faster-qwen3-tts-demo

This Space hosts the demo UI for **faster-qwen3-tts** with streaming audio, TTFA/RTF metrics, voice clone, custom voices, and voice design. It defaults to the G · running · License unknown

HuggingFaceM4 Open weights text to speech Needs account

Open demo

HF likes: 254
Status: running
License: Unknown
SDK: docker

Hugging Face

HuggingFaceM4/faster-qwen3-tts-demo

SpaceHuggingFaceM4/faster-qwen3-tts-demo
SDKdocker
Likes254
LicenseUnknown
Updated2026-08-25T13:46:00.000Z

Open on Hugging Face

Try this demo

Decision details

This Space hosts the demo UI for **faster-qwen3-tts** with streaming audio, TTFA/RTF metrics, voice clone, custom voices, and voice design. It defaults to the GGML/qwentts.cpp backend and includes a Settings toggle to switch between GGML and the Torch CUDA-graph backend for speed comparisons.

Availabilityrunning
Last checked2026-09-28T22:23:10.256092+00:00
LicenseUnknown
Publisher (host)HuggingFaceM4
Formatstext, audio
Requirementsaccount
Example inputText
Example outputSpoken audio

text to speech

Underlying models

5 · not the same as publisher

Qwen/Qwen3-TTS-12Hz-0.6B-Base underlying model
Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice underlying model
Qwen/Qwen3-TTS-12Hz-1.7B-Base underlying model
Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice underlying model
Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign underlying model