Downloads · 30 days
4.2K
47% of all-time downloads
Arm/qwen3-tts-0-6b-custom-voice-mix-precision
qwen3-tts-0-6b-custom-voice-mix-precision is a text-to-speech model from Arm. Use it when you need text read aloud. It is set up for litert. The card lists the license as apache-2.0.
This repository packages a mixed-precision, mixed-framework deployment bundle derived from Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice for multilingual text-to-speech using preset speakers. The bundle uses GGUF, ONNX, and Li…
Downloads · 30 days
4.2K
47% of all-time downloads
All-time downloads
9.1K
Public
Repo size
2.5 GB
Likes
6
Public
Click a slice to open those files.
.onnx1.3 GB · 51%
From the Hugging Face model README
This repository packages a mixed-precision, mixed-framework deployment bundle
derived from
Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice
for multilingual text-to-speech using preset speakers. The bundle uses GGUF,
ONNX, and LiteRT artifacts and supports streaming synthesis at 24 kHz.
Qwen3-TTS-12Hz-0.6B-CustomVoice is a multilingual text-to-speech model that synthesizes speech using a selected preset speaker.
Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoiceqwen3_tts_12hz_0.6b_custom_voiceFollow these steps to run Qwen3-TTS 0.6B Custom Voice on macOS® with Apple Silicon or on Linux® with an Arm®-based CPU, such as a Raspberry Pi® 5.
git clone https://github.com/ARM-software/ComputeFlow.git && cd ComputeFlow
uv sync && source .venv/bin/activate
compute-flow update
compute-flow model install tts:qwen3-tts qwen3-tts-0-6b-custom-voice
compute-flow build native
python -m cli_examples.tts_cli \
-m @manifest/qwen3-tts/qwen3-tts-0-6b-custom-voice/qwen3_tts_custom_voice_en_manifest.json \
-p "Hello, I am Aiden, your personal AI assistant. How can I help?" \
-t 1 \
-o qwen3_tts_output.wav
Quality was assessed using WER on a subset of the Seed-TTS corpus. No intelligibility regression was observed against the FP32 Qwen3-TTS CustomVoice implementation.
Settings: English, Aiden speaker, temperature 0.9, top-k 50, and seed 12345. Audio was converted to mono 16 kHz before ASR scoring.
Performance was measured under the following conditions:
The following methodology and definitions were used:
Compared with the FP32 Qwen3-TTS CustomVoice implementation, the optimized model provides the following uplift.
| Metric | MacBook Pro®: FP32 | MacBook Pro®: Optimized | Uplift | Android™ Vivo X300: FP32 | Android™ Vivo X300: Optimized | Uplift |
|---|---|---|---|---|---|---|
| RTF | 0.839 | 0.376 | 2.23× faster | 3.516 | 0.836 | 4.21× faster |
| End-to-end latency | 3.962 s | 1.684 s | 2.35× faster | 16.877 s | 3.878 s | 4.35× faster |
| Time to first chunk | 0.305 s | 0.139 s | 2.19× faster | 1.240 s | 0.314 s | 3.95× faster |
| Average memory | 4619.3 MiB | 2605.9 MiB | 43.6% lower | 5253.7 MiB | 2883.4 MiB | 45.1% lower |
Generate streaming 24 kHz speech from text using a selected preset speaker.
| Component role | Framework / format |
|---|---|
| Text tokenization | Hugging Face tokenizer |
| Embeddings | ONNX Runtime / ONNX |
| Talker, code prediction, and speech decoding | llama.cpp / GGUF and LiteRT / TFLite |
| Input | Description |
|---|---|
| Text prompt | UTF-8 string |
The model returns synthesized 24 kHz audio in streaming mode.
At runtime, users can select the synthesis language and preset speaker and set the number of active code-prediction groups. The packaged defaults are language- and speaker-specific and use 16 active code-prediction groups.
Supported languages are Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian. Automatic language selection is also available.
qwen3_tts_custom_voice_<language>_manifest.json — language-specific
configuration.tokenizer.json — tokenizer model.*.onnx, *.gguf, and *.tflite — model components.assets/ — language, speaker, and variant maps.SHA256SUMS — model-bundle checksums for reproducibility.SHA256SUMS was generated by recursively hashing every regular file in the
model bundle, including files in subdirectories, except the generated root
SHA256SUMS and paths with a dotfile component.
From the model bundle root, verify the checked-out files with:
shasum -a 256 -c SHA256SUMS