README.md

metadata

title: TuniSpeech-AI
emoji: 🗣️
colorFrom: indigo
colorTo: blue
sdk: static
pinned: false
license: afl-3.0

🎙️ TuniSpeech-AI

TuniSpeech-AI is a Tunisian research initiative dedicated to advancing open speech and language technologies for Tunisian Arabic (Derja).
Our mission is to build, share, and promote inclusive and high-quality resources for low-resource Arabic dialects through open datasets, reproducible benchmarks, and model development.

---

Current Projects

---

Research Focus

• Automatic Speech Recognition (ASR) for Tunisian dialect.
• Corpus creation and normalization for low-resource Arabic varieties.
• Domain adaptation (broadcast, conversational, musical speech).
• Dialect-aware language modeling and code-switch handling.
• Evaluation frameworks for underrepresented languages.

---

Recent Publication

---

License & Access

All datasets and models released under TuniSpeech-AI are made available for academic and non-commercial research use.
For inquiries regarding extended permissions or collaborations, please contact us directly.