0. 说明
- 用 M2VoC 的数据训练
- 可以加上 databaker, thucoss
- 并且可以在 RTVC-7 Voice Cloning Model 的 Tacotron Pretrained Model 上忽略 embedding layer, 进行 Tune
1. Git Clone
- git clone https://github.com/ruclion/Pretrained-EarSpeech.git Tuned-EarSpeech -b 81428b30a486314577bfbd59b272d7e4bd369ef4
- mv Pretrained-EarSpeech Tuned-EarSpeech
- cd Tuned-EarSpeech
- git init
- git add .
- git commit -m 'init'
- git remote add origin https://github.com/ruclion/Tuned-EarSpeech.git
- git push -u origin master
2. 训练数据
2.1. AISHELL-3_denoise_dereverb_
实验室降噪过的数据, 感谢大家. (不过没关注怎么降得~)
/ceph/dataset/AISHELL-3_denoise_dereverb_
2.2. M2VoC
M2VoC 比赛的数据, 目标也是合出来