Easily train a good VC model with voice data <= 10 mins!
For Nvidia GPU users: https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006Nvidia.7z
For AMD/Intel GPU users: https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006AMD_Intel.7z
中国用户可使用以下2个源加速下载 1、无需登录,免费满速下载链接https://www.123pan.com/s/5tIqVv-QHNcv.html 2、有度盘超级会员的可以用度盘https://pan.baidu.com/s/19530AOh2H3Feuti_D51cXw?pwd=reqy
Changelog (English verison):
我们制作了一个用于实时变声的界面go-realtime-gui.bat/gui_v1.py(事实上早就存在了),本次更新重点也优化了实时变声的性能。对比0813版:
注意输入输出设备应该选择同种类型,例如都选MME类型。
1006版本整体的更新为:
For Nvidia GPU users: https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC0813Nvidia.7z
For AMD/Intel GPU users: https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC0813AMD_Intel.7z
中国用户可使用以下2个源加速下载 1、无需登录,免费满速下载链接https://www.123pan.com/s/5tIqVv-QHNcv.html 2、有度盘超级会员的可以用度盘https://pan.baidu.com/s/19530AOh2H3Feuti_D51cXw?pwd=reqy
Changelog (English verison):
1-常规bug修复
2-重点更新
Please look forward to the pretrained base model of RVCv3, which has larger parameters, larger training data, better results, unchanged inference speed, and requires less training data for training.
https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC-beta.7z
0619update: If you're using small model extract feature, you should update this file because there is a small bug when config is v2-32k/48k
How to update from 0528v2 version: 1、download or clone updated codes from github, and replace 0528v2 version. 2、download new pretrained_v2 weights from https://huggingface.co/lj1995/VoiceConversionWebUI/tree/main/pretrained_v2 (32k and 48k weights needed. 40k has already been supported in 0528v2 version)
Changelog:
https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC-beta.7z
下过老完整包的用户可以只下更新包 Users that downloaded old base package can use the updated package.
见下方Assets
, 解压到RVC根目录覆盖完整包下的一些文件 Unzip it in RVC root and replace some files of old version.
对比上个0428版本,划重点(Compared to the previous 0428 version, the most significant updates are): 1、增加了v2版本模型支持 (Updated to v2 model) 2、保护呼吸、清辅音、齿音,削减电音 (Protect voiceless consonant and breath, less artifact) 3、增加crepe推理,音高哑音更少 (Crepe (deep-learning based method) pitch detect model supported) 4、人声伴奏分离新引入了UVR5中的去混响和去延迟模型 (Dereverb and de-echo model in UVR5 supported)
todolist:
https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC-beta.7z
下过老完整包的用户可以只下更新包 Users that downloaded old base package can use the updated package.
见下方Assets
, 解压到RVC根目录覆盖完整包下的一些文件 Unzip it in RVC root and replace some files of old version.
功能:
底模:
https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC-beta.7z
下过老完整包的用户可以只下更新包 Users that downloaded old base package can use the updated package.
见下方Assets
, 解压到RVC根目录覆盖完整包下的一些文件 Unzip it in RVC root and replace some files of old version.
修正训练参数,提升显卡平均利用率,A100最高从25%提升至90%左右,V100:50%->90%左右,2060S:60%->85%左右,P40:25%->95%左右,训练速度显著提升
修正参数:总batch_size改为每张卡的batch_size
修正total_epoch:最大限制100解锁至1000;默认10提升至默认20
修复ckpt提取识别是否带音高错误导致推理异常的问题
修复分布式训练每个rank都保存一次ckpt的问题
特征提取进行nan特征过滤
修复静音输入输出随机辅音or噪声的问题(老版模型需要重做训练集重训)
新增本地实时变声迷你GUI,双击go-realtime-gui.bat启动
训练推理均对<50Hz的频段进行滤波过滤
训练推理音高提取pyworld最低音高从默认80下降至50,50-80hz间的男声低音不会哑
WebUI支持根据系统区域变更语言(现支持en_US,ja_JP,zh_CN,zh_HK,zh_SG,zh_TW,不支持的默认en_US)
修正部分显卡识别(例如V100-16G识别失败,P4识别失败)
完整包(base package) https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC-beta.7z
20230410更新包(解压到RVC根目录覆盖完整包下的一些文件)已放出(20230410updated package: unzip it in RVC root and replace some files of old version.)
更新日志见(changelog:) https://github.com/liujing04/Retrieval-based-Voice-Conversion-WebUI/blob/main/docs/Changelog_CN.md