WavTokenizer

SOTA Discrete Codec Models With Forty Tokens Per Second for Audio Language Modeling

🎉🎉 with WavTokenizer, you can represent speech, music, and audio with only 40 tokens per second!

🎉🎉 with WavTokenizer, You can get strong reconstruction results.

🎉🎉 WavTokenizer owns rich semantic information and is build for audio language models such as GPT4-o.

🔥 News

• 2024.09.09: We release WavTokenizer-medium checkpoint on huggingface.
• 2024.08.31: We release WavTokenizer on arxiv.

Installation

To use WavTokenizer, install it using:

conda create -n wavtokenizer python=3.9
conda activate wavtokenizer
pip install -r requirements.txt

Infer

Part1: Reconstruct audio from raw wav

from encoder.utils import convert_audio
import torchaudio
import torch
from decoder.pretrained import WavTokenizer


device=torch.device('cpu')

config_path = "./configs/xxx.yaml"
model_path = "./xxx.ckpt"
audio_outpath = "xxx"

wavtokenizer = WavTokenizer.from_pretrained0802(config_path, model_path)
wavtokenizer = wavtokenizer.to(device)


wav, sr = torchaudio.load(audio_path)
wav = convert_audio(wav, sr, 24000, 1) 
bandwidth_id = torch.tensor([0])
wav=wav.to(device)
features,discrete_code= wavtokenizer.encode_infer(wav, bandwidth_id=bandwidth_id)
audio_out = wavtokenizer.decode(features, bandwidth_id=bandwidth_id) 
torchaudio.save(audio_outpath, audio_out, sample_rate=24000, encoding='PCM_S', bits_per_sample=16)

Part2: Generating discrete codecs

from encoder.utils import convert_audio
import torchaudio
import torch
from decoder.pretrained import WavTokenizer

device=torch.device('cpu')

config_path = "./configs/xxx.yaml"
model_path = "./xxx.ckpt"

wavtokenizer = WavTokenizer.from_pretrained0802(config_path, model_path)
wavtokenizer = wavtokenizer.to(device)

wav, sr = torchaudio.load(audio_path)
wav = convert_audio(wav, sr, 24000, 1) 
bandwidth_id = torch.tensor([0])
wav=wav.to(device)
_,discrete_code= wavtokenizer.encode_infer(wav, bandwidth_id=bandwidth_id)
print(discrete_code)

Part3: Audio reconstruction through codecs

# audio_tokens [n_q,1,t]/[n_q,t]
features = wavtokenizer.codes_to_features(audio_tokens)
bandwidth_id = torch.tensor([0])  
audio_out = wavtokenizer.decode(features, bandwidth_id=bandwidth_id)

Available models

🤗 links to the Huggingface model hub.

Model name	HuggingFace	Corpus	Token/s	Domain	Open-Source
WavTokenizer-small-600-24k-4096	🤗	LibriTTS	40	Speech	√
WavTokenizer-small-320-24k-4096	🤗	LibriTTS	75	Speech	√
WavTokenizer-medium-600-24k-4096	🤗	10000 Hours	40	Speech, Audio, Music	Coming Soon
WavTokenizer-medium-320-24k-4096	🤗	10000 Hours	75	Speech, Audio, Music	√
WavTokenizer-large-600-24k-4096	🤗	80000 Hours	40	Speech, Audio, Music	Coming Soon
WavTokenizer-large-320-24k-4096	🤗	80000 Hours	75	Speech, Audio, Music	Coming Soon

Training

Step1: Prepare train dataset

# Process the data into a form similar to ./data/demo.txt

Step2: Modifying configuration files

# ./configs/xxx.yaml
# Modify the values of parameters such as batch_size, filelist_path, save_dir, device

Step3: Start training process

Refer to Pytorch Lightning documentation for details about customizing the training pipeline.

cd ./WavTokenizer
python train.py fit --config ./configs/xxx.yaml

Citation

If this code contributes to your research, please cite our work, Language-Codec and WavTokenizer:

@misc{ji2024wavtokenizerefficientacousticdiscrete,
      title={WavTokenizer: an Efficient Acoustic Discrete Codec Tokenizer for Audio Language Modeling}, 
      author={Shengpeng Ji and Ziyue Jiang and Xize Cheng and Yifu Chen and Minghui Fang and Jialong Zuo and Qian Yang and Ruiqi Li and Ziang Zhang and Xiaoda Yang and Rongjie Huang and Yidi Jiang and Qian Chen and Siqi Zheng and Wen Wang and Zhou Zhao},
      year={2024},
      eprint={2408.16532},
      archivePrefix={arXiv},
      primaryClass={eess.AS},
      url={https://arxiv.org/abs/2408.16532}, 
}

@misc{ji2024languagecodec,
      title={Language-Codec: Reducing the Gaps Between Discrete Codec Representation and Speech Language Models}, 
      author={Shengpeng Ji and Minghui Fang and Ziyue Jiang and Rongjie Huang and Jialung Zuo and Shulei Wang and Zhou Zhao},
      year={2024},
      eprint={2402.12208},
      archivePrefix={arXiv},
      primaryClass={eess.AS}
}