✨│ai-help

[tough vine]

pip uninstall torch torchvision torchaudio

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

[arctic badge]

i guess i have the other one not pytorch

[tough vine]

i think this error happened because ur replaced gpu uses a different compute capability

[tough vine]

which

[arctic badge]

that pop out in ur browser

[tough vine]

what

oh

[arctic badge]

u understand me ?

[tough vine]

understood

okay

do the fix i told you about

and try again?

[arctic badge]

okayy

[tough vine]

if it doesn't work

it means the pytorch is baked in the mmv

so jsut reinstall mmv

[arctic badge]

can you please guide me if possible

which one

[tough vine]

just run the command in the cmd

it will download them

or just install torch torchvision and torchaudio

[arctic badge]

can you type it for me again ?

[tough vine]

sec

[tough vine]

.

scroll up a bit

two commands

one for uninstall

one for install

[arctic badge]

i think i ununstall but how do i unstall bc i typed the cmd and

@tough vine

[tough vine]

pip install

not install

just copy the command

[tough vine]

.

[arctic badge]

i did i got the same error

[tough vine]

ok

u dont have python installed

@arctic badge try uninstalling MMVCServerSIO.exe then installing it again from the website

[arctic badge]

will i lose all the mods i have ?

i mean the voices i had downloaded

[tough vine]

yes

u gotta manually save

they usually in a folder called model_dir

move it to ur desktop

they uinstall then reinstall

[arctic badge]

and when i put them back in the folder ill have them ?

[tough vine]

mhm

[arctic badge]

can i have the website again ?

[tough vine]

when u reinstall just put the folder back

sec

https://github.com/w-okada/voice-changer

[arctic badge]

which one should i download ?

bc i dont remember downloading it this way

i remember downloadin this@tough vine

[tough vine]

go for it

i didnt know what u download exactly so

go for that

[arctic badge]

@tough vine i fixed it but another problem when i select my gpu instead of cpu this happen

and now im back to the same problem

[tough vine]

Show me ur command line

[arctic badge]

where is that

is it this
?

i downloaded this

do i have to for this too ?

[tough vine]

Makes sense

Okay

go to ur W-Okada folder

click the adress bar

press delete

then type cmd

[tough vine]

this should open a black window like this one

tell me when u do that

[arctic badge]

done

[tough vine]

paste this (to basically explain it forces the program to use its internal python to download the nightly drivers for ur gpu (which is the 50-series))

._internal\python.exe -m pip install --pre --upgrade torch torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128

since ur gpu is kinda new

u should use the nightly buikd

[tough vine]

can u type this

dir /s python.exe

then show me what it says*

[tough vine]

tf*

wait

[arctic badge]

maybe i should do this so it wont be too much work for u

[tough vine]

fair that works, go for that, and also dont forget ur mods folder

[arctic badge]

@tough vine pffft another problem

[tough vine]

ok attendez svp

oh

[arctic badge]

how many should i download here

[tough vine]

MMVCServerSIO.2.z02

ur missing MMVCServerSIO.2.z02

[viral mason]

what are you downloading???

what gpu do u have

none of that looks like the up to date stuff

[tough vine]

downloading this

[tough vine]

.

[viral mason]

ty for explaining nothing

[arctic badge]

eehh

still same problem

[tough vine]

np

[tough vine]

same command error?

[arctic badge]

5000 serie mode

[viral mason]

just download vonovox

[arctic badge]

whats that

[viral mason]

whatever mike is telling you to download is not working because he is giving u random stuff

[tough vine]

i didnt tell him to download anything

[arctic badge]

:-:

[tough vine]

he had errors i helped fix

[viral mason]

realtime voice changer client that is the best for nvidia atm

[tough vine]

whats this passive aggressivness?

[arctic badge]

can i have the link and will the mods i have wont work on that ?

[viral mason]

mods?

u don't need mods for a voice changer

[arctic badge]

model dir

[viral mason]

https://github.com/dr87/Vonovox/archive/refs/tags/v1.6.9.zip

[tough vine]

🤦‍♂️

[viral mason]

do you also have vac lite?

[arctic badge]

never heard of it

[viral mason]

it's a virtual audio cable

I'll get the link

basically vb cable but better for windows

[arctic badge]

oh i have this

[viral mason]

hm

it says direct sound in front of it so idk

[arctic badge]

last time u told me to download i

before i change my gpu to 5070TI i had 3060TI

[viral mason]

I don't have the best memory sorry

but if u have it from me I probably gave u the right thing

[arctic badge]

what do i need to do now tho

[viral mason]

ok so for vonovox, extract and run setup, then run start

setup downloads the rest of what is needed to run vonovox automatically then start will launch it

[arctic badge]

what about those models i used to have they will be usable there ?

[viral mason]

sorry I just saw a bunch of errors and stuff I didn't recognize which kinda upset my brain, thought u were giving them old stuff

[viral mason]

all voice models u had on the one you had before will work with vonovox ^^

[tough vine]

yeah i think so

[tough vine]

its ok i was just helping them debug the errors

[arctic badge]

so no more owakada ?

[tough vine]

i mean

i can help u owakada but worm says vonovox is better

[viral mason]

vonovox is wokada, just way better

diff name but still does the same stuff

[arctic badge]

oh okay

thank u very much guys

[viral mason]

you're welcome

if u need more help u can message me here

[tough vine]

@viral mason did vonovox update for the 50 series yet

[viral mason]

I'm not sure tbh, it's for all nvidia drivers idk if there's a specific 50 series download

[tough vine]

hold on

[arctic badge]

does the setup take too long to finish or just few mins

[viral mason]

could join the vonovox discord server if you're not in it and ask around

[arctic badge]

oh yes actually good question

[tough vine]

its not for me, @arctic badge got a 50-

thats what i was trying to help them with

earlier

[viral mason]

I'll be gone for a bit but I'll send u an invite for the Vonovox Discord server

[tough vine]

meanwhile @arctic badge do u wanna get W-Okada up?

[tough vine]

can u show me ur cmd?

[arctic badge]

this ?

[tough vine]

yes

one sec

[arctic badge]

just got this error in the cmd after activating the

[tough vine]

okay great

go back to ur program window

look for the audio section

select the input from default to ur actual microphone

and ur output to ur actual headphones

then click start

[arctic badge]

where is that

[tough vine]

ur main window

ur wokada

[arctic badge]

EEEEH

THERE IS TOO many razer kraken

[tough vine]

wait

[MME] Microphone (Razer Kraken V4 X)

the one u got selected

[arctic badge]

already have that

[tough vine]

output for [MME] Headphones (Razer Kraken V4 X)

then click start

[tough vine]

[MME]Haut-parleurs (Razer Kraken V4 X)

one below

[arctic badge]

its like the mic not catching everything i say just the first words

[tough vine]

wait the error is gone right

[arctic badge]

yes

[tough vine]

in ur settings

look for Threshold

lower it to -50

[arctic badge]

okay one sec

[tough vine]

tyt

[arctic badge]

i don see it

[tough vine]

The In sens

try to play around with it and see if its better

and

the biggest issue is the vol

first screenshot

the "in"

the slider is at 10%

make it 100% and see

or 50%

[arctic badge]

also the voice its like cracking not smooth

[tough vine]

did the voice input get better?

[arctic badge]

yes i can hear myself clear but its like laggy not smooth

[tough vine]

2nd screenshot

the SR

make it 48000

if it doesnt work lmk

[arctic badge]

got better just little

[tough vine]

3rd screenshot

CHUNK

[arctic badge]

its 304 now

[tough vine]

make it 512ms

oh

atp just tweak around the settings

yw :D

[arctic badge]

okay its good now

[plucky crown]

Guys help, I'm gonna get my first PC for voice ai app here, rtx 3090 ti founders edition with CPU ryzen 9 5950x VS rtx 3090 CPU 9900x3d, which do I get here I'll max out the ram 256gb

@tough vine is my specs ok which one is best I want zero lagss

[hallow thistle]

You sure b2332 W-Okada fork could work with GeForce RTX 50 series by modifying it? The b2397 W-Okada fork and "Vonovox" are made work with RTX 50 in mind without having to "pip install --upgrade" that way.

[plucky crown]

Can someone recommend

Where's the support

[queen glen]

anyone free to vc so i can share my screen i got vonovox and okada didnt get any errors but nothing is picking up

[willow nova]

anybody know the right settings for vonovox its not working no audio is coming out of line 1

[willow nova]

figured it out but now my voice is echoing from vonovox itself

when i speak i can hear it how would i disable the self hearing

[tough vine]

The specs looking good

Especially the ram

[tough vine]

Yep it works

They have support for 50xx

Not sure about vonovox tbh

[hallow thistle]

[tough vine]

We didn’t modify the build itself though

They have 50 support :3

[hallow thistle]

You think?

[plucky crown]

What does that mean

[tough vine]

Unrelated

The specs are good

[plucky crown]

Btw the previous choice is sold so I only have this seller available: AMD Ryzen 9 5950x
ROG Crosshair VIII Hero (WiFi)
GSkill Tridentz 1x32GB 3600mhz (RGB)
NVIDIA GeForce RTX 3090 Ti Founders Edition
LG 32" UltraGear IPS Monitor - 32GQ950
Corsair Liquid Cooling RGB
NZXT Tower Case
2TB Samsung Evo NVME

Should I go with this or nah get a dual 3090 custom PC that means I'll find 2nd hand parts and build it

@tough vine

If I get that PC I can max out it's ram with the other left funds I saved specifically for a PC, or I can just use all this funds to get dual 3090 custom PC

[hallow thistle]

You still not giving up about this? Two other people have debunked about your budget case in #🧬│ai-chat.

[plucky crown]

I have a budget

@tough vine what do u think get a PC or dual gpu

4k plus USD is what I can't afford, only below it

[hallow thistle]

Haha, I'd consider it "not giving up".

[plucky crown]

Huh

[tough vine]

Don’t buy the reseller pc

Go for dual

[plucky crown]

Hopefully it won't put a huge dent in electricity bills

[tough vine]

Btw The Ryzen 5950X caps at 128GBs

[tough vine]

Depends how you’re going to use it

[plucky crown]

Voice train ai like you guys do here

[tough vine]

Yes but how often the training is and how good u want it will make a difference

But it won’t be that much

[plucky crown]

What do u mean

Won't more vram help the voice training be better

Because in chats here you guys put limits on what can work, like lower vram gpus and such being not recommended

I saw in chat histories here

[tough vine]

I’m talking about the electricity

[tough vine]

Of course

That’s why I recommended dual

[craggy bough]

how did you downgrade from 256gb ram to 32gb

[plucky crown]

There is no 256gb ram motherboard they said only for servers or very expensive ones

I just discovered

I'm upgrading from dell optiplex 9020 Its my first time getting to know these PC specs

Even if you have PC specs in mind if there are no 2nd hand sellers out there that matches this specs or they get sold out quickly then the plan constantly changes

@craggy bough

[craggy bough]

might as well go custom instead of prebuilt

[pure field]

can someone help me find the bext vovonox settings for me i'm only planning on running files thru it so the latency doesn't matter i'm also willing to dedicate as much of my gpu to it because i won't be running anything else while using it

[pure field]

anyway if any of u know about this it'd be nice if u could dm me so we can go over it in depth ^_^

[hallow thistle]

There's no absolute best setting for Vonovox. This voice changer program is made to give better audio quality in mind.

[pure field]

yeah but i mean for my specific circumstance

[hallow thistle]

https://cdn.discordapp.com/attachments/1159290139609137264/1446358776587489351/image.png?ex=69841d94&is=6982cc14&hm=498cd9ec50df616abf46cc28d7695154a561336672ff59868f3b947d7dbb58cd&

[pure field]

i guess a better question would be

[hallow thistle]

The recommended "Extra time" value would always be at "2.0". Set "Crossfade" to 0.15s for a bit better audio quality.

[pure field]

what settings should i change as somebbody who doesn't care about latency or extra time and just wants absolute max quality

[pure field]

oh tyyy

anything in advanced settings or whatever?

[hallow thistle]

As what I said earlier.

[elfin nebula]

Is there a locally hosted llm with api integration that works with a full AMD build?

I got an RX 7900xtx and a R9 7900x

[gloomy lynx]

What's the easiest way to use Qwen3-TTS without the use of ComfyUI? Shit sucks, every time I try to set up ComfyUI it fails completely

Voicebox does work out of the box; the problem with that though is that there's no X-vector only setting, and it doesn't allow you to leave the reference text empty; this basically means Japanese samples cannot be used for English output without very heavy Engrish (unless I'm missing something), attempts at other cross-language output would presumably have similar results

Meanwhile my friend who managed to set Comfy up did set it to X-vector only, and that's let him use Japanese samples in fluent English

[opaque hull]

how do i uninstall

[willow nova]

Delete all of the files and than delete the vac

[tough vine]

Download Pinokio

it’s a browser for AI apps that handles all the garbage for you

btw

for Japanese-to-English issue you wanna be looking for the standard WebUI usually under "Cross-Lingual" or "Zero-Shot" mode

unlike voicebox qwen3 seperates the speaker embedding from the content flow much better

[tame oracle]

(he was saying he was gonna get a pc with a 3090 in it for 500 dollars)

[gusty moth]

-colab

[patent trellis]

📒 Google Colab Notebooks

Google Colab is a Cloud (Remote Good PC) Service. While the Free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

• **Applio**

by IA Hispano
Google Colab

• **RVC Mainline**

by Hina
Google Colab

• **UVR5 NO UI**

by Eddy
Google Colab

• **UVR5 UI**

by Eddy
Google Colab

• **Wokada Deiteris Fork**

by Deiteris & Hina
Google Colab

• **Hina's Modified Original Wokada**

Google Colab

• **RVC-AI-Cover-Maker-WebUI**

by Shiro & Eddy
Google Colab

• **FaceFusion UI**

by Nick088
Google Colab

• **FaceFusion NO UI**

by Nick088
Google Colab

• **Music Source Separation Training (Inference)**

by Jarredou & Makidanye
Google Colab

[viral mason]

kaggle's better :p

[hallow thistle]

The insanity.

[hallow thistle]

ComfyUI can do TTS? I thought ComfyUI can only do Stable Diffusion models.

[tame oracle]

[hallow thistle]

[tame oracle]

[quasi condor]

nah bruh lightning ai is wayyyy better

with the nvidia l4 WITH 22 GB OF VRAM

[hallow thistle]

I see a big debate.

[tame oracle]

HEY WHAT

[quasi condor]

https://tenor.com/view/ok-oh-yes-yes-o-yeah-yes-no-yes-go-on-yea-yes-gif-14382673246413447193

[tame oracle]

Ive been missing out

[slender lotus]

is there a working google colab link for rvc model training ?

[rustic pumice]

i think i finished training a ai voice model, idk how to download it

[slender lotus]

How can I actually train from the rvc mainline by hina link

[rustic pumice]

give me pic perms so i can send what it looks like

[quasi condor]

why do u need pic perms to send a voice model ;-;

[rustic pumice]

dawg i meant how do i download a voice model that i trained in applio

[hallow thistle]

On Applio RVC UI, go to "training" tab, scroll down and you'll see this option.

[rustic pumice]

yeah thats what i am on

[quasi condor]

https://tenor.com/view/flashbang-gif-22572243

[rustic pumice]

it wont let me click anythign on dropdown or smthing

it says select pth file to be exported and i can type idk what the hell to type

[hallow thistle]

Or in a file manager, go to "/Applio/logs" and spot which folder where trained model is located.

[rustic pumice]

where can i find that folder

[hallow thistle]

You don't know?

[rustic pumice]

no

[rustic pumice]

i dont see that

[hallow thistle]

Then where did you run "Applio RVC" from?

[rustic pumice]

nvm

from downloads

right

[hallow thistle]

No stupid this time, just curious about your motive.

[rustic pumice]

please stop being rude

im actually mentally stupid

[quasi condor]

https://tenor.com/view/stop-whining-bru-hmm-hmm-hmm-hmm-hmm-hmm-hmm-hmmmmm-dog-whining-dog-whining-gif-17430511343641148586

[rustic pumice]

i dropped out of my school went to a alternative school

cause im dumb

[hallow thistle]

https://tenor.com/view/dog-walk-away-weirdo-touch-gif-26494423

[rustic pumice]

ash

also this is my first time making a rvc model

[slender lotus]

its ok dont feel stupid

[slender lotus]

they been there too

[rustic pumice]

what do i do now

[hallow thistle]

I study in college, and even if my college didn't taught me about AI, I still know how one program works. It's okay to no know things, but at least be clear about your steps.

[rustic pumice]

this is what u meant right

i found in downloads

[slender lotus]

its also okay to not be rude btw

[hallow thistle]

Literally everyone thinks the same about how I talk.

[slender lotus]

so?

being rude isnt cool either

[rustic pumice]

can someone help me or what

[slender lotus]

dont be proud of it

And I bet u dont talk like this to ppl irl actually

[rustic pumice]

im diagnosed with stupidity for real

i get bullied

for that

[tame oracle]

you also realize that some of us here aren't native english speakers, the way they talk might be different and come off as rude even if thats not what they were going for

[rustic pumice]

this is a ai help bro

[slender lotus]

Now yall know what yall saying dont even

[rustic pumice]

can someone help me

[slender lotus]

I saw the gifs etc. dont be making excuses helpp

[hallow thistle]

What is your step again?

[rustic pumice]

this is what im stuck on

i apparently finished training and idk what to do next

[hallow thistle]

How about this?

[rustic pumice]

doesnt show anything for me

can u give me the file location

to applio folder

[hallow thistle]

Are you sure you have even trained anything there? The "Output Information" part might show something "successfully" if model finished training.

[rustic pumice]

output infortmation doesnt say anything for me

[hallow thistle]

If you didn't click on "Start Training" button, it's likely the issue.

[rustic pumice]

i did bro

i think its actually training rn

i clciked start training many times

it didnt work

[hallow thistle]

Look up your "terminal" window.

[rustic pumice]

is ts good

WATTT i realizied this i gonna take time

than i expected

epoch = 2 to more more

lol

[hallow thistle]

I don't remember how I trained a voice model on Applio RVC earlier, but this looks like it's just the start.

[slender lotus]

Im gonna go crazy

https://colab.research.google.com/github/hinabl/RVC-Online/blob/main/Mainline_Colab_Full.ipynb#scrollTo=_wW3pERwGeDe it says I need to restart it everytime I run it, now I didnt restart its still running Im waiting for whats gonna happen

[slender lotus]

in the second step

[hallow thistle]

This one doesn't look like "Applio RVC".

[slender lotus]

yea its colab

[hallow thistle]

Yes, "the website" is Google Colab, but the notebook link isn't "Applio RVC", it's obviously "mainline RVC".

[slender lotus]

can I not train from there

bcs my gpu isnt working on applio

wait the second step just completed

Except I didnt restart

I hope it wont cause a problem

[hallow thistle]

I now have two instances of "Applio RVC"; one online, and another I run locally from my PC.

[slender lotus]

and

[hybrid cedar]

Hello there, just a quick question - I spent whole yesterday tweaking the Deiteris' W Okada fork, yet I can't get the voice to sound natural... Few sylabes are perfectly fine, but most of the voice just sounds scratchy enough to give unpleasant, artifical feel. Male -> Female, chunk size set a bit too high just to be sure it's not caused by lack of resources (yet it's the same from ~200ms to ~430ms). Running on 3060Ti without any game running in background (yet). I'll be happy for any suggestions

[hallow thistle]

And?

[slender lotus]

Wait can I not train from mainline

is it for something else

[hallow thistle]

The "mainline RVC" was last updated since like **few years **ago, meanwhile "Applio RVC" was last updated few weeks ago. While mainline RVC might work, it's sometimes buggy as of today.

[slender lotus]

hmm okay

I have no choice other than mainline so

[hallow thistle]

There's always another choice. "Applio RVC" also exists as Colab notebook, not always be a batch file.

[slender lotus]

Wheree

It might help me

The issue is that it doesnt see my gpu or whatever it is

But it might work different in colab

[hallow thistle]

How do I explain about this one?

[slender lotus]

Can I have the link

of applio colab one

[hallow thistle]

Since there's a problem running "Applio RVC GUI" with free tier Google Colab, you might wanna try Kaggle one instead.

-kaggle

[patent trellis]

📘 Kaggle Notebooks

Kaggle is a Cloud (Remote Good PC) Service that offers 30 hours of GPU weekly, but needs a phone number verification

• **Applio Notebook**

by IAHispano
Kaggle

• **Hina Mod Original Wokada**

by Hina
Kaggle

• **Wokada Deiteris Fork**

by Hina & Deiteris
Kaggle

• **UVR5 UI**

by Eddy, ArisDev & Nick088
Kaggle

• **UVR5 NO UI**

by Eddy
Kaggle

• **RVC AI Cover Maker UI**

by Shirou & ArisDev
Kaggle

• **Music Source Separation**

by Shirou
Kaggle

[slender lotus]

ahh its so complicated

what do I do now

[hallow thistle]

Neither Kaggle or Colab uses your PC resources. If "Applio RVC" as a local program doesn't detect your PC GPU, it can mean your PC doesn't have a dedicated GPU, or you have Apple Mac.

[slender lotus]

If I dont have a dedicated gpu how to train one..

model

[hallow thistle]

Simply, there are online websites. You get it?

[slender lotus]

no..

[hallow thistle]

This is how I run Applio RVC on Kaggle.

[slender lotus]

how do I get there

Im so dumb for this

[hallow thistle]

[slender lotus]

So I clicked

applio notebook

where to click after opening the link

[hallow thistle]

Look up this guide link on how to use Applio RVC on Kaggle. https://docs.aihub.gg/rvc/cloud/applio-kaggle/

[slender lotus]

Okayy hope I can handle it

where is the tunnel menu

I cant find it

[hallow thistle]

Replace Tunnel = "Ngrok" with Tunnel = "Gradio + LocalTunnel".

[slender lotus]

Im on the first step and there isnt a tunnel text for me

[hallow thistle]

Scroll down in your Kaggle.

[slender lotus]

So I clicked the link it gave me after running it, I entered the password I was given too and now it stucks in the loading

I submitted

[hallow thistle]

The actual link for Applio RVC's GUI would be "* Running on public URL:" then following the actual link to a Gradio tunnel website.

The other two links are of loca.lt/LocalTunnel tunnel website.

[slender lotus]

It only gave me localtunnel public urls

[hallow thistle]

Try screenshot your Kaggle interface.

[hallow thistle]

Then what about this?

[slender lotus]

I dont have the rest

[hallow thistle]

Click "cancel run" and try re run this code cell again.

[slender lotus]

python3: can't open file '/kaggle/working/app.py': [Errno 2] No such file or directory

it says this under the tunnel links

[hallow thistle]

I think you either did something wrong there. Try click "stop session", then "start session" and do all over from start again.

[dapper imp]

-rtc

[hallow thistle]

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[slender lotus]

Okay Im now training on colab applio

Omg its actually fast this time

[hallow thistle]

How do I verify that statement?

[slender lotus]

Idk

Im just happy cuz Ive been workin on this since yesterday loll

[hallow thistle]

I didn't have to fix anything in Applio RVC the first time I run it on Kaggle, and it still works. When I run again for 6-7 times, it still works for me regardless.

This server is not where you promote your thing, poopy.

[solemn blaze]

whats the minimal specs to run w-okada smoothly? no stutters no nun

[hallow thistle]

At least "NVIDIA GeForce RTX 2060" or the comparable "AMD Radeon RX 5600 XT".

[solemn blaze]

wt abt cpu

vram ram

[hallow thistle]

At least eighth generation Intel Core CPU or second gen AMD Ryzen; 8 GB main RAM is the most minimum not just for the voice changer program, but any RAM equals or above 16 GB is more prefered.

[cunning tree]

im so mad, I have rx 6750xt and its not getting recognised by the progam. I tried both the ML and normal version and it just always uses the cpu. Is there really no solution?

[solemn blaze]

thanks

[cunning tree]

@hallow thistle is there definitely no solution like 100% say yes or no please.

[tough vine]

dont worry

ur rx is good

ur problem is software not hardware

so u must use ONNX, so go to w-okada, edit, model settings for the character

then find the option that says export to ONNX

once u got the .onnx ur gpu will spike a bit but ur latency will drop

another issue could be version of ur w-okada

so make sure u downloaded the directml w-okada

[hallow thistle]

Are you using "Tg Develop's W-Okada fork"? Because the original version of W-Okada like v.1.5.3.18a, especially its DirectML variant, is so buggy that many people here have been asking the same. More recent W-Okada forks like those made by Deiteris and Tg Develop can at least detect your AMD GPU, even though slightly different interfaces.

[hallow thistle]

[tough vine]

hello, namari

[hallow thistle]

Why did you tell them like that? All W-Okada versions (b2332, b2397 and v.1.5.3.18a) generally have their own** DirectML variants**, so you would be specific about which version supposed to work.

[tough vine]

true true however for his amd 6750xt he wont need to hunt for specific versions, just grab the zip file that says directml or dml, if he grabs the standard version he wont have directml backend

the version number matters less than the directml tag

[cunning tree]

I did all you said, generated the onnx file and then uploaded it as a next character and the difference is crazy. Before, the voice was cracking every word, now its pretty much perfect, thank you.

Also ive noticed, as you said, a big spike in gpu usage for a while, but then basically 0.

Cpu is still doing the work

[hallow thistle]

[tough vine]

Of course, anytime let me know if u got more issues

[tough vine]

Dw about it

[cunning tree]

Im using the v1.5.3.18 ML version, how do i download the new one?

[hallow thistle]

Deiteris' W-Okada fork looks like this. https://cdn.discordapp.com/attachments/1159290139609137264/1467535508019347671/image.png?ex=69860272&is=6984b0f2&hm=1e3fb96ec58b893d53c96185724efe79c0025cc93252bfa865e2b4436c6af8a9&

[hallow thistle]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE

⚔️ Wokada Tg-Develop Fork vs Vonovox

For Windows Nvidia, Both Wokada Tg-Develop 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 Tg-Develop 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

[tough vine]

if u really wanna see the actual performance go to ur task manager, performance, gpu, from 3D click compute 0 or compute 1 and u'll see the activity there

[hallow thistle]

The "cause" has already been identified.

[tough vine]

thank you namari

[hallow thistle]

This is Tg Develop's W-Okada fork. https://cdn.discordapp.com/attachments/1159290139609137264/1467527309958447166/image.png?ex=6985fad0&is=6984a950&hm=e9695c9f6f8b86d632350c7afad103b122e8873e100827ccf48e174ada479a70&

[cunning tree]

Thank you, ive checked the compute 0 and compute 1, and still, when im speaking its on 0%, the cpu does all the work.

[tough vine]

fair enough

[hallow thistle]

Both W-Okada forks from Tg Develop and Deiteris will work; Tg Develop one has more recent features, whereas Deiteris one is older but presumed to be stable for specific cases.

[cunning tree]

Is the cpu supposed to have all the workload (60%) and gpu be on stable 1% while im talking though?

Or do i still have to fix something

[tough vine]

yes its normal

ur cpu still has to handle the audio stream, ui, data transfer

[gusty moth]

-colab

[patent trellis]

📒 Google Colab Notebooks

Google Colab is a Cloud (Remote Good PC) Service. While the Free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

• **Applio**

by IA Hispano
Google Colab

• **RVC Mainline**

by Hina
Google Colab

• **UVR5 NO UI**

by Eddy
Google Colab

• **UVR5 UI**

by Eddy
Google Colab

• **Wokada Deiteris Fork**

by Deiteris & Hina
Google Colab

• **Hina's Modified Original Wokada**

Google Colab

• **RVC-AI-Cover-Maker-WebUI**

by Shiro & Eddy
Google Colab

• **FaceFusion UI**

by Nick088
Google Colab

• **FaceFusion NO UI**

by Nick088
Google Colab

• **Music Source Separation Training (Inference)**

by Jarredou & Makidanye
Google Colab

[viral mason]

<@&1159293140440723499> burn him alive or something

[tropic steppe]

i have a problem that when i speak the ai voice will talk for 1 second and stop then talk for 1 second and stop and its repeating everytime, how can i fix that? (im using RVC voice's)

[celest cipher]

anyone know why its not working for me

[viral mason]

what's not working?

[wind egret]

Does anyone know any meta work flows for making the best novelai gens by img2img inpainting and general base gens? I don't wanna "experiment" all day because I don't have enough time to make up one, I just wanna make a good Gen with the Opus membership
I am noob to SD but i have used it alot
it never really appeased me to the same extend as seeing other people's ai gens
and yet they never disclose their workflow out of some elitiist secret kids club shtick
please i want to know

[soft eagle]

so I got a question, does the voice changer stop working if the graphics card changes?

[slender lotus]

I have a question, on applio, I have a 6 minute long vocals, how many epochs would be the best for me

[viral mason]

Epoch isn't important, train to around 300 save every 5 or 10 whichever you like, and test each one until it sounds good, and if it still doesn't sound good it could be due to many reasons like audio not properly being cleaned, or bad luck

[random star]

how do i get a models url💔

[golden yew]

Hello, Ive Been Trying To Get My MMVCServerSIO To Work But Im At A Stand Still

I Have A Windows Pc, And Once Had It Working, Then I Changed Pcs And I Used An External Harddrive To Transfer My Data, And When I Tried Loading MMVCServerSIO, I Get A
**The process cannot access the file because it is being used by another process" PermissionError, When I Follow The C:\User Line Of Direction, I Hit A Brick Wall Because I Dont Have Said Files To Follow

I Have Deleted MMVCServerSIO And Re-Installed It From The Website But After Reinstalling It, It Still Gives Me The Same Error

I Dont Ask The Server Often So I Believe Im In The Right Place, I Miss Using This Fun Voice Mod

[viral mason]

what?

[random star]

on the gradio ui uploading python and index isnt working and i cant find where the urls for models are anymore

[viral mason]

what link did you use to download this?

you may have an old software

[golden yew]

I Used The Link From AI Hub I Believe, The Server Has Moved Around Since I Last Been Here

But It Was In One Of The Channels, Like RVC? , Sorry Im Not Good At This Kind Of Stuff

[viral mason]

I wouldn't know what you have just based off that tbh

[golden yew]

I've Been Looking, I Cant Find It, If You Think The Issue Is Its An Old Website Or An Old Verison I've Redownloaded It On, Can You Show Me The Link Where The Most Recent Updated Version Is?

[viral mason]

what is your graphics card? Nvidia or AMD?

[golden yew]

The Folder Name Is, "voice-changer-windows-nvidia-b2332"

Nvidia

[viral mason]

this could be Wokada deiteris fork

unsure tho

but for the best I would say use wokada tg fork, I'l lget the link and help with it

[golden yew]

Thanks Thanks < 3

[viral mason]

-rt

[patent trellis]

💻 Local Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE

⚔️ Wokada Tg-Develop Fork vs Vonovox

For Windows Nvidia, Both Wokada Tg-Develop 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 Tg-Develop 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

[viral mason]

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

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.001

https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-cuda.zip.002

srry I took a million years

the first one is a virtual audio cable like VB cable but better

and the other two are the voice changer

first one is the main one and other one u just need to put in the same place as the first

[golden yew]

Oh No, Thank You For Helping! May I Ask, What Do You Mean By That, So 001 Is The One I Use, What Exactly Do I Do With 002?

[viral mason]

so basically just drag and drop 002 into 001 after u extract 001

[golden yew]

Extract? I Thought It Would Be A Zip File, Im Not Sure How To Extract These

I Already Had The Voice Cable Voice Installed Though!

[viral mason]

try reinstalling the first one

[golden yew]

Its Still The Same File Type

[viral mason]

rename the 001 file to remove anything after it says .zip

[golden yew]

Oh Wow! That Worked! So Put Th .002 In .001 After I Extract Both?

[viral mason]

yep! u cannot extract 002 tho

just drag it into 001

[golden yew]

Oh, Maybe It Doesnt Work

One Sec, I Think I Gotta Do This And Fix It

[viral mason]

wdym?

[golden yew]

Extracting It I Had 2 Errors, Is This Fine?

[viral mason]

errors how?

[golden yew]

Sorry, I Got The Voice Changer 1 Unzipped But The Second One Isnt Unzipping

[viral mason]

ye second one cannot be extracted

just place it in 001

[golden yew]

Oh? So I Just Put That Into The New Folder?

Ok

All Right!~ Whats Next

[viral mason]

yea just drag the 002 zip into the folder for 001

then run mmvcserversio.exe in the folder of the same name

[golden yew]

It Doesnt Launch

It Pulls Up A Command Prompt For Less Than A Second Then It Closes

[golden yew]

I Managed To Get A Look At What It Saysk It Pops Up For A Half A Second And Says
File "socket.py:, line 52, in <module>
moduelNotFoundError: No Module Named'_SOCKET'
[PYI-6772:Error] Error - Failed To Execute Script 'pyi_rth_multiprocessing' Due To Unhanded Exeption!

Thats The Last 3 Lines Before It Closes

[viral mason]

weeird

[golden yew]

I Redownloaded Both And The 001 One, I Got This

[hallow thistle]

Does your MMVCServerSIO folder look like this?

[golden yew]

It Looks Like This

[hallow thistle]

Make sure you don't extract anything on your desktop folder.

[golden yew]

Oh? I Shouldnt? I Didnt Know That

[hallow thistle]

[golden yew]

Still Looks Like Before Hand

Click Into Mmv

[hallow thistle]

Try extract to somewhere like "D:\MMVCServerSIO" or "C:\Users\your username\Downloads\MMVCServerSIO". Make sure you have "Hide extensions for known file types" disabled in your Explorer to make every file showing their file extension and to avoid mistaking ".zip.001" for a ".zip" file.

[golden yew]

What Are These By The Way> They're Like 2Gbs Each And I Cant Delete Them From My Downloads

[hallow thistle]

Did you just click download files of something else on repeat? These files are downloading, and are not what supposed to happen unless otherwise.

Check your browser download (history) section.

[golden yew]

Ok That Did It! Thanks

Its Still Not Working, Is There A Different Version, Or What Am I Doing Wrong

Sorry For Struggling, I Havent Used The Voice Mod In A Year And I Really Miss The Goofy Voices

[hallow thistle]

I run the same program for like 19 times, and still have not get any problem. What do you mean you can't run the program? If it possible, make sure one of your antivirus programs won't block the program, or try run it through CMD and screenshot that error message if there any.

If neither W-Okada "fork" programs work, it usually indicates something is wrong with your PC system or you have missed few important steps. Review your steps.

[golden yew]

Hm... I Didnt Think About That, I Guess When I Tranfered Some Of The Data To My New Pc, I May Of Not Installed Anything That WASNT Just The W-Okada, I Will Look Into It, Can You Send Me The Link Of The Steps To Install Everything Before The W-Okada

[golden yew]

Wait, But Would That Explain Why Im Missing Those Files Though?

[hallow thistle]

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

Some other files and folders than MMVCServerSIO.exe and a folder _internal don't always mean the program won't gonna work; they will appear after you run the program MMVCServerSIO.exe the first time.

[golden yew]

Ok, Thank You! I'll Try To Make Time Tomorrow And See If That Works, Thanks For Helping The Two Of You!

[slender lotus]

Guys I need help

so I train on colab Applio and, first time I was training a model, colab disconnected mid process but I still had the model, Now another training and it disconnected midprocess to I think it was at like 210 epochs at least, but the model isnt here in my drive, But I can still see the name and the steps it has on colab, Can I continue training it somehow?

[slender lotus]

Guys?

[chilly furnace]

if u can still download the steps .pth file in your colab folder tree you can continue it from there

otherwise

[slender lotus]

I start to train it but then it stops midprocess idk why

[chilly furnace]

train on kaggle

or use a colab disconnection code idk

[viral mason]

Kaggle

[slender lotus]

I just need the pth file I already trained but it only gave me index..

[viral mason]

Experience has taught me you will be able to make many models with 30 hours for free over colab giving you a random number between 2 and like 4

[plucky crown]

Guys

I'm so disappointed rn, everything gets sold so quickly even the 2nd hands in fb marketplace

I inquired again and told me it's reserved to other buyers, other sellers say they took some parts out,others forgot to update the listing saying it's already sold

[chilly furnace]

??

[hallow thistle]

As of 2026, I trained the first RVC voice model in Applio RVC on Kaggle as of 2026. The first attempt once made then-current Kaggle environment to run out of storage because I forgot to change "Save Every Epoch" from 10 to 50 and enable "Save Only Latest" there, eventually making Applio RVC to give too many files during training. That one was hilarious. But then the second attempt, I finally got few perfect pth files of the same model project of different epoch numbers ranging between 50 to 300.

[hallow thistle]

Try look up his past messages here and in #🧬│ai-chat, the hint is that he only talks about the same thing and nothing else.

[hybrid fractal]

Hi!

is it worth to train model on spin v2 and refinegan?

on applio V3.6.1

[sharp fractal]

Can anyone suggest me a good tts on which I can use tbe rvc models

[hallow thistle]

Applio RVC has its own TTS feature.

[hardy yew]

From what I've heard, no. Due to lack of decent pretrains

Supposedly for now hifigan + contentvec is the way to go

[hybrid fractal]

😮

thank you!

for letting me know

[hallow thistle]

Nuh uh.

[elder coral]

how do you approve models in foreign languages

[elfin nebula]

Is there a locally hosted llm with api integration that works with a full AMD build?

I got an RX 7900xtx and a R9 7900x

[indigo geyser]

.

[tough vine]

hello!
yes, there is.
the best two option you have with amd support are:

1. ollama (the standard choice), u run it and it will be on http://localhost:11434
2. lm studio (if u want gui) it has a specific version (AMD ROCm) and on http://localhost:1234/v1

[nimble pawn]

https://openclaw.ai/
at the bottom

[vivid remnant]

Can someone help me with vonovox please?

[tough vine]

of course

how can i help

[vivid remnant]

I don't know how to make it lol

[tough vine]

like setting it up?

[vivid remnant]

yep

[tough vine]

need a tutorial video?

okay one sec

[vivid remnant]

yes please

[viral mason]

extract it, run the file called setup then run the file start after

[tough vine]

@vivid remnant

https://www.youtube.com/watch?v=pHhjg2JwdPI

[vivid remnant]

Thank you guys. I really appreciate it

[tough vine]

If you got stuck with the tutorial let me know

Just ping

[faint sable]

i got the file for the voice changer

but idk how to install

[viral mason]

why arr you suggesting a reallllllllllllly old voice changer

[tough vine]

😭sorry

[viral mason]

he was asking for help for vonovox

[tough vine]

do u have a more updated one?

swy

[faint sable]

yo

idk how to install i got the file but i cant find a install

[viral mason]

remove whatever u just downloaded from the yt video and keep vonovox, run the file called Setup, the run the file named start

[faint sable]

vonobox

latest one on sep 7?

alright its downloading

[viral mason]

ur good man, just use the command -rt

it shows the three recommended voice changers

anything on yt is all outdated for rvc realtime voice changers

[tough vine]

okayy learning with you guys really, im not into these voice changers alot

@vivid remnant

-rt

[patent trellis]

💻 Local Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE

⚔️ Wokada Tg-Develop Fork vs Vonovox

For Windows Nvidia, Both Wokada Tg-Develop 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 Tg-Develop 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

[faint sable]

alright once i got it idk what to do lmao

let me get it first

[vivid remnant]

Yeah?

[tough vine]

ignore the tutorial i sent you

its outdated

[vivid remnant]

oh ok

[tough vine]

here follow this

[viral mason]

ahem

@faint sable

read please

[faint sable]

yh i did that i mean

the models

[viral mason]

o

[faint sable]

like the models loading n stuf

gotta have it install first its doing torch installs n stuff

also @viral mason will it ever like break the voice changer or nah

before when i tried a voice changer it like swapped to my regular voice for a few seconds

its installing torch bits

[quartz flame]

Hello friends, I have a problem. I have a rather weak computer and my voice seems robotic. I want to use a female voice so that people don't notice the AI. I need a couple of tips. (9950x3d and 4080s 64gb ddr5)

[viral mason]

download one from https://discord.com/channels/1159260121998827560/1175430844685484042 extract the zip file ,then click one of the preset squares in vonovox

[tough vine]

gotchu

[faint sable]

is MAE good? it apparently is reccommended from a yt

[tough vine]

first that isn't a weak computer 😭

[faint sable]

im a boy tho and i need a boy voice but like if im a egirl

ill get spoiled

type shi type shi

[viral mason]

please don't be using it to be weird..

[viral mason]

I will no longer be helping you sorry

[faint sable]

im jokingl mao

i need a good boy voice

i dont know any

[viral mason]

use your own voice lol

[tough vine]

the robotic voice isn't from the hardware, it's from ur software
switch your processing method (f0 detector) to RMVPE
drop your Index Rate to roughly 0.3 or 0.4
make sure your pitch/tune is set to +12 (or -12)

[faint sable]

ill try find one

[tough vine]

bro gave up 😭

[viral mason]

I don't help catfishers

or scammers

[tough vine]

same

[viral mason]

they should be burned/j

[tough vine]

@faint sable shame on you!

[faint sable]

im not catfishing💔😭

if i was catfishing i would use a girl voice

and im looking for a boy one

[viral mason]

gulp

[faint sable]

im joking chillllll

[tough vine]

bro wants to get groomed(?)

why do u wanna pretend to be a girl bruh

[faint sable]

i dont

[viral mason]

he wants to be a boy

but he already is a boy

[tough vine]

ah sp

[tame oracle]

the neighbor kid nobody does

[faint sable]

acctually

[tough vine]

deeper voice boy

or

[faint sable]

i want to be a man

and im a boy

so like

[tough vine]

lighter voice boy

like femboy voice

[tame oracle]

heh

[faint sable]

im already a femboy

[tough vine]

UNDERAGE ALERT ⚠️

[tame oracle]

EYAH

[faint sable]

cough cough

[tame oracle]

YEAH

[faint sable]

must have been the wind

[faint sable]

chill

[viral mason]

dear god..

[faint sable]

im 14

[viral mason]

EW

sorry I don't like children

[tame oracle]

😭

then you dont like me

[faint sable]

so when you have kids you will hate them

[tough vine]

HUH

[tame oracle]

yeah

[faint sable]

HUH

[viral mason]

I like dudes

[faint sable]

yeah

[tame oracle]

im 14 gonna be 15 in March

[faint sable]

i got vonovox

now i jst gotta get a vc

[tough vine]

good luck xbio

next time dont tell us what will you do with it

[faint sable]

use it to lure in dudes

[tough vine]

bro

[faint sable]

yes

[tough vine]

https://media.discordapp.net/attachments/873293933566890084/990047177919307816/received_728789665112499.gif

you rn

[viral mason]

<@&1159293140440723499> should like, actually kill this dide

bro wants to be groomed

[tough vine]

alos hes underage

also hes a catfisher

[faint sable]

brother

are we deadass

[tame oracle]

what am i watching

😭

[viral mason]

this server has the worst people ever

legit

[faint sable]

ikr like

weirdoes fr

[viral mason]

https://media.discordapp.net/attachments/722780815087501363/1182647658989486080/Tree_eater_speech_bubble.gif

[faint sable]

no clue what im doing

lowk

anyone got a tut

YO CHILL

[viral mason]

@tough vine recommend bro the ancient unknown tutorials u got

trust

[faint sable]

hall of shame the day i joined

what a flex

[viral mason]

bc ur weird

so u get a medal for being weird

[faint sable]

thanks

[tough vine]

😭 im not unc

[faint sable]

so like

[tough vine]

https://www.youtube.com/watch?v=pHhjg2JwdPI

[faint sable]

that was not the wind buddy

[tame oracle]

(it's alright, ive been in the #🥵│hall-of-shame more times than i can count)

[viral mason]

not the croissant tutorial 😭

[faint sable]

damn

[faint sable]

im using vovonox tho

[viral mason]

https://tenor.com/view/mario-madness-fnf-discord-mod-gif-10408089198531628400

[faint sable]

lowk lost

[tame oracle]

😭

[faint sable]

am i extra autistic

i have no clue what im doing

[tough vine]

IM SORRY 😭😭I tried to delete it asap

[faint sable]

do i put the voice i want in a preset

[tough vine]

someone asked for a croissant tutorial

so i helped her

[faint sable]

im deadass so confused

[faint sable]

i got told to use vonovox so im using vonovox

ts looks nothing like mine

[tough vine]

xbio

[faint sable]

yes

[viral mason]

I like garlic bread

[faint sable]

i think i got it

i just dont know pitch format

output volume n shit

ill just mess around with it

stfu Sapphire

i hate you

[tough vine]

ill make u

[viral mason]

I would appreciate the bread good sir

[tough vine]

i bake good

[viral mason]

you are my friend now

[faint sable]

damn

[tough vine]

YAY

[faint sable]

stfu sapphire

they are tryna talk

always ruining things

[viral mason]

it's automatic

[faint sable]

@viral mason how do i transfer it to discord

@tough vine

[faint sable]

its not working errr

idk what im doing

bro

do i need to install some virtual cable driver or something

man

nvm got it

[deft berry]

how do i load rvc pretrained models? (newbie here)

[tough vine]

go to vonvox, click edit, menu appears, drop ur .pth file into "Model" slot, and ur .index file into "Index" slot

[deft berry]

i just downloaded rvc from hugging face zip and extracted it and ran the web-ui
i have an amd gpu

and downloaded the voices

[tough vine]

open your rvc folder, look for assets then weights, drop ur .pth file there

For the index file make a folder inside logs with the model name and put it there, or just putting the .pth file in weights

once the file is moved click refresh

[deft berry]

kk thx

[deft berry]

does applio support DirectMl?

[ember coral]

Hi i have a problem on arch linux where my gpu doesn't show up in MMVCServer i have a amd gpu could really use some help

[raw raft]

would anyone know why as time goes on, it seems like the voice is degrading?

it was super good earlier in the day, and then as time goes on you can hear it get more roboty / tinny

(didnt change any settings)

[violet nacelle]

when i launch ai voice changer my cmd says client closed

and it wont work i tried everything

[deft berry]

i think its an issue with the vram leakage

[deft berry]

amd gpu or nvidia gpu?

[raw raft]

nvidia rtx 5080

[deft berry]

appolio or rvc project from hugging face?
and the question about the gpu wasnt for you 🙂

[hardy pollen]

how would I create an ai agent to read news financial news sites like bloomberg, finviz and yahoo finance daily and give me insights on whats going on without me actually reading each site

[raw raft]

ah sorry my bad. its rvc

i reinstalled rvc to a new folder and its working good rn

so im not sure, seems just an "over time" thing

[golden yew]

Can Someone Explain To Me Which One I Download?

I Was Here Yesterday, Followed The Instructions Of Redownloading On A New Pc, I Already Have A Virtual Audio Cable

Im A Little Confused

"Download All The Cuda:

So Do I Not Download The Cpu.Zip And Dml.Zip?

[violet nacelle]

nvidia

[violet nacelle]

it was working fine but i did tweaks to my pc and it didnt start anymore

[deft berry]

if u are using normal rvc from hugging face and not appolio
you can copy your weights and indexes aka back them up

then reinstall the rvc again

otherwise without a descriptive log file i cant know the issue specifically

[violet nacelle]

i deleted everything and reinstalled it and it wont work

[deft berry]

if its disconnecting only then probably its a port issue
have you tried changing the port?

in the bat file there is a number consisting of 4 digits
now idk which system you're on but maybe just maybe something else is listening on that port and preventing the web user interface from working

try adding 1 to that 4 digit number or subtracting 1 and run the bat file again

or a firewall issue idk

[violet nacelle]

the app is not working

[deft berry]

idk bro sorry

[violet nacelle]

np ty anyways

[hallow thistle]

RVC or W-Okada voice changer? And what is your PC GPU?

[hallow thistle]

Try Tg Develop or Deiteris W-Okada fork instead.

[hallow thistle]

RVC (retrieval-based voice conversion) or W-Okada voice changer?

[hallow thistle]

You could use @ helper for this one, just saying.

[raw raft]

the W-okada one
Real time voice changer client.

[violet nacelle]

rvc-nvidia

[hallow thistle]

RVC (e.g. Applio RVC) and W-Okada are two different programs of different purposes, despite both commonly confused as initials "RVC". Tg Develop's W-Okada fork and Vonovox are realtime voice changers, and are known to work with GeForce RTX 50 series GPU. Older versions (like v.1.5.3.18a) won't work with that.

[raw raft]

Yea I was using the w-okada one for sure

Not the beta version, I wa siding the 1.5 I believe

[hallow thistle]

RVC1006Nvidia.7z (retrieval-based voice conversion; non-realtime) or MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a.zip (the actual realtime voice changer)? To check your PC GPU, go to Task Manager, go to Performance tab, and spot GPU 0 or GPU 1 in the left panel.

[hallow thistle]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Deiteris' fork (modified version) of wokada that doesn't get updates anymore. GUIDE

⚔️ Wokada Tg-Develop Fork vs Vonovox

For Windows Nvidia, Both Wokada Tg-Develop 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 Tg-Develop 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

[raw raft]

Yea I was using that mmcssrversio one

In the spot on settings it did recognize the 5080

It defaulted to it

[hallow thistle]

"W-Okada", "VCClient" and "MMVCServerSIO" refer to the same program, just named in different places.

[raw raft]

Sorry I’m not home atm so I’m trying to go off memory 😭

https://huggingface.co/wok000/vcclient000/blob/main/MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a.zip

I used this

[hallow thistle]

Yes, that's the original W-Okada version (v.1.5.3.18a). Not always recommended to run for most scenarios.

[raw raft]

Ah ok, what would you recommend?

I’ve only ever used that one, and came back after like a year. So I’m not aware of other stuff

[violet nacelle]

its MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a. my gpu is 1650gtx

[hallow thistle]

Try Tg Develop's W-Okada fork instead of this version, like I said to other user earlier.

[violet nacelle]

is it also a realtime?

[raw raft]

Gotcha, thank you so much for your help

[hallow thistle]

This is the guide for Tg Develop's W-Okada fork https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/, and this is the GitHub for the program https://github.com/tg-develop/voice-changer/releases/tag/b2397, you download voice-changer-windows-amd64-cuda.zip.001 and voice-changer-windows-amd64-cuda.zip.002 from there.

[hallow thistle]

All W-Okada versions are generally the realtime voice changer.

[violet nacelle]

imma try it rn ty

[hallow thistle]

Pretrained models (the base models for training a voice model) and pretrains (F0/pitch extraction models) are two different models of different purposes. Those F0 models (e.g. rmvpe, crepe-tiny) are found in most RVC softwares not just W-Okada/Vonovox. What you described is simply about "to upload an already-trained RVC voice model to Vonovox", which I don't think it aligns with the user's initial query.

[violet nacelle]

it worked tysm

but is there best settings?

[hallow thistle]

In my screenshot, this is "Tg Develop's W-Okada voice changer fork". https://cdn.discordapp.com/attachments/1159290139609137264/1467527432457424969/image.png?ex=6987f52d&is=6986a3ad&hm=4746756b1532909271b58fab623ed1727996c085eee9846fe829812053d82ac7&

And this one is Deiteris' W-Okada fork. https://cdn.discordapp.com/attachments/1159290139609137264/1467535508019347671/image.png?ex=6987fcb2&is=6986ab32&hm=2cbc94f18349c650c5479456bb717cafb29b5b5e143a726440d59ae108d04cd2&

Yours looks nothing like any of these.

[violet nacelle]

theres this page also

[hallow thistle]

[raw raft]

[hallow thistle]

What you use** looks nothing like mine.** That one might be older than either more recent W-Okada fork for sure.

[viral mason]

what ancient technology is that

[violet nacelle]

this one

[viral mason]

yeah that's really old

you should switch to either wokada tg fork or deiteris

Namari can help if they wish as I'm busy rn

[violet nacelle]

can u give me the newest one but not this program cuz its not working with me

[hallow thistle]

You should've read this. #✨│ai-help message

[raw raft]

@hallow thistle works pretty good, id assume delay from talking to output is chunk size

[hallow thistle]

Set extra to "2.7 s", and set chunk to 128 ms. If you encounter an issue where "VB-Cable doesn't output any sound, there's "Virtual Audio Cable lite" as a backup plan.

[raw raft]

yea cause its making me use VB cable 16ch

if i use the normal one, it says "Error opeing output stream: Device unavailale [PaErrorCode -9985]

[hallow thistle]

Use:
Sample rate: 48000
Input: (WASAPI) microphone
Output: (WASAPI) Line 1 (Virtual Audio Cable)
Monitor: (WASAPI) your speakers/headphones.

[raw raft]

so doing that, it auto changes it back to 44100

i may redo vb cable just to be sure

cause heres why my devices are

if i use the bottom one, it just doesnt work. i get hit with that error again

[hallow thistle]

The same audio settings work for me.

[raw raft]

ok for sure, i will go ahead and redo vb cable real quick

[hallow thistle]

Try close your voice changer, and then force the voice changer to use "48000" by do this, replace everything "44100" with "48000". "Virtual Audio Cable" and "VB-Cable" are two differnet programs made by different authors.

[raw raft]

i mean since virtual cable works for you, i should just uninstal vb and go your route

[wild wagon]

What's the Google website link where I import the vocal audio then convert it to the AI voice?

[hallow thistle]

-colab

[patent trellis]

📒 Google Colab Notebooks

Google Colab is a Cloud (Remote Good PC) Service. While the Free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

• **Applio**

by IA Hispano
Google Colab

• **RVC Mainline**

by Hina
Google Colab

• **UVR5 NO UI**

by Eddy
Google Colab

• **UVR5 UI**

by Eddy
Google Colab

• **Wokada Deiteris Fork**

by Deiteris & Hina
Google Colab

• **Hina's Modified Original Wokada**

Google Colab

• **RVC-AI-Cover-Maker-WebUI**

by Shiro & Eddy
Google Colab

• **FaceFusion UI**

by Nick088
Google Colab

• **FaceFusion NO UI**

by Nick088
Google Colab

• **Music Source Separation Training (Inference)**

by Jarredou & Makidanye
Google Colab

[elder coral]

help me i'm making a model in our language

[chilly furnace]

nbd wanna kno brother 😭😭😭

[elder coral]

idc

[hallow thistle]

13 is the minimum age requirement for most social medias, including Discord, but then certain features might be limited for minors until 18. The question is, why did you reveal your age like this?

[elder coral]

i deleted it

it's an advanced april fools joke

[hallow thistle]

This month is not April, brother.

[elder coral]

Okay i mean does this model sound robotic

i understand this, but i can translate it for you

[elder coral]

uvr ui can't be loaded

[red harbor]

help

who can help me

[hallow thistle]

!howtoask

[patent trellis]

❓ 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 4060 8gb vram desktop)
• 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.

🚫 Prohibited Topics (We Will NOT Help With These)

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

• ANY illegal activities.
• NSFW/Porn.
  Requests for these topics may be ignored, not helped and result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.
• Don't Ask To Ask.

[paper topaz]

No free AIs available that make videos based off of a photo ?

[obtuse summit]

hi guys so i'm using Deiteris' W Okada Fork on the last version of windows 11 but i get an error saying: an error occurred during voice conversion check the command line for more details and cmd says this
Traceback (most recent call last):
File "voice_changer\VoiceChangerManager.py", line 212, in change_voice
audio, vol, perf = self.vc.on_request(receivedData)
File "torch\utils_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
File "voice_changer\VoiceChangerV2.py", line 159, in on_request
raise VoiceChangerIsNotSelectedException("Voice Changer is not selected.")
Exceptions.VoiceChangerIsNotSelectedException: 'Voice Changer is not selected.'

how do i fix? i have an nvidia 5060 and 32gb ddr4 of ram

anyone that can help me?? plzz

plzzzz

help me

[hallow thistle]

Deiteriss W-Okada fork wasn't compiled to work with NVIDIA GeForce RTX 50 series GPU. Try Tg Develop's W-Okada fork instead.

[hollow tusk]

Does anyone know the steps to install VVC into Linux? the instructions on the Github page are not clear

I understand I need to clone the git and then also download std linux(aarch64) Beatrice but from there the github does not explain

Im using a NVIDIA GeForce RTX 3060, 32RAM, using Arch CachyOS

[cosmic epoch]

Okay sooo, i submitted to this model making i think a week or a few days ago (don't remember exactlt when), why still can't i share models in the voice model secrio ? How much longer is this going to take? Is there a simpler and quicker way to get a permission to share a voice model,

[hallow thistle]

Do you mean "W-Okada realtime voice changer" or some another RVC (non-realtime) fork?

[distant lagoon]

Hello

can I ask you guys about an image if it is ai generated or not?

from the looks of it

[vale sparrow]

go ahead

[distant lagoon]

I cannot send images, can I reach out in private?

[vale sparrow]

yes ofc

[opal jewel]

hello everyone

How can I make a high quality cover for a song with an another voice?

[hallow thistle]

Try talk in here until your username turns blue, and you'll be able to send image in this channel, as well as #🏙│ai-images and #🌇│ai-videos.

[eternal galleon]

can someone help me setup this?

rtx 3080 10gb

idk what are the best options, as of chunk etc

[obtuse ferry]

Any luck with HeartMuLa? What are the kids using nowadays for music locally?

[golden viper]

yo something is wrong with my thing

can someone help me

[compact panther]

https://nohello.net/en/

give the details right away

[compact panther]

using more chunk and more extra increases VC quality, but also delay and hardware load

on low-end, i'd go for about 0.75s both, but on your gpu, i think 1.5s of chunk and 5.0s extra is fine

if you don't care about real-time VC, and you just want to record some audio with VC, it's better to use max on both... just mind the delay

[golden viper]

so basically it cuts off repeats itself like 3 times and the voice is just ass

[compact panther]

at least the following info is needed:

• your hardware: GPU, RAM, maybe CPU
• your operating system (e.g. Windows 11)
• the VC software you're using

[golden viper]

i have a amd ryzen 5 4600H

Windows 10

onnxdirect MLcuda

no i mean windows 11

mb

[compact panther]

that's just the runtime branch

what's the actual app called? (e.g. are you using w-okada)

and also, tell me your GPU

[golden viper]

w okada

[golden viper]

how can i find that

[compact panther]

task manager, performance tab

[golden viper]

i have 16gb ram

if thats what u mean

[compact panther]

that's something, but give me your GPU name nonetheless

and also, describe your problem a little more, like, e.g. how to reproduce it step by step

[golden viper]

the voice is just very high it cuts out and yeah

[compact panther]

9 out of 10 times, it's because of hardware load

[golden viper]

im on a laptop

this might be the reason

[compact panther]

that's usually a compounding factor yes

[compact panther]

@golden viper does your VC look anything like this

as in, can you see this section