Logo Lanfrica

kitoDev24/shat

Domaine:

natural language processing

Type de record:

dataset
Créateur:
kit
Hôte:
is a dataset for hassaniya chat with a web editor and trnaslator SHAT — الشَّت Hassaniya Arabic dataset + web editor + translator --- ## What is SHAT? **SHAT** (الشَّت) is an open-source dataset and toolchain for **Hassaniya Arabic** — the dialect of Mauritania and parts of the Sahara. It includes: - **A growing dictionary** (~700+ English↔Hassaniya entries and growing) - **A web-based editor** for managing translations, prompts, and stories - **A grammar-aware translator** that applies real Arabic linguistic rules - **Training data** for building chat models in Hassaniya --- ## What is Hassaniya? **Hassaniya** (حسانية) is the Arabic dialect spoken by millions across: | Region | Notes | |--------|-------| | 🇲🇷 **Mauritania** | Official language, spoken by most of the population | | 🇪🇭 Western Sahara | Widely spoken | | 🇩🇿 Southwest Algeria | Tindouf area | | 🇲🇱 Northern Mali | Azawad region | | 🇳🇪 Western Niger | Border areas | | 🇸🇳 Northern Senegal | River region | Hassaniya preserves many features of classical Arabic that were lost in other dialects, including the full use of Arabic's consonantal root system and gender/number agreement. ### Example | English | Hassaniya (Latin script) | |---------|--------------------------| | Hello | Slm | | How are you? | Ch7alak / Chmassy | | Good night | Leyle zeine | | I don't know | Ane man3rav | | See you later | Nchovak ba3din | --- ## About Mauritania Islamic Republic of Mauritania جمهورية موريتانيا الإسلامية **Mauritania** is a country in **Northwest Africa**, bordered by the Atlantic Ocean, Western Sahara, Algeria, Mali, and Senegal. It is a crossroads between Arab North Africa and Sub-Saharan Africa, and its culture reflects both identities. - **Capital**: Nouakchott (نواكشوط) - **Official language**: Arabic (Hassaniya is the spoken dialect) - **Population**: ~4.5 million - **Known for**: The Sahara desert, the Senegal River, ancient trading cities like Chinguetti and Ouadane --- ## Features ### 📖 Dictionary - **700+ entries** covering greetings, daily life …

Languages