VieNeu-TTS-v2 TTS
A Vietnamese-first, CPU-only model with en-vi code-switching, 7 regional preset voices, and zero-shot cloning.
Wrap your text in SSML tags for precise control:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Tags déi d'gewielt Modell verstinn - klickt fir eng an Ärem Text ze setzen wou se geschitt:
Dëse Modell liest einfache Text, sou datt Inline-Tags ignoréiert ginn. Fir Tag-baséiert Emotiounen, wielt e expressiven Modell wéi Orpheus oder Bark.
Eegen Aussproochen definéieren (Wuert = Aussprooch):
Iwwer VieNeu-TTS-v2
VieNeu-TTS-v2 is a 300M-parameter Vietnamese-first model built on a Qwen3 backbone and trained on more than 10,000 hours of bilingual data. It handles seamless English-Vietnamese code-switching, ships 7 preset voices spanning Northern and Southern accents, and clones a voice instantly from just 3-5 seconds of reference audio. Notably it runs entirely on CPU — using GGUF Q4 inference plus an ONNX audio decoder — with no GPU required, finishing a generation in about 7 seconds. It's purpose-built for Vietnamese content and bilingual en-vi narration, an underserved niche in open TTS.
Bescht fir: Vietnamese content and bilingual en-vi narration
All sichen VieNeu-TTS-v2 StimmenOp ee Bléck
- Entwéckler
- Phạm Nguyễn Ngọc Bảo
- Lizenz
- Apache 2.0
- Tier
- standard
- Geschwindegkeet
- fast
- Sprooche-Klonen
- Ja
- Sproochen
- Vietnamese, English
- Maximal Zeichen
- 1000