UM6P-EMINES/Awesome-Darija-Arabic-NLP-Resources

Awesome Darija Arabic NLP Resources # Awesome Darija Arabic NLP Resources A curated list of resources for Natural Language Processing (NLP) in Moroccan Darija Arabic. ## 📋 Contents - Datasets - Text Datasets - Speech Datasets - Multimodal Datasets - Models - Language Models - Translation Models - Speech Models - NER & Classification Models - Summarization Models - Frameworks & Tools - Benchmarks - General Language Understanding - Translation Benchmarks - Summarization Benchmarks - LLM Evaluation Frameworks - Tools & Libraries - Learning Resources - Research Papers - Contributing - License ## 📊 Datasets ### Text Datasets #### Translation & Lexical Resources | Dataset | Description | Size | Year | Link | |---------|-------------|------|------|------| | **Darija Open Dataset (DODa)** | The largest open-source collaborative project for Darija-English translation. Includes words with different spellings, verb-to-noun and masculine-to-feminine correspondences, conjugations, and translated sentences. Available in both Latin and Arabic alphabets. | ~150,000 entries | 2021 | GitHub / Website | | **Moroccan Dialect Darija Open Dataset** | Open-source contributions in Darija. | >13K entries | 2021 | Website | | **darija_english** | Darija-English pairs for translation tasks. | 10k-45.1k rows | - | Hugging Face | | **Moroccan Darija Wikipedia dataset** | Dataset derived from Moroccan Darija Wikipedia. | - | - | Hugging Face | #### News & Media | Dataset | Description | Size | Year | Link | |---------|-------------|------|------|------| | **Goud.ma** | News dataset for summarization in Darija. | 158k articles | 2022 | GitHub / Hugging Face | | **MNAD (Moroccan News Articles Dataset)** | Collection of documents from Moroccan news websites. | 418,563 documents | 2021 | Kaggle / Hugging Face | #### Sentiment Analysis & Content Classification | Dataset | Description | Size | Year | Link | |---------|-------------|------|------|------| | **Moroccan Arabic Sentiment Analysis Corpus** | Twitter dataset …