Public beta version address 公测版地址: 趣丸千音

MaskGCT (Masked Generative Codec Transformer) is a fully non-autoregressive TTS model that eliminates the need for explicit alignment information between text and speech supervision, as well as phone-level duration prediction. MaskGCT is a two-stage model: in the first stage, the model uses text to predict semantic tokens extracted from a speech self-supervised learning (SSL) model, and in the second stage, the model predicts acoustic tokens conditioned on these semantic tokens. MaskGCT follows the mask-and-predict learning paradigm. During training, MaskGCT learns to predict masked semantic or acoustic tokens based on given conditions and prompts. During inference, the model generates tokens of a specified length in a parallel manner. Experiments with 100K hours of in-the-wild speech demonstrate that MaskGCT outperforms the current state-of-the-art zero-shot TTS systems in terms of quality, similarity, and intelligibility. Audio samples are available at demo page.

• 2024/10/19: We release MaskGCT, a fully non-autoregressive TTS model that eliminates the need for explicit alignment information between text and speech supervision. MaskGCT is trained on Emilia dataset and achieves SOTA zero-shot TTS perfermance.

If you encounter any issue when using MaskGCT, feel free to open an issue in this repository. But please use English to describe, this will make it easier for keyword searching and more people to participate in the discussion.

To run this model, you need to follow the steps below:

1. Clone the repository and install the environment.
2. Run the Jupyter notebook / Gradio demo / Inference script.

This parts, follow the steps below to clone the repository and install the environment.

1. Clone the repository, you can choose (a) partial clone or (b) full clone.
2. Install the environment follow guide below.

Since the whole Amphion repository is large, you can use sparse-checkout to download only the needed code.

# download meta info only
git clone --no-checkout --filter=blob:none https://github.com/open-mmlab/Amphion.git

# enter the repositry directory
cd Amphion

# setting sparse-checkout
git sparse-checkout init --cone
git sparse-checkout set models/tts/maskgct

# download the needed code
git checkout main
git sparse-checkout add models/codec utils

If you prefer to download the whole repository, you can use the following command.

git clone https://github.com/open-mmlab/Amphion.git

# enter the repositry directory
cd Amphion

Before start installing, making sure you are under the Amphion directory. If not, use cd to enter.

Since we use phonemizer to convert text to phoneme, you need to install espeak-ng first. More details can be found here. Choose the correct installation command according to your operating system:

# For Debian-like distribution (e.g. Ubuntu, Mint, etc.)
sudo apt-get install espeak-ng
# For RedHat-like distribution (e.g. CentOS, Fedora, etc.) 
sudo yum install espeak-ng

# For Windows
# Please visit https://github.com/espeak-ng/espeak-ng/releases to download .msi installer

Now, we are going to install the environment. It is recommended to use conda to configure:

conda create -n maskgct python=3.10
conda activate maskgct

pip install -r models/tts/maskgct/requirements.txt

If no error occurs, the environment is successfully installed. Now, you can choose you preferred way to run the model.

• Run Gradio 🤗 Playground Locally
• Jupyter Notebook
• Inference Script

You can run the following command to interact with the playground:

python -m models.tts.maskgct.gradio_demo

We provide a Jupyter notebook to show how to use MaskGCT to inference.

You can open this notebook and start running.

If you do not want to use Juptyer notebook or Gradio, you can use the python inference script.

python -m models.tts.maskgct.maskgct_inference

Running this will automatically download the pretrained model from HuggingFace and start the inference process. The result audio is by default saved in generated_audio.wav, you can change this in the script.

We provide the following pretrained checkpoints:

You can download all pretrained checkpoints from HuggingFace or use huggingface API.

from huggingface_hub import hf_hub_download

# download semantic codec ckpt
semantic_code_ckpt = hf_hub_download("amphion/MaskGCT", filename="semantic_codec/model.safetensors")

# same for other models

By running the Jupyter/Gradio/inference script, it will automatically download the pretrained model from HuggingFace.

If you having trouble connecting to HuggingFace, you try switch endpoint to mirror site:

export HF_ENDPOINT=https://hf-mirror.com

We use the Emilia dataset to train our models. Emilia is a multilingual and diverse in-the-wild speech dataset designed for large-scale speech generation. In this work, we use English and Chinese data from Emilia, each with 50K hours of speech (totaling 100K hours).

If you use MaskGCT in your research, please cite the following paper:

@article{wang2024maskgct,
  title={MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformer},
  author={Wang, Yuancheng and Zhan, Haoyue and Liu, Liwei and Zeng, Ruihong and Guo, Haotian and Zheng, Jiachen and Zhang, Qiang and Zhang, Xueyao and Zhang, Shunsi and Wu, Zhizheng},
  journal={arXiv preprint arXiv:2409.00750},
  year={2024}
}

@inproceedings{amphion,
    author={Zhang, Xueyao and Xue, Liumeng and Gu, Yicheng and Wang, Yuancheng and Li, Jiaqi and He, Haorui and Wang, Chaoren and Song, Ting and Chen, Xi and Fang, Zihao and Chen, Haopeng and Zhang, Junan and Tang, Tze Ying and Zou, Lexiao and Wang, Mingxuan and Han, Jun and Chen, Kai and Li, Haizhou and Wu, Zhizheng},
    title={Amphion: An Open-Source Audio, Music and Speech Generation Toolkit},
    booktitle={{IEEE} Spoken Language Technology Workshop, {SLT} 2024},
    year={2024}
}