Support:
- ✅ multi-speaker TTS
- ❓ single-speaker TTS (not tested)
- ❌ Voice Conversion
- ❌ HuBERT-VITS
- ❌ W2V2-VITS
See the conversion script:
ONNXVITS_to_onnx.py
See the inference notebook:
ONNXVITS_inference.ipynb
Run MoeGoe.exe
Path of a VITS model: path\to\model.pth
Path of a config file: path\to\config.json
INFO:root:Loaded checkpoint 'path\to\model.pth' (iteration XXX)
TTS or VC? (t/v):t
Text to read: こんにちは。
ID Speaker
0 XXXX
1 XXXX
2 XXXX
Speaker ID: 0
Path to save: path\to\demo.wav
Successfully saved!
TTS or VC? (t/v):v
Path of an audio file to convert:
path\to\origin.wav
ID Speaker
0 XXXX
1 XXXX
2 XXXX
Original speaker ID: 0
Target speaker ID: 6
Path to save: path\to\demo.wav
Successfully saved!
Path of a hubert-soft model: path\to\hubert-soft.pt
Path of an audio file to convert:
path\to\origin.wav
ID Speaker
0 XXXX
1 XXXX
2 XXXX
Target speaker ID: 6
Path to save: path\to\demo.wav
Successfully saved!
Path of a w2v2 dimensional emotion model: path\to\model.onnx
TTS or VC? (t/v):t
Text to read: こんにちは。
ID Speaker
0 XXXX
1 XXXX
2 XXXX
Speaker ID: 0
Path of an emotion reference: path\to\reference.wav
Path to save: path\to\demo.wav
Successfully saved!