Skip to content

DeepMind's Tacotron-2 Tensorflow implementation

License

Notifications You must be signed in to change notification settings

wuzl/Tacotron-2

 
 

Repository files navigation

Tacotron-2:

原文档: README_.md

原仓库:Tacotron-2

此项目为 Windows(10)修正版,修正Windows中路径、创建目录等问题,并添加了测试服务(demo_server.py)。

数据集为清华大学开源普通话语料:data_thchs30百度云

项目结构:

Tacotron-2
├── datasets
├── data_thchs30	(0)
│   └── data
│   └── dev
│   └── lm_phone
│   └── lm_word
│   └── test
│   └── train
│   └── README.TXT
├── logs-Tacotron	(2)
│   ├── eval_-dir
│   │ 	├── plots
│ 	│ 	└── wavs
│   ├── mel-spectrograms
│   ├── plots
│   ├── pretrained
│   └── wavs
├── logs-Wavenet	(4)
│   ├── eval-dir
│   │ 	├── plots
│ 	│ 	└── wavs
│   ├── plots
│   ├── pretrained
│   └── wavs
├── papers
├── tacotron
│   ├── models
│   └── utils
├── tacotron_output	(3)
│   ├── eval
│   ├── gta
│   ├── logs-eval
│   │   ├── plots
│   │   └── wavs
│   └── natural
├── wavenet_output	(5)
│   ├── plots
│   └── wavs
├── training_data	(1)
│   ├── audio
│   ├── linear
│	└── mels
└── wavenet_vocoder
	└── models

The previous tree shows the current state of the repository (separate training, one step at a time).

  • Step (0): 下载data_thchs30数据集(百度云),并将其解压至 data_thchs30,如上所示。(preprocess时将会处理其子文件夹 data 中的数据,若想更改请修改 datasets/preprocessor.py 或者替换文件夹 data 。)。
  • Step (1): 数据预处理。 将会生成 training_data 目录。
  • Step (2): 训练模型。其中生成的模型、对齐图等将保存至 logs-Tacotron 目录内。
  • Step (3): 生成音频。生成的音频将保存至 tacotron_output 目录内。(或者运行demo_server.py,指定模型并且在线输入中文和生成普通话音频,内部使用pypinyin转换汉字。)
  • Step (4): Train your Wavenet model. Yield the logs-Wavenet folder.
  • Step (5): Synthesize audio using the Wavenet model. Gives the wavenet_output folder.

About

DeepMind's Tacotron-2 Tensorflow implementation

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 98.4%
  • Jupyter Notebook 1.6%