✨│ai-help

[next geyser]

windows 11

no tut link

[rigid tartan]

thats whats happening to me so ill answer this too
Nvidia RTX 3050
Windows 11
Roleplay in games and vc
https://www.youtube.com/watch?v=SxdnGxicJOg&t=673s&pp=ygUSdm9pY2UgY2hhbmdlciBnaXJs

also i deleted the vc to find if i downloaded the wrong one

voice changer*

would be very helpful if somene guided me though everything

[toxic isle]

Hey anyone i downloaded VBCable and when i downloaded the driver i cant even hear anything like yt i only hear people on discord and i also cant talk

[next geyser]

use vac cable lite ver

[toxic isle]

wdym

im new to this

i dont understand anything

[next geyser]

https://www.youtube.com/watch?v=SxdnGxicJOg&t=673s&pp=ygUSdm9pY2UgY2hhbmdlciBnaXJs

watch this

[toxic isle]

i downloaded this

when i was on the step

with the vbcable

i downloaded driver like him

it told me to restart

and i couldnt hear anytinh

[next geyser]

screenshot

rvc

and make sure ur actual mic works

[toxic isle]

what is rvc

it does

it was working

[next geyser]

voice changer client

[toxic isle]

and after i downloaded rvc it doesnt

[rigid tartan]

also happened the same with me after all the config when i tried to use in discord there was no sound

[next geyser]

is it started

[toxic isle]

no

[next geyser]

ok start it

[toxic isle]

i did

[next geyser]

and check in discord

with mic

[toxic isle]

it got delay asf

[next geyser]

set it as cable

[toxic isle]

how

[next geyser]

yea no shit get the settings right

for ur gpu

thats some dog settigns

[toxic isle]

lol

[next geyser]

https://docs.aihub.gg/

🍏 Applio Guide

‎ Deiteris' W-okada Guide

‎ Vonovox Guide

which one for vc male voice to male voice rp

[toxic isle]

so how to do it

so it doesnt have delay

[next geyser]

bro just look in the link for a settings for gpu tab

i cant hold your hand through everything you lazy bum

[toxic isle]

😭

you can

[next geyser]

hell na

[toxic isle]

nega you think i understand those things

[next geyser]

😂

[toxic isle]

thats some fucking high levels for me

[next geyser]

then this isnt for you

go back to playing in the sand

[toxic isle]

sybau dog shit

[next geyser]

just find the settings tab in the link i gave you

it isnt that hard

[next geyser]

u found?

[next geyser]

🍏 Applio Guide
‎ Deiteris' W-okada Guide
‎ Vonovox Guide
which one for vc male voice to male voice rp

[knotty moth]

bro left

[candid rune]

guys when i try and put meta data file in the index thing in the voicechanger it doesnt work can someone help asap

i cant send images

but it says

extonsion of file something somethng

if someone helps ill buy them nitor

<@&1159293204038955078>

[thick ferry]

send a ss gng

[candid rune]

@thick ferry

[thick ferry]

whats the extension of the file

ur trying to put

[candid rune]

metadata.json in index

[thick ferry]

🙏🏻

what are all the files in the folder

name them

[candid rune]

just that and model.pth

[thick ferry]

ok then the voice changer has no index

[candid rune]

but when i try and like

use it without index its not working

@thick ferry

[thick ferry]

js use a different vc

[candid rune]

i need this one in specific

and its used by 100s of people

its not broken

[thick ferry]

can u send the file

[candid rune]

https://www.weights.com/models/clxt6tdhw00316h2e7vaz1bpv

[thick ferry]

shows a index for me

[candid rune]

so what od i do

reinstall it?

[thick ferry]

did u extract the folder

[candid rune]

yesd @thick ferry

[latent kettle]

shoul i use 40khz or 30khz for sample rate

[thick ferry]

hm thats weird

it doesnt show in my file explorer

js my w-okada

[candid rune]

yea

so what do i do

@thick ferry how do i make it show up in my w-okada

or just send me the index file

[thick ferry]

let me try if it works

1 second

[candid rune]

okay thank you

[latent kettle]

@viscid moss help me please

[analog obsidian]

32k

[latent kettle]

i downloaded the files with 20khz that means it should be 40khz but when i extracted it using UVR it went to 15khz idk whyyy😭

[analog obsidian]

some uvr models apply a frequency cutoff of 15k

like uvr de echo

mel roformer models doesnt do that

[rigid tartan]

when i try to download the voice changer files it says this

no its not my network and i tried changin browsers

ts is not helping

[merry forge]

how do i fix this? RuntimeError: CUDA error: no kernel image is available for execution on the device
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

im very new to ts so sorry if its a dumb question

[dark depot]

you're so pretty shady

[viral mason]

[upper wedge]

How many epoch does it normally take to train a model? i'm using replay now (im now at 9 epoch)

[floral quiver]

i need help

its about my quality with the client and it sounds supper chunky any suddgestions?

[simple ore]

hard to say without the whole error stack. 1) out of vram 2) incompatible device

[merry forge]

i just bought a rtx 5060 TI so i dont think im out of VRAM

weights worked before without a dedicated graphics card

[simple ore]

so what program is failing?

[merry forge]

applio

[simple ore]

https://github.com/IAHispano/Applio/issues/1020

[merry forge]

so i js run env\python -m pip install torch==2.7.0 torchvision==0.22.0 torchaudio==2.7.0 --upgrade --index-url https://download.pytorch.org/whl/cu128 ?

[simple ore]

from Applio's folder

you can even run it as X:\Applio\env\python -m pip install torch==2.7.0 torchvision==0.22.0 torchaudio==2.7.0 --upgrade --index-url https://download.pytorch.org/whl/cu128

just change the path to env

[merry forge]

running it rn

[merry forge]

omg tysm

[floral quiver]

so can i still get help?

[thorny plaza]

Oh got dang it, now i see "CABLE input (VB audio virtual cable)" facepalms, if you still have that problem, try using that one. Could be it. 🫡

Edit, or it could had been discord-browser setting, scroll bit down from this message.

If you had w-okada tool downloaded and set up, now you can change your voice real time based of what you say to your mic, or out of sound file you load into the software. W-okada by default have few voices included, but if you want more voices, you can use this site MODELS section to download new voices from huggingface & weights.com. After you downloaded model you click this in w-okada (1st pic) then upload (2nd pic) and then you select .pth file for model and .index file for index and click upload, and that's it, you got new model to use 🤗 (3rd pic). You can download any picture from internet and if you click here (4th pic) (after uploading the model) you can give your new sound an visual icon for easier navigation in future when you will have couple dozens of models 🫡

[trail spruce]

im new to this ai thing

how do u make it sound like the person u want it to be?

[thorny plaza]

w-okada tool give me sec

Wokada will do you both real time and from a file (voice recording)(those links are to the newer wokada fork)
"How to use
Running locally on Windows
Before you start

1 [If not installed] Download and install 7-Zip or WinRAR. (https://www.7-zip.org/)

2 [If not installed] Download and install VAC Lite by Muzychenko. (https://software.muzychenko.net/freeware/vac470lite.zip) (or the https://vb-audio.com/Voicemeeter/banana.htm)

3 Navigate to the releases section." (https://github.com/deiteris/voice-changer/releases)"
And then you can use this discord (⁠🎧│voice-models) to get voice models, or use online tool like (https://colab.research.google.com/github/IAHispano/Applio/blob/main/assets/Applio_NoUI.ipynb#scrollTo=0pKllbPyK_BC) to make your very own voices (easy to use, no set up required) or you can install appolio locally (https://docs.aihub.gg/rvc/local/applio/) but that might not be easy for someone who is green

Whats above is for speeding up process, but you might want to read things, if so, go with link under (i realized not everybody has windows 😅 )
.
.
Here link for the site https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/

[rigid tartan]

No like

The virtual cable

Wont show on discord

[thorny plaza]

Hmm

as i dont use VAC by muzychenko, i can't really help, i don't know why it does not work. That is what i got, and how its set up, both set as default devices, microphone is as input (in wokada) and B1 as output and it works for me

The out: is what the PC should "hear" and the mon (monitor) is not necessary to be enabled (as it will make you hear the changed voice with second of delay - very distracting)

You need to set virtual cable out put to be default comunication device, for system to by default use it in all aplications, instead of microphone

Did you set that up with browser?

[cursive gyro]

h

[thorny plaza]

Check your browser microphon setting

this is very likely the cause why the tool is not working for you

click this

[low shard]

!howtoask

[vernal bone]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1405252622734065816> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[rare folio]

Hi! Okada suddenly stopped working for me at all today, it initializes properly but doesn't register my voice at all

[rare folio]

<@&1159293204038955078>

[cold shoal]

Hey I'm on bandlab and I'm trying to add the beat to my track can someone help me

[low shard]

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • E girl trolling
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[blazing panther]

anyone know why audio picked up on my pc is relaying to w-okada?

[lethal basin]

nick, which one am i suppose to download if i dont have amd?

[low shard]

you need to elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • E girl trolling
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[lethal basin]

uhh i use a nvidia rtx 3060
windows 11

which one do i download

😭

[gloomy lynx]

Is there a way to use AICoverGen locally? I feel like I've had way more luck with it than trying to separate vocals manually, but Idk if I'd risk running cloud AI via Drive, IIRC Google's not very happy with using it for that

[low shard]

AICoverGen is outdated
elaborate your pc gpu and windows version

[lethal basin]

also wait the ss

[low shard]

this is a general server, elaborate everything:

[lethal basin]

roleplay

and tutorial link im using is ur forum for realtime

[low shard]

everything, not just a bit:

• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • E girl trolling
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

that forum has multiple links for different guides for different programs 😭

[gloomy lynx]

Windows 11, Geforce 3070 Ti (laptop version with 6 GB VRAM); doesn't run a whole lot AI-wise as far as I'm aware

[lethal basin]

https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork

[low shard]

for ai covers, you'll be fine locally:
check https://docs.aihub.gg/rvc/local/aicovermaker/

this is only for making ai covers btw

[gloomy lynx]

In theory UVR5 runs relatively quick on it, but a lot of the time the isolation just isn't really good

[low shard]

you shouldn't be on github at all, don't click the first links in the introduction, read to the VAC Lite part and windows nvidia (normal, not rtx 50 serie) part

[low shard]

AICoverMaker automatically separates (like UVR5) and processes and mixes all together

it's the easiest automatic u can run locally

[lethal basin]

is there any more recent one?

[low shard]

yes that one, that's the latest one

[lethal basin]

okay, is there any like uhh setitngs i should disable? if i wanna use it for games?

[gloomy lynx]

Like AICoverGen yeah, for some reason IIRC I had a better success rate just letting it do all the work, even if I ran the whole line of instrumental -> de-reverb -> de-noise on UVR5

[gloomy lynx]

And yeah I know, would probably still need to stick with Applio for something like training a model

[low shard]

yeah aicovermaker is more updated, id suggest it over aicovergen

[low shard]

yup, and know that 6gb vram is good for ai covers, but not that much for training

[low shard]

wdym? which settings are you talking about? you downloaded it already and did everything?

[lethal basin]

uhh for the uh voice changer, ill tkae a ss and show u when it runs

expect it wont run

😡

[gloomy lynx]

Tends not to be much for anything AI-related, but I dunno much of an alternative, not sure if Google still cracks down on using its thing for AI training

[low shard]

be sure to follow the guide for how to run it

[lethal basin]

yes, is there any way to not make it open in ym broswer?

[low shard]

no https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/#why-does-it-run-in-a-browser-and-not-its-own-window

[lethal basin]

okay

[low shard]

if you're talking about google colab, the T4 free gpu would have more vram which is good, but it has limited free time, kaggle would be better as a cloud alternative

[gloomy lynx]

Not sure why, but I remember having a really bad time trying to set up Kaggle

[low shard]

it is harder and requires phone number, it offers more things for free

else there's lightning.ai, which is also another great alternative but still requires phone number

[gloomy lynx]

I'd be down for either, if there's guides/resources that use those anywhere

[low shard]

https://docs.aihub.gg got everything

let me know ofc

[low shard]

lmk

[lethal basin]

for these settings, is there anything i should do to make it sound better

[gloomy lynx]

Oh wait, no wonder I couldn't find these anywhere, I was looking at outdated docs I think

I was using this link https://ai-hub-docs.vercel.app/

[low shard]

i forgot i hosted that for over a year in vercel when the docs were temporarely down..

i wonder how u even found that link

anyways ima take it down

[gloomy lynx]

Good question, you would've had to ask me the first time I found it

[lethal basin]

nick, why doesnt my input and output not show anything, and if i select any of the options it jst crahses

[low shard]

crazy uptime lol

never expected so many things would change in 10 months

[low shard]

are u trying to do e girl trolling by ur models?

[valid violet]

#1406770205211889825 im looking for help

nick

[lethal basin]

I mean sure, i stated i jst liked to roleplay

[valid violet]

like fivem?

[gloomy lynx]

Just to be sure about Kaggle, it just needs a phone number, and the 30 free hours are presumably a one-time thing, right?

[low shard]

so like e girl trolling or roleplay? or both?

[lethal basin]

perferably id only do it for roleplay

[gloomy lynx]

Don't think I'd need more than thirty, but I should have ways to deal with that, just making sure

[lethal basin]

but how do i set my mic and headset?

[low shard]

great then, as e girl trolling is catfishing and banned

[low shard]

what browser u using? u gave it microphone perms?

[low shard]

30 hours for free weekly

[lethal basin]

yes my mic works, on discord, and other stuff but ill check again

[lethal basin]

ohlol

[gloomy lynx]

Oh, weekly sounds wild, that's definitely gonna be way more than enough

[low shard]

they don't stack over eachother btw

[gloomy lynx]

Yeah that's fine

I should probably get around to updating my model, current one sorta works? I guess? But IIRC it was Colab training that shut off in the middle of it

[lethal basin]

oh weirdly it was blocked

[low shard]

and also you technically shouldn't use web uis in neither kaggle nor google colab free tier, they didn't get detected as the code is encrypted but if they would ever, using it could give u trouble, which is stated in the docs glossary
lightning.ai is the only free tier that allows web uis

lightning.ai offers around 80 hours freely monthly, i mean depends on how u spend ur credits and which provider, u can even get extremely good gpus like the best for free but with fewer time

[low shard]

that explains

[lethal basin]

2 questions first why cant i change any of the advanance settings
secondly do u prefer server or client?

[low shard]

first why cant i change any of the advanance settings
did you click stop first?
secondly do u prefer server or client?
that depends on ur needs

• Client: easy to use, can use echo, sup1, sup2 cancellation
• Server: harder to use, can significally reduce delay if used with Wasapi or Asio

[gloomy lynx]

I might give that a go yeah, but a 2-3 day verification seems spooky

[low shard]

i mean, it's all ur choice

u got like 4 options to run applio: locally, colab, kaggle & lightning.ai

[lethal basin]

thanks, but i mean if i wanted to voice for talking while perferably having little to no voice breaks what would u do?

[low shard]

voice breaks? are u trying to scream or laugh? play on lowest graphics 1080p 60fps cap, show a screenshot of the program running

[lethal basin]

yeah like that, okay well ill try

[low shard]

rvc can't do always perfectly super realistic laughs, or especially screams

[lethal basin]

well thank you, u were amazing !!

[low shard]

you're welcome!
need any other help or is this solved?

[lethal basin]

check ur dms

😡

[gloomy lynx]

Okay yeah, one inference in CoverMaker and the result is immediately a hundred times better than half an hour of screwing around in UVR5

[blazing panther]

4060, windows 10, let me know some settings to use

using nvidia i believe, its for either e girl trolling or roleplay but its mostly just to mess around with fun voices

[timid zinc]

Hey, currently training a spin model for the first time and its very time consuming 😆

Is there any way to pause the training process and continue later?

[simple ore]

ctrl-c when it saves the 10th epoch

before it starts more steps for next epoch

[timid zinc]

And is it normal that it takes like 20-30min on average for each epoch ?
20min dataset with 12 batch size and a 2060ti SUPER

[simple ore]

no

unless you ran out of vram

[timid zinc]

So if i stop before, progress is gone ?

[simple ore]

yes

show task manager/performance/gpu

[timid zinc]

cant post screenshots here
but i have 8gb vram and it seems to be 100% utilized

[simple ore]

yeah, you probably spilled it into shared memory

and that kills the peformance

can lower batch size / use checkpointing

[timid zinc]

ohhh, is the batch size that affects vram usage ? or how do i calibrate the configs

[simple ore]

larger batch size, more memory it uses

[timid zinc]

ahhh i see the setting

people said the quality is better with batch size 8-12, so i figured lets go with 12

[simple ore]

you can probably manage with 8, for 12 you'll need to use checkpointing

[timid zinc]

whats the drawback of using checkpointing ?

[simple ore]

lil slower, but faster than using shared memory

[timid zinc]

understood tyvm ❤️

[low shard]

E girl trolling is catfishing, which is illegal. You have been warned both verbally and via sapphire and continuing to ask will lead to further actions possibly like a ban

[low shard]

how's it going btw? any issues?

-# btw u okay? i checked ur profile

[gloomy lynx]

I mean, it works better than trying to do it manually, that's for sure, I'll probably try Kaggle/Lightning whenever I feel like doing something related to training

As far as the latter, not the right place to talk about venty shit, so I'll pass on answering

[low shard]

well that's great, u can always ask here or make a #1192011222023950368

[low shard]

sorry, I just try to be nice and help people, i do know life can be hard

[gloomy lynx]

I mean, an AI server probably isn't the right place to go extensively into that sorta stuff, I'd imagine

[low shard]

btw i replied here

[low shard]

understandable 😭

this somehow reminds me that some people use chatbots for mental health and therapist

which isn't great as they are just predicting text and could just say some bad shi, like it happened already

it is ironic tho, because the first chatbot, Eliza, was made as a therapist chatbot

[gloomy lynx]

Already tried, they're ass; they suck up to you too much, constantly blame themselves when you don't get something, and completely refuse to say anything that someone might deem as being too mean/accusatory/etc.

Last time I tried, for some reason it got stuck constantly responding in a 2-point format

[timid zinc]

if you want to vent to a real stranger DM me
its 3am and im bored enough :P

[low shard]

don't trust chatbots for such delicate topics, they could either glaze you saying bullshit and not actually help you, or say bad things which makes you worse

it already happened that a kid took his life because his ai gf venting chatbot or smt told him to..
-# source

[gloomy lynx]

Already had both happen, probably not to that much of an extreme, though

[low shard]

I'd really suggest you to vent to vent to someone's close to you, such as a family member, a friend you trust, a partner, or whoever you feel comfortable with irl/online that think they can help you

[low shard]

sorry for the question but what chatbot did u try 😭 c.ai? gemini?
i thought they kinda made their security for atleast the 2nd thing to not happen

[gloomy lynx]

Don't have anyone I'd trust or feel comfortable with for that sorta thing ngl

[gloomy lynx]

Basically all the big LLMs, local ones seem ass and even if they weren't, I'm sure not running them on 6 GB VRAM

[low shard]

btw u could try ollama which uses both ur normal ram and vram (iirc) and try running something like gemma3n or qwen3 which got pretty good small models

ofc do NOT do it for those topics tho

[gloomy lynx]

Haven't had luck with using that

[low shard]

I mean, I do understand that, but just know to not like keep it urself and die inside, u could feel worse overtime yk, sometimes spitting it out makes u feel better

but hey, everyone's different

[gloomy lynx]

Not gonna feel any worse than I have for ages now, but that's getting into venty shit and, again, this isn't the right time or place

[low shard]

weird, maybe try lm studio which has a warning for which models and which ur hardware can?

i don't like that its GUI isn't open source tho, only the cli,
personally

oh right, u could also try GGUF model files, like Q4_K_M, which are efficient file formats

[gloomy lynx]

Already tried both LM Studio and GGUF models too, local textgen just doesn't really work with this setup

[low shard]

but yeah u can deffo run some nice for its size models locally, i did it even on my phone (ofc the ones on my phone is way worse than my rtx 4060 ti 16gb, but u get what i mean)

[low shard]

that's weird, may i ask a specific model u tried?

[gloomy lynx]

I couldn't remember

[low shard]

i literally ran qwen3:0.6b_Q4_K_M yesterday on my phone cpu via termux and gpt-oss 20b on my rtx 4060 ti 16gb (yes i test random shit for fun lol), maybe u tried too big models

anyways, goodluck with ur life and i hope u get better
feel free to ask any ai help here, like for llms or if applio gives issues

[sleek burrow]

hi

[granite spruce]

How do I train my own model?

[blazing panther]

i don’t do that, but you asked what i wanted to do and i don’t have a specific reason but that’s the closest thing

[timid zinc]

How do I know my model is finished training?
Im at epoch 200 now and the loss/g/total lowest point was at epoch 130 so far.
Does this mean i can stop now and after 130 its been overtraining?
The value doesnt really go up or down by a lot anymore

Using spin embedder, if thats relevant

[minor isle]

how to fix okada repeating peoples voice from headset

change it to gpu usage in the menu

its under chunks

[minor isle]

anyone?

[lone vale]

C:\Users\saudi\Downloads\ai voice changer\MMVCServerSIO>MMVCServerSIO.exe -p 18888 --https false --content_vec_500 pretrain/checkpoint_best_legacy_500.pt --content_vec_500_onnx pretrain/content_vec_500.onnx --content_vec_500_onnx_on true --hubert_base pretrain/hubert_base.pt --hubert_base_jp pretrain/rinna_hubert_base_jp.pt --hubert_soft pretrain/hubert/hubert-soft-0d54a1f4.pt --nsf_hifigan pretrain/nsf_hifigan/model --crepe_onnx_full pretrain/crepe_onnx_full.onnx --crepe_onnx_tiny pretrain/crepe_onnx_tiny.onnx --rmvpe pretrain/rmvpe.pt --model_dir model_dir --samples samples.json
Booting PHASE :main
PYTHON:3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)]
Activating the Voice Changer.
[Voice Changer] download sample catalog. samples_0004_t.json
[Voice Changer] download sample catalog. samples_0004_o.json
[Voice Changer] download sample catalog. samples_0004_d.json
[Voice Changer] model_dir is already exists. skip download samples.
Internal_Port:18888
protocol: HTTP
-- ---- --
Please open the following URL in your browser.
http://<IP>:<PORT>/
In many cases, it will launch when you access any of the following URLs.

[VCClient] Access http://127.0.0.1:18888/
[VCClient] wait web server...0 http://127.0.0.1:18888/

It's just stuck like that

ive been waiting for 20 minutes

It used to work but I had deleted it a few months back

[floral quiver]

need help

Hot do i make my ai sound smooth?

[queen osprey]

how to fix the voice changer not working like its not talking

[candid rune]

where is the lastest w okada version

[rich brook]

i have a problem with google collabs w/okada voice changer

after start a first step it is written here "Restart session. The following packages were previously imported in this runtime"

but when I click reset the GPU can't reconnect

any solution?

[glossy ingot]

real

[flint vapor]

Which is your voice changer? And which is your hardware?

[flint vapor]

-wokada

[vernal bone]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[flint vapor]

https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/ Follow this guide 😄

[queen osprey]

w-okada and i have a 3050

[flint vapor]

Do you have deiteris fork of okada or an old version?

[queen osprey]

how do i check

[flint vapor]

You have on right up the name of the voice changer version

[queen osprey]

so i open the voicechanger and se

[flint vapor]

Yes 😄

[queen osprey]

v1.5.3.18a

[flint vapor]

Ok you have mainline outdated okada, i will give you new version and a guide, you have to download version for nvidia, not 5000 series

-wokada

[vernal bone]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[flint vapor]

https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/

Follow this 😄

Uninstall the current version you have now

[queen osprey]

but for some reason when i first used it it worked

i didnt change anything big

[flint vapor]

Oh but the mainline okada is not so much updated and can have some issues

[flint vapor]

On voice changer?

[flint vapor]

However if i can ask, which is your main purpose for using it? E-trolling? Catfishing? Testing? Roleplay? or other

[queen osprey]

idk tbh it wasnt really trolling

i was wanted to sound like ichigp

[flint vapor]

Ohh ok ok, however all you need is written in the guide, but if you need more help or something don't work i'm always here for help:D

[queen osprey]

where do i install

[queen osprey]

what do i click so i can install

[flint vapor]

Wait i give you link

[flint vapor]

https://huggingface.co/Shadicti/deiteris-Fork/resolve/main/voice-changer-windows-nvidia-b2332.zip?download=true

[queen osprey]

so after i unzip

do i put the model and call it a day or do i need to install that cable again

[flint vapor]

You installed before vac or vb cable?

[queen osprey]

https://vb-audio.com/Cable/

this one the one i used

[flint vapor]

Ohh you have to uninstall this

https://software.muzychenko.net/freeware/vac470lite.zip

Install this

[glossy ingot]

anyway to make the voice changer sound smoother and shi?

[queen osprey]

it wont uninstall

[flint vapor]

You can uninstall the driver

[flint vapor]

Which voice changer are you using

[queen osprey]

can i just keep it

[flint vapor]

Open Device Manager. Expand the "Sound..." device category by clicking on the "+" sign. Right-click the Virtual Audio Cable device and choose Uninstall.

[flint vapor]

It can causes some conflicts

[lone vale]

@humble dust its just stuck like this for some reason

i dont know what im doing wrong

[glossy ingot]

not sure

what its name is

just says voice changer native

[queen osprey]

@flint vapor it finished downloading now how do i open it and can you show me how to set it up

[flint vapor]

Sorry for late

Have you run the .exe after unzipt it?

Hi, didn't you tried also other models of killua?

[flint vapor]

Do you have series of the version?

on right up

[queen osprey]

yea got it working

also is there a english dub ichig

[flint vapor]

Ohh i really don't know you can try search it on weights.gg

[flint vapor]

https://www.weights.com

[high spoke]

yes

[low shard]

E girl trolling is catfishing, which is illegal. You have been warned both verbally and via sapphire and continuing to ask will lead to further actions possibly like a ban

[viral mason]

there should be a wall of shame for those kind of users to be displayed on

[elfin nebula]

When voice training an ai model, would it be better to use a word list with all the english phonetic sounds?

[autumn orchid]

Hey, I was testing out the Albert / Flamingo (V4) (RVC V2) (400 Epochs) model, and what I heard from the MP3s he sent sounds completely different from what I have. On my end, the voice changer sounds like the "ai/model" has rocks in its mouth. No offense to the person who made it, but either way, it's an issue on my end, and I was wondering if an admin could help me out.

[simple ore]

to say plainly, a voice of someone reading from a dictionary is better than 100 hours or someone saying 'Hello' non-stop

[young halo]

i'm using kaggle for training but i can't do anything because of this ModuleNotFoundError: No module named 'gradio'

[simple ore]

that means gradio library somehow did not get installed (and probbaly some others)

[young halo]

how can i fix that?

[simple ore]

it would depends on which notebook you're using

[young halo]

how can i see about that? sorry if there are too many questions

[viral mason]

they're asking if you're using like applio or something else

[young halo]

yeah, i'm using the applio notebook on kaggle

i tried re running and it didn't work

[simple ore]

in kaggle create a new notebook, then file/import notebook,

as I see it starts just fine

[still holly]

can anyone tell me some good settings?

[young halo]

i get the same error

it's weird, my training was going just fine

[simple ore]

did you run install cell? 🙂

I just tried importing the notebook code as explained and it was fine

[young halo]

yep

Unless i didn't understand properly 😭

I installed the Notebook

And then re runed the one i was using

But i don't know if that's what i had to do

[simple ore]

go to kaggle, +Create, Notebook

File/Import Notebook -> screenshot I gave you

[young halo]

Done

[simple ore]

after that you can rename your notebook so you know what to use next time

[young halo]

So I'll do my work here?

[simple ore]

you should not see this screen

if you follow my instructions

DO NOT DO +CREATE/IMPORT

because it immediately runs something

or you can make a copy of https://www.kaggle.com/code/aznamir/applio-latest

[young halo]

Idk what I'm doing wrong 🙁

I never had this issue before

Thanks for the help though

[primal anchor]

who knows why custom voice models arent working on okada voice changer?

[simple ore]

click +Create, select Notebook (not import notebook)

it opens a blank notebook

use File/Import Notebook

[sand heart]

How to download voice changer on my mac air m3 apple silicon chip

[young halo]

Okay, i'll try

[sand heart]

How to download voice changer on my mac air m3 apple silicon chip

??

[young halo]

Like this?

[glossy ingot]

sum help me download w-okada

[young halo]

It's working, thank you @simple ore !

[young halo]

This froze, the tensorboard and applio show the training, but here it's frozen

[low shard]

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • E Girl trolling
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

voice changer is too vague

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[timid fox]

I keep changing my mic and output in okada voice changer but none are working

[dusk jewel]

my mic dont work when i use the Voice Changer, why?

[patent gazelle]

keeps crashing

[prisma grove]

hey, I wanna train some models but I don't know which pretrain to use

I know

but I want to know which pretrain specifically I should use

I want to do 2 character voices, one has a dataset of about 1 hour with very clear but repetetive japanese singing audio, and the 2nd one is of a character that has an entirely computer-generated voice so it sounds very muffled and noisy, the dataset has about 30 minutes of mostly japanese speech and some singing

should I go with TITAN? KLM? Snowie?

I know, I've read the guidelines

[viral mason]

Original is best

[prisma grove]

original?

[viral mason]

Mhm

[prisma grove]

as in, the default one?

[viral mason]

yup

[prisma grove]

okay, thanks

[viral mason]

no problem!

[low shard]

!howtoask

[vernal bone]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1405252622734065816> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[low shard]

!howtoask

[vernal bone]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1405252622734065816> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[low shard]

!howtoask

[vernal bone]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1405252622734065816> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[viral mason]

NickBot9000

[low shard]

it is pretty hard to understand what someone means by only "not working" or "keeps crashing", that's why we got that command lol

[viral mason]

Ik I was just poking fun lol

[low shard]

[daring sail]

What should I use to install different apps that require different versions of Python and CUDA so that they don't conflict with each other?

[timid fox]

750 GTX TI
windows 11
Cant see to get mic and output working
https://www.youtube.com/watch?v=YGyUgvx1J_Y

[low shard]

that tutorial is outdated

delete everything

i don't think ur gpu meets the bare minimum for ai tasks

u still want to try locally? but i feel like ur gpu wont be recognized

[timid fox]

its ancient but can it do stuff?

[low shard]

i have a feeling it won't be recognized

what are u trying to do? roleplay in vc? e girl trolling? or roleplay in games?

[timid fox]

shits and giggles with my friedns?

it would be funny

[low shard]

in games or a vc?

[timid fox]

vc

[low shard]

less demanding on ur gpu for sure, but ehh not sure if it would be recognized, the bare minimum is a gtx 900

u still want to try locally? or want to use cloud?

[timid fox]

what does trying locally and cloud mean?

[low shard]

using something:

• locally: runs on your hardware, like running the software on your gpu
• cloud: remote good pc, running the software on a service that allows u to use way better gpus, with limited time and subscriptions, and as it's a realtime voice changer, ud need a good connection too for this

cloud would have better performance in ur case ofc, if ur gpu can even get recognized as it's super ancient

[timid fox]

i live in a third world country with 20/mbps max internet lol

[low shard]

😭 i mean u could use cloud but this would just add more delay

ofc u can't expect to run things with poor hardware and poor connection

u still want to try locally?

and then check cloud if it doesn't work?

[timid fox]

this seems too much work bro my normal voice is okay 🥀

[low shard]

yeah AI is complex and intensive lol

only products expensive made by companies are 1 click

[timid fox]

cloud would be faster, idw reinstall everything from scratch

[low shard]

but AI isn't 1 click at all

this is an open source program driven by the community

[timid fox]

so how do u use cloud?

[low shard]

About Cloud, there are different services:

• Kaggles (30 hours weekly of better GPUs, T4x2 & P100, harder to use, requires an account and a phone number, doesn't allow a Web User Interface for free it usually doesn't get detected but if it does you could get in trouble):
  • W-Okada's Deiteris' Fork Voice Changer Kaggle (the most free gpu hours)
• Lightning.AI (75 hours monthly of free T4 gpu, harder to use, requires an account and a phone number, allows web user interfaces) :
  • W-Okada's Deiteris' Fork Voice Changer Lightning.AI (the safest for free)
• Google Colabs (4 hours daily of free T4 gpu, easy to use, require only a google account) :
  • W-Okada's Deiteris' Fork Voice Changer Google Colab (currently works only on google colab PAID tier)

[low shard]

by choosing one of the cloud guides to read

[timid fox]

maybe one day...

thanks for the help though

[low shard]

i mean if u find trouble u can just ask there

[trail sleet]

How do I lower my input buffer in wokada

And output buffer

Sometimes my output is at 124s sometimes input is 124s

[unreal kettle]

what chunk values have u guys found that works good without being too delayed? using Deiteris w-okada fork for rtx 50 series (5070ti) paired with 7800x3d

[viral mason]

Try chunk at 360.0 and extra at 2.7

If that's not good for you lower the chunk a bit until the delay is good for you ^^

[unreal kettle]

ty for the suggestion, i tried it out and delay was a bit higher than id like so i changed chunk to 301.3 and extra to 3.7, and it seems a lil more tolerable for me

[viral mason]

You're welcome! Have fun with the voice changer ^^

[unreal kettle]

im using rtx voice as well for noise reduction since my mic picks up background noise a lot

[unreal kettle]

also i keep hearing myself even though i have no monitor device set and none of my mics have "listen to this device" enabled?

[viral mason]

I'm not sure how to fix that issue, you can ask one of the helpers or mods who may know more

[unreal kettle]

its all good, just messed around with the input and output until i didn't hear myself anymore

[mortal kite]

hel

p

[viral mason]

https://tenor.com/view/anya-forger-taking-notes-oomfie-spy-x-family-memo-gif-8287618218761733944

[rocky smelt]

Why when I use AI voice, sometimes it records and translates my voice well, sometimes it records and pronounces poorly, so it can't speak clearly.

[mortal kite]

Mine doesnt work

I get errors

But i can't send 9mgs

Immages

[glossy ingot]

real

[fallow lava]

Since Okada is no longer getting updated. Which would you recommend that's on the lower end when it comes to specs requirement?

I tried Vonovox yet the results are inconsistant

[dusk jewel]

!howtoask

[vernal bone]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1405252622734065816> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[dusk jewel]

GPU NAME : NVIDIA GEFORCE RTX 4060

Operating System : Windows 11

Detailed Description:** My mic wont work on discord when i try to use the voice changer**

Tutorial Used: https://www.youtube.com/watch?v=SxdnGxicJOg&list=WL&index=2

Screenshot: Dont show any errors

[latent kettle]

Is it a good idea to set batch size of 4 for 24 minutes of dataset

[simple ore]

4..8

[latent kettle]

What do you mean ? 4 or 8 ?

Or does it mean between 4 to 8 is good?

[simple ore]

4, 5,6,7,8

give it a try with 4 and with 6 and with 8

see which one comes out better

there may be some differences in results

[latent kettle]

😭 it takes hours

[simple ore]

it depends entirely on the dataset

well, I did test once with 2,4,6,8,10,12,16,24 batches

2 was bad

but here was no much noticeablew difference between 4 and 6

[latent kettle]

I see the applio ui says a setting of 4 offers improved accuracy

[simple ore]

well, it is not exactly that

[latent kettle]

Why it is written there 😶

[simple ore]

#🔊│ai-development message

because a proper explanation would take pages and pages

[latent kettle]

Where can I find a proper explanation?

[latent kettle]

What does the smoothed and Value means here ?

[simple ore]

it is just the total loss.. scroll down for examples.

👉 How many chunks of audio the model looks at before it adjusts itself.

Differences in Batch Sizes

Small batch size (like 2, 4, 6)

The model only looks at a few audio chunks at a time.

Pros:

Works even on weaker GPUs (less memory used).

Sometimes can capture more subtle details of a voice.

Cons:

Training is slower because the model updates too often.

Can be a bit “noisier” — results may vary more between training steps.

Medium batch size (like 8, 12, 16)

A balance: the model sees a fair number of chunks per step.

Pros:

Training is smoother and faster.

Usually good quality and stability.

Cons:

Needs more GPU memory than very small batches.

Large batch size (32, 64, 128 …)

The model sees lots of chunks before updating itself.

Pros:

Training is very stable and efficient (less “noise”).

Often used when you have a big, powerful GPU.

Cons:

Needs lots of GPU memory.

Sometimes can “average out” too much, losing some finer details of a specific voice.

A Simple Analogy

Imagine you’re learning to sing a song:

Small batch size: After every 2–3 notes, your teacher stops you and corrects you. It’s detailed, but it takes a long time and can feel jumpy.

Medium batch size: Your teacher lets you sing a full line, then corrects you. It’s a good balance.

Large batch size: You sing the whole verse before getting feedback. It’s smoother and efficient, but little mistakes might get overlooked.

For RVC training, most people use batch sizes between 8 and 16 if their GPU allows it. If the GPU is weak, go smaller (2–4). If it’s strong, you can experiment with larger ones, but medium is usually best.```

not a perfect explanation but there you have it

[latent kettle]

I see. Now I know how it works

[simple ore]

you have an upper limit (VRAM) how big of a batch you can use, that can be cheated a bit by using BFloat16 precision (if your gpu allows it, 3000series+), or by using checkpointing

unless you're using 10+ hour dataset it is not an issue, you should not go over like 16 for <1hr set anyway

[low shard]

Outdated tutorial,
Delete everything
Are you trying to do e girl trolling like in the video?

[low shard]

Show a screenshot of ur entire wokada, tutorial link, PC gpu and Operating System

[dusk jewel]

yes

[low shard]

Elaborate more pls, show an entire screenshot of the program

[low shard]

E girl trolling is catfishing, which is illegal. You have been warned both verbally and via sapphire and continuing to ask will lead to further actions possibly like a ban

[dusk jewel]

coold

[low shard]

It's written both in the rules and help guidelines

[dusk jewel]

hard

[low shard]

Huh?

[fallow lava]

I messed up, RVC Okada is now outputting gargled delay sounds

I only tinkered with Chunk & Extra. Now it's just a mess even if I try to place it back

[glossy ingot]

ya'll think this good

[viral mason]

yikes

that's an over year version

[weary minnow]

im using deiteris fork with vac lite and put my input as my regulkar mic and myoutput as the line 1 and input as line 1 on discord and output as my headphones but i only hear my regular voice in the mic test

[grizzled heart]

is there any tts website that i could upload my voice model and tell the ai to generate it for me?

[rose arrow]

GPU: nvidia rtx 4050
OS: Windows 11
Detailed Description: Everytime i open Deiteris fork, my system audio becomes bad. Every sound is lower and a bit distorted; spotify, youtube, media player, etc.
Tutorial: https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/

[rose arrow]

I think it happened after i allowed access to my mic because i tried blocking access after and my audio became normal again. What seems to be the problem here and possible fix?

[low shard]

RVC doesn't mean realtime voice changer

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

nope, the settings are completely wrong, and that's an extremely old program
delete everything
Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

show an entire screenshot of the program

[low shard]

11labs?

it depends alot, there isn't just a single version
Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

are u sure it isn't https://learn.microsoft.com/en-us/answers/questions/4104553/all-audio-lowers-when-people-(in-call-or-in-game)?forum=windows-all&referrer=answers

u couuld also send a screenshot

[glossy ingot]

gimme da new program

[glossy ingot]

roleplay

[low shard]

there's different programs for different things, elaborate everything in #✨│ai-help message

[glossy ingot]

vc

mostly rp

ig

[low shard]

that isn't enough, you need to elaborate everything

[glossy ingot]

oh

[low shard]

don't just say only roleplay, elaborate everything asked in that message, there isn't just 1 program and 1 version

[glossy ingot]

win 11, NVIDIA GeForce GTX 1660 SUPER, Roleplay Vc, No tut,

good?

[low shard]

you seem to be using e girl models, are you looking for e girl trolling or e girl rp?

[glossy ingot]

girl rp

4 family

since we got no females

lol

e girl trolling is against rules anyway

these were only good 1s i could find

[glossy ingot]

you still here

[hallow thistle]

Roleplaying in which place?

[glossy ingot]

roblox maybe gta

but roblox rn

[hallow thistle]

Recently, the whole Roblox platform itself has been accused of uh. Would you still go for it?

[glossy ingot]

like?

its dc vc but brookhaven

with my friends

[cold cave]

AttributeError: 'Namespace' object has no attribute 'normalization_mode'

[brittle minnow]

yo

my voicechanger don t work

can somebody hellp me

[viral mason]

https://cdn.discordapp.com/attachments/1159290139609137264/1406678293574324357/Buzzspeechbubble.gif

[viral mason]

if you look up "schlep" on youtube that's all u need to know

[lime siren]

how do I fix the voicechanger hearing other people in the discord call and speaking for them?

do I gotta use a noise gate

[glossy ingot]

oh lol

[glossy ingot]

headset?

[lime siren]

realtek

[glossy ingot]

microphone or headset tho

[lime siren]

wdym?

[glossy ingot]

or

these lol

somtimes headsets do that

[lime siren]

left headset

[glossy ingot]

might be ur headset loudness like too high

or use noise gate

🤷 tho im not smart so prop ask helper

[clear crater]

where did you get that version?

[low shard]

-realtime

[vernal bone]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[low shard]

either vonovox or wokada deiteris fork

[low shard]

u shouldn't, it's old

[glossy ingot]

friend sent it

[clear crater]

ahh okay im just asking bc thats the one i used to use and it worked but the version im trying to use now isnt working for me

Currently trying to use RVC

[cold cave]

Did you have this issue?

[clear crater]

but when trying to hear the converted voice I only hear myself not the voice model

[low shard]

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

rvc doesn't mean realtime voice changer

RVC = Retrieval-based-Voice-Conversion, the best Few Shots Speech To Speech AI Models (on v2), Inferences (use models) pre-recorded audio (ai covers) and train (make) models. Technically, Mainline RVC does have a go-realtime.bat (aka RVC-GUI), but it's pretty messy and outdated so it's extremely not suggested for realtime. There also updated forks with extra features like Applio.

Wokada = uses RVC for realtime inference. There's 2 main versions, Original made by Wok, and the most suggested one is Deiteris Fork (modified version)

Vonovox = Another Realtime Voice Changer based on RVC, with similar quality and performance to wokada deiteris fork but other perks

[low shard]

!howtoask

pls elaborate

[vernal bone]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1405252622734065816> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[clear crater]

So if im wanting realtime which would be best to go for?

[low shard]

there's different programs and versions, you have to elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[clear crater]

• GPU Nvidia RTX 3060
• Windows 10
• Roleplay in games
  The link im using is https://docs.aihub.gg
  I ofc dont have a screenshot as I'm still figuring out which programme I would need for what Im trying to do

[low shard]

be sure to play lowest graphics 1080p 60 fps cap

also be aware windows 10 support is ending in like 2 months

-realtime

[vernal bone]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[low shard]

you can either use vonovox or wokada deiteris fork

it's better you read the pros&cons of both

[lime siren]

how do I fix the voicechanger hearing other people in the discord call and speaking for them? should I use noisegate?

[clear crater]

Okay perfect thank you :D and thats okay I plan to change to windows 11 soon anyway due to the support for windows 10 ending anyway

[low shard]

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

btw vonovox gets more still updates, unlike wokada deiteris fork

and yeah, for ur safety id suggest windows 11

tbh windows 11 isn't as bad as everyone says, been using it since a year

[lime siren]

Nvidia GTX 1060
Windows 11
AI Covers
Im using the deteris fork
I need pic perms to show screenshot

[low shard]

oh you are using the completely wrong program then

RVC = Retrieval-based-Voice-Conversion, the best Few Shots Speech To Speech AI Models (on v2), Inferences (use models) pre-recorded audio (ai covers) and train (make) models. Technically, Mainline RVC does have a go-realtime.bat (aka RVC-GUI), but it's pretty messy and outdated so it's extremely not suggested for realtime. There also updated forks with extra features like Applio.

Wokada = uses RVC for realtime inference. There's 2 main versions, Original made by Wok, and the most suggested one is Deiteris Fork (modified version)

Vonovox = Another Realtime Voice Changer based on RVC, with similar quality and performance to wokada deiteris fork but other perks

wokada deiteris fork isn't for ai covers at all

you can delete everything

for ai covers you can use either https://docs.aihub.gg/rvc/local/aicovermaker/ or https://docs.aihub.gg/rvc/local/applio/

[lime siren]

theres a record button that I use all the time I just say what I want to say and put it into another site

[low shard]

it's not suggested and meant for that at all, realtime inference may sacrifice quality

it's way better you use either ai cover maker or applio for ai covers

aicovermaker even automatically separates the song and mixes it back with converted vocals, it's the easiest way locally

there's a difference between voice changer and realtime voice changer

and also this way others voices won't be converted, since you'll automatically upload the audio :D

[lime siren]

I shouldve said I have multiple purposes but AI covers are the main reason and it gives me an easy way to do that, my question was when I am in a discord call with the voice changer on and not talking how do I make it so it doesnt pick up on anyone elses voices?

I leave my computer and apparently it repeats what everyone else is saying

[low shard]

!give-media-perms 1h @lime siren

forgot to give it

you can send a full screenshot of the entire program without cropping

[lime siren]

hold on let me restart my pc

and open it

[low shard]

alr

[clear crater]

Yeah honestly i just didnt upgrade due to partial lazyness and the fact I just like windows 10 , I do remember people were shidding on windows 11 at the start something about it taking alot of your pc resources etc

[low shard]

i mean tbf it is kinda heavier/bloated, but not to the point that it ruins my experience, it runs fine even on my years old laptop

i dont mention my desktop bc that's ofc more powerful than my old laptop

i think part of it is that not everyone can upgrade if they got an ancient pc

[clear crater]

trueee i mean if needed its always possible to debloat anyway to help lessen the load

[lime siren]

what I dont know is if the issue is in the voice changer or my mic

it should only work when im speaking

[viral mason]

luigi

btw you have no mic settings set up

[lime siren]

my settings clear everytime I close it

[viral mason]

that's not normal

[lime siren]

well just the inputs

not all my settings

[viral mason]

still not normal

it shouldn't do that

[low shard]

use sup2
input: microphone
output: line 1
monitor: headphones optionally to hear urself

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Advanced Settings -> Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Advanced Settings -> Crossfade Lenght: Controls how smoothly the AI stitches different processed parts "chunks" of your voice back together. 0.1 for fastest voice, 0.15 for improved quality but increases delay by ~50 ms
• Reduce the delay on Windows via the Wasapi / Asio Guide

You could also use vonovox btw, it's a windows nvidia realtime voice changer based on RVC which still gets updates, but ur choice

[lime siren]

Ok so my issue is, when i'm on a discord call the voice changer sometimes speaks from the voices of the other people in the discord call

I already use all of these settings and i dont know why its picking up on that

[low shard]

u tried sup2 on, echo on and in sens to the right max?

[lime siren]

I've never used echo but yeah I use in sens and sup2

[low shard]

i meant "in sens" all the way to the right

try echo too

else it might just be ur headphones

[lime siren]

I have it at -65 which is 5 away from max ill try that

[chilly kestrel]

my realtime voice changer wont work

like when i try and hear my voice nothing comes out

and when i record as well there is no audio from my mic

[low shard]

This is a General AI Server, AI has many fields, so we can't know your issue with little info

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • E girl trolling
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[chilly kestrel]

gpu is a rtx 5070 12gb
windows 11
egirl trolling

and i cant upload a screenshot

[low shard]

E girl trolling is catfishing, which is illegal. You have been warned both verbally and via sapphire and continuing to ask will lead to further actions possibly like a ban

[chilly kestrel]

alr

i actually didnt know

sorry about that

[fallow lava]

-Graphics card is NVIDIA GeForce GTX 1060 3GB
-OS is Windows 10
-Just VC roleplays
-The only guide I used is this: https://docs.aihub.gg/

[low shard]

Use rmvpe without onnx, be sure the input is ur microphone, and optionally set ur headphones as monitor to hear urself

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Advanced Settings -> Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Advanced Settings -> Crossfade Lenght: Controls how smoothly the AI stitches different processed parts "chunks" of your voice back together. 0.1 for fastest voice, 0.15 for improved quality but increases delay by ~50 ms
• Reduce the delay on Windows via the Wasapi / Asio Guide

[tired timber]

Hi, I just want to set up RVC for character voices.
Can someone guide me on what I need to download first?

[low shard]

RVC doesn't mean realtime voice changer

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using (if any)
• a screenshot of the program (if any)

[tired timber]

I have an RTX 4070, Windows 11.
I want to use RVC for real-time character voices in OBS and recording software.
Which program should I download first?

[low shard]

For recording software? Do you want to use them on pre-recorded audios of yours?

Because realtime isn't meant for that, it's meant for discord VCs or games

[tired timber]

I want to use it in real-time through OBS while recording.
Not only for pre-recorded audios.

want to record gameplay in OBS, but I don’t want to use my real voice.
I just want to use RVC in real-time to make a character voice while recording.

[deft pewter]

is there any other better fork of wokada

because honestly i just launched wokada latest version today and suddenly im talking but i dont hear any output even tho its set to the right input and monitor

[simple ore]

probably did not give access to mic to the browser 😛

[blissful flint]

Please tell me how to use a sample attached to a specific voice?

[low shard]

-realtime

[vernal bone]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[low shard]

You can use either vonovox or wokada deiteris form

[tired timber]

so which one Vonovox or Wokada

[low shard]

Read each pros&cons and choose it depending on ur needs

[tired timber]

ok thx

[deft pewter]

i've been using it for multiple days bro

it just stopped itself

also it picks up my voice but never converts it

[simple niche]

Hello, how do I not have delay?

[opaque tinsel]

hello, when i start rvc my ping goes crazy like 2.5K and stay like this even if i close rvc. Only restart of my PC helps. Im not sure is it using local or cloud. Can someone help please?
RTX 3070 TI laptop
Win 11

[valid spruce]

Help

This happened after I continued training from a model

[simple ore]

it seems that you've selected a wrong sample rate, not the one you used originally to preprocess files

[low shard]

hows it going

[low shard]

This is a General AI Server, AI has many fields, so we can't know your issue with little info

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

rvc doesn't mean realtime voice changer

Please Elaborate:

• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • E girl trolling
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[opaque tinsel]

Hello i dont have access to my PC rn so i can't give you screenshot of it, but here's link(https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/releases), I've got it from all working rvc collection on this server(it's literally first in working local rvcs). Didn't really used one special guide but did it by myself probably I missed something or etc. but before i used another one rvc but deleted it because crazy ping and here its again

[knotty moth]

make sure the sample rate setting should match the one used in preprocess and also match the pretrain used
if you try to resume training with Applio on dataset preprocessed/trained with older RVC, unfortunately it might not work

[low shard]

that's original/mainline RVC

RVC = Retrieval-based-Voice-Conversion, the best Few Shots Speech To Speech AI Models (on v2), Inferences (use models) pre-recorded audio (ai covers) and train (make) models. Technically, Mainline RVC does have a go-realtime.bat (aka RVC-GUI), but it's pretty messy and outdated so it's extremely not suggested for realtime. There also updated forks with extra features like Applio.

Wokada = uses RVC for realtime inference. There's 2 main versions, Original made by Wok, and the most suggested one is Deiteris Fork (modified version)

Vonovox = Another Realtime Voice Changer based on RVC, with similar quality and performance to wokada deiteris fork but other perks

I'm not sure what you mean by ping, Are you trying to do ai covers, training rvc models, e girl trolling, roleplay in vc or roleplay in games?

[opaque tinsel]

I with friends trying to do some kind of RP. When i speak in discord there is very big latency and there is connection check of call and it shows big ping numbers like 2500ms but if i dont start voicechanger its 60ms

[low shard]

ohh, so roleplay like peter griffin / Dungeons & Drugons in discord VC, or like e girl roleplay/trolling?

[opaque tinsel]

Its something like manga reading together

Idk how to explain

[low shard]

you seem to be using the realtime.bat in original/mainline rvc, don't trust video tutorials, that's extremely outdated

It's better you explain this situation along with ur pc gpu and everything in #1192011222023950368 , since there are different versions and programs depending on ur gpu and OS

[opaque tinsel]

I already wrote there but it's been ignored so i installed that which i gave link, but it didn't help

Do i need to create a new one thread or?

[low shard]

it would be better you make a new post with all the info and ping me, i will be able to help you

[opaque tinsel]

Ok

[rose arrow]

nope. I tinkered a bit more last night and i found out that it only happens when i set my input device into any microphones, the audio's normal when i set it to "none"

I also tried the original w-okada before which runs in its own client instead of the browser and it didnt have this problem so im not sure what the issue is

[low shard]

@pastel oak u got any idea?

[foggy ice]

Yo i need help

[low shard]

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
• E girl trolling
• TTS
• Roleplay in VC
• Roleplay in Games
• etc
• what tutorial link are you using
• a screenshot of the program

[foggy ice]

I am trying use program for these girl voices but idk what to use i heard about w-okada but idk

I have nvidia 2060

Ryx

Rtx

I am. Windows 11

[low shard]

oh like e girl trolling?

[foggy ice]

No

Like roleplaying as girl in game

I just need voice for it and program to use it in

Because i don’t know really what to use it kinda confusing

?

[low shard]

mm, sorry but there are different programs for different things, a voice changer is different than a realtime voice changer, it's best you explain so i know which program to help you with

are you trying to roleplay as peter griffin in a game, do ai covers, or trying to troll/mess with others as a girl lol?

[foggy ice]

U know roblox?

[low shard]

yeah
are u like trying to roleplay or troll or do ai covers in roblox?

[foggy ice]

Why u keep saying trolling?

[low shard]

because I'm trying to understand which program you need, if it's a program for like family roleplay or mess with people as a girl, or do ai covers

[foggy ice]

There is map for role playing i am trying use it to play with my friends

As anime girl or smth yk

[foggy ice]

So which program and voice suits it

[tired timber]

@low shardCan I also use this model for making songs, or is there a separate tool for that?

[tired timber]

great 🤍

[viral mason]

you're part of the problem

[hallow thistle]

What model for making song?

[tired timber]

IDK

im asking

[viscid moss]

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
• E girl trolling
• TTS
• Roleplay in VC
• Roleplay in Games
• etc
• what tutorial link are you using
• a screenshot of the program

[rose arrow]

GPU: nvidia rtx 4050
OS: Windows 11
Detailed Description: Everytime i open Deiteris fork, my system audio becomes bad. Every sound is lower and a bit distorted; spotify, youtube, media player, etc. It happens when I set my input device into any microphones, audio becomes normal again when it's set to "none".
Tutorial: https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/

[knotty moth]

so are you trying to use it for catfishing/trolling?

[fallow lava]

The advanced settings kinda got rid of the delay but it's still a gargled mess.
Also how do I get the non-onnx version?

[deft pewter]

bro

man

wokada is broken as hell for me

i even allowed mic and all its legit picking up on my voice

but its not even working even tho it is picking up on my voice

i trieds switching to multiple other voice models (i even checked settings its using my mic and i changed to other seperate mics i have still the same) and the output is the same as my speakers

im just using this to sound like arthur morgan from rdr2 lol

i've been js makin ppls days in rdr2 online

until this issue happened

can i also have help

[rocky gate]

What are your thoughts on Vonovox? I’m not sure what to get. Currently running on a 2 year old Okada download. I use an Nvidia 4060

On the GitHub Vonovox is shown to have more perks/better performance but I’m not sure how true that is—

[viral mason]

Tbh I recommend Vonovox although I don't use it as of now but I've tested it, I only don't use it because of not having enough slots for models as of the current update

Besides that it'll all positives

[rocky gate]

Ah… awesome thanks. How many slots are there on estimate?

[viral mason]

From what I remember 8

Or 6

I have kind of a spotty memory

[rocky gate]

What! That’s bull… I guess I’ll go with Okada until they get that situated LOL

Thanks for the help, seriously!

[viral mason]

Deiteris has over 100 slots so I recommend that for now if you want a bunch of models!

[rocky gate]

Now that’s impressive, definitely what I’m looking for.

[viral mason]

Soon Vonovox will update to have a slot system that continues to increase dynamically but not sure when

Dr did say that'd be added tho eventually

[deft pewter]

but the thing is i love the new ui

[pastel oak]

When you set your mic as input, is it in client mode or server mode? Did you test both

[rose arrow]

client mode, server mode doesnt seem to work.

[pastel oak]

Do you know how to use server mode?

[rose arrow]

with WASAPI right?

[pastel oak]

Yes

[pine mist]

what is better than rvcgui

[rose arrow]

yea, it doesnt connect

[pastel oak]

Then there could be an issue with your audio devices but can't pinpoint it atm. Client is MME by default and seems to cause issues and wasapi doesn't work at all even though it's a newer version
Have you tried the vonovox voice changer on the guide?

[rose arrow]

ill try it later

[minor isle]

how to fix voice rechocing on discord

like others ppl voice are coming through okada

[faint loom]

Hello 🙂 where is the right place to ask for okada help?

[faint loom]

Thank you! I cant get the vcclient to run properly with my 7900XTX. cuda 2078 beta works only with cpu (5800x3d) and stutters (it recognizes it as cpu-1). 214 alpha seems to recognize my gpu but is even slower (takes 10-20secs to voice mod) and dml does not work at all.

[low shard]

wokada deiteris fork and vonovox are just for realtime

for making songs, check https://docs.aihub.gg/rvc/local/aicovermaker/

[low shard]

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • E girl trolling
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[low shard]

rvc gui is super outdated, it depends on what program ur even talking about

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • E girl trolling
  • TTS
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[rose arrow]

GPU: RTX 4050
OS: Windows 11
Description: Im using Vonovox and it just says this:

[low shard]

Be sure to not run it as admin, you sure you ran setup.bat before?

[rose arrow]

yes and it said Setup Complete

[low shard]

Could you please check if ur windows and PC GPU Drivers are up-to-date, then restart ur PC and run it again?

Nvidia drivers came out yesterday

[rose arrow]

oh yea i just checked, i have windows and nvidia driver updates

[low shard]

Be sure to do them

:D

[charred drum]

what program can i use to separate a choir?

[rose arrow]

it works now, thanks.

[low shard]

you're welcome!
need any other help?

[rose arrow]

last question, do the settings save when you close the client? I dont see any "save settings" buttons like the other RVCs

[low shard]

RVC doesn't mean realtime voice changer, it means Retrieval-based-Voice-Conversion
This is a realtime voice changer program based on running RVC models

and yes the settings should automatically close

[rose arrow]

oh yea i read it from the thread and i assumed they were all the same mb. Tysm, have a good day

[low shard]

you're welcome and you too :D

[pastel oak]

old ass program
whats your gpu name and what are you trying to do with the voice changer, like trolling egirl catfishing etc

https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/

virtual audio cable
windows-> download nvidia (first link, not rtx 5000)
windows-> opening on windows

Thats all, you can read how to upload voice models on voice models, audio setup on how to route the virtual cable and settings to find your own good ones, i recommend f0 rmvpe ; chunk 200 ; extra 2.7

[median drum]

what's the best AI website to do text to speech with a imported file that sounds the best without paying

[oak edge]

@median drum

you can ask here

[median drum]

thanks

[median drum]

need help on how to make voice clone like this for free

[simple ore]

that looks like incompatible torch and torchaudio libraries installed

2.2.0 cuda12.1 is so old

requirements are

[pine mist]

Rx6600
Windows 11
AI Covers
idk

[low shard]

Your AMD GPU is good enough to do inference (use models) locally (on ur pc), not sure about training

You can:

• Locally (runs on your pc so the speed depends on that, you will have to set it up with the guides):
  • Applio (AMD Windows) : A fork of RVC with some extra features like Applio TTS, kinda faster and simpler but same quality tho
• Cloud (remote good pc, easier and faster than ur PC but it's limited):
  • Weights.com: Partnered with AI Hub, lets u do them easily but u may be in a queue
  • Ilaria RVC Zero: pretty easy and fast
  • Applio Colab: max 4 hours daily, not granted, of GPU
  • RVC-AI-Cover-Maker-U Colab & Kaggle: Automatically separates the vocals and instrumentals, converts the voice and mix all together back

Easiest possible (automatically separates vocals & instrumentals) : weights.com & rvc-ai-cover-maker-ui colab/kaggle
easiest cloud: Ilaria rvc zero
easiest local: Applio

[simple ore]

probably because python there is 3.12

faiss-gpu is

it is very old

you can change requirements and install faiss-cpu instead

on windows it installs cpu, on lunux it installs gpu

faiss is for index

if you want to use that, then you need to downgrade python env to 3.10

generally no

[viral mason]

it's done this 3 damn times in a row, I can't train anything like this

---

NameError Traceback (most recent call last)
Cell In[1], line 18
3 rot_47 = lambda encoded_text: "".join(
4 [
5 (
(...)
14 ]
15 )
17 new_name = rot_47("kmjbmvh_hg")
---> 18 branch_name = requests.get(rot_47(codecs.decode("pbbxa://oqbpcj.kwu/QIPqaxivw/Ixxtqw/zmtmiama/tibmab", "rot_13")), allow_redirects=True).url.split("/")[-1]
19 findme = rot_47(codecs.decode("pbbxa://oqbpcj.kwu/Dqlitvb/qurwg-mtnqvlmz.oqb", "rot_13"))
20 uioawhd = rot_47(codecs.decode("pbbxa://oqbpcj.kwu/QIPqaxivw/Ixxtqw.oqb", "rot_13"))

NameError: name 'requests' is not defined

[simple ore]

what are you trying to do?

[viral mason]

just train a model ion kaggle applio

I haven't done anything differently but now it's bugged

[simple ore]

the code from above looks old

[viral mason]

could this be outaded somehow?https://www.kaggle.com/code/deiant/applio

I used it yesterday just fine

[simple ore]

you can import the latest notebook from github

create new notebook, import, then search

make sure you do create new notebook first, do not use import as 1st step

[viral mason]

could I have the link?

[simple ore]

did you read what I said

click create, notebook, file, import notebook, then serch 'applio' in github

maybe run each line manually and see what fails

you got them in order

[viral mason]

got it, someone should let them know to update the applio link soon bc that's kinda annoying to do

[prisma grove]

is 1 epoch per 2 minutes a normal speed for training

on a 2080 Ti

the docs don't say anything about training speed I checked

[simple ore]

depends on batch size

and dataset size

[prisma grove]

already moved to applio colab

idk where to put the dataset zip in my drive

and if I should put ".zip" at the end of the path (I'm using the noUI version)

no clue what do do with any of these options either

[simple ore]

does it say .zip?

[burnt hull]

this is an ai server we are real men who prefer real woman

if you will troll, you must leave or be destroyed

he is a troll

ban him we dont need trolls ruining our servers

[viral mason]

I agree with this guy

we need more ppl like you Promtgod

[marsh stratus]

I have an amd gpu and didnt want to dual boot linux to do Ai/ML stuff. I managed to get rocm working on wsl but it was slower than using my cpu.

[simple ore]

there's rocm for windows build available, depending on your gpu

[marsh stratus]

There is? I looked but couldnt find anything

[simple ore]

https://github.com/ROCm/TheRock/blob/main/RELEASES.md#installing-pytorch-python-packages

there are nightly builds

[marsh stratus]

interesting, im failing to import torch now due to som os error. saying some dll couldnt be loaded. Might just be an unstable build?

OSError: [WinError 126] The specified module could not be found. Error loading "C:\Users\[User]\AppData\Roaming\Python\Python312\site-packages\torch\lib\shm.dll" or one of its dependencies. to be very specific.

Thanks for the info, ill keep an eye on the builds and see if one works with my setup

[long chasm]

why does the voiice changer repeat everything iit liiistens to even the video

[rocky gate]

Hiya! How is the latency for y’all using WOkada Deiteris Fork? It is different for AMD / Nvidia users?

[red crescent]

guys for the voice ai when i upload the rvc it doens't work

[rocky gate]

The last time I used WOkada I was on an older AMD laptop with a RX 6700M, and the latency was pretty high, now I am on a Gaming PC with an Nvidia RTX 4060. Wondering if there will be much of a difference!

[quasi condor]

dont watch the tutorials of w okada on youtube

those are outdated

[viral mason]

yay nvidia

[rocky gate]

Yippie indeed!

[quasi condor]

how the laptop has 10gb vram on a gpu

maybe u put the latency like 500 or higher thats why the latency was doo doo

[rocky gate]

I just did whatever I could to keep it from glitching. It was an MSI Delta 15. I was just wondering how the latency is for my newest build.

[quasi condor]

also in the laptop was the f0 on fcpe_onnx?

[rocky gate]

I honestly have no clue, that was two(ish) years ago. I know I had to convert the files within WOkada after uploading the voice.

[quasi condor]

oh the year that fcpe didn't exist yet

or rmvpe

tho make sure u use the latest on okada on ur new pc

detries w okada

[rocky gate]

Yup, I have it installed along with my voices, just haven’t gotten the chance to try it yet. Just worried about the latency, I hope it’s close to when I’m speaking because before it would come through at least 2 seconds after…

[quasi condor]

yea that's normal

[rocky gate]

Oh okay, that sucks! Oh well, better than nothing

[simple ore]

you may need to check dependencies

[simple ore]

most likely you're missing this file

[frozen tendon]

how to make voice changer chunk sec

[marsh stratus]

Just for my knowledge, how did you figure this out?

ill try again after putting that file in system32 but that seems random to me

[simple ore]

ran a dependency checker, I've seen this before, on some system by some reason microsoft did not deploy this math library

it usually comes with VC++ resist

[low shard]

This is a General AI Server, AI has many fields, so we can't know your issue with little info

Please Elaborate:

• your PC GPU (Such as Nvidia RTX 3060, AMD RX 9070 XT..)
• your operating system (Like Windows 10, Windows 11, Linux, MacOS 26..)
• what you want to do? There isn't a program that does everything, there's a program for each thing:
  • AI Covers
  • Train RVC Models
  • TTS
  • E girl trolling
  • Roleplay in VC
  • Roleplay in Games
  • etc
• what tutorial link are you using
• a screenshot of the program

[marsh stratus]

sorry to bother you again, but im running into an import error and you seem very knowledgablefrom torch._C._distributed_c10d import FakeProcessGroup ModuleNotFoundError: No module named 'torch._C._distributed_c10d'; 'torch._C' is not a package Everything online suggests to install another version of torch, but that wont help in this case

[simple ore]

what are you trying to run?

dont follow online answers, this is brand new shit

experimental shit

if you're trying to run training in applio there's another fix needed

some things are not implemented yet

[marsh stratus]

its not applio, its image inpainting

[simple ore]

show the full error message

[quasi condor]

waht 😭

wdym by that

[marsh stratus]

redacted some file names for privacy

[simple ore]

🤦‍♂️

whatever, looks like it is blowing up on attempt to init a distributed process group for multi-gpu

that has not been implemented in those wheels yet

[marsh stratus]

could this be because i also have integrated graphics?

[simple ore]

no

[marsh stratus]

anyway, thanks for the help

ill keep an eye on these wheels and use them when thay are more developed

[simple ore]

probably can edit the file

Python312\site-packages\transformers\generation\utils.py

and set the value to False

synced_gpus = (is_deepspeed_zero3_enabled() or is_fsdp_managed_module(self)) and dist.get_world_size() > 1 -> synched_gpus = False

[quasi condor]

https://tenor.com/view/intel-dead-inside-logo-gif-16821712

[red crescent]

i need egirl trolling what shud i get

[cerulean fiber]

heya, is ~3 minutes a normal time for an RVC epoch to take with a 45 minute dataset?
On a 8gb 4060

[rocky gate]

Having trouble getting WOkada to play through Voicemod…

[drifting folio]

do you use original wokada or the fork

[rocky gate]

Fork!

Just upgraded

Bit of a noob with this since it’s been a couple years… not sure what I should have set for the input and output—

[drifting folio]

do you use a virtual cable to get it into voicemod?

[rocky gate]

This is currently how it looks, and the input in voicemod is CABLE Output (VB-Audio Virtual Cable)

[rocky gate]

I downloaded the newest VB Cable on the GitHub:)

[flint vapor]

I recommend you to uninstall vb cable and install vac lite

*i suggest you

[fast crest]

why are most of my f0s N/A? i cant set them

[quasi condor]

also change the chunk like 100 or 200

512 has alot of latency

and dat boi to fcpe

cuz onnx is for amd

[rocky gate]

You all are so helpful, thank you so much!

[queen fjord]

Which AI cover are you using now?

Could you please tell me?

[vast bough]

Hi there I am a live streamer on twitch. I am roleplaying as a mafia person and want a real time voice as as a mob boss.

Is there a real time AI voice on voicemod replicate a mobster

If there is i will get the pro version today

Could some1 plz help with this requests 🙏

[broken urchin]

you should use W okada deiteris fork