README.md

metadata

datasets:
  - Thomcles/Czech-Speech-Monospeaker
language:
  - cs
base_model:
  - rhasspy/piper-voices
pipeline_tag: text-to-speech

🇨🇿 Piper TTS Czech 🇨🇿

• Base model : https://huggingface.co/rhasspy/piper-voices/tree/main/cs/cs_CZ/jirka/medium

• Finetuning dataset : https://huggingface.co/datasets/Thomcles/Czech-Speech-Monospeaker

• Piper Github : https://github.com/OHF-Voice/piper1-gpl/tree/main

---

Use the model with Python 🐍

Install piper with :

pip install piper-tts

Install medium quality or high quality model :

Medium (recommended) :

!wget -O cs_medium.onnx https://huggingface.co/Thomcles/Piper-TTS-Czech/resolve/main/medium/model.onnx
!wget -O cs_medium.onnx.json https://huggingface.co/Thomcles/Piper-TTS-Czech/resolve/main/medium/model.onnx.json

Python API :

import wave
from piper import PiperVoice

voice = PiperVoice.load("cs_medium.onnx", "cs_medium.onnx.json")
with wave.open("test.wav", "wb") as wav_file:
    voice.synthesize_wav("Dobrý den, vítáme vás v našem testu syntézy řeči.", wav_file)