s2-pro-gguf

> **ALPHA — EXPERIMENTAL** > The inference engine ([s2.cpp](https://github.com/rodrigomatta/s2.cpp)) is an early-stage, community-built project. Expect rough edges and breaking changes. Not production-ready. GGUF-quantized weights of [Fish Audio S2 Pro](https://fish.audio), a high-quality multilingual text-to-speech model with voice cloning support, packaged for local inference with [s2.cpp](https://github.com/rodrigomatta/s2.cpp) — a pure C++/GGML engine with no Python dependency. > **License:** Fish Audio Research License — free for research and non-commercial use. Commercial use requires a separate license from Fish Audio. See [LICENSE.md](LICENSE.md) and [fish.audio](https://fish.audio). All GGUF files contain both the transformer weights and the audio codec in a single file. - GPU with Vulkan support (AMD/NVIDIA/Intel) **or** CPU with enough RAM - [s2.cpp](https://github.com/rodrigomatta/s2.cpp) built from source (C++17 + CMake) git clone --recurse-submodules https://github.com/rodrigomatta/s2.cpp.git cd s2.cpp cmake -B build -DCMAKE_BUILD_TYPE=Release -DS2_VULKAN=ON cmake --build build --parallel $(nproc) huggingface-cli download rodrigomt/s2-pro-gguf s2-pro-q6_k.gguf tokenizer.json --local-dir . ./build/s2 \ -m s2-pro-q6_k.gguf \ -t tokenizer.json \ -text "Hello, this is a test." \ -v 0 \ -o output.wav ./build/s2 \ -m s2-pro-q6_k.gguf \ -t tokenizer.json \ -pa reference.wav \ -pt "Transcript of the reference audio." \ -text "Text to synthesize in that voice." \ -v 0 \ -o output.wav Reference audio: 5–30 seconds, clean recording, WAV or MP3. The GGUF files in this repository were re-quantized using a modified `llama-quantize` tool with improved quantization routines. S2 Pro uses a **Dual-AR** architecture (~4.56B parameters total): - **Slow-AR** — 36-layer Qwen3 transformer (4.13B params), GQA (32 heads / 8 KV heads), RoPE 1M base, persistent KV cache - **Fast-AR** — 4-layer transformer (0.42B params) generating 10 acoustic codebook tokens per semantic step - **Audio codec** — convolutional RVQ encoder/decoder (10 codebooks × 4096 entries) The model weights are licensed under the **Fish Audio Research License**. - **Research and non-commercial use:** free under this license - **Commercial use:** requires a separate written license from Fish Audio Attribution: *"This model is licensed under the Fish Audio Research License, Copyright © 39 AI, INC. All Rights Reserved."* Full terms: [LICENSE.md](LICENSE.md) · Commercial: [fish.audio](https://fish.audio) · [business@fish.audio](mailto:business@fish.audio)