Qwen3-TTS 12Hz 1.7B Base — GGUF (CrispASR)
GGUF / ggml conversions of
Qwen/Qwen3-TTS-12Hz-1.7B-Base for use with the
qwen3-tts backend in
CrispStrobe/CrispASR.
Qwen3-TTS 12Hz 1.7B Base is the larger sibling of
cstr/qwen3-tts-0.6b-base-GGUF, with the same multilingual voice-cloning behaviour:
- 10 supported languages:
zh en ja ko de fr ru pt es it
- discrete multi-codebook LM architecture with a separate 12 Hz tokenizer / codec
- runtime voice cloning from
(ref_audio, ref_text) or pre-baked voice-pack GGUFs
- Apache-2.0 licence
The 1.7B variant has a wider talker (hidden=2048, vs 1024 for the 0.6B), a matching ECAPA enc_dim=2048, and an additional small_to_mtp_projection bridge from the talker into the 1024-d code predictor. The CrispASR runtime handles all three differences automatically — same CLI, same GGUF schema.
This repo contains the
talker / code-predictor / speaker-encoder model. It must be used together with the separate tokenizer / codec GGUF from
cstr/qwen3-tts-tokenizer-12hz-GGUF.
Files
| File | Size | Notes |
|---|
qwen3-tts-12hz-1.7b-base.gguf | 3.6 GB | F16 |
qwen3-tts-12hz-1.7b-base-q8_0.gguf | 1.9 GB | Q8_0, recommended quantised talker |
Quick Start
Build CrispASR:
1git clone https://github.com/CrispStrobe/CrispASR
2cd CrispASR
3cmake -B build -DCMAKE_BUILD_TYPE=Release
4cmake --build build -j$(nproc) --target crispasr
Download the talker + tokenizer:
1huggingface-cli download cstr/qwen3-tts-1.7b-base-GGUF \
2 qwen3-tts-12hz-1.7b-base-q8_0.gguf --local-dir .
3
4huggingface-cli download cstr/qwen3-tts-tokenizer-12hz-GGUF \
5 qwen3-tts-tokenizer-12hz.gguf --local-dir .
Voice clone from a reference WAV:
1./build/bin/crispasr \
2 --backend qwen3-tts-1.7b-base \
3 -m qwen3-tts-12hz-1.7b-base-q8_0.gguf \
4 --codec-model qwen3-tts-tokenizer-12hz.gguf \
5 --voice clone.wav \
6 --ref-text "Exact transcript of clone.wav" \
7 --tts "Hello there" \
8 --tts-output hello.wav
Or let CrispASR pull both files for you on first run:
1./build/bin/crispasr \
2 --backend qwen3-tts-1.7b-base -m auto \
3 --voice clone.wav \
4 --ref-text "Exact transcript of clone.wav" \
5 --tts "Hello there" \
6 --tts-output hello.wav
When --voice points to a .wav, --ref-text is required. When --voice points to a .gguf, it is treated as a baked voice pack.
Quantisation Notes
qwen3-tts-12hz-1.7b-base.gguf
- reference baseline (3.6 GB)
qwen3-tts-12hz-1.7b-base-q8_0.gguf
- recommended quantised deployment (1.9 GB)
- ASR-roundtrips word-exact on English prompts in current CrispASR testing
Lower-bit talker quants (q6_k, q5_k, q4_k) can still load but are not numerically faithful to the F16 reference and should be treated as experimental.
The companion tokenizer / codec should stay at F16.
How this was made
- The upstream HF safetensors checkpoint was converted to GGUF F16 by
models/convert-qwen3-tts-to-gguf.py.
- Quantised variants are produced with CrispASR's GGUF quantiser.
- Inference is implemented in
src/qwen3_tts.cpp, using ggml graphs for the talker / code-predictor path and the companion tokenizer GGUF for codec encode/decode.
Reference implementation
Architecture and behaviour were checked against the official Qwen release:
The CrispASR runtime is a clean C++ / ggml re-implementation for this repo's backend stack.
Related
License
Apache-2.0, inherited from the base model.
Provenance and EU AI Act Art. 53 note
- Upstream model: Qwen/Qwen3-TTS-12Hz-1.7B-Base — published by
Qwen.
- Upstream licence:
apache-2.0. This repository redistributes under the same terms; it grants no rights the upstream licence does not.
- What was done here: format conversion and/or quantisation only (GGUF/GGML). No training, no fine-tuning, no merging, no distillation, no change to architecture, vocabulary or capability. Only the numeric representation of the upstream weights differs.
- Training data: documented — where it is documented at all — by the upstream provider; see the upstream model card. No training data was used, added or selected by this repository. No training-content summary was found on the upstream model card at the time of writing; that documentation gap is upstream's and is not filled here.
- Provider status: under Regulation (EU) 2024/1689 the upstream authors remain the provider of this model. Converting the serialisation format does not make this repository the provider of a new general-purpose AI model, and no such claim is made. Questions about training content, copyright policy or model capability belong upstream.
