基于 RTVC-7 Voice Cloning Model 的 Cross-Lingual TTS 的三步走: 第一步 Pretrained-EarSpeech 搭建

0. 说明

RTVC-7 Voice Cloning Model

demo: https://blue-fish.github.io/experiments/RTVC-7.html

  1. 跑通 Git 上的 RealTime VoiceClone: https://github.com/CorentinJ/Real-Time-Voice-Cloning
  2. 简化它, 写在 Git 中: https://github.com/ruclion/Pretrained-EarSpeech

1. Git Clone

  1. git clone https://github.com/CorentinJ/Real-Time-Voice-Cloning.git
  2. mv Real-Time-Voice-Cloning Pretrained-EarSpeech
  3. cd Pretrained-EarSpeech
  4. git init
  5. git add .
  6. git commit -m 'init'
  7. git remote add origin https://github.com/ruclion/Pretrained-EarSpeech.git
  8.  git push -u origin master

2. Install Requirements

  • librosa>=0.8.0
  • matplotlib>=3.3.0
  • numpy==1.19.4
  • scipy>=1.0.0
  • SoundFile
  • Unidecode
  • inflect
  • numba

3. Download Pretrained Models

3.1 下载页面

https://github.com/CorentinJ/Real-Time-Voice-Cloning/wiki/Pretrained-models

3.2. Pretrained 说明

https://blue-fish.github.io/experiments/RTVC-7.html

3.3. 步骤

  1. 下载到 Pretrained-EarSpeech 文件夹
  2. unzip pretrained.zip

人家设计的比较好, 直接就解压在了对应的位置

4. Test Configuration

  1.  python demo_cli.py --no_sound
  2. /ceph/home/hujk17/Pretrained-EarSpeech/samples/1320_00000.mp3
  3. I like Pretrained Ear Speech and China, salute

5. Commit 版本

https://github.com/ruclion/Pretrained-EarSpeech/commit/81428b30a486314577bfbd59b272d7e4bd369ef4

 

已标记关键词 清除标记
相关推荐
©️2020 CSDN 皮肤主题: 大白 设计师:CSDN官方博客 返回首页