Awras
Toward an ecosystem dedicated to preserving and empowering the Algerian dialect (Darija) through advanced Artificial Intelligence.
Native vs. The Rest
See the difference between standard Arabic LLMs and Awras when handling authentic Algerian Darija.
The Prompt
"واش هي مكونات الشوربة الوهرانية؟"
fanar
"تتكون الشوربة الوهرانية من حمص، فول، عدس، بصل، ثوم، طماطم، توابل، وملح وفلفل حار، مع إمكانية إضافة لحم أو خضروات أخرى حسب الرغبة."
Awras
"الشوربة الوهرانية، يا حصراه على البنة! لازمك لحم غنمي، فريك، طماطم مصبرة، بصل، ثوم، قصبر، كرافس، حمص طايب، وتوابل كيما فلفل كحل، قرفة، وزعفران. وما تنسايش شوية زيت وسمن باش تجي ميدمة وبنينة."
Silma 9B
"أجيبك، المكونات هما الدجاج، البرز، الكرمط، اللوز، الطعمه، البهارات وما لوشو غزل."
* Comparison based on cultural nuance and dialect authenticity.
The Vision
We are building the foundational layer for Algerian AI applications. A complete suite of tools designed to digitize our dialect.
The ultimate goal is to build a AI ecosystem that truly comprehends our language and culture.
Large Language Model fine-tuned on vast datasets of Algerian Darija to understand context, nuance, and cultural references.
Automatic Speech Recognition capable of transcribing spoken Darija with high accuracy, handling various regional accents.
Text-to-Speech synthesis that produces natural-sounding speech with authentic Algerian intonation and pronunciation.
The Process
Our continuous loop of improvement ensures our AI models evolve with the language, becoming more accurate and culturally relevant through each iteration.
Gathering diverse datasets of Algerian Darija from various sources including text, audio, and cultural content to build a comprehensive foundation.
Fine-tuning advanced AI models on the collected data, ensuring they understand context, cultural nuances, and regional variations of Algerian dialect.
Rigorous testing and validation of model performance followed by deployment to production environments for real-world usage.
Continuous learning from user interactions and feedback to refine models, improve accuracy, and adapt to evolving language usage patterns.
Frequently Asked Questions
Find answers to common questions about Awras
Still have questions? Contact us