Downloads · 30 days
104
6% of all-time downloads
qwen3-tts-rk3588
kautism/qwen3-tts-rk3588
qwen3-tts-rk3588 is a machine learning model from kautism. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
This repo is qwen3-tts model split 3 part, You can use python to load them by hand.
All-time downloads
1.8K
Public
Repo size
6.4 GB
Likes
2
Public
Repo makeup
Click a slice to open those files.
.gguf2.6 GB · 41%
At a glance
- Access
- Public
- Created
- Feb 27, 2026
- Updated
- Oct 12, 2026
- SHA
- ad3679d7
Model card
From the Hugging Face model README
Qwen3-tts
This repo is qwen3-tts model split 3 part, You can use python to load them by hand.
Or use my project Qwen3-tts to running on low cost sbc.
It can even running on 4G ram device.
Voice cloning (speaker encoder)
speaker_encoder/spk_enc.gguf is the ECAPA-TDNN speaker encoder extracted from the
official Qwen/Qwen3-TTS-12Hz-0.6B-Base checkpoint (76 tensors, weights F16 +
biases F32, 17.7 MB). It lets encode-voice store a 1024-dim speaker_embedding
in the voice profile, which the talker uses in QWEN3_TTS_CLONE_MODE=speaker
(x-vector) mode:
qwen3-tts encode-voice --audio ref.wav --ref-text "reference text" -o voice.json
qwen3-tts serve --port 19986 # workers read the voice JSON via the `voice` field
encode-voice resolves this file automatically (env QWEN3_TTS_SPK_ENCODER_GGUF
overrides; otherwise <hf-model>/speaker_encoder/spk_enc.gguf, else this repo).
- Created
- Feb 27, 2026
- Updated
- Oct 12, 2026