Darija LID Benchmark V1: Moroccan Arabic Dialect Identification Dataset

A curated dataset for binary text classification focused on identifying Moroccan Darija (Arabic script or Arabizi) versus other dialects, Modern Standard Arabic, or languages like English. The dataset was built from multiple public sources, cleaned, and taxonomized by author atlasia. It was last updated on 2026-06-07.

Use Cases

• Train a binary classifier to detect Moroccan Darija text based on the described classification task.
• Benchmark dialect identification models on a cleaned and taxonomized dataset built from multiple sources.
• Study the linguistic features distinguishing Moroccan Arabic from other Arabic dialects or Modern Standard Arabic.

Strengths

• Dataset is specifically curated, cleaned, and taxonomized for the dialect identification task.
• Supports binary classification between Moroccan Darija and other dialects/MSA/English.

Limitations

• Row count, file formats, and column-level documentation beyond 'src' and 'text' are unknown.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Multiple public sources, aggregated by atlasia on Hugging Face.

Collection Method

Curated, cleaned, and taxonomized from source datasets.

Freshness

Last updated 2026-06-07 15:16:09; freshness should be verified.

Geography

Focus on Moroccan Darija (Moroccan Arabic).