ZeroTTS
Copyright (c) 2026 ZeroWeight AI
Licensed under the MIT License (see LICENSE).

------------------------------------------------------------------------------
This product redistributes third-party components.
------------------------------------------------------------------------------

MOSS-Audio-Tokenizer-Nano
Copyright the OpenMOSS Team.
Licensed under the Apache License, Version 2.0.
Full text: LICENSES/MOSS-Audio-Tokenizer-Apache-2.0.txt

The ZeroTTS weights repository (huggingface.co/zeroweight-ai/ZeroTTS) bundles
the ONNX **decoder** graphs of MOSS-Audio-Tokenizer-Nano under onnx/codec/, so
that ZeroTTS has no runtime dependency on an external model repository. The
encoder graphs are not redistributed. The graphs are unmodified except for the
removal of the encoder entries from the accompanying metadata file.

  Upstream: https://github.com/OpenMOSS/MOSS-Audio-Tokenizer
            https://huggingface.co/OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX

------------------------------------------------------------------------------

soe-vinorm
Copyright (c) vinhdq842.
Licensed under the MIT License.
Full text: src/zerotts/text_norm/data/LICENSE.soe-vinorm

src/zerotts/text_norm/vi_normalizer.py adapts soe-vinorm's Vietnamese number,
date, time, version and fraction expansion rules, and vendors its abbreviation
table as data/abbreviations.txt. None of soe-vinorm's machinery is carried over
— that project tags tokens with a CRF and disambiguates acronyms with an ONNX
scorer, both downloaded at runtime; this module is pure stdlib regex.

  Upstream: https://github.com/vinhdq842/soe-vinorm
