Arabic Dialect Classifier (5-Class Identification Model)

License: MIT Transformers

📌 Model Overview

Fine-tuned transformer model specializing in identifying 5 primary Arabic dialect clusters:

  • EGY: Egyptian Arabic
  • GLF: Gulf Arabic
  • LEV: Levantine Arabic
  • MAG: Maghrebi Arabic
  • MSA: Modern Standard Arabic

📊 Evaluation & Accuracies

Tested against the Yousef1258852/arabic-dialect-eval-1k benchmark:

Dialect Class Precision Recall F1 Score
Egyptian (EGY) 0.91 0.90 0.905
Gulf (GLF) 0.88 0.87 0.875
Levantine (LEV) 0.89 0.89 0.890
Maghrebi (MAG) 0.86 0.87 0.865
Modern Standard (MSA) 0.94 0.93 0.935
Macro Average 0.896 0.892 0.894

💻 Usage Example

from transformers import pipeline

classifier = pipeline("text-classification", model="Yousef1258852/arabic-dialect-classifier")
print(classifier("واش راك خويا لاباس عليك كلشي مزيان"))
# [{'label': 'MAG', 'score': 0.978}]

👤 Author & Maintainer

Yousef Khaled
Senior AI Systems Architect

Downloads last month
12
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Evaluation results