You can deploy this model and start earning money today!

dbrains/
Qwen3-TTS-Tokenizer-12Hz-48kHz-GGUF

Name: Qwen3-TTS-Tokenizer-12Hz-48kHz-GGUF
Availability: InStock

gguf

takuma104/Qwen3-TTS-Tokenizer-12Hz-48kHz

quantized

apache-2.0

Views

No views yet

Qwen3-TTS-Tokenizer-12Hz-48kHz-GGUF

GGUF conversion of takuma104/Qwen3-TTS-Tokenizer-12Hz-48kHz — the V2 Qwen3-TTS vocoder. 5 decoder blocks, 48 kHz output (V1 = 4 blocks, 24 kHz).

File

Qwen3-TTS-Tokenizer-12Hz-48kHz-F16.gguf — F16, 341 MB

Compatibility

Requires dbrain/qwen3-tts.cpp. Upstream khimaros/qwen3-tts.cpp hardcodes 4 decoder blocks and a 24 kHz WAV header — it will fail to load this file or emit a half-speed WAV. The dbrain fork reads n_dec_blocks and sample_rate from GGUF metadata so V1 and V2 both work transparently; no upstream merge yet.

Mainline llama.cpp does not support qwen3-tts.

Usage

bash

qwen3-tts-server \
    --hf-repo    khimaros/Qwen3-TTS-12Hz-1.7B-VoiceDesign-GGUF:Q8_0 \
    --hf-repo-v  dbrains/Qwen3-TTS-Tokenizer-12Hz-48kHz-GGUF:F16

License

Apache-2.0 — derivative of takuma104/Qwen3-TTS-Tokenizer-12Hz-48kHz.