Spaces:
Running on Zero
Running on Zero
File size: 1,252 Bytes
21e8a45 9528c5e 21e8a45 789e868 21e8a45 9528c5e 21e8a45 9528c5e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 | ---
title: KasbahTTS V0 — Algerian Dardja TTS
emoji: 🎙️
colorFrom: red
colorTo: green
sdk: gradio
sdk_version: 6.20.0
app_file: app.py
pinned: false
license: mit
short_description: Zero-shot Algerian Dardja (الدارجة) text-to-speech
models:
- MenaVoice/KasbahTTS-V0
tags:
- text-to-speech
- arabic
- algerian-arabic
- dardja
- f5-tts
- voice-cloning
---
# KasbahTTS V0 — Algerian Dardja TTS
Interactive demo for [`MenaVoice/KasbahTTS-V0`](https://huggingface.co/MenaVoice/KasbahTTS-V0),
the first open-source text-to-speech model built for **Algerian Dardja** (الدارجة الجزائرية).
Give it a few seconds of any voice and some Dardja text, and it speaks that text in that voice.
## How it works
Built on **F5-TTS** (DiT-based flow matching), fine-tuned from **Habibi-TTS** on real Algerian
conversational speech. Weights are pulled at runtime from the model repo, so this Space does not
carry its own copy.
## Limitations
- Arabic script only — no French code-switching
- No digits — spell numbers out (`ثلاثة`, not `3`)
- No diacritics (تشكيل)
- May occasionally repeat words; try `nfe_step=64` or shorter segments
---
*من القصبة للعالم — From the Kasbah to the world*
|