---
license: apache-2.0
language:
- ar
pipeline_tag: text-to-speech
library_name: omnivoice
base_model: oddadmix/lahgtna-omnivoice-v2
tags:
- text-to-speech
- tts
- arabic
- egyptian-arabic
- arz
- omnivoice
- voice-cloning
- multi-speaker
---

# Lahgtna-OmniVoice Egyptian — **v3** (multi-voice, non-diacritized)

Multi-voice **Egyptian-Arabic (`arz`)** TTS. v3 continues from v2 on **raw, non-diacritized
text** to fix pronunciation problems caused by MSA diacritics on Egyptian colloquial. It speaks
two trained voices (Eqkawkab + Noselleel) and clones any reference voice zero-shot.
**Recommended model for a multi-voice product** — best multi-voice intelligibility, best on
numbers, and lowest latency (no diacritizer at inference).

## 👪 Model family
| | repo | voices | text | notes |
|---|---|---|---|---|
| base | [lahgtna-omnivoice-v2](https://huggingface.co/oddadmix/lahgtna-omnivoice-v2) | — | diacritized | upstream |
| v1 | [egyptian](https://huggingface.co/ehabnegm/lahgtna-omnivoice-egyptian) | 1 | diacritized | best CER, single voice |
| v2 | [egyptian-v2](https://huggingface.co/ehabnegm/lahgtna-omnivoice-egyptian-v2) | 2 | diacritized | multi-voice |
| **v3** ← you are here | [egyptian-v3](https://huggingface.co/ehabnegm/lahgtna-omnivoice-egyptian-v3) | 2 | **raw** | **recommended** |
| 🎛️ Compare GUI | [Space](https://huggingface.co/spaces/ehabnegm/lahgtna-egyptian-tts-compare) | — | — | 4-model A/B |
| 📚 Data | [eqkawkab](https://huggingface.co/datasets/ehabnegm/eqkawkab-egyptian-tts) · [noselleel](https://huggingface.co/datasets/ehabnegm/noselleel-egyptian-tts) | — | — | ~34h |

## Usage — feed RAW Egyptian text (no diacritics)
```python
import torch, soundfile as sf
from omnivoice.models.omnivoice import OmniVoice
m = OmniVoice.from_pretrained("ehabnegm/lahgtna-omnivoice-egyptian-v3", device_map="cuda", dtype=torch.float16)
audio = m.generate(text="أهلاً بحضرتك، معاك خدمة العملاء، ممكن أساعدك إزاي؟", language="arz",
                   ref_audio="voice.wav", ref_text="نص الكليب المرجعي")
sf.write("out.wav", audio[0], 24000)
```
Verbalize numbers (e.g. `٢٥ → خمسة وعشرين`) and **strip/transliterate English words** before
synthesis (v3 attempts untrained English otherwise).


## ▶️ Run on Colab
Free GPU, step-by-step: **[COLAB.md](./COLAB.md)** · runnable notebook: **[lahgtna_colab.ipynb](./lahgtna_colab.ipynb)** (open it in Colab → Runtime → T4 GPU). Log in with a HF read token (repo is private).

## Training
Combined **~34 h, 10,828 clips** (Eqkawkab ~8h + Noselleel ~26h), **raw + orthography-normalized**
text. Continued from v2 `checkpoint-8000`, 5,000 steps, SDPA, bf16, 1× L4.

## Benchmark (ASR round-trip CER, 13 pure-Arabic sentences — lower = clearer)
| model | CER ↓ | |
|---|---|---|
| v1 | 0.056 | single-voice |
| **v3** | **0.063** | **best multi-voice** |
| v2 | 0.102 | multi-voice, diac |
| base | 0.109 | |
**v3 beats v2 by ~38%** on pure-Arabic — removing the wrong diacritics worked. Full benchmark +
audio: [v2 repo `/benchmark`](https://huggingface.co/ehabnegm/lahgtna-omnivoice-egyptian-v2/tree/main/benchmark).

## 321-word long-form stress test

A full Egyptian-Arabic article (~321 words, ~2:50) synthesized by each model, then transcribed
back with Whisper large-v3 (ASR round-trip CER — lower = stays on-text).

| model | audio | RTF | CER ↓ |
|---|---|---|---|
| **v3** | 173 s | 0.25 | **0.21** — stays coherent ✅ |
| base | 170 s | 0.35 | 0.69 — drifts ⚠️ |
| v1 | 170 s | 0.35 | 0.71 — drifts ⚠️ |
| v2 | 170 s | 0.43 | 0.71 — drifts ⚠️ |

**Finding:** on paragraph-length input the diacritized models (base/v1/v2) lose coherence and
babble; **v3 (non-diacritized) is the only version that stays on-text** — decisive for long-form
and agent replies. Listen: `base.wav`, `v1.wav`, `v2.wav`, `v3.wav`.

Audio: [`long300/`](https://huggingface.co/ehabnegm/lahgtna-omnivoice-egyptian-v2/tree/main/long300).

## Limitations & responsible use
Code-switched English is not trained (strip/transliterate it). Audio derived from public YouTube
channels (@Eqkawkab, @noselleel) — credit the creators, get permission before commercial use, and
don't use the voices to impersonate or mislead.


## Latency — time-to-first-chunk (streaming, raw text)
Dropping the diacritizer makes v3 **~250 ms faster than v2** at every setting:

| num_step | TTFC |
|---|---|
| 32 | 1.82 s |
| 16 | 0.96 s |
| **8 (recommended)** | **0.54 s** |
| 4 (fastest) | **0.32 s** |