: string
    splits:
      - name: validation
        num_bytes: 192772707
        num_examples: 1600
      - name: train
        num_bytes: 185043632
        num_examples: 1600
    download_size: 377189001
    dataset_size: 377816339
  - config_name: Palestine
    features:
      - name: ID
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: transcription
        dtype: string
    splits:
      - name: validation
        num_bytes: 278026988
        num_examples: 1600
      - name: train
        num_bytes: 280740886
        num_examples: 1600
    download_size: 558565234
    dataset_size: 558767874
  - config_name: UAE
    features:
      - name: ID
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: transcription
        dtype: string
    splits:
      - name: validation
        num_bytes: 214892944
        num_examples: 1600
      - name: train
        num_bytes: 216164999
        num_examples: 1600
    download_size: 428864450
    dataset_size: 431057943
  - config_name: Yemen
    features:
      - name: ID
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: transcription
        dtype: string
    splits:
      - name: validation
        num_bytes: 242916166
        num_examples: 1600
      - name: train
        num_bytes: 231792726
        num_examples: 1600
    download_size: 472439578
    dataset_size: 474708892
configs:
  - config_name: Algeria
    data_files:
      - split: validation
        path: Algeria/validation-*
      - split: train
        path: Algeria/train-*
  - config_name: Egypt
    data_files:
      - split: validation
        path: Egypt/validation-*
      - split: train
        path: Egypt/train-*
  - config_name: Jordan
    data_files:
      - split: validation
        path: Jordan/validation-*
      - split: train
        path: Jordan/train-*
  - config_name: Mauritania
    data_files:
      - split: validation
        path: Mauritania/validation-*
      - split: train
        path: Mauritania/train-*
  - config_name: Morocco
    data_files:
      - split: validation
        path: Morocco/validation-*
      - split: train
        path: Morocco/train-*
  - config_name: Palestine
    data_files:
      - split: validation
        path: Palestine/validation-*
      - split: train
        path: Palestine/train-*
  - config_name: UAE
    data_files:
      - split: validation
        path: UAE/validation-*
      - split: train
        path: UAE/train-*
  - config_name: Yemen
    data_files:
      - split: validation
        path: Yemen/validation-*
      - split: train
        path: Yemen/train-*
license: bigscience-openrail-m
task_categories:
  - automatic-speech-recognition
language:
  - ar
tags:
  - ASR
  - Arabic
  - NLP
  - Speech
  - ArabicNLP
  - ArabicSpeech
  - ArabicDialects
  - Dialects

Dataset Summary

This datasets is related to NADI-2015 Subtask-2: Multidialectal Arabic ASR.

This subtask aims to develop ASR systems that transcribe Arabic speech across dialects. This subtask is hosted on Codabench. Participants must handle phonetic and dialectal variation using the provided training and validation data. A private test set will be released soon. Models can be evaluated zero-shot on validation data or fine-tuned using the training set. For the SER subtask, participants will evaluate directly on the test set without access to training or validation data.

---

Access

Please register for the shared task.

• Register here: https://forms.gle/WHsyFMtyaewufN7E8
• Codabench → https://www.codabench.org/competitions/8587
• Website → https://nadi.dlnlp.ai/2025/

---

Contact & Links

• 🌐 Website: https://nadi.dlnlp.ai/2025/
• 📝 Registration Form: https://forms.gle/WHsyFMtyaewufN7E8
• 📣 Google Group: https://groups.google.com/u/4/g/nadi-shared-task-2025
• 📩 Email: [email protected]

---

Citation

Kindly cite the following paper when using this dataset:

@article{talafha2024casablanca,
  title={Casablanca: Data and Models for Multidialectal Arabic Speech Recognition},
  author={Talafha, Bashar and Kadaoui, Karima and Magdy, Samar Mohamed and Habiboullah, Mariem and Chafei, Chafei Mohamed and El-Shangiti, Ahmed Oumar and Zayed, Hiba and Alhamouri, Rahaf and Assi, Rwaa and Alraeesi, Aisha and others},
  journal={arXiv preprint arXiv:2410.04527},
  year={2024}
}

License: BigScience OpenRAIL-M

This dataset is released under the BigScience OpenRAIL-M license and is intended for academic research and non-commercial use only. Commercial use, redistribution, or integration into products is prohibited without prior written permission.