Production First and Production Ready End-to-End Text-to-Speech Toolkit

pip install git+https://github.com/wenet-e2e/wetts.git

Command-line usage (use -h for parameters):

wetts --text "今天天气怎么样" --wav output.wav

Python programming usage:

import wetts

# TODO

We suggest to install WeTTS with Anaconda or Miniconda.

Clone this repo:

git clone https://github.com/wenet-e2e/wetts.git

Create the environment:

conda create -n wetts python=3.8 -y
conda activate wetts
pip install -r requirements.txt

We mainly focus on end to end, production, and on-device TTS. We are going to use:

• backend: end to end model, such as:
  • VITS
• frontend:
  • Text Normalization: WeTextProcessing
  • Prosody & Polyphones: Unified Mandarin TTS Front-end Based on Distilled BERT Model

We plan to support a variaty of open source TTS datasets, include but not limited to:

• Baker, Chinese Standard Mandarin Speech corpus open sourced by Data Baker.
• AISHELL-3, a large-scale and high-fidelity multi-speaker Mandarin speech corpus.
• Opencpop, Mandarin singing voice synthesis (SVS) corpus open sourced by Netease Fuxi.

We plan to support a variaty of hardwares and platforms, including:

• x86
• Android
• Raspberry Pi
• Other on-device platforms

export GLOG_logtostderr=1
export GLOG_v=2

cd runtime/onnxruntime
cmake -B build -DCMAKE_BUILD_TYPE=Release
cmake --build build
./build/bin/tts_main \
  --frontend_flags baker_bert_onnx/frontend.flags \
  --vits_flags multilingual_vits_v3_onnx/vits.flags \
  --sname baker \
  --text "hello我是小明。" \
  --wav_path audio.wav

For Chinese users, you can aslo scan the QR code on the left to follow our offical account of WeNet. We created a WeChat group for better discussion and quicker response. Please scan the personal QR code on the right, and the guy is responsible for inviting you to the chat group.

Or you can directly discuss on Github Issues.

1. We borrow a lot of code from vits for VITS implementation.
2. We refer PaddleSpeech for pinyin lexicon generation.