Home Creators Posts Import Register Favorites Logout

Downloads

Content

RVC WebUI Premium converts and trains voices, with real-time microphone voice changing, voice merging and batch folder processing, using under 4 GB VRAM.

1-click installers for Windows, Linux, RunPod, Massed Compute and SimplePod support all GPUs including RTX 1000, 2000, 3000, 4000 and 5000 with the fastest libraries: Torch 2.9.1, CUDA 13, Flash Attention, Sage Attention and xFormers. Downloads 30+ demo voices and separates vocals from music for full-song conversion.

Patreon exclusive posts index to find our scripts easily, Patreon scripts updates history to see which updates arrived to which scripts and amazing Patreon special generative scripts list that you can use in any of your task.

Join discord to get help, chat, discuss and also tell me your discord username to get your special rank : SECourses Discord

Please also Star, Watch and Fork our Stable Diffusion & Generative AI  GitHub repository and join our Reddit subreddit and follow me on LinkedIn (my real profile)

=======

Latest installer zip file : RVC_Installer_V6.zip

  • This app was requested so I found the most up to date fork and forked and improved it

    • Using this app is not trivial so look for guides and tutorials

  • I have downloaded pre-trained audio models from here and tested and works

  • There are 450 GB of pre-trained voice models pth files in above repo that you can use

    • Put those pth files into Q:\RVC_Installer_V1\RVC_Premium\assets\weights

    • Put those index files into G:\RVC_Installer_V1\RVC_Premium\assets\indices

    • You will see auto downloaded models so you will understand how it works

  • Our installer will auto download all of the necessary models for voice training and other tasks automatically

  • Our installer will also download over 30 demo voices

  • I don't have much experience with this app but you can find lots of tutorials to learn more about it and this is the most up to date version forked from here : https://github.com/fumiama/Retrieval-based-Voice-Conversion-WebUI

  • Our installer will generate a Python 3.10 venv and install all the libraries inside it

    • Torch 2.9.1 with CUDA 13, Flash Attention, Sage Attention, xFormers, compatible with all GPUs starting from RTX 1000 to RTX 5000 GPUs

  • We also have installers for RunPod, Massed Compute and SimplePod

  • This is a very advanced version of original RVC Webui Premium developed by SECourses

1 May 2026 V6

  • Some training related files were not auto downloaded and this issue fixed with newest zip file

  • Overwrite older files and run install bat file again

16 April 2026 V5.1

  • Training related errors fixed and tested

  • Zip file is same just run Windows_Install_and_Update.bat to update

10 April 2026 V5

  • Full tutorial video published that shows how to install and use on Windows, RunPod, SimplePod and Massed Compute : https://youtu.be/ZRrzvD4wNys

  • Upon request, merge voice feature implemented to generate custom mixed voices

  • Supports merging more than 2 voices with rates as well and including indexes (not mandatory)

  • Fairseq library installation was extremely hard on Windows therefore I hace pre-compiled it and now it is auto installed with support to all GPUs out ther

  • Download latest zip file, extract and overwrite all files and then Just run Windows_Install_and_Update.bat file

18 March 2026 V2

  • I have upgraded the app significantly

  • Now we are moved to newest Torch 2.9.1 and CUDA 13

  • Workin on all GPUs, Windows, Linux, RunPod, Massed Compute and SimplePod

  • Cloud installation tested and verified too

  • Now the installer will download 30+ pre-trained demo voices

  • Full preset save and load system implemented

  • Loop all models feature implemented thus you can loop and generate with all models and see how they work

  • Now all generations will be saved inside outputs folder with metadata

  • Now batch folder processing works perfect just give input and output folder path, it will process every file in given folder and save with same file name in output folder

  • Auto separate vocals + music and remix after conversion implemented thus with 1 click you can voice change full music audio files

  • Save as MP3 feature added

  • The parameters on the interface set to best as default

  • Uses below 4 GB VRAM and ultra fast

  • I have tested Windows_Start_Realtime_Voice_Changing_Desktop_GUI.bat and realtime voice change working too - taking speaking from your microphone and real time

  • For offline inference and training use Windows_Start_Voice_Change_And_Train_WebUI.bat

  • Command_To_Test_RunPod_SimplePod_GPU.txt added because out of 3 pods i rented, 2 had broken hardware so best to test

  • Make sure to select CUDA 12.9 and 13 filter in RunPod selection due to outdated NVIDIA drivers

  • Make a fresh installation with Windows_Install_and_Update.bat

Windows Requirements

Massed Compute (Recommend Cloud) :

RunPod (Cloud):

App Screenshots

Related posts

All posts in this topic: Voice, Text-to-Speech, Transcription and Music AI · New to SECourses? Start Here: Best and Most Popular SECourses Posts

Files

Comments

Nerdlicht

Wow, super cool! Thank you so much for the effort of providing an updated and improved RVC app! <3 But... QWEN3-TTS has 'Voice Design'! :D It would be absolutely amazing if we had a QWEN3-TTS Suite Gradio app. Or even better: a trainer for LTX2 Character LoRAs that we could then use with LTX2 and QWEN3-TTS in a single pipeline. Then you could generate a video of any character with their original (or a different) voice and even have granular control over how the voice speaks and sounds. Mind-blowing.

SECourses

ye i am working on those. currently working ltx2 presets + audio to video feature in swarmui

Mig Test

FWIW, I had to turn on "developer mode" in system settings on Win 11 to get this to install properly. I didn't see that mentioned anywhere. Personally, I'd feel better if I didn't have to install it that way, but in any case, I hope this comment helps someone out. The that I got (and resolved) was this: build_meta.py", line 313, in run_setup exec(code, locals()) File "", line 246, in OSError: [WinError 1314] A required privilege is not held by the client: '..\\examples' -> 'fairseq\\examples'

SECourses

no such thing needed. you just need to follow requirements tutorial and i am using windows 11 too : https://youtu.be/DrhUHnYfwC0 do not run anything as administrator just double click

Richard Nagy

Cool, thanks! But I miss one thing in the description: What is this stuff? 😊

Ahmet Inceelli

I got this error below when I try to install. Please help. 😸 WARNING. For this auto installer to work you need to have installed Python 3.10.11, Git, FFmpeg, cuDNN 9.17+, CUDA 13.0, MSVC and C++ tools This tutorial shows all step by step : https://youtu.be/DrhUHnYfwC0?si=UAAVyZ8_QUPAjy7a Cloning into 'RVC_Premium'... remote: Enumerating objects: 218, done. remote: Counting objects: 100% (218/218), done. remote: Compressing objects: 100% (181/181), done. remote: Total 218 (delta 28), reused 216 (delta 28), pack-reused 0 (from 0) Receiving objects: 100% (218/218), 994.55 KiB | 742.00 KiB/s, done. Resolving deltas: 100% (28/28), done. HEAD is now at 0a55fa4 2.0 Already up to date. Python launcher is available. Generating Python 3.10 VENV Requirement already satisfied: pip in d:\rvc\rvc_premium\venv\lib\site-packages (23.0.1) Collecting pip Using cached pip-26.0.1-py3-none-any.whl (1.8 MB) Installing collected packages: pip Attempting uninstall: pip Found existing installation: pip 23.0.1 Uninstalling pip-23.0.1: Successfully uninstalled pip-23.0.1 Successfully installed pip-26.0.1 Collecting uv Downloading uv-0.11.4-py3-none-win_amd64.whl.metadata (12 kB) Downloading uv-0.11.4-py3-none-win_amd64.whl (25.2 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 25.2/25.2 MB 6.1 MB/s 0:00:04 Installing collected packages: uv Successfully installed uv-0.11.4 Using Python 3.10.11 environment at: RVC_Premium\venv Updating https://github.com/One-sixth/fairseq.git (HEAD) x Failed to download and build `fairseq @ git+https://github.com/One-sixth/fairseq.git` |-> Git operation failed `-> process didn't exit successfully: `C:\Program Files\Git\cmd\git.exe submodule update --recursive --init` (exit code: 128) --- stderr Submodule 'fairseq/model_parallel/megatron' (https://github.com/ngoyal2707/Megatron-LM) registered for path 'fairseq/model_parallel/megatron' Cloning into 'C:/Users/Sep/AppData/Local/uv/cache/git-v0/checkouts/debb76fe88cbae63/44800430/fairseq/model_parallel/megatron'... fatal: unable to access 'https://github.com/ngoyal2707/Megatron-LM/': Failed to connect to github.com port 443 after 21111 ms: Could not connect to server fatal: Fetched in submodule path 'fairseq/model_parallel/megatron', but it did not contain adb23324c222aad0aad89308e70302d996a5eaeb. Direct fetching of that commit failed. Traceback (most recent call last): File "D:\RVC\Download_RVC_Models.py", line 1, in from huggingface_hub import HfApi, list_repo_files, hf_hub_url, get_hf_file_metadata ModuleNotFoundError: No module named 'huggingface_hub' Press any key to continue . . .

SECourses

hello. please use v4 i have pre compiled fraiseq for you. get v4 zip file overwrite all and install and let me know.

Heisenberg

When using the training section, the first two steps work well except the training itself. It fails with this error: RuntimeError: makeDeviceForHostname(): unsupported gloo device on RVC_Premium\infer\modules\train\train.py", line 135, in run The main workaround of posts online seem to be downgrading torch to 2.7.1. The version installed in the environment is pytorch 2.9.1+cu130. I've also tried modifying the train script to use ncll, but it fails with RuntimeError: Distributed package doesn't have NCCL built in I guess I need to keep tweaking around and testing stuff to make it work, but flagging in case you know a quick workaround that doesn't require messing around too much or downgrading torch. Thanks

SECourses

hello. sorry for late reply. i just made a fix. please run Windows_Install_and_Update.bat to update . zip file is same. if you get error please email me full error logs : monstermmorpg@gmail.com . make sure that you see 5.1 on gradio top left

Heisenberg

No worries, thanks for checking that out. The update went well (5.1), though the training this time was failing with the following error FileNotFoundError: [WinError 3] The system cannot find the path specified: 'C:\Users\alexi\Retrieval-based-Voice-Conversion-WebUI-main/logs/mute/0_gt_wavs/mute48k.wav' My directory is different, this is from a thread here: https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/issues/2176 I've solved locally by putting the zipped archive including that wav that one of the users shared here: https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/issues/557, into the /logs/ folder. My understanding from what they say is that it's supposed to manage silence in the training? Anyway, not sure if it's the proper fix, but I was able to complete a couple of trainings (EPOCH 500 and 800) with good results. I don't notice any irregularities in the output, but I can send a full log to your eMail during the weekend in case it's useful for bug fixing. Thanks!

JAV Pomoe

Hello, Doctor. I'm getting the following error during model training. It seems like some files are missing. There were no errors during installation. D:\RVC_Installer_V5\RVC_Premium\venv\lib\site-packages\fairseq\modules\multihead_attention.py:15: FutureWarning: xformers.components is deprecated and is not maintained anymore. It might be removed in a future version of xFormers from xformers.components.attention import build_attention WARNING:root:Using insecure weight loading for fairseq dictionary DEBUG:faiss.loader:Environment variable FAISS_OPT_LEVEL is not set, so let's pick the instruction set according to the current CPU INFO:faiss.loader:Loading faiss with AVX2 support. INFO:faiss.loader:Successfully loaded faiss with AVX2 support. D:\RVC_Installer_V5\RVC_Premium\venv\lib\site-packages\fairseq\modules\multihead_attention.py:15: FutureWarning: xformers.components is deprecated and is not maintained anymore. It might be removed in a future version of xFormers from xformers.components.attention import build_attention WARNING:root:Using insecure weight loading for fairseq dictionary DEBUG:faiss.loader:Environment variable FAISS_OPT_LEVEL is not set, so let's pick the instruction set according to the current CPU INFO:faiss.loader:Loading faiss with AVX2 support. INFO:faiss.loader:Successfully loaded faiss with AVX2 support. INFO:000sizuha:{'data': {'filter_length': 2048, 'hop_length': 480, 'max_wav_value': 32768.0, 'mel_fmax': None, 'mel_fmin': 0.0, 'n_mel_channels': 128, 'sampling_rate': 48000, 'win_length': 2048, 'training_files': './logs\\000sizuha/filelist.txt'}, 'model': {'filter_channels': 768, 'gin_channels': 256, 'hidden_channels': 192, 'inter_channels': 192, 'kernel_size': 3, 'n_heads': 2, 'n_layers': 6, 'p_dropout': 0, 'resblock': '1', 'resblock_dilation_sizes': [[1, 3, 5], [1, 3, 5], [1, 3, 5]], 'resblock_kernel_sizes': [3, 7, 11], 'spk_embed_dim': 109, 'upsample_initial_channel': 512, 'upsample_kernel_sizes': [24, 20, 4, 4], 'upsample_rates': [12, 10, 2, 2], 'use_spectral_norm': False}, 'train': {'batch_size': 1, 'betas': [0.8, 0.99], 'c_kl': 1.0, 'c_mel': 45, 'epochs': 20000, 'eps': 1e-09, 'fp16_run': True, 'init_lr_ratio': 1, 'learning_rate': 0.0001, 'log_interval': 200, 'lr_decay': 0.999875, 'seed': 1234, 'segment_size': 17280, 'warmup_epochs': 0}, 'model_dir': './logs\\000sizuha', 'experiment_dir': './logs\\000sizuha', 'save_every_epoch': 5, 'name': '000sizuha', 'total_epoch': 20, 'pretrainG': 'assets/pretrained_v2/f0G48k.pth', 'pretrainD': 'assets/pretrained_v2/f0D48k.pth', 'version': 'v2', 'gpus': '0', 'sample_rate': '48k', 'if_f0': 1, 'if_latest': 0, 'save_every_weights': '0', 'if_cache_data_in_gpu': 1, 'author': ''} Process Process-1: Traceback (most recent call last): File "C:\Python310\lib\multiprocessing\process.py", line 314, in _bootstrap self.run() File "C:\Python310\lib\multiprocessing\process.py", line 108, in run self._target(*self._args, **self._kwargs) File "D:\RVC_Installer_V5\RVC_Premium\infer\modules\train\train.py", line 170, in run train_dataset = TextAudioLoaderMultiNSFsid(hps.data.training_files, hps.data) File "D:\RVC_Installer_V5\RVC_Premium\infer\lib\train\data_utils.py", line 32, in __init__ self._filter() File "D:\RVC_Installer_V5\RVC_Premium\infer\lib\train\data_utils.py", line 46, in _filter lengths.append(os.path.getsize(audiopath) // (3 * self.hop_length)) File "C:\Python310\lib\genericpath.py", line 50, in getsize return os.stat(filename).st_size FileNotFoundError: [WinError 3] 指定されたパスが見つかりません。: 'D:\\RVC_Installer_V5\\RVC_Premium/logs/mute/0_gt_wavs/mute48k.wav'

SECourses

thanks for report. with v6 zip file now those missing files will be auto downloaded . get latest zip file, overwrite previous and run installer again and let me know