未验证 提交 9d0ff7a7 编写于 作者: 小湉湉's avatar 小湉湉 提交者: GitHub

Merge pull request #1890 from yt605155624/docs

[doc]add pptts readme
([简体中文](./PPTTS_cn.md)|English)
- [1. Introduction](#1)
- [2. Characteristic](#2)
- [3. Benchmark](#3)
- [4. Demo](#4)
- [5. Tutorials](#5)
- [5.1 Training and Inference Optimization](#51)
- [5.2 Characteristic APPs of TTS](#52)
- [5.3 TTS Server](#53)
<a name="1"></a>
## 1. Introduction
PP-TTS is a streaming speech synthesis system developed by PaddleSpeech. Based on the implementation of [SOTA Algorithms](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/docs/source/released_model.md#text-to-speech-models), a faster inference engine is used to realize streaming speech synthesis technology to meet the needs of commercial speech interaction scenarios.
#### PP-TTS
Pipline of TTS:
<center><img src=https://ai-studio-static-online.cdn.bcebos.com/ea69ae1faff84940a59c7079d16b3a8db2741d2c423846f68822f4a7f28726e9 width="600" ></center>
PP-TTS provides a Chinese streaming speech synthesis system based on FastSpeech2 and HiFiGAN by default:
- Text Frontend: The rule-based Chinese text frontend system is adopted to optimize Chinese text such as text normalization, polyphony, and tone sandhi.
- Acoustic Model: The decoder of FastSpeech2 is improved so that it can be stream synthesized
- Vocoder: Streaming synthesis of GAN vocoder is supported
- Inference Engine: Using ONNXRuntime to optimize the inference of TTS models, so that the TTS system can also achieve RTF < 1 on low-voltage, meeting the requirements of streaming synthesis
<a name="2"></a>
## 2. Characteristic
- Open source leading Chinese TTS system
- Using ONNXRuntime to optimize the inference of TTS models
- The only open-source streaming TTS system
- Easy disassembly: Developers can easily replace different acoustic models and vocoders in different languages, use different inference engines (Paddle dynamic graph, PaddleInference, ONNXRuntime, etc.), and use different network services (HTTP, WebSocket)
<a name="3"></a>
## 3. Benchmark
PaddleSpeech TTS models' benchmark: [TTS-Benchmark](https://github.com/PaddlePaddle/PaddleSpeech/wiki/TTS-Benchmark)
<a name="4"></a>
## 4. Demo
See: [Streaming TTS Demo Video](https://paddlespeech.readthedocs.io/en/latest/streaming_tts_demo_video.html)
<a name="5"></a>
## 5. Tutorials
<a name="51"></a>
### 5.1 Training and Inference Optimization
Default FastSpeech2: [tts3/run.sh](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/csmsc/tts3/run.sh)
Streaming FastSpeech2: [tts3/run_cnndecoder.sh](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/csmsc/tts3/run_cnndecoder.sh)
HiFiGAN:[voc5/run.sh](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/csmsc/voc5/run.sh)
<a name="52"></a>
### 5.2 Characteristic APPs of TTS
text_to_speech - convert text into speech: [text_to_speech](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/text_to_speech)
style_fs2 - multi style control for FastSpeech2 model: [style_fs2](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/style_fs2)
story talker - book reader based on OCR and TTS: [story_talker](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/story_talker)
metaverse - 2D AR with TTS: [metaverse](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/metaverse)
<a name="53"></a>
### 5.3 TTS Server
Non-streaming TTS Server: [speech_server](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/speech_server)
Streaming TTS Server: [streaming_tts_server](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/streaming_tts_server)
For more tutorials please see: [PP-TTS:流式语音合成原理及服务部署
](https://aistudio.baidu.com/aistudio/projectdetail/3885352)
(简体中文|[English](./PPTTS.md))
# PP-TTS
- [1. 简介](#1)
- [2. 特性](#2)
- [3. Benchmark](#3)
- [4. 效果展示](#4)
- [5. 使用教程](#5)
- [5.1 模型训练与推理优化](#51)
- [5.2 语音合成特色应用](#52)
- [5.3 语音合成服务搭建](#53)
<a name="1"></a>
## 1. 简介
PP-TTS 是 PaddleSpeech 自研的流式语音合成系统。在实现[前沿算法](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/docs/source/released_model.md#text-to-speech-models)的基础上,使用了更快的推理引擎,实现了流式语音合成技术,使其满足商业语音交互场景的需求。
#### PP-TTS
语音合成基本流程如下图所示:
<center><img src=https://ai-studio-static-online.cdn.bcebos.com/ea69ae1faff84940a59c7079d16b3a8db2741d2c423846f68822f4a7f28726e9 width="600" ></center>
PP-TTS 默认提供基于 FastSpeech2 声学模型和 HiFiGAN 声码器的中文流式语音合成系统:
- 文本前端:采用基于规则的中文文本前端系统,对文本正则、多音字、变调等中文文本场景进行了优化。
- 声学模型:对 FastSpeech2 模型的 Decoder 进行改进,使其可以流式合成
- 声码器:支持对 GAN Vocoder 的流式合成
- 推理引擎:使用 ONNXRuntime 推理引擎优化模型推理性能,使得语音合成系统在低压 CPU 上也能达到 RTF<1,满足流式合成的要求
<a name="2"></a>
## 2. 特性
- 开源领先的中文语音合成系统
- 使用 ONNXRuntime 推理引擎优化模型推理性能
- 唯一开源的流式语音合成系统
- 易拆卸性:可以很方便地更换不同语种上的不同声学模型和声码器、使用不同的推理引擎(Paddle 动态图、PaddleInference 和 ONNXRuntime 等)、使用不同的网络服务(HTTP、Websocket)
<a name="3"></a>
## 3. Benchmark
PaddleSpeech TTS 模型之间的性能对比,请查看 [TTS-Benchmark](https://github.com/PaddlePaddle/PaddleSpeech/wiki/TTS-Benchmark)
<a name="4"></a>
## 4. 效果展示
请参考:[Streaming TTS Demo Video](https://paddlespeech.readthedocs.io/en/latest/streaming_tts_demo_video.html)
<a name="5"></a>
## 5. 使用教程
<a name="51"></a>
### 5.1 模型训练与推理优化
Default FastSpeech2:[tts3/run.sh](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/csmsc/tts3/run.sh)
流式 FastSpeech2:[tts3/run_cnndecoder.sh](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/csmsc/tts3/run_cnndecoder.sh)
HiFiGAN:[voc5/run.sh](https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/csmsc/voc5/run.sh)
<a name="52"></a>
### 5.2 语音合成特色应用
一键式实现语音合成:[text_to_speech](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/text_to_speech)
个性化语音合成 - 基于 FastSpeech2 模型的个性化语音合成:[style_fs2](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/style_fs2)
会说话的故事书 - 基于 OCR 和语音合成的会说话的故事书:[story_talker](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/story_talker)
元宇宙 - 基于语音合成的 2D 增强现实:[metaverse](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/metaverse)
<a name="53"></a>
### 5.3 语音合成服务搭建
一键式搭建非流式语音合成服务:[speech_server](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/speech_server)
一键式搭建流式语音合成服务:[streaming_tts_server](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/demos/streaming_tts_server)
更多教程,包括模型设计、模型训练、推理部署等,请参考 AIStudio 教程:[PP-TTS:流式语音合成原理及服务部署
](https://aistudio.baidu.com/aistudio/projectdetail/3885352)
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册