Project Overview

This project is an extension of PaddleSpeech's TTS (Text-to-Speech) streaming service, supporting mixed Chinese and English TTS as well as speaker ID selection. With our extension, users can utilize text-to-speech functionalities more flexibly.

File Description

This project includes the following main files:

1. pretrained_models.py: This is a modified model list file sourced from the PaddleSpeech library. You need to replace this file in your PaddleSpeech installation directory.

2. tts_engine: This is another modified file that should be replaced in a specific PaddleSpeech path to ensure the smooth operation of the streaming TTS service.

Usage Instructions

Install NLTK DATA

1、make sure you have NLTK support, and you need install necessary data operator

ntlk.down('averaged_perceptron_tagger')

Replace Files

1. Find your PaddleSpeech installation path by using the following command:

   pip show paddlespeech

2. Replace the pretrained_models.py file in the following path:

   <your PaddleSpeech installation location>/site-packages/paddlespeech/resource/
   

3. Replace the tts_engine file in the following path:

   <your PaddleSpeech installation location>/site-packages/paddlespeech/server/engine/tts/online/onnx/
   

Run Streaming TTS Service

Use PP-TTS_mix_streaming's tts_online_application to run your streaming TTS service. You can start it with the following command:

python -m paddlespeech.server    ./tts_online_application.yaml

Environment Requirements

Ensure your environment meets the following requirements:

• PaddleSpeech version: 1.2.0
• PaddlePaddle version: 2.3.2

For more environment configurations, please refer to the requirements.txt file.

Contributing

Feel free to submit Issues or Pull Requests to improve and enhance the project. Your contributions are highly appreciated!

License

This project is licensed under the Apache License 2.0. For more details, please refer to the license file.