Qwen3-TTS 12Hz 0.6B Base — GGUF (q4_k / q5_K)
GGUF conversions of Qwen3-TTS-12Hz-0.6B-Base for GGUF-based TTS runtimes,
so the model can be loaded from a single file instead of the upstream safetensors
layout.
| File | Quantization | Size |
|---|---|---|
qwen3-tts-12hz-0.6b-base-q5_k.gguf |
q5_K | 1,491,039,968 bytes (~1.49 GB) |
qwen3-tts-12hz-0.6b-base-q4_k.gguf |
q4_k | 1,412,003,552 bytes (~1.41 GB) |
Each file is self-contained: talker, predictor, speech tokenizer, text tokenizer and
the model spec are embedded, so no extra config.json / vocab.json / merges.txt
files are needed next to it.
Provenance and license
- Source checkpoint:
Qwen/Qwen3-TTS-12Hz-0.6B-Base(official release). - Conversion:
audiocpp_gguf(audio.cpp 0.8.2) from the officialmodel.safetensorsandspeech_tokenizer/weights, with--type q4_k/--type q5_k. No fine-tuning, no additional training, no voice cloning: these are the upstream weights, requantized and repackaged. - License: Apache-2.0, same as the upstream model.