This is the code for MOSS-Audio-Tokenizer presented in the MOSS-Audio-Tokenizer: Scaling Audio Tokenizers for Future Audio Foundation Models. MOSS-Audio-Tokenizer is a unified discrete audio tokenizer based on the Cat (Causal Audio Tokenizer with Transformer) architecture. Scaling to 1.6 billion parameters, it functions as a unified discrete interface, delivering both lossless-quality reconstruction and high-level semantic alignment.
Key Features:
- Extreme Compression & Variable Bitrate: It compresses 24kHz raw audio into a remarkably low frame rate of 12.5Hz. Utilizing a 32-layer Residual Vector Quantizer (RVQ), it supports high-fidelity reconstruction across a wide range of bitrates, from 0.125kbps to 4kbps.
- Pure Transformer Architecture: The model features a "CNN-free" homogeneous architecture built entirely from Causal Transformer blocks. With 1.6B combined parameters (Encoder + Decoder), it ensures exceptional scalability and supports low-latency streaming inference.
- Large-Scale General Audio Training: Trained on 3 million hours of diverse audio data, the model excels at encoding and reconstructing all audio domains, including speech, sound effects, and music.
- Unified Semantic-Acoustic Representation: While achieving state-of-the-art reconstruction quality, Cat produces discrete tokens that are "semantic-rich," making them ideal for downstream tasks like speech understanding (ASR) and generation (TTS).
- Fully Trained From Scratch: Cat does not rely on any pretrained encoders (such as HuBERT or Whisper) or distillation from teacher models. All representations are learned autonomously from raw data.
- End-to-End Joint Optimization: All components—including the encoder, quantizer, decoder, discriminator, and a decoder-only LLM for semantic alignment—are optimized jointly in a single unified training pipeline.
Summary: By combining a simple, scalable architecture with massive-scale data, the Cat architecture overcomes the bottlenecks of traditional audio tokenizers. It provides a robust, high-fidelity, and semantically grounded interface for the next generation of native audio foundation models.
This repository is the official implementation of MOSS-Audio-Tokenizer.
Architecture of MOSS-Audio-Tokenizer
- [2026/2/12] 🌟 We released the technical report: MOSS-Audio-Tokenizer: Scaling Audio Tokenizers for Future Audio Foundation Models. Read the paper.
- [2026/2/10] 🎉 Released MOSS-TTS Family. Please refer to our blog for details; models and docs can be found in the MOSS-TTS GitHub repository.
- [2026/2/9] 🔥 We released code and checkpoints of MOSS-Audio-Tokenizer. Check out the GitHub repository, Hugging Face weights, and ModelScope weights.
For MOSS‑TTS Family models and docs, visit the GitHub repo: https://github.com/OpenMOSS/MOSS-TTS
| Model | Hugging Face | ModelScope |
|---|---|---|
| MOSS-Audio-Tokenizer |  |
|
| Model | Hugging Face | ModelScope |
|---|---|---|
| MOSS-TTS | |
|
| MOSS-TTS-Local-Transformer | |
|
| MOSS-TTSD | |
|
| MOSS-TTS-Realtime | |
|
| MOSS-VoiceGenerator | |
|
| MOSS-SoundEffect | |
|
# Clone the repository
git clone https://github.com/OpenMOSS/MOSS-Audio-Tokenizer.git
cd MOSS-Audio-Tokenizer
# Install dependencies
conda create -n moss-audio-tokenizer python=3.10 -y
conda activate moss-audio-tokenizer
pip install -r requirements.txtimport torch
from transformers import AutoModel
import torchaudio
repo_id = "OpenMOSS-Team/MOSS-Audio-Tokenizer"
model = AutoModel.from_pretrained(repo_id, trust_remote_code=True).eval()
wav, sr = torchaudio.load('demo/demo_gt.wav')
if sr != model.sampling_rate:
wav = torchaudio.functional.resample(wav, sr, model.sampling_rate)
wav = wav.unsqueeze(0)
enc = model.encode(wav, return_dict=True)
print(f"enc.audio_codes.shape: {enc.audio_codes.shape}")
dec = model.decode(enc.audio_codes, return_dict=True)
print(f"dec.audio.shape: {dec.audio.shape}")
wav = dec.audio.squeeze(0)
torchaudio.save("demo/demo_rec.wav", wav, sample_rate=model.sampling_rate)
# Decode using only the first 8 layers of the RVQ
dec_rvq8 = model.decode(enc.audio_codes[:8], return_dict=True)
wav_rvq8 = dec_rvq8.audio.squeeze(0)
torchaudio.save("demo/demo_rec_rvq8.wav", wav_rvq8, sample_rate=model.sampling_rate)MossAudioTokenizerModel.encode and MossAudioTokenizerModel.decode support simple streaming via a chunk_duration
argument.
chunk_durationis expressed in seconds.- It must be <=
MossAudioTokenizerConfig.causal_transformer_context_duration. chunk_duration * MossAudioTokenizerConfig.sampling_ratemust be divisible byMossAudioTokenizerConfig.downsample_rate.- Streaming chunking only supports
batch_size=1.
import torch
from transformers import AutoModel
repo_id = "OpenMOSS-Team/MOSS-Audio-Tokenizer"
model = AutoModel.from_pretrained(repo_id, trust_remote_code=True).eval()
audio = torch.randn(1, 1, 3200) # dummy waveform
# 0.08s @ 24kHz = 1920 samples, divisible by downsample_rate=1920
enc = model.encode(audio, return_dict=True, chunk_duration=0.08)
dec = model.decode(enc.audio_codes, return_dict=True, chunk_duration=0.08)from transformers import AutoModel
model = AutoModel.from_pretrained("OpenMOSS-Team/MOSS-Audio-Tokenizer", trust_remote_code=True).eval()conda activate moss-audio-tokenizer
cd MOSS-Audio-Tokenizer
python demo/test_reconstruction.pyThe table below compares the reconstruction quality of open-source audio tokenizers with MOSS-Audio-Tokenizer on speech and audio/music data.
- Speech metrics are evaluated on LibriSpeech test-clean (English) and AISHELL-2 (Chinese), reported as EN/ZH.
- Audio metrics are evaluated on the AudioSet evaluation subset, while music metrics are evaluated on MUSDB, reported as audio/music.
- STFT-Dist. denotes the STFT distance.
- Higher is better for speech metrics, while lower is better for audio/music metrics (Mel-Loss, STFT-Dist.).
- Nvq denotes the number of quantizers.
Reconstruction quality comparison of open-source audio tokenizers on speech and audio/music data.
The plots below compare our MOSS-Audio-Tokenizer model with other open-source speech tokenizers on the LibriSpeech dataset, evaluated with SIM, STOI, PESQ-NB, and PESQ-WB (higher is better). We control the bps of the same model by adjusting the number of RVQ codebooks used during inference.
If you use this code or result in your paper, please cite our work as:
@misc{gong2026mossaudiotokenizerscalingaudiotokenizers,
title={MOSS-Audio-Tokenizer: Scaling Audio Tokenizers for Future Audio Foundation Models},
author={Yitian Gong and Kuangwei Chen and Zhaoye Fei and Xiaogui Yang and Ke Chen and Yang Wang and Kexin Huang and Mingshu Chen and Ruixiao Li and Qingyuan Cheng and Shimin Li and Xipeng Qiu},
year={2026},
eprint={2602.10934},
archivePrefix={arXiv},
primaryClass={cs.SD},
url={https://arxiv.org/abs/2602.10934},
}MOSS-Audio-Tokenizer is released under the Apache 2.0 license.
