A basic Flutter demo app for Arabic TTS with models from tts_arabic / tts-arabic-pytorch.
The app runs ONNX models using the package onnxruntime_flutter.
Audio samples can be found here.
- (Set up Flutter)
- Download the onnx models from Google drive. By default,
mixer128.onnxandvocos22.onnx(and the optional vowelizershakkelha.onnx) are activated. - Put the onnx files into the
assets/modelsfolder. - (Modify the following lines in
pubspec.yamlandmain.dartto use other models thanmixer128.onnxandvocos22.onnx)
in pubspec.yaml
assets:
- assets/models/mixer128.onnx
- assets/models/vocos22.onnx
- assets/models/shakkelha.onnxin main.dart
_ttsModel!.initSessions(
modelPath: "assets/models/mixer128.onnx",
vocoderPath: "assets/models/vocos22.onnx",
);
| Model | Model ID | Type | #params | File | Paper |
|---|---|---|---|---|---|
| FastPitch | fastpitch | Text->Mel | 46.3M | fp_ms.onnx | arxiv |
| MixerTTS | mixer128 | Text->Mel | 2.9M | mixer128.onnx | arxiv |
| MixerTTS | mixer80 | Text->Mel | 1.5M | mixer80.onnx | arxiv |
| Vocos | vocos | Vocoder | 13.4M | vocos22.onnx | arxiv |
A few resources to get you started if this is your first Flutter project:
For help getting started with Flutter development, view the online documentation, which offers tutorials, samples, guidance on mobile development, and a full API reference.
