Please first follow README for webui_preprocess.py of pytorch backend pipeline first, after downloading all the needed pytorch model files, run webui.py and then follow the README below.
First give some Chinese example input for Language ZH pipeline, then click 生成音频, you will get Error but it's okey, you will also get the converted IR model BERT_ZH under folder ov_models:
Then rerun webui.py and give some English example input for Language EN pipeline, then change Language selection to EN and click 生成音频, you will get Error but it's okey, you will also get the converted IR model BERT_EN under folder ov_models:
Then rerun webui.py and give some Chinese example input for Language ZH pipeline, then click 生成音频, you will still get Error but it's okey, you will also get the converted IR model BERTVits2 under folder ov_models:
Then rerun webui.py and give some Chinese and English example input Language mix pipeline, then click 检测语言，并整理为MIX格式 and click 生成音频, you will get the result and you can listen.
Following are the original README:
VITS2 Backbone with multilingual bert
For quick guide, please refer to webui_preprocess.py.
简易教程请参见 webui_preprocess.py。
FishAudio下的全新自回归TTS Fish-Speech现已可用，效果为目前开源SOTA水准，且在持续维护，推荐使用该项目作为BV2/GSV的替代。本项目短期内不再进行维护。
Demo Video: https://www.bilibili.com/video/BV18E421371Q
Tech slides Video: https://www.bilibili.com/video/BV1zJ4m1K7cj
请注意，本项目核心思路来源于anyvoiceai/MassTTS 一个非常好的tts项目
MassTTS的演示demo为ai版峰哥锐评峰哥本人,并找回了在金三角失落的腰子
- anyvoiceai/MassTTS
- jaywalnut310/vits
- p0p4k/vits2_pytorch
- svc-develop-team/so-vits-svc
- PaddlePaddle/PaddleSpeech
- emotional-vits
- fish-speech
- Bert-VITS2-UI