SAC 16k 62 5Hz

--- license: apache-2.0 language: - en - zh pipeline_tag: audio-to-audio tags: - pytorch - codec --- <div align="center"> <h1> SAC: Neural Speech Codec with Semantic-Acoustic Dual-Stream Quantization </h1> <p> <a href="https://github.com/Soul-AILab/SAC"> <img src="https://img.shields.io/badge/SAC-GitHub-black?logo=github&logoColor=white" alt="GitHub Repo"> </a> <a href="https://sac-codec.github.io/"> <img src="https://img.shields.io/badge/🌐%20Demo-Page-brightgreen" alt="Demo Page"> </a> <a href="https://arxiv.org/abs/2510.16841"> <img src="https://img.shields.io/badge/arXiv-2510.16841-blueviolet?logo=arxiv&logoColor=white" alt="arXiv"> </a> <a href="https://huggingface.co/collections/Soul-AILab/sac-68f1df9572a6314d1dc1f91e"> <img src="https://img.shields.io/badge/🤗%20SAC-Models-yellow" alt="Hugging Face"> </a> </p> <p align="center"> <i>A semantic–acoustic dual-stream speech codec achieving state-of-the-art performance in speech reconstruction and semantic representation across bitrates.</i> </p> </div> ## 🛠️ Environment Setup ```bash conda create -n sac python=3.10 conda activate sac pip install -r requirements.txt # pip version == 24.0 ``` ## 🧩 Model Checkpoints To use SAC, you need to prepare the pretrained dependencies, including the [GLM-4-Voice-Tokenizer](https://huggingface.co/zai-org/glm-4-voice-tokenizer) for semantic tokenization and the [ERes2Net](https://modelscope.cn/models/iic/speech_eres2net_sv_en_voxceleb_16k) speaker encoder for speaker feature extraction (during codec training). Make sure the corresponding model paths are correctly set in your configuration file (e.g., `configs/xxx.yaml`). The following table lists the available SAC checkpoints: | Model Name | Hugging Face | Sample Rate | Token Rate | BPS | |:-----------:|:------------:|:------------:|:-----------:|:---:| | SAC | [🤗 Soul-AILab/SAC-16k-37_5Hz](https://huggingface.co/Soul-AILab/SAC-16k-37_5Hz) | 16 kHz | 37.5 Hz | 525 | | SAC | [🤗 Soul-AILab/SAC-16k-62_5Hz](https://huggingface.co/Soul-AILab/SAC-16k-62_5Hz) | 16 kHz | 62.5 Hz | 875 | ## 🎧 Inference To perform audio reconstruction, you can use the following command: ```bash python -m bins.infer ``` We also provide batch scripts for [audio reconstruction](./scripts/batch/reconstruct.sh), [encoding](./scripts/batch/encode.sh), [decoding](./scripts/batch/decode.sh), and [embedding extraction...

Source context: 50 downloads · 2 likes · Pipeline audio-to-audio · Repo Soul-AILab/SAC-16k-62_5Hz