Vosk-Zundamon

Connect Vosk (Speech Recognition) and Zundamon (Text-to-speech application, VOICEVOX).

Installation

1. Download & extract vosk-model-small-ja-0.22 from https://alphacephei.com/vosk/models and save it to resource/
2. Download & extract open_jtalk_dic_utf_8-1.11 from http://open-jtalk.sourceforge.net/ and save it to resource/
3. pip install -r requirements.txt
4. Follow the instruction of voicevox_core and install 0.12.0-preview.3 or later.

Usage

python mic.py

Note

GPU may accelerate the speed of translation. See README in voicevox_core and change the arguments of the initialize function.

License

See LICENSE