README.md

metadata

title: VoTSpeech
emoji: 🗣️
colorFrom: indigo
colorTo: blue
sdk: gradio
sdk_version: 6.30.0
app_file: app.py
short_description: Design a voice from a text description, then speak (zh)
python_version: '3.12'
license: cc-by-nc-4.0
startup_duration_timeout: 30m
models:
  - Ywbn16/VoTSpeech

VoTSpeech — instruction-following voice-design TTS

Demo of VoTSpeech (Decoupling Voice Design from Speech Generation in Instruction-Following TTS, code: ywbn/VoTSpeech). Given text and a natural-language voice description, it designs a continuous voice latent and generates 48 kHz speech in that voice. Built on dots.tts-soar. Trained for Mandarin Chinese.

The dots_tts/ package is the authors' inference code (Apache-2.0). Model weights: CC BY-NC 4.0 — non-commercial use only. Example prompts come from the authors' project page. Generated audio is synthetic speech; do not impersonate identifiable people without permission.