Qwen3-TTS GGUF (for qwen3-tts.cpp)

This repository provides GGUF-converted weights of the Qwen3-TTS model for use with qwen3-tts.cpp.

Purpose

The upstream project qwen3-tts.cpp loads GGUF weights, but the official model is distributed in the original framework format.
To reduce the user burden of converting weights locally, this repo publishes pre-converted GGUF files.

• Upstream runtime: https://github.com/predict-woo/qwen3-tts.cpp
• My customized fork (target use): https://github.com/endo5501/qwen3-tts.cpp
• Original model source: https://github.com/QwenLM/Qwen3-TTS
• Original model on Hugging Face: https://huggingface.co/Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice

What’s included

• GGUF files converted from the original Qwen3-TTS model weights (no architectural changes intended)
• (Optional) conversion logs / scripts (if you decide to publish them)

License

The original Qwen3-TTS model weights are released under the Apache License 2.0.
This repository redistributes the weights only in a converted format (GGUF).

• License: Apache-2.0 (see LICENSE)
• Upstream attribution: QwenLM / Qwen3-TTS
• Changes: converted to GGUF format for compatibility with qwen3-tts.cpp

If the upstream repository provides a NOTICE file, this repo includes it as well (see NOTICE).

Disclaimer

This repository is not affiliated with or endorsed by QwenLM.
All trademarks and model names belong to their respective owners.