Nofated095
/
so-vits-svc
mirror of https://github.com/svc-develop-team/so-vits-svc.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
SoftVC VITS Singing Voice Conversion
2 Commits 13 Branches 0 Tags 16 MiB
Python 64.1%
Jupyter Notebook 35.9%
Go to file
Ναρουσέ·μ·γιουμεμί·Χινακάννα a7da244dda
Update LICENSE 		2023-06-14 00:17:08 +08:00
configs upload code 2023-03-10 19:56:17 +09:00
harmof0 upload code 2023-03-10 19:56:17 +09:00
monotonic_align upload code 2023-03-10 19:56:17 +09:00
samples upload code 2023-03-10 19:56:17 +09:00
text upload code 2023-03-10 19:56:17 +09:00
.gitignore upload code 2023-03-10 19:56:17 +09:00
LICENSE Update LICENSE 2023-06-14 00:17:08 +08:00
README.md upload code 2023-03-10 19:56:17 +09:00
app.py upload code 2023-03-10 19:56:17 +09:00
attentions.py upload code 2023-03-10 19:56:17 +09:00
commons.py upload code 2023-03-10 19:56:17 +09:00
data_utils.py upload code 2023-03-10 19:56:17 +09:00
losses.py upload code 2023-03-10 19:56:17 +09:00
mel_processing.py upload code 2023-03-10 19:56:17 +09:00
models.py upload code 2023-03-10 19:56:17 +09:00
modules.py upload code 2023-03-10 19:56:17 +09:00
preprocess.py upload code 2023-03-10 19:56:17 +09:00
preprocess_wave.py upload code 2023-03-10 19:56:17 +09:00
requirements.txt upload code 2023-03-10 19:56:17 +09:00
train.py upload code 2023-03-10 19:56:17 +09:00
train_ms.py upload code 2023-03-10 19:56:17 +09:00
transforms.py upload code 2023-03-10 19:56:17 +09:00
utils.py upload code 2023-03-10 19:56:17 +09:00

README.md

SoftVC VITS Singing Voice Conversion

重大BUG修复

  • 断音问题已解决，音质提升了一个档次
  • 过几天更新

模型简介

歌声音色转换模型，通过SoftVC内容编码器提取源音频语音特征，与F0同时输入VITS替换原本的文本输入达到歌声转换的效果。

目前模型修使用 coarse F0 ，尝试使用HarmoF0 进行f0提取但效果不佳，尝试使用icassp2022-vocal-transcription提取midi替换f0输入但效果不佳

模型推理、训练、一键脚本汇总整理仓库 sovits_guide