✨│ai-help

[rocky star]

yoo

nickkk

nickkkk

[low shard]

what?

[rocky star]

nickk i need help

[low shard]

No need to ping people

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[rocky star]

i downloaded a ai voice changer

[low shard]

Elaborate your request

[rocky star]

how to send ss ?

i dont have perm

[low shard]

• what's your pc gpu
• what's your operative system?
• what you want to do?
• what tutorial link did you follow
• what is the issue

[low shard]

^

The more context, the better

[rocky star]

what's your pc gpu : 2gb ddr5
what you want to do? its working but idk what to doooo like i am trying to setup but its not working
what tutorial link did you follow : https://www.youtube.com/watch?v=YGyUgvx1J_Y
what is the issue : audio setting like input output

[low shard]

2gb ddr5
What? That's not even a GPU

[rocky star]

gtx 660

[low shard]

That seems your ram? I hope not, because 2gb are so little

[low shard]

that gpu is ancient

it's released in 2012

[rocky star]

its not brother

[low shard]

it is.. it can't run the voice changer at all

[rocky star]

its workingggggg on my pccccccc

[low shard]

it doesnt meet bare minimum requirements

[rocky star]

i jus need help w settingsss

can yall hellpppp ?

[low shard]

it wont work, the pc is too old

[rocky star]

it is working tho

[low shard]

!give-media-perms 1h @rocky star

show a screenshot

[low shard]

also, the version you're using is old

youtube tutorials are outdated asf

there's no single video tutorial that is updated

[rocky star]

hmm

but i think the setting matters

my friend using the same version

[low shard]

that's just opening the program, it's not even running and it wont work with that gpu

its like forcing to run cyberpunk on that pc, it wont work

[rocky star]

can ya join vc ?

[low shard]

I cant VC

[low shard]

that version is outdated asf

[rocky star]

how it need to load

[low shard]

dont use video tutorials

they are old asf

[rocky star]

bro i didnt copied form yt

[low shard]

simply, you cant your pc is too old

[low shard]

you can uninstall everything you got off that tutorial, its outdated asf

[rocky star]

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

[low shard]

also vb audio cable causes issues on windows as many users reported

[rocky star]

i am using this one

[low shard]

that is extremely outdated.. it's over a year old

[rocky star]

aghhhhhhhhhh

my brain already braining

[rocky star]

can ya help or no ?

[low shard]

simply forget everything you got off youtube

You got 3 options:

• Buy a better pc
• Run it locally (on ur pc) using the CPU mode of the wokada fork which has better performance https://rentry.co/ForkVoiceChangerGuide (but this isn't suggested as it could be unstable)
• Use **cloud **(remote good pc):

About Cloud, there are different services:

• 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)
  • How to use Original W-Okada's Voice Changer Google Colab (has a Guide) (currently broken)
• Kaggles (30 hours weekly of better GPUs, T4x2 & P100, harder to use, requires an account and a phone number):
  • W-Okada's Deiteris' Fork Voice Changer Kaggle (the best and only working one currently for free)
  • Original W-Okada's Voice Changer Kaggle (currently broken)

[rocky star]

hmmm

[low shard]

AI is an intesive task

you can't expect it to run on ancient pcs

chatgpt runs on cloud, not locally like this program

[rocky star]

can we atleast try to fix the problem without buying new pc ?

[low shard]

i told you, that version is over a year old and bugged

and your pc is too old

either buy a new one or use cloud (remote good pc with limited time)

AI isn't a 1 click program

[rocky star]

my friend using the same

how u can say thattt

and it is working for himm

[low shard]

he shouldn't use that version, it got performance issues

tell him to join the server, I can help him up too

[rocky star]

he made me download that one

[low shard]

he's missing on alot of performance updates

[low shard]

he shouldn't have, it has way worse quality and performance

[rocky star]

so what i do ?

[low shard]

read this #✨│ai-help message

Simply, either use cloud or buy a better pc

[rocky star]

i hate u haha

[low shard]

there's no other way, you could try cpu but that's just unstable and laggy, and considering you got such an old gpu, you surely got a bad cpu too

[low shard]

what? you can't expect every single program to run on a 2012 gpu

it's 13 years old dude

[rocky star]

lmao ty for letting me know

[low shard]

so, are you choosing cloud?

[rocky star]

is it free ?

[low shard]

read carefully, there's only updated written guides, all video tutorials are outdated

[rocky star]

can i get the link

and how to uninstall my ai voice changer

[low shard]

I already gave you it above: #✨│ai-help message

[rocky star]

ty sirr

[low shard]

yw and lmk, tell also your friend to join the server, he shouldn't use that messy old version

[rocky star]

he is in it he sended me this server link

):

[low shard]

You may want to @ him into this convo, so he knows he's using outdated software

[rocky star]

no

[low shard]

I mean it would help him have better quality and performance but ig

[rocky star]

i think he knows better he's been here for too long

[low shard]

he surely doesn't know better if he's using that outdated tutorial lol, might just be he joined here like a year ago and didn't realize programs updated

[rocky star]

i jus sended him that link idek if he is using this one

ima jus confirm that

haha

[uncut dust]

One message removed from a suspended account.

One message removed from a suspended account.

[solemn dawn]

Output is not working, I tried 3 different virtual cables.

[low shard]

alr

[low shard]

elaborate:

• your pc gpu
• a screenshot of the error

[low shard]

Applio is an RVC Fork

[low shard]

uninstall vb audio cable from windows app settings

https://support.microsoft.com/en-us/windows/uninstall-or-remove-apps-and-programs-in-windows-4b55f974-2cc6-2d2b-d092-5905080eaf98

then get vac lite https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#online-alternatives-colabkaggle

[brittle wing]

Is there an RVC app for Android?

Like I can import an RVC and use it offline

[solemn dawn]

I used an actual android voice changer two years ago and It actually worked, but it was sketchy so Im not gonna link it

[solemn dawn]

I followed these steps, now which output do I pick?

[uncut dust]

One message removed from a suspended account.

[low shard]

No, the only way to use rvc locally on Android is via termux (basically running Linux stuff), but it's very very low performing, not suggested at all

It can take mins even for a few seconds audios

[low shard]

Uninstall vb audio cable, uninstall voice meeter

Use Line 1 (vac lite)

You need only vac lite, no other vac

[uncut dust]

One message removed from a suspended account.

[stiff dome]

i wanted to learn how to make voice models for W-okada can anyone explain it to me or give me advice were to look for more info

[simple ore]

voicemeeter is fine

if it is used for something other than wokada virtual cable that is

[low shard]

Elaborate:

• your PC GPU
• your operative system

[stiff dome]

oh yea duh my falt
gpu: 4060
operrating system: windows 11

[low shard]

4060 laptop or desktop?

[stiff dome]

desktop

[brittle wing]

Hi, why do ai vocals get delayed w instrumental when mixing

I use applio colab what can I do to avoid delay/latency

[low shard]

As you got a good PC, you can use RVC locally, you can choose between:

• Applio: A fork of RVC with some extra features like Applio TTS, kinda faster and simpler but same quality tho
• Mainline: The original RVC

[brittle wing]

Hi, I asked smthg

[stiff dome]

im realy new to it so idk what woud be best but i think Applio or what doas the other one have in its faver going

[low shard]

I'm not sure why this happens, @simple ore is this a bug related to applio colab?

[low shard]

It's just better to use Applio for model training

[stiff dome]

thx than that i plan on making models of my frends

[stiff dome]

were do i install it gethub or huggingface?

[low shard]

read up the guide, get it from there

[simple ore]

I'm not sure.. the inferred audio should match the original

[stiff dome]

thx is the guide on the dc server or doas it have its own site? sry if this sounds dumb

[stiff dome]

found it i just googled thx for all the help

[low shard]

I literally sent you it

Hyperlink: blue text that when clicked redirects you to a link

[low shard]

Don't get it from google, click the blue Applio

[simple ore]

compare the source audio to inferred output

[stiff dome]

oh sry im dumb i thout it was just a highlight

[simple ore]

I dont see much difference

[brittle wing]

There's still a millisecond difference source aligns with instrumental

[simple ore]

the end is a bit off

I'll check, but it is likely because of the hop size used for f0/features

0.02s

[brittle wing]

Yes yes

Will you fix it.

[simple ore]

I need to check

[brittle wing]

Niceee

[low shard]

Dw and yw

[low shard]

Btw meanwhile you can manually cut that extra part

[stiff dome]

thx again

[brittle wing]

I don't cut I just shift em

But it's not always accurate so I'm worried

[sharp pagoda]

Hey anyone using n8n

[patent gazelle]

How can I use the ai voice changer in vr chat and has a really bad delay

[solemn dawn]

done but still no output, but atleast now i can hear my voice with monitor i wasnt able to before

[low shard]

Elaborate:

• your PC gou
• your operative system
• what you want to do
• what tutorial link did you follow

[patent gazelle]

Omen gaming desktop

Windows

Use voice ai im vrchat

Not sure

[low shard]

you should set the input voice setting of whatever program you're gonne use the vc in, to line 1

[low shard]

Omen gaming desktop
what's your pc gpu?

Windows
which?

Not sure
Did you install anything yet?

[solemn dawn]

i turned it off and on and it worked, over and out
😎👆
🗿🍷

[pallid olive]

Problem:
RuntimeError: CUDA error: an illegal instruction was encountered
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.

[soft stratus]

I had a problem, when I made a text to speech with any language in Elevenlabs, i found Elevenlabs with errors ; i want text to speech like Elevenlabs without errors

[pallid olive]

Can't I open the voice changer in a window on its own? It doesn't save my settings on the web version

[cyan tapir]

.mp3, .wav or .ogg format audio file. for Character chat on weights?

[tacit ice]

d

[brittle wing]

Yeah it is 😬

[soft stratus]

I had a problem, when I made a text to speech with any language in Elevenlabs, i found Elevenlabs with errors ; i want text to speech like Elevenlabs without errors

[low shard]

Elaborate:

• your PC GPU
• what's your operative system
• what you want to do
• what did you do
• what tutorial link are you using

[low shard]

Elaborate it more

[soft stratus]

What do you mean

[low shard]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#why-does-it-run-in-a-browser-and-not-its-own-window

Also it should save your settings

[low shard]

"errors" is too vague

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[pallid olive]

Nvidia RTX 3070 Laptop
Windows 10
I want to change my pitch
Changed my pitch
Tutorial? I'm just using the program like usual, after installing it from your guide

[low shard]

Oh yeah I just remembered we talked before

Be sure your Nvidia drivers and windows versions are updated, then retry opening the program

Does that illegal error appear when you do a specific thing?

[pallid olive]

Yeah, when I try to change my pitch in the program.

[low shard]

The program isn't just for changing pitch, it's an ai realtime voice changer

You have to get a model, select it, click start, then you can play around with the pitch of the ai voice model

[pallid olive]

You are an AI, huh... well, I AM doing that. I chose a model, I selected it and played with the pitch, and that's when the errors started happening.

[low shard]

I'm not an AI 😭

NickGPT Is Just as a joke

Could you please show a screen recording of the issue?

[low shard]

Changing it since I don't wanna confuse people lmao

[solemn dawn]

bro but voicemeeter helps make my mic better i dont use the VACs of voicemeeter my mic trash without it, are there alternatives?

[pallid olive]

😭 I'm sorry, but this felt so random to ask to specify
This pops up when booting it up:
2025-06-24 02:00:35.4311056 [W:onnxruntime:, transformer_memcpy.cc:74 onnxruntime::MemcpyTransformer::ApplyImpl] 2 Memcpy nodes are added to the graph main_graph for CUDAExecutionProvider. It might have negative impact on performance (including unable to run CUDA graph). Set session_options.log_severity_level=1 to see the detail logs before this message.

And- of course, now the issue isn't even there. Nvm, no idea what caused it before.

[low shard]

It's alright if you don't just use it for wokada, I was saying in case you get confused

[low shard]

Lmaoo

[simple ore]

I can not solve the last chunks mismatch, especially if they are not ending at some even point

but it should not longer lose frames in between chunks

[low shard]

Was this a known RVC issue? Or does this happen only when you enable split audio on Applio?

I don't remember having much of that issue, maybe slightly

[analog obsidian]

applio only issue

[simple ore]

no

dont think so

[analog obsidian]

time to open good old mainline

[simple ore]

all this BS

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/7ef19867780cf703841ebafb565a4e47d1ea86ff/infer/modules/vc/pipeline.py#L319

it chunks audio into 41s pieces

to stay under 6GB vram

but it loses 1 frame on each piece because hubert returns 1 less frame than 41*16000/320

[analog obsidian]

lol

[simple ore]

I threw all that bullshit padding nonsense away

it was not done properly

[analog obsidian]

how can this be fixed in applio?

[simple ore]

50 lines instead of 120+

now can process 5 hour audio in one go without splitting

if that's what you want

that's how you do it properly

[analog obsidian]

[brittle wing]

Finally is it fixed on the colabs too?

[simple ore]

no, just testing it in the experimental fork

[brittle wing]

does anyone know how to clone your own voice?

for context, i want to make subliminals for personal use without spending hours reading out affirmations.
Thank you 🙂

[brittle wing]

Oh brother.
im tryna download applio ai TTS but it doesnt look the same as the guide.
do you know what happened by any chance?

[simple ore]

Applio is not a TTS

[brittle wing]

Oh.

[simple ore]

it has TTS as a demo for the audio source

[brittle wing]

sorry

[simple ore]

it applies voice replacement on top of it

[brittle wing]

this is all just really confusing

im really bad with computers LOL

[simple ore]

the included tts is just microsoft edge screen reader

very lame one

[brittle wing]

oh

well thanks u saved me an hour of stressing

downloading okada was like 4-6 hours of my day yesterday

becuz of the weird guides

[brittle wing]

do u know a good ai tts?

or any that can clone your own voice?

you seem knowledgable

[simple ore]

what's the language you need?

[simple ore]

.

[brittle wing]

just english

sorry i was writing affirmations

thank you again

[simple ore]

new one is chatterbox

[brittle wing]

Ah okay. is it something local / that doesnt have a limit?

[simple ore]

git clone https://github.com/resemble-ai/chatterbox
cd chatterbox
python -m venv venv
venv\scripts\activate
uv pip install .
uv pip install torch torchaudio --upgrade --index-url https://download.pytorch.org/whl/cu128
uv pip install gradio
gradio_tts_app.py```

[brittle wing]

all the free web ones want me to sign up, and limit me

[simple ore]

with Python 3.11 and git installed

[brittle wing]

ah. i think i have python

[simple ore]

you can run it on CPU if you dont have nvidia card

[brittle wing]

if its not a bother, what is git?

😭

[simple ore]

git is a source control

[brittle wing]

what is that

[simple ore]

you can just download the zip file https://github.com/resemble-ai/chatterbox/archive/refs/heads/master.zip

and unzip into chatterbox folder

[brittle wing]

what should i click first after downloading these??

there is different ones

[simple ore]

you download this master.zip

you unzip it

then you open the command line (cmd.exe)

and you run the following command pip install uv python -m venv venv venv\scripts\activate uv pip install . uv pip install torch torchaudio --upgrade --index-url https://download.pytorch.org/whl/cu128 uv pip install gradio gradio_tts_app.py

[brittle wing]

does it have to be on 3.11 or can i use a newer version?

[simple ore]

3.11

[brittle wing]

sorry for so many quetions im just so lost

ok thanks ill go get that

i downloaded python but it just brings me to the set up where it says repair uninstall etc and no prompt

[simple ore]

show the screenshot

or actually open cmd.exe

and run where python

[brittle wing]

what is cmd exe i searched in in my files but nothing showed up

[frozen whale]

Hi, I'm Eduardo and I'm from Argentina. I need to know what I have to do now to be able to put an AI voice.

??

[brittle wing]

OH

thanks

LOL

and i copy and paste the text u sent earlier?

[simple ore]

where python

[brittle wing]

i feel like a hacker muhehehehe

it says in programs

ah

[simple ore]

if you have it installed properly you'll see C:\Users\user\AppData\Local\Programs\Python\Python311\python.exe

[brittle wing]

alrighty it says python 3.11 64 bit

i used the text and said invalid syntax

https://tenor.com/view/cyberpunk2077-edgerunners-lucy-hacking-linux-users-gif-26750192

[simple ore]

one line at a time

and so on

[sharp juniper]

can i get some help with the voice changer

[cursive vault]

uhh you got it yet?

[sharp juniper]

ye

i just cant open it

[sharp juniper]

the start bat things not working

[knotty moth]

try this recommended version

that may solve your problem

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

[sharp juniper]

ok ty

[grim harness]

there's like hella delay like 2-3s w amd radeon does anyone have a fix

[brittle wing]

Sorry for asking but is the error fixed in the colab too?

[simple ore]

no

[brittle wing]

Applio?

[simple ore]

f0_spin exp branch

the fix in #🔊│ai-development

[low shard]

what's your pc gpu and operative system? and what do you want to do?

[low shard]

do you need it?
what's your pc gpu and operative system? and what do you want to do?

[low shard]

never use video tutorials for rvc and wokada, they are old

you had an old original wokada prob off youtube

[sharp juniper]

i got it to work

[low shard]

elaborate:

• your pc gpu
• your operative system
• what you want to do
• what tutorial link did you follow
• a screenshot of the program

This is a general ai server, not just voice anymore, so you need to elaborate specifically

[low shard]

do you may want me to check your settings and give you suggestions for the best quality and performance?

[sharp juniper]

ngl custom voices are mad laggy

[brisk trail]

After re downloading and reinstalling everything I still get this error

[low shard]

!give-media-perms 1h @sharp juniper

show a screenshot of the program settings

also, wdym with custom voice? did anyone do you a model commission?

[sharp juniper]

no voice models be laggy

[low shard]

oh right we talked before, you got an rtx 4060, are on chrome and on windows

Could you elaborate what you did exactly to get this screen? maybe a screenrecording would be better

did you perchance rename the models?

[low shard]

show a screenshot of your program, i can help you fix that

it depends on your settings

[low shard]

lemme guess, you used a youtube tutorial?

that is the old original wokada, outdated asf, and vb audio cable can cause issues on windows

[sharp juniper]

ohhhhh

i see

[low shard]

forget everything off youtube, video tutorials are all old for wokada, uninstall everything

what's your pc gpu and operative system?

only written guides are updated

[brisk trail]

It happened after I installed a bunch of vpns but I deleted those and the problem still persists

[sharp juniper]

how do i check

[low shard]

You can check your pc gpu on Windows via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU

And you can check your operative system via settings > system

[low shard]

oh damn that's pretty old, 8 years old, it's basically the bare minimum, it will work but not super low delay and not on intensive games like marvel rivals

fortunately the updated wokada deiteris fork has improvements for amd gpus, just dont expect it to work on such intensive games

are you gonna use it for discord vc or games? if games, which?

[sharp juniper]

roblox and dc shi

[upbeat carbon]

@low shard heya i finally manage to download it!

in case if your wondering ;_;

[low shard]

vpns should have nothing to do with wokada? that's weird, does it do this from startup? could you show a screen recording?

[low shard]

it seriously took 2 days if not more man 😭 i even forgot what was your pc gpu, it would be better you tell me and then show me a screen recording of your program settings (guessing you already opened the program)

[brisk trail]

Yeah sure wait

[upbeat carbon]

yup! and here it is!

[brisk trail]

I'll send it to you in dms cuz where I live I lack privacy

[low shard]

close every other program in the background

set chunk to 200
uncheck sup1

on wokada, you can optionally:

• Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Reduce the delay via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

[low shard]

huh?

[brisk trail]

Nvm it fixed itself

Apparently my c drive was full

[low shard]

oh lmao

[upbeat carbon]

force what? disable what? and what delay!?!?! i'm so confused i'm sorry😭

[low shard]

force fp32 mode on is the name of a setting you find in advanced settings, if you turn it on, there will be a bit more delay (like a bit more time till the voice comes out), but it will have better quality

disable jit compilation is also the name of another setting in advanced settings

if you turn it on, you will have less delay

if you turn it off, your program will load slightly faster

and this part is just linking you to the part of the guide that explains you how to have as much less delay as possible

delay: the amount of time till your voice gets converted and comes out, it's the perf value at the top left of the voice changer, it's in milliseconds

@upbeat carbon was i clear enough or is there something you don't understand?

[upbeat carbon]

yes you were clear

i'm just silly i guess

anyways i forced the thingy

[low shard]

great, you could also put "disable jit compilation" on, so the program loads some seconds later, but you get 10ms less of delay

I would also suggest you the reduce delay part of the guide, But that's more complex, not sure if you wanna do it

[limber venture]

Did we ever get a better ai voice than RVC?

[lone cedar]

Im from America

[low shard]

no, rvc is basically dead, its original developers left it to rot
there's still nothing that beats rvc, but our engineers are trying to experiment with it via forks like applio and codename
this is also why the server is becoming more general ai related rather than just RVC

[low shard]

Hello, do you need any help? if so, it would be better you elaborate your help request

[lone cedar]

I need help with Setting up discord server that I'm making for my games I made im not very familiar with discord mod engaging and was wondering if there's any Ai server related building apps

[proven mortar]

guildedgoatgang's Slot Machine

Pocket: ⏣ 24,900
Winnings: ⏣ 0
-# Net: ⏣ -100

[ ]

[ ] 0x

[ ]

Bet: ⏣ 100 | Min: ⏣ 100 | Max: ⏣ 100,000

[upbeat carbon]

i can give it a shot

so how do i do that?

[low shard]

it's more complex, you'd have to read a kinda long part of guide in https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

it's not as easy as just changing a setting and would also have to mess a bit with your windows audio settings

[low shard]

This channel is meant for helping users to use AI Programs, rather than helping in discord moderation

[upbeat carbon]

why it forcing me to download opera gx so much?!?!😭

[low shard]

wdym??

whats forcing you?

[upbeat carbon]

wait nvm it's gone now

[low shard]

wokada doesn't even work well on opera gx

and wokada doesn't have ads

nor sends analytics or shit

so what forced u exactly?

[upbeat carbon]

okay so i got flexASIO and when i opened it, it said:are you sure you want to open this app from unkown publisher? :fear:

[cursive vault]

just get an ad blocker or something

[low shard]

that's just windows general warning for any app without a known publisher (like microsoft)

[low shard]

🙏

[upbeat carbon]

ah understanable!

[low shard]

yeah we aren't microsoft lol

even discord gives a general warning when you send any direct download link through it

so they cant get their ass sued and warn people about scams, but that doesn't even check if the download link is something like whatsapp or a scam

btw Wokada is FOSS, Free and Open Source Software, meaning it will always stay free and everyone can check its code

[upbeat carbon]

wheres audio server?

can't seem to find it on the page

[low shard]

on wokada

you should have currently selected client, you have to change it to server

also the image is wrong, it should be 48000

[upbeat carbon]

hm it's only this

[low shard]

wdym? can you show a full screenshot

[upbeat carbon]

now?

[low shard]

try it now

[upbeat carbon]

nope, dosen't work

[distant turtle]

-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

[dark ginkgo]

Is this accurate?

I was having issues not being able to fix my rtx 5060ti from being used with most ai softwares.

Every software demands that version of python.

[simple ore]

not true

you need pytorch cu128

[brittle wing]

Did you fix it in like the applio colab

[simple ore]

python version depends on the included libraries, usually 3.11 is fine

[simple ore]

not gonna happen any time soon

[brittle wing]

I'm sad

[simple ore]

it is a reality.. every other rvc has the same issue

[brittle wing]

I know

I have used them all

[low shard]

huh? is line 1 the input in the other program you're tryna use wokada in?

[simple ore]

if you want an immediate fix, you can just copy-paste the code

[low shard]

just manually shift or cut it up tbh

[brittle wing]

Send the code and tell me where to paste it

[simple ore]

it is a little tricky, it is just one frame every 41s

[simple ore]

#🔊│ai-development message

[brittle wing]

Wait

[dark ginkgo]

I get this error log

[brittle wing]

Where do I paste it

[simple ore]

are you really trying to use old as shit automatic11111?

[dark ginkgo]

The only ai software that hadn't had this issue or rather run at all is autoamatic 1111

all the other ones dont work

[simple ore]

you have python 3.12, that's an issue

[brittle wing]

Where do I paste it

[simple ore]

and it also tries to install

instead of cu128 you need

[brittle wing]

Do I paste that in a separate cell

[low shard]

I used to just manually fix it myself if that happened to me when I made ai covers

[simple ore]

please just chill the f out

[brittle wing]

Uh okay

I'm gonna

Chill

[simple ore]

this is the code you can use locally

I've never said it is for colab

for colab you need to use experimental branch

instead of 3.2.9

if you use the experimental branch, you can replace the file in rvc/infer

[brittle wing]

Uh I don't have q computer

[simple ore]

dont ask me how, figure it yourself

[brittle wing]

I'm not a coder like you to understand everything

Uh so I'll just shift then

Whatever

[brittle wing]

How-

[low shard]

it's just better you wait till they have time to push a new applio colab update or just shift it manually

experimental is not meant for public usage, and is not stable

[brittle wing]

I'll just shift atp

[simple ore]

you can edit the install cell

[brittle wing]

Idk

[simple ore]

!git clone --depth 1 {decoded_url} --branch exp/f0_spin --single-branch

[low shard]

there could be other changes that could be not stable, I don't work for applio team like noobies but yeah its just better to shift manually

[brittle wing]

is tencent hunyuan 3d safe to download from?

[brittle wing]

Yes it's easier

[simple ore]

and then you can double click pipeline.py

[brittle wing]

Yes and replace which file

[simple ore]

and paste the code from that file over

[brittle wing]

I understand now

Can I delete & then upload the other file in the folder like replace the file in the folder

[simple ore]

yes, you can delete the old file, then use upload button, and move the uploaded file into that folder

[brittle wing]

Exactly

Will that stuff work.

I hope it won't break-

[simple ore]

use at your own risk

someone needs to test that I did not break this shit

loaded audio (79096077,) 32000
after resampling (39548039,) 16000
chunk before padding (188039,)```

[thorny folio]

how to fix it was working yesterday

and its not working now

[simple ore]

so the resulting audio may be shorter if the input file is weird

[thorny folio]

😔

[brittle wing]

Okay I'll try for you

[dark ginkgo]

Ok, This explains why although automatic 1111 works and all other ai software doesn't, it also was a bit slow despite being a new card.

[brittle wing]

Okayyy

[simple ore]

stop asking chatgpt about 5000 series card, it does not know what the problem is

I told you what the problem is - you need to fix the requirements so it installs cu128 torch

[low shard]

This is a general AI server, not just voice anymore

Elaborate:

• your pc gpu
• your operative system
• what you want to do
• what is the issue
• what tutorial link did you use

[low shard]

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[low shard]

We can't know what program you're using, we aren't an ai voice server, we are a general ai server and do many programs

[thorny folio]

3060 ti
windows 11
i want to make it working again
the ai voice is not working on my mic
i used https://www.youtube.com/results?search_query=how+to+sound+like+as+a+girl+voice+changer

[low shard]

all video tutorials are old for wokada and rvc

you're just wasting time using youtube for ai

ai changes at sonic speed

[simple ore]

this is the problem

[thorny folio]

any new tutorials

[low shard]

uninstall everything you got off youtube, vb audio cable and the old original wokada

[low shard]

there's no single updated video tutorial, only written ones

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

read wokada deiteris fork, the 1st link

dont get used to 1 click programs with a video tutorial, since Open Source AI isn't like that

AI is a complex and intensive task

[low shard]

@thorny folio uninstall everything you got off youtube, read that guide and let me know

[dark ginkgo]

So do I use this one then?https://download.pytorch.org/whl/nightly/cu128 As that's the only cu128 I can find

[simple ore]

just change that file, cu121 -> cu128

and remove version on the next line

it will install 2.7.1 most likely, that's the only version that is compatible

[dark ginkgo]

Also mine looks like this

[simple ore]

remove ==2.3.1 and ==0.18.1

change that

[dark ginkgo]

so like this or the whole bottom line?

[simple ore]

yes

you may need to nuke the existing venv

[dark ginkgo]

already did that

[simple ore]

if you got stuff installed there already

though I'm not quite sure why you're bothering with auto1111

it is morally outdated

vs ComfyUI or SD.Next

[low shard]

ComfyUI for the win 🙏

[dark ginkgo]

I have a TON of things and custom preset configured on automatic 1111. But I'm not trying to use automatic 1111 or rather thats not what I'm trying to fix as I'm using forge ui. I want to use flux. Comfyui was having the same error as well with wan and all other software like gradio, appolio, ect

like base a1111 is the ONLY ai program that works somehow

[brittle wing]

@simple ore tried and that's what I got from the gradio link:

[dark ginkgo]

I tried sd.next and that didn't work either, same error logs as forge ui

[simple ore]

what's in the colab log?

[brittle wing]

Wait

[simple ore]

if you tried any time recently, it should have all the fixes for 5000 series

dev branch especially

[dark ginkgo]

last week

[simple ore]

I said log

as an output text

but I guess you did not fix the install cell

[brittle wing]

"An error occurred connecting to Discord: could not find discord installed and running on this machine"

[brittle wing]

I changed it

The branch

[dark ginkgo]

This is the error log I get on forgeui, comfyui, sd.next

I installed this version of sd.next btw https://github.com/vladmandic/sdnext

[simple ore]

no kernel image = using torch that is not cu128

I pray you did not buy 8GB 5060Ti

[low shard]

you would be shocked by how many people still have 2gb/6gb vram..

while nowdays you need atleast more than 8gb

[simple ore]

@dark ginkgo if you want to use flux-dev 1, you need to use quantization in sd.next

especially if you got 8GB card

[upbeat carbon]

it's in output

[simple ore]

that's not the actual error, but it seems gradio.live does not work, you need to use ngrok tunnel

[brittle wing]

What

[low shard]

#📰│dev-updates message

[low shard]

thx for saying it, just warned users

[low shard]

average day of a cloud user, cloud is unstable asf

[haughty shuttle]

https://open.spotify.com/intl-tr/track/1k5qY9w9drauPjRGlDWrAw?si=a57cb915949842bf

[dark ginkgo]

16gb, who tf is using 8gb card still in 2025? XD

[haughty shuttle]

which character belongs to

[dark ginkgo]

I upgraded from rtx 3060 12gb to rtx 5060 ti because I needed more vram for wan 2.1

in the process, I need to fix the ai systems I already had.

[dark ginkgo]

ok this is really getting on my nerves

[simple ore]

if you got the right torch, this is running out of memory then, check task manager/performance tab

you are using forge, I dont think there's aproper automatic quantization

[low shard]

flux.1-dev q8 is so goated

99% identical results, a bitfaster, slightly less vram usage, slightly less storage

[simple ore]

in sd next, only uses 8GB vram

[low shard]

are u using flux.1-dev q8? or regular flux?

[simple ore]

dev1 regular

the devs are making fun quantizations

1-bit, 3-bit

uint8 used 12GB

red gala

[distant turtle]

Does anyone know why when I open Applio with colab and press on public url the page loads without letting me go to the site?

thanks

[simple ore]

gradio live is offline

[unreal hinge]

does anyone know why models without indexe files dont work anymore?
i was using a model and it randomly stopped working and now i cant use any that dont have index files

[glacial pollen]

What do you use specifically
( which fork, framework, app, whatev - with rvc models

[dark ginkgo]

prob not

[glacial pollen]

Also, do you get any errors ? ( any logs ?

[unreal hinge]

using this: https://huggingface.co/Shadicti/deiteris-Fork/blob/main/voice-changer-windows-nvidia-b2332.zip

w rvc models

[glacial pollen]

ah, w-okada

[dark ginkgo]

Is ai just compatible with the current blackwells card?

or does the rtx5060 ti use a different version or something

[glacial pollen]

well, models should work with and without index, regardless of what it is
whether static stuff such as rvc or applio or real-time, such as w-okada n so on

[glacial pollen]

So something must be off

Does the console give you any info?

any errors, warnings etc

[unreal hinge]

2025-06-24 10:00:12,350 WARNING  [PipelineGenerator] Index file not found. Index will not be used.
2025-06-24 10:00:12,350 INFO     [Pipeline] GENERATE INFERENCER<voice_changer.RVC.inferencer.RVCInferencerv2.RVCInferencerv2 object at 0x000001ECC97A5520>
2025-06-24 10:00:12,350 INFO     [Pipeline] GENERATE EMBEDDER<voice_changer.embedder.OnnxContentvec.OnnxContentvec object at 0x000001ECC9A3EA20>
2025-06-24 10:00:12,351 INFO     [Pipeline] GENERATE PITCH EXTRACTOR<voice_changer.pitch_extractor.CrepePitchExtractor.CrepePitchExtractor object at 0x000001ECC9AAF200>
2025-06-24 10:00:12,359 INFO     [RVCr2] Initialized.```

when i load the ones that dont work

[glacial pollen]

And that's all there is?

[unreal hinge]

yeah

[simple ore]

this error is because you're using torch that is not cu128

sm_120 is only supported by torch 2.7.x+cu128

[dark ginkgo]

yeah but I did this thing so why didn't it install the new one then?

[glacial pollen]

Have you tried normal w-okada or vonovox?

also, your hardware? ( gpu )

Another thing I wanna ask ~ Have you done anything specific so then models stopped working?

[simple ore]

what exactly did you download?

repo/zip file

[glacial pollen]

that one

2332

[dark ginkgo]

her is the python version I have on system

[simple ore]

not asking you

[dark ginkgo]

I assume I need a newer one

[simple ore]

I'm asking how did you download stable diffusion

[glacial pollen]

ah

mb

[unreal hinge]

the normal w-okada worked for a bit than stopped working with those specific voice models out of nowhere

, so i switched from the one from the guide, and then it stopped working completely for those specific models

and i didnt touch anything related to the models specifically, just some audio settings that are related to how its mixed

AMD Ryzen 7 8845HS, RTX 4070

[glacial pollen]

Try vonovox then

https://github.com/dr87/Vonovox

See if you get the same issue there

[dark ginkgo]

automatic installer, 2 years ago. But again, stable diffusion isnt the issue, EVERYTHING ELSE IS so forge, the same official forge install method, so is comfyui, so is applio and gradio

before the upgrade to the rtx5060ti, it was working just fine

after the upgrade, it broke everything

except for the original, old a1111

[simple ore]

okay, I'm just gonna take https://github.com/lllyasviel/stable-diffusion-webui-forge/archive/refs/tags/latest.zip

and install it

[unreal hinge]

wdym

[simple ore]

what model did you download?

[unreal hinge]

e-woman in the voice models channel

[dark ginkgo]

how new is that repoo? did they have an update cause last month, it didn't work

[simple ore]

I made the fix in the lauch file

    torch_index_url = os.environ.get('TORCH_INDEX_URL', "https://download.pytorch.org/whl/cu128")
    torch_command = os.environ.get('TORCH_COMMAND', f"pip install torch torchvision --extra-index-url {torch_index_url}")

[dark ginkgo]

ah, cause whenI try to run the same thing you did, I got this

ok, I assume I need python 318 correct? or whatver is the newset one that works with it

[simple ore]

3.10 or 3.11 python should work fine

[unreal hinge]

yep, vonovox worked

[simple ore]

3.12 may not if the project is old, 3.13 definitely no go

[unreal hinge]

tysm

[dark ginkgo]

so I can stick with 3.10 then

[simple ore]

yes

[dark ginkgo]

in that case, why it's still not working I dont understand. I inputted the same code replacement as you showed

[glacial pollen]

Np, glad it worked

enjoy

[simple ore]

as I see the right torch got installed, so it should work

[dark ginkgo]

hold on, it is working finally! Not sure why the 5th time of doing the exact same thing fixed it but sure. I'll take it.

[simple ore]

for all other projects the fix is the same - activate venv, run pip install torch torchvision torchaudio --upgrade --index-url https://download.pytorch.org/whl/cu128

may break very old projects that did not have torch.load fixed

[dark ginkgo]

cool, depending on how fast these setup is, this is may or may not be a 1-3 day projects. I have. A lot of ai projects across many drives

honestly should buy a new dedicate ai drive like 2 or 4tb of it

[brittle wing]

It says error for ngrok token now

[simple ore]

did you get the token?

did you paste it in the cell?

[shadow roost]

i need help

[latent kettle]

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[pastel solstice]

I want a new pfp of a realistic giant panda that only has black and orange fur bur every ai I use still leaves white fur on the face

[low shard]

uint4? I think last time I ran it it was fp16

dunno if im just being dumb, since I use comfyui lol

[simple ore]

fp16 is not possible

"fp16 - uses 2x less vram and 2x faster speed then fp32 while being same quality but only works in gpu and unstable in training(Flux.1 dev will take 24gb vram at the least with this)"

fp8 takes 12GB

[dark ginkgo]

ay, flux is working again! Passed my red apple test so good enough for me. Now gotta do the same with, comfyui, gradio, appolio, and also test to see if a1111 can be updated too or not. I have a tester rig for it.

with how cheap the 5060ti 16gb is, I wish there was more documentation regarding ai usage and optimization for it.

[low shard]

sorry I only remember using it in fp8 and q8, i dont use it since months tho but its kinda different on comfyui, i remember it was around 12gb vram

[simple ore]

it was q8 gguf I bet

[low shard]

btw keep your drivers updated

[shadow roost]

i am on what do i get i have nvidia card

[dark ginkgo]

when was this? I havent seen any new driver as of recent ever since buying it. (which on the gaming side of things, really needs it. cant play cyberpunk 2077 still)

[kindred quail]

hello guys, is this the voice ai serv ? I just installed the app and get stuck in a page. Ping me if u can help ty

[shadow roost]

i am on what do i get i have nvidia card

[dark ginkgo]

oh cool, it just came out few days ago

and also oh cool, the update is broken lol!

[low shard]

June 17

[low shard]

This is a general ai server, we don't do just voice anymore
We use multiple programs

Elaborate:

• your pc gpu
• your operative system
• what you want to do

[low shard]

huh

[low shard]

This is a general ai server, we don't do just voice anymore
We use multiple programs

Elaborate:

• your pc gpu
• your operative system
• what you want to do
• what tutorial link did u use

[dark ginkgo]

it requires deleting entry from registry! Lol Nvidia, you messed up!

[low shard]

also don't spam your requests, be patient

[low shard]

wtff? where did you read that

[dark ginkgo]

I tried to run the Nvidia app, it said, new version needs to be installed! Mind you, this app is the latest prior to this update.

meaning, they messed something up on their end with the rtx 5060ti, or even just the software iteself

team green's track record on the driver side of things has been bad but god does this not help it's pr

[low shard]

atp i dont regret not waiting for 50 serie 😭

[glacial pollen]

same

everywhere I go it's just issues with 50xx series

[low shard]

I thought i was dumb for getting an rtx 4060 ti 16gb in december but maybe its not that dumb

the rtx 50 serie initially was expensive asf

and they still have issues after 6 months

[simple ore]

wrong, there were mspr 5070tis

[glacial pollen]

Yeah, both the price, availability and reliability
at least in my region

[simple ore]

not any more

[low shard]

u sure? when i firstly checked online i could find rtx 50 serie at like 4k

maybe that was just the 5090

[glacial pollen]

imo AMD should wake the fuck up and start investing in AI sector, that includes proper pytorch and all other libraries support

and then? hell, I might change my mind and switch back to em

[simple ore]

5090 were ridiculous from the start

but there a few 5070TI models at MSRP for a few weeks

[low shard]

yeah, they got good deals but their support sucks

[glacial pollen]

yup

As much as I never had " famous amd driver issues " the AI support most certainly sucks ass

and even if one was to tell me to just simply switch to linux, no thanks.
Not only it's, imo, half-baked but I ain't switching OS

doesn't justify it for me

[simple ore]

Nvidia was worse driver issues this year than AMD ever had

[glacial pollen]

Wonder if similar issues happen for their AI / powerhouse line

[simple ore]

most of AMD's issues were due to ultra low power mode that did not work properly and caused random crashes

[low shard]

I think the drivers issues were mostly just for the 50 serie, since I never had them , tho this doesn't make it any better

[glacial pollen]

cause if it's just ' standard consumer ' aka gaming ones, welp

[simple ore]

AMD's pytorch is coming soon(tm)

[glacial pollen]

I do hope that's the case

tbf, still holding back from updating

having heard of all the craps and black screens

[low shard]

my drivers are up to date and never had a single issue

[glacial pollen]

Last thing I'd want is rolling back and forth, fkin around ddu

[glacial pollen]

ah, but then, you have 40 series

[low shard]

yeah thats why surely

50 serie is cooked

such a delusional serie

[glacial pollen]

yuh

[low shard]

wonder how the 60 serie will go

[glacial pollen]

Like, I don't have any issues with " wuaaa, AI over raster " but honestly, that driver bs is a mess and biggest turn-off

[low shard]

tho atleast dlss4 is great

previous series are getting more benefits atp

[glacial pollen]

Oh yea, about that

wut bout 30 series?

getting any benefits or nah?

[low shard]

i think so but not related to frame gen

[glacial pollen]

haven't honestly looked much into gaming lately so, ain't entirely sure on the status

ye, that's fine, frame gen sucks ass imo

I'm one of those people that can sense few extra ms of latency and cry over it ( ye, ex osu player )

bothering as hell

[rancid island]

Error
unhandledrejection
no error stack
TypeError: (0 , Zz.removeDB) is not a function
TypeError: (0 , Zz.removeDB) is not a function
at http://127.0.0.1:18888/index.js:2:3293292
at m (http://127.0.0.1:18888/index.js:2:1536451)
at Generator.<anonymous> (http://127.0.0.1:18888/index.js:2:1537797)
at Generator.next (http://127.0.0.1:18888/index.js:2:1536880)
at e (http://127.0.0.1:18888/index.js:2:1543652)
at s (http://127.0.0.1:18888/index.js:2:1543855)

[glacial pollen]

Not sure how I endured cp2077 on rx 570 with frame gen 💀

[rancid island]

how to fix this?

[low shard]

Elaborate:

• your pc gpu
• your operative system
• what you want to do
• what tutorial link u used
• what you did

[glacial pollen]

Oh yea, in that case, looks sweet
tho, wonder how it's gonna go with the latency / input lag thingy they announced

[low shard]

This is a general ai server, not just voice anymore, we use multiple programs, you need to elaborate

[low shard]

ehh I use frame gen and don't really find it bad at all, but im a controller player not those who use everything wired for the 1ms delay

[crude flame]

never had this issue and im on a 5060ti 16gb

[rancid island]

RTX 4060, just downloaded it and whenever select input then this error pops up , I used the forum but it didnt help cause my antivirus is disabled

[low shard]

downloaded it
Downloaded what?

[simple ore]

dlss + gsync

[low shard]

There's thousands of AI programs

[dark ginkgo]

yeah but it was so CHEAP!

[rancid island]

downloaded the client from github

[low shard]

also what's your operative system? and what's your anti virus? and what do you want to do? what tutorial link did you follow?

[crude flame]

ah

[glacial pollen]

Yeah, might be
Not so much if you live in Poland

tech prices always sucked ass in here so

esp Ngreedia

[dark ginkgo]

I got mine AT MSRP of $429

[low shard]

client, in computing means:

(in a network) a desktop computer or workstation that is capable of obtaining information and applications from a server.

You need to elaborate what you mean

reply to all my questions please

[dark ginkgo]

16gb vram, with better hardware then my rtx 3060 12gb, for only $100 more. That's a steal!

[low shard]

AI is a complex and intensive task, don't expect any 1 click program

[rancid island]

Windows 11, Windows defender (disabled ) , I just wanted to enable the voice changer to use the E-Woman model

[low shard]

you shouldn't have to disable windows defender for it generally, where did you get the program?

what tutorial or download link did you use?

[rancid island]

Github

[low shard]

be aware all video tutorials are old

[low shard]

send the link

[rancid island]

I used the forum in the server but it didnt help

[dark ginkgo]

That's 4 whole GB of vram! When your most "demanding" games uses 9gb of vram, (hence why 8gb card dont cut it) then the 4g for $100 with other better newer stuff is just no brainer.

[low shard]

send the link of wherever you downloaded the program off

[rancid island]

https://github.com/w-okada/voice-changer/blob/v.2/docs_i18n/README_en.md

[low shard]

that's original wokada

its not suggested anymore, delete it

[rancid island]

the program ?

[low shard]

uninstall vb audio cable too if you have it, it causes issues on windows

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[crude flame]

lucky 😔

i paid 70 bucks over msrp for my gpu

[low shard]

1st link, read wokada deiteris fork

[low shard]

yes

[rancid island]

am I supposed to download it from here?

[low shard]

you're supposed to read the 1st link tutorial to know how to download the updated program

be sure to not miss any steps, missing a single step can cause to fuck up your whole audio system, the only exception is when a step is unrelated to your situation, like Mac

[rancid island]

thank you I will let you know

[low shard]

you're welcome, be sure to deleted the outdated program you had before, and let me know

[rancid island]

so its like opening in the browser now unlike the previous one...is it normal ?

[low shard]

yes, read why on https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#why-does-it-run-in-a-browser-and-not-its-own-window

[rancid island]

but the problem is same ...I am on Edge btw

[simple ore]

look at the error in the console window

not in the browser

[low shard]

did you do anything or did it instantly crash

[rancid island]

while changing the input

[low shard]

#✨│ai-help message

also to which input<'

i hope you aren't using vb audio cable

[rancid island]

to litreally anything at this point

[rancid island]

checked multiple times

[low shard]

show the error in the console

[low shard]

also, you got vac lite right? https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#virtual-audio-cable

[rancid island]

pretrain/rmvpe.onnx
2025-06-24 22:32:43,861 INFO [Downloader] Verified pretrain/fcpe.pt
2025-06-24 22:32:43,870 INFO [Downloader] Verified pretrain/fcpe.onnx
2025-06-24 22:32:43,871 INFO [WeightDownloader] All weights are loaded!
2025-06-24 22:32:43,871 INFO [main] protocol: HTTP
2025-06-24 22:32:45,179 INFO [loader] Loading faiss with AVX2 support.
2025-06-24 22:32:45,194 INFO [loader] Successfully loaded faiss with AVX2 support.
2025-06-24 22:32:45,560 INFO [VoiceChangerManager] Initializing...
2025-06-24 22:32:45,575 INFO [DeviceManager] Initialized DeviceManager. Backend statuses:
2025-06-24 22:32:45,575 INFO [DeviceManager] * DirectML: False, device count: 0
2025-06-24 22:32:45,575 INFO [DeviceManager] * CUDA: True, device count: 1
2025-06-24 22:32:45,575 INFO [DeviceManager] * MPS: False
2025-06-24 22:32:45,578 INFO [DeviceManager] Switched to CPU (cpu). FP16 support: False
2025-06-24 22:32:45,578 INFO [IORecorder] -------------------------- - - - C:\Users\ravia\AppData\Local\Temp\tmpa09vhswg\tmp_dir\in.wav, C:\Users\ravia\AppData\Local\Temp\tmpa09vhswg\tmp_dir\out.wav
2025-06-24 22:32:45,580 INFO [VoiceChangerV2] Allocated SOLA buffer size: 4800
2025-06-24 22:32:45,580 INFO [VoiceChangerManager] Initialized.
2025-06-24 22:32:45,580 WARNING [VoiceChangerManager] Model slot is not found -1
2025-06-24 22:32:45,580 INFO [MMVC_Rest] Initializing...
2025-06-24 22:32:45,594 INFO [MMVC_Rest] Initialized.
2025-06-24 22:32:45,594 INFO [MMVC_SocketIOApp] Initializing...
2025-06-24 22:32:45,594 INFO [MMVC_SocketIOApp] Initialized.
2025-06-24 22:32:45,875 INFO [main] --------
2025-06-24 22:32:45,875 INFO [main] The server is listening on http://127.0.0.1:18888/
2025-06-24 22:32:45,875 INFO [main] --------
2025-06-24 22:32:47,065 INFO [MMVC_Namespace] Connected SID: wLZwFQ74Ep-bvW7rAAAB

its now happening instatly

without selecting anything

[rancid island]

yea

any fixes?

[low shard]

try using chrome or firefox

[rancid island]

2025-06-24 22:40:11,981 INFO [PipelineGenerator] Loading index...
2025-06-24 22:40:11,981 WARNING [PipelineGenerator] Index file not found. Index will not be used.
2025-06-24 22:40:11,982 INFO [Pipeline] GENERATE INFERENCER<voice_changer.RVC.inferencer.RVCInferencerv2.RVCInferencerv2 object at 0x00000234EFE8BDA0>
2025-06-24 22:40:11,982 INFO [Pipeline] GENERATE EMBEDDER<voice_changer.embedder.OnnxContentvec.OnnxContentvec object at 0x00000234EFEFCCE0>
2025-06-24 22:40:11,982 INFO [Pipeline] GENERATE PITCH EXTRACTOR<voice_changer.pitch_extractor.RMVPEOnnxPitchExtractor.RMVPEOnnxPitchExtractor object at 0x00000234EFB433B0>
2025-06-24 22:40:12,045 INFO [RVCr2] Initialized.
2025-06-24 22:40:12,045 INFO [RVCr2] Allocated audio buffer size: 9792
2025-06-24 22:40:12,045 INFO [RVCr2] Allocated convert buffer size: 18080
2025-06-24 22:40:12,045 INFO [RVCr2] Allocated pitchf buffer size: 114

[rancid island]

will let you know

[rancid island]

didnt help...same problem exists

now its happening when I am trynna select the input device

2025-06-24 22:43:04,347 INFO [MMVC_Namespace] Disconnected SID: xjSHtDqddepJGBSUAAAF
2025-06-24 22:43:05,122 INFO [MMVC_Namespace] Connected SID: SnA7uRKU35Qiuj81AAAH
2025-06-24 22:44:34,234 INFO [MMVC_Namespace] Disconnected SID: SnA7uRKU35Qiuj81AAAH
2025-06-24 22:44:41,018 INFO [MMVC_Namespace] Connected SID: _5tAtH8YxGjmaCwYAAAJ

[low shard]

you sure you gave your browsers microphone permissions?

aloso, show all the input options before selecting one

!give-media-perms 1h @rancid island

[simple ore]

switch to server mode

[rancid island]

yup

[rancid island]

Solved it...I am so sorry to waste your time and I feel really bad about it ..thanks for helping me tho..it was the permission issue

last quick question should I continue using this one or the older one?

[low shard]

oh lmao

[low shard]

this one, the older one can be deleted

[low shard]

set extra to 2.7

chunk to 100

input to microphone

output to line 1

monitor to headphones optionally to hear urself

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

• Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Reduce the delay via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

[low shard]

this program has way better quality and performance

you can also try the things i said above

[rancid island]

like I used in at 0.3 but here its in percentage...so how should I like select

same goes with out

[low shard]

are you talking about in and out?

play with them till they are good, there isn't an universally good setting lol

[rancid island]

yea what are those exactly

[low shard]

idk why u used 0.3

[rancid island]

saw on youtube smwhere

are those volume? oh alr

[low shard]

all video tutorials have outdated info, forget everything you got off youtube

did they also tell you to use crepe lmao

that's old asf

[rancid island]

yea

what should I use then

[low shard]

hell nah, thats over 2 years old outdated info

[low shard]

rmvpe

pls just forget everything they teach you in those tuts

they are as updated as using windows xp in 2025

[low shard]

yeah... idek why the guy told you to do that.. its just volume

[rancid island]

damn you are roasting em now haha

[rancid island]

whats the difference?

[low shard]

lol, I also gave you optional settings above to get better quality and less delay, if u wanna check them out #✨│ai-help message

[low shard]

better quality and robust to noise

[rancid island]

whats formant?

[low shard]

Formant Shift: Alters harmonic frequencies and changes the voice timbre without affecting the pitch

every setting is explained in the guide, or also if you hover your move over the setting

[zenith spade]

i want to export and use my voice model i created as text to speech, is that possible with weights?

[rancid island]

got ya thanks again mate

[low shard]

need any other help?

[low shard]

wdym with export and use it as tts? please elaborate more

[rancid island]

its working fine now thanks

[zenith spade]

like go to my model, export it (which is possible) and then make a text to speech that uses that voice model

idk if that helped at all

[low shard]

Voice models on Weights are made via RVC, RVC is natively Speech to Speech, not TTS

the only way you'd use it via tts, is via using another tts first, and use the output as input

[zenith spade]

awww

[low shard]

you can just directly use weights' TTS

it uses another tts, and use the output as input for rvc

[zenith spade]

its rvcv2 weights uses right?

orrrrrrr am i braindead

[low shard]

yes

[zenith spade]

YAY

thats good for vonovox

teehee

[low shard]

vonovox is a realtime voice changer btw, not TTS

[zenith spade]

ik

[low shard]

then why did you ask about exporting it for TTS?

[zenith spade]

bc i was gonna try to use for tts, if i couldnt do that id ask abt vonovox

[low shard]

There are different Text To Speech (TTS) AIs:

GPT So Vits: RVC isn't as good as GPT So Vits for tts, but gpt so vits (few shot tts, which means needs just a lil training for models) can't use rvc models (and viceversa), and its only limited to: english, chinese, Cantonese, japanese & korean, if you wanna check gpt so vits instead, read https://docs.aihub.gg/tts/gpt-sovits/

Freemium 11labs: Easy way to do TTS is https://elevenlabs.io/, you can't use RVC model on this but its a mostly premium easy way for good quality TTS

FishSpeech: FishSpeech is a 0 shot (no explicit training needed) TTS, if you got a good pc you can use it locally else use their site

You can check TTS in our tts index

With RVC Models:

RVC is natively for Speech To Speech, but forks such as Applio have built in tts (using Microsoft Edge TTS to make a generated tts audio, which i suggest you to choose a tts model that is the same gender and language of the rvc model you wanna use, and then convert it with rvc)

If you wanna do tts locally with RVC Voice Models (if you got a good pc):

• You can get Applio in our docs

If you don't got a good pc you can do tts with RVC Voice Models on cloud:

• Ilaria RVC Zero (Running on A100 GPU, free fasted rvc on cloud) and the guide

• Use Applio UI Colab (with google colab T4 free daily limit gpu)

• You could try another tts from our tts index and use the output as an input in rvc

this could help you out

[forest elk]

hey can u help me?

i downloaded the deiteris fork got it working

then i decided to try the realism page, downloaded lighthost and voicemeter

but now i can't hear myself in w-okada

i did the voicemeter setup and ended up getting this:

and i got the phase distortion on:

but when i try this: (input being my real mic), monitor being my real headphones, and output being line 1

nothing is played back to me, even though i see the b2 meter changing when i set it on discord

[zenith spade]

im using vonovox (a realtime voice changer for rvcv2 models) with my real mic being connected through wasapi and its not playing back to me, output being my headphones

[fluid crater]

whats a good girl model

i need one fast

[forest elk]

@low shard found the error, maybe updatwe in the guide but u need this to be there (a1 needs to be selected as well)

[prime shell]

how i can download applio?

[simple ore]

https://huggingface.co/IAHispano/Applio/resolve/main/Compiled/Windows/ApplioV3.2.9.zip

if you have nvidia card

[dark ginkgo]

btw, does the version matter?

oh also, more importantly, is the new version of koyah ss flux compatible?

I'm very familiar with koyah training but not so much with the others. I noticed koyah gui ss had an update about a week ago.

[simple ore]

game ready = has optimization shims that improve gaming performance, may crash randomly
studio = stable driver without any optimizations

[dark ginkgo]

I looked in to it. Both version this time around is actually identical apparently.

[prime shell]

Where can I find templates for this page that work well?

@simple ore

[dark ginkgo]

Hey @simple ore the method that works for forge ui doesnt seem to work for gradio and applio

how do I make it work for them?

[dark ginkgo]

alternatively, if you have rtx 50 series, and would be able to tell how you made it work, that works too

[low shard]

I didn't make that guide, @crude flame did

[low shard]

How to (unofficially) use Applio for RTX 50 serie cards

Follow to download it as said it in https://docs.aihub.gg/rvc/local/applio/

After you extracted the precompiled, go to the path in Windows explorer, write "CMD" and press enter, then in CMD write 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

If you get any already satisfied requirement issue, run env\python -m pip uninstall torch torchvision torchaudio then the command said above

[low shard]

yes, the screenshot I sent was only for the studio driver version

[dark ginkgo]

before I start fixing applio and gradio, is there other projects that has surpassed these yet?

actually now thinking about it, gonna just focus on applio and it's replacement since I dont actually use gradio. It's faster for me to make custom sound effects using audacity or just record it in a daw like ableton

[halcyon lynx]

Question, is it possible to create custom AI models using Chrome OS?

Like creating models locally?

[brittle wing]

@simple ore sorry for not telling you I was taking a nap (until gradio is back) & what you told me to do actually worked

The misalignment is still here.

It's in the input audio maybe cause of uvr online

It's from the input

[low shard]

wdym

[low shard]

if you got a chomebook, ofc not, chromebook are shitty asf cheap computers meant just for googling

u can only do cloud

[halcyon lynx]

I see, okay

[low shard]

do you may have another laptop or just a chromebook?

do you want me to tell you how to use cloud?

[halcyon lynx]

nope, just a chromebook and I already know how to do it via cloud method

[low shard]

alr

[glacial pollen]

This is quite possible because uvr or even mvsep aren't exactly 100% aligned perfectly

smaller or bigger hiccups can happen depending on many factors

[brittle wing]

That's the case

[glacial pollen]

I see

[simple ore]

is there a misalignment between the input audio and the output audio from the inference?

[low shard]

just shift manually, it's some milliseconds thing 🙏

[brittle wing]

Actually the problem is from the audio separation sites it stays the same

[brittle wing]

That's the solution

[simple ore]

It is likely a similar problem - the separator does not process the whole song at once, it likely takes small pieces and joins them together later

and it may be losing some frames in between

[viral mason]

getting some error message which is wasting my zero gpu when I try imputting this sound in with these settings in the huggingface uvr space

yes I'm bored enough to try and make a slenderman model

[low shard]

@viscid moss

[viscid moss]

I'm like 75% sure that the problem is audio length

too short

I'll test anyways

[low shard]

also why a slenderman model

you could go just watch a serie or play a gamr if ur bored 😭

Try Cuphead or Breaking Bad lol

[viscid moss]

fr 😭

[low shard]

yeah it's just wasting time making models out of a 5 seconds weird glitchy audio

[viscid moss]

yep that's the problem

[low shard]

you saved him precious time from wasting it 😭

i still dunno why so many people waste time on meme models like planet sounds or gif laughs

[viscid moss]

For some weird reason some models doesn't works with -10sec audio

[viscid moss]

😭

[viscid moss]

ye idk

[low shard]

do people still do meme models? i thought they got bored after 2 years

[low shard]

maybe put a check returning an error telling the user to use a longer audio

or do u think its worth to make it work on 5 seconds audio?

[viscid moss]

Not so often

[viscid moss]

dat's a good idea, but some models are working with dat audio length

[viscid moss]

the thing is idk how to fix it :p

[low shard]

maybe try asking that python package you use

audio infer or smt

[viscid moss]

ye python-audio-separator dev also don't know

:p

This is what we have atm:
https://github.com/nomadkaraoke/python-audio-separator/issues/189

[low shard]

just return an error atp

[viscid moss]

yeah

[viral mason]

weird

[low shard]

you saved a lot of time 🔥

[dark ginkgo]

Brand new install, new software, same stupid cuda issue, and I doubt different fix method.

Zonos

[viscid moss]

For my app this way works:
https://github.com/Eddycrack864/UVR5-UI/issues/44

https://github.com/Eddycrack864/UVR5-UI/blob/main/UVR5-UI-installer.bat#L84

[simple ore]

same fix

[dark ginkgo]

I'm currently trying this version of the project https://www.youtube.com/watch?v=Aj18HEw4C9U

if it works, cool, if not, back to square one

[simple ore]

zonos is complete ass to install

[glacial pollen]

^

[dark ginkgo]

Ok, cool, it worked! Also the same guy had an auto blackwell ai system wide setup tool kit which seemed to configure more stuff better as well

demo audio

this is fun

[boreal flint]

do anyone know how to config runpod and template? any tutorial or resources?

[low shard]

i dont think anyone pays for runpod here btw

[upbeat carbon]

should i use this setting for talking/games?

[knotty moth]

just try whether you experience lag or voice cutting on it

[upbeat carbon]

alrighty thx!

[lusty nest]

Alright so I've just downloaded Okada and the delay I'm having (especially when talking to other people) is just a pain and I want to fix it but I have no idea how. At the moment I have my mic running into Okada, then a VB-audio cable input running out of Okada into Voicemeeter so I can EQ my microphone to my liking. Any help would be greatly appriciated!

[low shard]

u got an rtx 3050 so thats the suggested settings #✨│ai-help message
generally speaking, just be sure your chunk is a bit higher than ur perf value at the top left

[boreal flint]

then how people rent gpus here

whats the best way?

[upbeat carbon]

a bit higher you say? how high?

[low shard]

This is a general AI Server, not just voices, you will have to elaborate

Elaborate:

• your pc gpu
• your operative system
• what you want to do
• what tutorial link did you follow
• a screenshot of the program

VB-audio cable
UNINSTALL IT, it causes issues on windows, DO NOT TRUST VIDEO TUTORIALS, THEY ARE ALL OLD

[low shard]

like if your perf value is 300, put chunk to 400

[low shard]

most people here just use the free tier of google colab, or Kaggle, or more rarely Lightning.ai

or just do it locally

[boreal flint]

i see

[polar ridge]

Hi is the Applio still the best ai voice cloning and training? Or Pinokio with f5 tts?

[low shard]

Applio and F5 tts are 2 completely different AIs

Applio is an RVC Fork (modified version), RVC is Speech to Speech natively

STS, not TTS

[polar ridge]

So which one is best for cloning and training

[low shard]

it can do tts only because it uses microsoft edge tts and then use that output as input in rvc, but still that's not like f5 tts

[low shard]

Speech to Speech and Text To Speech are 2 different things

depends on what you wanna do

[polar ridge]

But can RVC models be used for tts?

right?

[low shard]

natively no,

There are different Text To Speech (TTS) AIs:

GPT So Vits: RVC isn't as good as GPT So Vits for tts, but gpt so vits (few shot tts, which means needs just a lil training for models) can't use rvc models (and viceversa), and its only limited to: english, chinese, Cantonese, japanese & korean, if you wanna check gpt so vits instead, read https://docs.aihub.gg/tts/gpt-sovits/

Freemium 11labs: Easy way to do TTS is https://elevenlabs.io/, you can't use RVC model on this but its a mostly premium easy way for good quality TTS

FishSpeech: FishSpeech is a 0 shot (no explicit training needed) TTS, if you got a good pc you can use it locally else use their site

You can check TTS in our tts index

With RVC Models:

RVC is natively for Speech To Speech, but forks such as Applio have built in tts (using Microsoft Edge TTS to make a generated tts audio, which i suggest you to choose a tts model that is the same gender and language of the rvc model you wanna use, and then convert it with rvc)

If you wanna do tts locally with RVC Voice Models (if you got a good pc):

• You can get Applio in our docs

If you don't got a good pc you can do tts with RVC Voice Models on cloud:

• Ilaria RVC Zero (Running on A100 GPU, free fasted rvc on cloud) and the guide

• Use Applio UI Colab (with google colab T4 free daily limit gpu)

• You could try another tts from our tts index and use the output as an input in rvc

This should explain you how it works

I also suggested some other TTS

[polar ridge]

wierd i remember training my own RVC models from google docs hugging face or whatever and my model gets created in google drive then i use the model on Applio i think to create tts

[low shard]

It does TTS, but not natively

If you're looking for TTS only, it's better you try F5 or eleven labs or fish tts

[polar ridge]

yuh so your telling me im looking for speech to speech then?

[low shard]

Applio won't be expressive bc it uses Microsoft edge tts as input

[lusty nest]

My apologies for not elaborating.

GPU - Radeon RX 6650XT
Operating system - Windows 10
What I want to do - I want to use a voice changer to be able to sound like a female with minimal delay and a final output that doesn't sound "robotic". I got recommended W-Okada by a friend of mine and so I looked up a tutorial on how to install it.
Tutorial link - https://www.youtube.com/watch?v=_JXbvSTGPoo&t=514s
Screenshot - https://cdn.discordapp.com/attachments/1364583446407680070/1387401021386395688/image.png?ex=685d3564&is=685be3e4&hm=ccce1e8406bee687dbb7a6697a5984e2015fc0c0eb0200ef505d00b869220f2b&

[low shard]

RVC is STS