Tts Arabic Onnx
This is a model card for the GitHub repo ttsarabic
Model source
- Kind
- Other
- File size
- 8.03–179 MB
- Version
- vRevision ab2743a
- Updated
- Sep 25, 2026
- Downloads
- 0
- Source
- Hugging Face
Source description
This is a model card for the GitHub repo tts_arabic
Arabic TTS models (FastPitch, MixerTTS) from the tts-arabic-pytorch repo in the ONNX format — usable as a Python package for offline speech synthesis.
Audio samples can be found here.
Install with
pip install git+https://github.com/nipponjo/tts_arabic.git
Examples
# %%
from tts_arabic import tts
# %%
text = "السَّلامُ عَلَيكُم يَا صَدِيقِي."
wave = tts(text, speaker=2, pace=0.9, play=True)
# %% Buckwalter transliteration
text = ">ls~alAmu Ealaykum yA Sadiyqiy."
wave = tts(text, speaker=0, play=True)
# %% Unvocalized input
text_unvoc = "القهوة مشروب يعد من بذور البن المحمصة"
wave = tts(text_unvoc, play=True, vowelizer='catt_eo')
Pretrained models
| Model | Model ID | Type | #params | Paper | Output |
|---|---|---|---|---|---|
| FastPitch | fastpitch | Text->Mel | 46.3M | arxiv | Mel (80 bins) |
| MixerTTS | mixer128 | Text->Mel | 2.9M | arxiv | Mel (80 bins) |
| MixerTTS | mixer80 | Text->Mel | 1.5M | arxiv | Mel (80 bins) |
| HiFi-GAN | hifigan | Vocoder | 13.9M | arxiv | Wave (22.05kHz) |
| Vocos | vocos | Vocoder | 13.4M | arxiv | Wave (22.05kHz) |
| Vocos | vocos44 | Vocoder | 14.0M | arxiv | Wave (44.1kHz) |
The sequence of transformations is as follows:
Text → Phonemizer → Phonemes → Tokenizer → Token Ids → Text->Mel model → Mel spectrogram → Vocoder model → Wave
The Text->Mel models map token ids to mel frames. All models use the 80 bin configuration proposed by HiFi-GAN. This mel spectrogram contains frequencies up to 8kHz. The vocoder models map the mel spectrogram to a waveform. The vocoders with vocoder_id hifigan and vocos artificially extend the bandwidth to 11025Hz, and vocos44 to 22050Hz. Samples for comparing the models can be found here.
Manuscript
More information about how the models were trained can be found in the manuscript Arabic TTS with FastPitch: Reproducible Baselines, Adversarial Training, and Oversmoothing Analysis (arXiv | ResearchGate).
TTS options
from tts_arabic import tts
text = "السَّلامُ عَلَيكُم يَا صَدِيقِي."
wave = tts(
text, # input text
speaker = 1, # speaker id; choose between 0,1,2,3
pace = 1, # speaker pace
denoise = 0.005, # vocoder denoiser strength
volume = 0.9, # Max amplitude (between 0 and 1)
play = True, # play audio?
pitch_mul = 1, # pitch multiplier
pitch_add = 0, # pitch offset
vowelizer = None, # vowelizer model
model_id = 'fastpitch', # Model ID for Text->Mel model
vocoder_id = 'hifigan', # Model ID for vocoder model
cuda = None, # Optional; CUDA device index
save_to = './test.wav', # Optionally; save audio WAV file
bits_per_sample = 32, # when save_to is specified (8, 16 or 32 bits)
)
Vowelizer models
| Model | Model ID | Paper | Repo | Architecture |
|---|---|---|---|---|
| CATT | catt_eo | arxiv | github | Transformer Encoder |
| Shakkelha | shakkelha | arxiv | github | Bi-LSTM |
| Shakkala | shakkala | - | github | Bi-LSTM |
References
The vocoder vocos44 was converted from (patriotyk/vocos-mel-hifigan-compat-44100khz).
The vowelizer catt_eo was converted from https://github.com/abjadai/catt/releases/tag/v2 best_eo_mlm_ns_epoch_193.pt (License: Apache-2.0)
DALL·E 2025-03-14 18 56 01 - A surreal digital painting of a camel in a vast desert, with a futuristic speaker embedded in its mouth, symbolizing text-to-speech technology The ca
Sources
1 source
Sources
1 sourceVerified Sep 6
Model artifacts
10 artifacts
Model artifacts
10 artifactsSource excerpts
2 excerpts
Source excerpts
2 excerpts--- language: - ar tags: - ONNX pipeline_tag: text-to-speech datasets: - halabi2016/arabic_speech_corpus --- This is a model card for the GitHub repo [tts_arabic](https://github.com/nipponjo/tts_arabic) Arabic TTS models (FastPitch, MixerTTS) from the [tts-arabic-pytorch](https://github.com/nipponjo/tts-arabic-pytorch) repo in the ONNX format — usable as a Python package for offline speech synthesis. Audio samples can be found [here](https://nipponjo.github.io/tts-arabic-speakers). **Install with** ``` pip install git+https://github.com/nipponjo/tts_arabic.git ``` **Examples** ```python # %% from tts_arabic import tts # %% text = "السَّلامُ عَلَيكُم يَا صَدِيقِي." wave = tts(text, speaker=2, pace=0.9, play=True) # %% Buckwalter transliteration text = ">ls~alAmu Ealaykum yA Sadiyqiy." wave = tts(text, speaker=0, play=True) # %% Unvocalized input text_unvoc = "القهوة مشروب يعد من بذور البن المحمصة" wave = tts(text_unvoc, play=True, vowelizer='catt_eo') ``` **Pretrained models** |Model|Model ID|Type|#params|Paper|Output| |-------|---|---|------|----|----| |FastPitch|fastpitch|Text->Mel|46.3M|[arxiv](https://arxiv.org/abs/2006.06873)|Mel (80 bins)| |MixerTTS|mixer128|Text->Mel|2.9M|[arxiv](https://arxiv.org/abs/2110.03584)|Mel (80 bins)| |MixerTTS|mixer80|Text->Mel|1.5M|[arxiv](https://arxiv.org/abs/2110.03584)|Mel (80 bins)| |HiFi-GAN|hifigan|Vocoder|13.9M|[arxiv](https://arxiv.org/abs/2010.05646)|Wave (22.05kHz)| |Vocos|vocos|Vocoder|13.4M|[arxiv](https://arxiv.org/abs/2306.00814)|Wave (22.05kHz)| |Vocos|vocos44|Vocoder|14.0M|[arxiv](https://arxiv.org/abs/2306.00814)|Wave (44.1kHz)| The sequence of transformations is as follows: *Text* → Phonemizer → *Phonemes* → Tokenizer → *Token Ids* → **Text->Mel** model → *Mel spectrogram* → **Vocoder** model → *Wave* The `Text->Mel` models map token ids to mel frames. All models use the 80 bin configuration proposed by [HiFi-GAN](https://github.com/jik876/hifi-gan). This mel spectrogram contains frequencies up to 8kHz. The `vocoder` models map the mel spectrogram to a waveform. The vocoders with `vocoder_id` `hifigan` and `vocos` artificially extend the bandwidth to 11025Hz, and `vocos44` to 22050Hz. Samples for comparing the models can be found [here](https://nipponjo.github.io/tts-arabic-speakers/#models-cmp). **Manuscript** More information about how the models...
Source context: 0 downloads · 2 likes · Pipeline text-to-speech · Repo nipponjo/tts-arabic-onnx
