license: cc0-1.0
language:
- lv
tags:
- piper
- tts
pipeline_tag: text-to-speech
widget:
- src: >-
https://huggingface.co/RaivisDejus/Piper-lv_LV-Aivars-medium/resolve/main/0.wav
example_title: Varavīksne
- src: >-
https://huggingface.co/RaivisDejus/Piper-lv_LV-Aivars-medium/resolve/main/1.wav
example_title: Vilciens
- src: >-
https://huggingface.co/RaivisDejus/Piper-lv_LV-Aivars-medium/resolve/main/2.wav
example_title: Balsu talka
Latvian Piper TTS voice "Aivars"
Voice model is built on audio books of Latvijas Neredzīgo bibliotēka. For privacy reasons the original voices in the recordings have been cloned to a LibriVox voice that you hear in the model and sample recordings. Audio book recordings are used with permission from Latvijas Neredzīgo bibliotēka.
| Voice samples |
 |
|---|---|
| Varavīksne | |
| Vilciens | |
| Balsu talka |
Aivars TTS voice model is licenced under CC0 1.0 Universal (CC0 1.0) Public Domain Dedication. You are free to use it for any purpose, including commercial, no strings attached. However, it would be nice if you would choose to help the latvian open speach data to grow by participating and spreading the word about the Balsu talka project.
For information about use of this voice model see Piper TTS GitHub repository.
Home Assistant integration
To add this voice model to your Home Assistant instance, download model files to /share/piper/ and rename them to aivars.onnx and aivars.onnx.json.
mkdir -p /share/piper/
cd /share/piper/
wget -O aivars.onnx https://huggingface.co/RaivisDejus/Piper-lv_LV-Aivars-medium/resolve/main/lv_LV-aivars-medium.onnx
wget -O aivars.onnx.json https://huggingface.co/RaivisDejus/Piper-lv_LV-Aivars-medium/resolve/main/lv_LV-aivars-medium.onnx.json
Then install Piper addon and configure Voice assistant to use aivars (medium) voice model. Our custom voice will not show up in the voice list in Piper configuration, but it will be available in Voice assistant configuration.
For latvian speech recognition in Home assistant use latvian whisper small model