CoolFace
Modelpublic

Xenna/cielvox2-tts-1.7b-base-GGUF

sourceHugging Faceapache-2.0updated 1mo agoView on Hugging Face
0likes3downloads
Model Card

CielVox 1.7B Base — GGUF

Local-first multilingual voice-cloning TTS model for the cielvox2-tts-1.7b-base backend in [stelnetxcis-create/cielvox2](https://github.com/stelnetxcis-create/cielvox2).

  • —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

This repo contains the talker / code-predictor / speaker-encoder model. It must be used together with the separate tokenizer / codec GGUF from `Xenna/cielvox2-tokenizer-12hz`.

Files

FileSizeNotes
cielvox2-tts-12hz-1.7b-base-q8_0.gguf2.0 GBQ8_0, recommended quantised talker

Quick Start

Build the runtime:

bash
git clone https://github.com/stelnetxcis-create/cielvox2
cd cielvox2
cmake -B build -DCMAKE_BUILD_TYPE=Release
cmake --build build -j$(nproc) --target stelnettts-lib

Download the talker + tokenizer:

bash
huggingface-cli download Xenna/cielvox2-tts-1.7b-base-GGUF \
    cielvox2-tts-12hz-1.7b-base-q8_0.gguf --local-dir .

huggingface-cli download Xenna/cielvox2-tokenizer-12hz \
    cielvox-tokenizer-12hz.gguf --local-dir .

Voice clone from a reference WAV:

bash
./build/bin/stelnettts \
    --backend cielvox2-tts-1.7b-base \
    -m cielvox2-tts-12hz-1.7b-base-q8_0.gguf \
    --codec-model cielvox-tokenizer-12hz.gguf \
    --voice clone.wav \
    --ref-text "Exact transcript of clone.wav" \
    --tts "Hello there" \
    --tts-output hello.wav

Use a baked voice-pack GGUF:

bash
./build/bin/stelnettts \
    --backend cielvox2-tts-1.7b-base \
    -m cielvox2-tts-12hz-1.7b-base-q8_0.gguf \
    --codec-model cielvox-tokenizer-12hz.gguf \
    --voice my-voice-pack.gguf \
    --tts "Hello there" \
    --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

  • —cielvox2-tts-12hz-1.7b-base-q8_0.gguf — recommended quantised deployment

Architecture

ComponentDetails
Talker LM28 layers, 2048 hidden, 16 heads, 8 KV heads, head_dim=128
Output head16 codebooks × 2048 (RVQ)
Code predictor5L + 15 separate codec embedding/lm_head pairs
CodecCielVox-Tokenizer-12Hz (separate GGUF, 12.5 fps RVQ)
Audio24 kHz mono float32 PCM

Implementation

The C++ runtime is implemented in `src/cielvox2_tts.cpp` using ggml graphs for the talker / code-predictor path and the companion tokenizer GGUF for codec encode/decode.

Related

  • —Companion tokenizer / codec GGUF: `Xenna/cielvox2-tokenizer-12hz`
  • —C++ runtime: [stelnetxcis-create/cielvox2](https://github.com/stelnetxcis-create/cielvox2)

License

Apache-2.0.