llasa-multi-1b

A Llasa model by HKUST Audio, from HKUSTAudio/Llasa-1B, packaged for phoonnx, our ONNX text-to-speech library.

Model details

Architecture	Llasa
Upstream	HKUSTAudio/Llasa-1B
Author	HKUST Audio
Languages	en-US, zh-CN
Text input	Unicode codepoints of the text

How to use

pip install --pre phoonnx

import wave
from phoonnx.model_manager import TTSModelManager

manager = TTSModelManager()
manager.load()
manager.merge_default_voices()

voice = manager.voices["llasa/HKUST/en/en_female_a"].load()

with wave.open("output.wav", "wb") as wav_file:
    text = "..."  # a sentence in en-US
    voice.synthesize_wav(text, wav_file)

In OVOS, phoonnx installs the ovos-tts-plugin-phoonnx plugin. Select the voice in mycroft.conf:

{"tts": {"module": "ovos-tts-plugin-phoonnx", "ovos-tts-plugin-phoonnx": {"voice": "llasa/HKUST/en/en_female_a"}}}

Text front end

phoonnx hands the text to scriptconv, our text conversion library, which normalises it and turns it into the model's input:

• unicode: UnicodeCodepointPhonemizer from scriptconv.phonemizers.base, in scriptconv itself.

phoonnx voices

• llasa/HKUST/en/en_female_a
• llasa/HKUST/en/en_male_a
• llasa/HKUST/zh/zh_female_a
• llasa/HKUST/zh/zh_male_a

Licence

The upstream declares the licence cc-by-nc-4.0 in the model card of HKUSTAudio/Llasa-1B.

Related projects

• phoonnx: runs this model, in Python and in OVOS.
• scriptconv: text normalisation, script conversion and phonemizers for phoonnx.
• The Llasa collection: every Llasa model packaged for phoonnx.