2019-11-13 22:22:46 +08:00
# Parakeet
Parakeet aims to provide a flexible, efficient and state-of-the-art text-to-speech toolkit for the open-source community. It is built on Paddle Fluid dynamic graph, with the support of many influential TTS models proposed by [Baidu Research ](http://research.baidu.com ) and other academic institutions.
2020-02-06 12:42:00 +08:00
< div align = "center" >
< img src = "images/logo.png" width = 450 / > < br >
< / div >
2019-11-13 22:22:46 +08:00
## Installation
2020-02-18 10:29:54 +08:00
### Install Paddlepaddle
See [install ](https://www.paddlepaddle.org.cn/install/quick ) for more details. This repo requires paddlepaddle's version is above 1.7.
### Install Parakeet
2019-11-13 22:22:46 +08:00
```bash
2020-02-18 10:29:54 +08:00
# git clone this repo first
cd Parakeet
pip install -e .
2019-11-13 22:22:46 +08:00
```
2020-02-19 11:00:17 +08:00
### Setup
Make sure libsndfile1 installed:
```bash
sudo apt-get install libsndfile1
```
2020-02-18 10:29:54 +08:00
### Install CMUdict for nltk
CMUdict from nltk is used to transform text into phonemes.
```python
import nltk
2020-02-19 11:00:17 +08:00
nltk.download("punkt")
2020-02-18 10:29:54 +08:00
nltk.download("cmudict")
```
2020-02-18 17:42:24 +08:00
2019-11-13 22:22:46 +08:00
## Supported models
2020-02-18 11:32:14 +08:00
- [Deep Voice 3: Scaling Text-to-Speech with Convolutional Sequence Learning ](https://arxiv.org/abs/1710.07654 )
2020-02-18 17:42:24 +08:00
- [Neural Speech Synthesis with Transformer Network ](https://arxiv.org/abs/1809.08895 )
- [FastSpeech: Fast, Robust and Controllable Text to Speech ](https://arxiv.org/abs/1905.09263 ).
2020-02-18 10:29:54 +08:00
## Examples
- [Train a deepvoice 3 model with ljspeech dataset ](./parakeet/examples/deepvoice3 )
2020-02-18 17:42:24 +08:00
- [Train a transformer_tts model with ljspeech dataset ](./parakeet/examples/transformer_tts )
- [Train a fastspeech model with ljspeech dataset ](./parakeet/examples/fastspeech )
