A sound cloning tool with a web interface, using your voice or any sound to record audio / 一个带web界面的声音克隆工具,使用你的音色或任意声音来录制音频
为减小编译包体积,语音->语音模型未包含在内,需要单独下载后放到项目根目录下的tts里,即和 app.exe 同目录的 tts 文件夹里,然后 "解压到当前文件夹"
编译版主文件clone-voice下载链接(已包含文字到语音模型text-to-speech): https://pan.baidu.com/s/1lQaosvD1DNLWjA5e3QA27g?pwd=mvgt 提取码:mvgt
文字到语音模型(text-to-speech)下载 链接:https://pan.baidu.com/s/1LA3JFIb0MnCgoF0Q1sW5dQ?pwd=5k7c 提取码:5k7c
语音到语音模型(speech-to-speech)下载 链接:https://pan.baidu.com/s/1vIYzxnlmx2_4prahufoEEw?pwd=hgh2 提取码:hgh2
下载后在该项目根目录下的tts文件夹内解压,将压缩包内的文件夹复制到 tts 文件夹内,解压后该tts文件夹内会新增3个文件夹,分别是
tts_models--multilingual--multi-dataset--xtts_v2
voice_conversion_models--multilingual--vctk--freevc24
wavlm
必须确保解压正确,如图
百度网盘下载链接(2个模型): https://pan.baidu.com/s/1CBnPmNElB3LHG7r_McAIwQ?pwd=61jc
也可单独下载每个模型,同样在tts文件夹内解压,得到3个文件夹/You can also download model-tts.7z and model-sts.7z from Assets below and unzip them to the tts directory. After unzipping, the structure is shown in the above figure
GitHub下载载文字->声音模型(download text->speech model ): https://github.com/jianchang512/clone-voice/releases/download/v0.0.1/model-tts.7z
GitHub下载声音->声音模型(download speech->speech model): https://github.com/jianchang512/clone-voice/releases/download/v0.0.1/model-sts.7z
使用预编译时,如果模型均已下载,并正确解压,解压后目录结构如上图所示,但仍提示连接 github 网站失败,请 下载 extra-to-tts_cache.zip 解压后得到2个文件,将这2个文件复制到 tts_cache 目录下
下载训练文件后,在软件根目录下解压,然后双击 runtrain.bat
百度网盘下载:https://pan.baidu.com/s/1_9n-4GJjvUoeVYhkHllJOA?pwd=1k1g
huggingface下载:https://huggingface.co/spaces/mortimerme/s4/resolve/main/train-extract.7z?download=true
本项目所用模型为coqui.ai出品的xtts_v2,模型开源协议为Coqui Public Model License 1.0.0,使用本项目请遵循该协议,协议全文见 https://coqui.ai/cpml.txt
The model used in this project is xtts_v2 produced by coqui.ai, and the model open source license is Coqui Public Model License 1.0.0 , please follow this agreement when using this project. The full text of the agreement can be found at https://coqui.ai/cpml.txt