{"full_name":"RVC-Project/Retrieval-based-Voice-Conversion-WebUI","owner":"RVC-Project","name":"Retrieval-based-Voice-Conversion-WebUI","url":"https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI","description":"Easily train a good VC model with voice data <= 10 mins!","homepage":"","topics":[],"language":"Python","license":"MIT","avatar":"https://avatars.githubusercontent.com/u/131328897?v=4","created_at":"2023-03-27T09:59:10Z","pushed_at":"2026-08-04T07:47:32Z","metrics":{"stars":39144,"forks":5351,"watchers":232,"issues":2078,"prs":0,"releases":8,"issues_open":null,"issues_closed":null,"prs_open":null,"prs_closed":null,"prs_merged":null,"discussions":0},"growth":{"day":{"stars":16,"forks":2,"base_day":"2026-11-09"}},"history":{"dates":["2026-11-09","2026-11-10"],"stars":[39128,39144],"forks":[5349,5351],"watchers":[232,232],"issues_open":[null,null],"prs_open":[null,null]},"boards":{},"same_owner":[],"issues":[{"number":272,"title":"AMD Support - Segmentation Fault","comments":42,"created_at":"2023-05-12T07:15:42Z"},{"number":136,"title":"Support for Apple's MPS [?]","comments":38,"created_at":"2023-04-23T18:42:44Z"},{"number":2109,"title":"Some historical legacy issues","comments":37,"created_at":"2024-06-05T11:43:42Z"},{"number":278,"title":"Does rvc support training with cpu?","comments":19,"created_at":"2023-05-13T07:25:40Z"},{"number":840,"title":"Very long epoch generation","comments":18,"created_at":"2023-07-23T13:34:44Z"},{"number":1445,"title":"AttributeError: 'NoneType' object has no attribute 'pipeline'","comments":17,"created_at":"2023-10-19T09:21:57Z"},{"number":1211,"title":"Perhaps it's time to find an alternative to google colab?","comments":17,"created_at":"2023-09-09T01:19:46Z"},{"number":1054,"title":"AttributeError: 'NoneType' object has no attribute 'tobytes'","comments":17,"created_at":"2023-08-19T04:50:11Z"}],"discussions":[],"latest_release":{"name":"2.3.260718","tag":"2.3.260718","published_at":"2026-07-21T02:28:49Z","body":"### 完整包 Complete package\r\n\r\nFor Nvidia GPU users (<50x0):\r\nhttps://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC20260718Nvidia.7z\r\n\r\nFor Nvidia GPU users (>=50x0):\r\nhttps://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC20260718Nvidia50x0.7z\r\n\r\nFor AMD/Intel GPU users:\r\nhttps://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC20260718AMD_Intel.7z\r\n\r\nRVC Real-Time VST2/VST3 Plugin (For user instructions, please refer to the README included in the archive. You should unzip the complete package first.)\r\nhttps://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVCRealtime-Win64VST.zip\r\n\r\n中国用户可使用以下2个源加速下载\r\n1、免费满速下载直链合集：https://www.yuque.com/flowercry/hxf0ds\r\n2、度盘需要超级会员才可不限速下载\r\n链接：https://pan.baidu.com/s/16BHeTXRq2aGUzDZo1BY1Lw \r\n提取码: tvy6 \r\n\r\nChangelog [(English verison)](https://github.com/liujing04/Retrieval-based-Voice-Conversion-WebUI/blob/main/docs/en/Changelog_EN.md):\r\n\r\n降低网页使用难度（推理输入音频无需输入路径，前端实时日志优化，保存推理用小模型后自动刷新模型列表，推理首页选择模型后自动检测匹配的index路径等）\r\n\r\n数据特征提取、训练模型、索引阶段可中途掐断，重启继续进度\r\n\r\nGPU自动识别判断推理模式和精度\r\n\r\n修复依赖问题（例如pyworld，fairseq，matplotlib，pyav，onnxruntime-gpu环境修复等）\r\n\r\n解决编码问题\r\n\r\n网页启动自动探寻可用端口\r\n\r\n单卡训练不用DDP\r\n\r\n基础模型不变，若干影响效果的修复，使用现有微调模型更新整合包/代码包版本重新推理，或者使用新版本重新训练再推理，均有可能效果会有提升。\r\n\r\n支持FCPE音高提取算法\r\n\r\n实时变声模式gui支持cuda graph推理加速，算法推理延时指标最高加速4.7倍（4090D测试结果），同时显卡占用相对降低20~30%左右\r\n\r\n移除了已明显无用或过时的功能\r\n\r\n人声与伴奏分离后端由 UVR5 更换为 PyMSS，并支持推理时中断 （更好的人声分离算法能显著提高推理效果）\r\n\r\n优化输入音频的读取与重采样流程，在条件允许时使用 GPU 处理，以提高推理效率并降低 CPU 占用\r\n\r\n修复 Windows 环境下训练进程的 libuv 兼容问题"},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-10T09:35:25Z","history_start":"2026-09-06"}