✨│ai-help

[barren hatch]

bro come on

again restart

...

i just want to fix the delay bro

[low shard]

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

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

lemme check rq

[barren hatch]

Bro

is not from vcable

is working

i just have the delay

[pastel oak]

You wont have 0 delay there will always be some delay based on gpu

Audio server mode then using [windows wasapi] on all cuts down some more delay

[barren hatch]

ok so can you help me

i think is from here

[pastel oak]

Advanced setting: crossfade length reducing this redues delay but also redues quality, 0.10 is sweetspot but if u dont csre you can go down to 0.05

[pastel oak]

Nick alreadytold u to do chunk

90*

[barren hatch]

this is my proccesor

this is my gpu now

[pastel oak]

...ok?

[barren hatch]

bro

[barren hatch]

??

[pastel oak]

We already gave you 3 things to do for less delay

[barren hatch]

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

yes very good

[pastel oak]

.

[barren hatch]

Bro i am not gonna

uninstall

to install it again

[pastel oak]

The chunk part

[barren hatch]

then to restart my computer

again

[pastel oak]

You ask about chunk then ignore what he said about chunk

Ok ignore the cable part dont ignore the rest of his message?

[pastel oak]

And this 3rd factor

[quiet flint]

we need someone to distill kimi-k2 into mythomax

[tough fiber]

yea i tried to install cu128 but when i run rvc mainline it says 120

[viral mason]

first ever non egirl screenshot

meant to reply oops lol

[simple ore]

show me how you did it

[somber tusk]

There's a cloudflare error on the Weights website

[worthy lance]

guys what do I do with the MMVC app because every time I start it, no audio at all will come out when monoring, an when recording, no audio either

[worthy lance]

I actually used the guide in the server and it worked

[wicked mica]

the voice got smoother and clear but the delays still happen , mostly when a game is also open

also when i reduce the chunk, why does the delay increase

[wicked mica]

my sound quality lowers whenever i use it through browser,

[distant idol]

It's been like that all night long... started again but its not doing anything

[viscid bluff]

i am trying to run tensorvenv however i keep on getting this message.

note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed

I am currently just following the AIHUB guide and I am now stuck here.

[viscid bluff]

I fixed it. If you have numpy 1.19.5, then you want to downgrade your python. I had 3.10 version when numpy works well with 3.9 version.

[shy summit]

Hey folks. I'm trying to create a voice model with Elevenlabs for a cartoon character, but I'm gonna want to remove background noise from most of the audio since there's ambience. Is there any specific site/service that's good for that?

(Also, how many samples do I usually need/how many minutes? And do I need a range of emotions?)

Hopefully I won't need to dig for scenes with no ambience lol

[distant idol]

@simple ore and yet after like 2 hours i did nothing

[distant idol]

again half hour and still nothing

[simple ore]

so what driver version do you have?

[distant idol]

Of?

[simple ore]

video card, adrenalin

[simple ore]

okay, i'm gonna update my spare pc, it will take some time

[jagged owl]

what are the best mel roformer models for isolating vocals from rock and metal?

[deft terrace]

I have some files where I’ve isolated a single character’s voice from an animation, but there’s still background music, vocal harmonies from the music, clanking sounds, and various other effects mixed in.
What’s the most effective method I could try?
Even with UVR, I couldn’t get a clean result.

[jagged owl]

bandit plus and melband karaoke (becruily) on mvsep

[deft terrace]

Is it great?

[jagged owl]

bandit will get rid of effects, melband karaoke will get rid of the vocal harmonies and music

do mel karaoke first

[cold shoal]

Hey sorry nick about last time something had came up anyway I need your help with something

[simple ore]

@distant idol

[distant idol]

i'm still stuck

[deft terrace]

Thanks, carnage Ill try it

[simple ore]

lemme try with updated zluda.py

i used the original

nope, works fine too

try closing Applio window, delete C:\Users\user\AppData\Local\ZLUDA

and re-start training, it will recompile again

works fine for me

[terse halo]

wha is this passthru

[distant idol]

starting now

something is happening haha

[simple ore]

well, you've deleted what it had compiled before, so it starting to compile again

see ya in 30

anyway, I've tested 25.6.1 driver with 6.2.4 hip sdk method on my old 6700xt and it works fine

[viral mason]

Yucky scammer (it's gone yay)

[distant idol]

Hour later

[simple ore]

can you go inside C:\Users\user\AppData\Local\ZLUDA and see the date stamp of the file there

technically you dont have enough compile lines shown yet

there are usually more

[simple ore]

right click, properties, see last update date?

[simple ore]

okay, so it is just stuck for some reason

[distant idol]

so delete the appdata, and try again?

[simple ore]

did you install the latest VC++ redist?

[distant idol]

where can i check?

[simple ore]

https://aka.ms/vs/17/release/vc_redist.x64.exe

[distant idol]

yeah, it says uninstall and repair

so i guess i'm up-to-date

[simple ore]

weird

@distant idol make a new thread on help forum so we dont spam it here

[distant idol]

I made one

[obsidian vector]

nah i just wanna try out a girl voice

lol

see how it sounds

maybe use it in roleplay games on roblox

[viral mason]

[hushed hazel]

.

i cant send screenshot

there

why

my device stuck in Activating the VC screen

[placid storm]

is aigencover working??

@flint vapor

[flint vapor]

Hi, i can't verify at the moment if it is working

[flint vapor]

Which VC do you have?

[smoky summit]

Hi guys! I have a problem. I have a pretty good pc (rtx 4070 + ryzen 5 7500f) but any sound model I tried sounds very robotic and the sound quality is bad

[astral harbor]

same here but with amd gpu

and its not running on my gpu no matter what i do

[hushed hazel]

MMVC

[jovial owl]

guys i cannot pick harvest i can only pick crepe_tiny and rmvpe_onnx why

[true minnow]

русские есть?

[simple ore]

#🌏│русский

[winged garnet]

hey, I'm new, I've never used rvc before. I'm trying to train a model for myself, but I always get the message "Unfortunately, there is no compatible GPU available to support your training". I've updated the driver to the latest version, downloaded cuda, downloaded torch, updated rvc to the latest version. deleted rvc and reinstalled. deleted the vga driver and reinstalled but it still doesn't work. My computer has: i5 14400f, rtx 5060ti 16gb. (running the latest game ready driver).

[simple ore]

which rvc?

5000 series nvidia requires cu128 torch, depends on the software you're trying to use there are different methods

[winged garnet]

I tried 2, the first one is Retrieval-based-Voice-Conversion-WebUI. the second one is mangio rvc. however both give error "Unfortunately, there is no compatible GPU available to support your training"

anyway, which rvc is better to use? or are they all the same?

[simple ore]

both are severely outdated

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

download, unzip, then open cmd.exe in the folder where you've unzipped it to and do 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

[winged garnet]

it work ;D tysm

[hot lotus]

how do i fix it glitching

when i talk

and it cuts out

24/7

[weary osprey]

Can anyone tell me if anyone knows the website and prompt/template that was used for this gif:

https://tenor.com/view/happy-birthday-gif-4446960541527894948

It s been popular on tiktok and I wanna use it too

Please ping me if u can help

[pastel oak]

• whats your gpu
• did you follow an old yt tutorial
• whats the name of the program you downloaded, full name of the file you downloaded
• screenshot of the program

[hot lotus]

send new tut

please

idk

what is a new one

[pastel oak]

you managed to avoid answering all 4 questions thats insane

-rt

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

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

• Vonovox

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

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even 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

[pastel oak]

first link

[weary osprey]

I found it

[gray dagger]

how long should i train a 30 minute dataset

[viscid bluff]

Question for pretrain. I am using mainline rvc and I filled in everything like the typical. I clicked One-click training and I only got a index file and no pth. I am doing something wrong? I have also loaded the G and D model path too. I have the used the KLM 4 pretrained. I have set the version to v2. I have made sure the sample rate was the same. Everything is in logs folder and I don't see nothing in the weights folder. I don't really understand what I am doing as the guides are very vague

[simple ore]

do you expect "One-click training" to do some unbelievable magic instead of doing those steps one by one and verifying each finishes properly?

@viscid bluff

[carmine lion]

can someone let me know wether Im basically done training a model or if I should train it more, I can send photo of the tensorboard

[viscid bluff]

i said i did everything like what the tutorial says. i created other models but not pretrained

also it says everything has finished which makes make confused

or in terminal it says it's complete.

[carmine lion]

what are you looking for the pth file?

[viscid bluff]

yes the pth file i don't understand where it is

i thought it'll be in the weights folder after finishing but it's not

[simple ore]

assets/weights

show the ui screenshot

[carmine lion]

doesnt command prompt show a message of where the last saved weights pth file was saved

[viscid bluff]

I clicked one-click training and I also can't send pictures

It says it's in my logs folder which is where the index stuff are

[knotty moth]

pth files are produced every particular epochs during the training session

[viscid bluff]

I'm only getting index things every time time I click the one-time training

[viscid bluff]

How do I start a training session with a pre trained model?

I thought it was the one-click training button as it is for other models when training it with the default base models

[simple ore]

do it step by step

[viscid bluff]

As I said I have done all of that. However I do not get pth file at the end

[simple ore]

I did ask you to show this screenshot

[viscid bluff]

I said I can't send pictures

[simple ore]

what do you have in logs/ your model name?

[viscid bluff]

This is me when I click both index and train model

This is in my logs

[simple ore]

and in assets/weights?

[viscid bluff]

the chipsterling50v was a model that i have trained before. it's not the same one as the one i am trying to train with the pretrained model

[simple ore]

and what is in train.log?

[simple ore]

indicates you trained nothing

[viscid bluff]

\

[simple ore]

just click train model

[viscid bluff]

how do I train it then I have clicked everything single thing

i have cliekced both train model and onen click training

it is perhaps the pretrain model needs more sounds?

this is why it's confusing me currently

[simple ore]

what do you see in the terminal window when you click Train Model?

how many slices are in 0_gt_wavs?

[viscid bluff]

hmm it's saying
FileNotFoundError: [WinError 3] The system cannot find the path specified: 'C:\Users\Local User\Documents\RVC/logs/mute/0_gt_wavs/mute40k.wav'

[simple ore]

you've deleted the mute folder

did you?

[viscid bluff]

ooo probably

i may have deleted by accident when deleteing my old folders

[simple ore]

that will do

[viscid bluff]

thanks for the help i was actually stumped on this

❤️

[edgy rapids]

anyone can help?

[simple ore]

are you running it as admin?

[edgy rapids]

nope

but lt worqed flrst tlme

sorry lf am belng dumb thls ls my flrst tlme

[knotty moth]

try this version that shouldn't have that kind of issue

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

[edgy rapids]

sorry but can you tell me what to do exactly

[slender jetty]

I have a 5070 and for some reason my audio sounds bizarre I have heard the audio sound wonderful on worse gpu's

[edgy rapids]

hey you now any vldeos?

that show lt out

@steel forge

[craggy bough]

for trolling or catfishing

[edgy rapids]

can you help me please?

[craggy bough]

nope i have no idea

[edgy rapids]

lt stays lqe thls

whlte

[edgy rapids]

anyone that can help?

cuz lt worqed flrst tlme

second restart lt stayed lqe thls

[craggy bough]

it says ur missing a file

[edgy rapids]

ban teh dude under me scam btw

[edgy rapids]

you now the dlls needed?

[edgy rapids]

l can send the whole cmd

lf needed

[craggy bough]

nope

[craggy bough]

all that ik is thats the error

[edgy rapids]

alr lg l cant use lt

can you help me man

where dld u lnstall lt

wlch verslon

l meant wlch verslon of hte volce changer

can u send lt?

@north lava

[knotty oyster]

for some reason in discord it repeats my normal voice and also the ai voice, how do i fix it?

[knotty oyster]

in game also

[viral mason]

https://cdn.discordapp.com/attachments/1344423628954927126/1393351946877796473/ezgif.com-animated-gif-maker_2.gif

[viral mason]

https://tenor.com/view/buzz-lightyear-gif-21327280

[sage lagoon]

what is an index file

[viral mason]

Ew

I don't think that's the kinda milk you want

[jovial owl]

guys when i speak into okada it doesnt pick up my voice

like its radio silent

how do i fix that btw i tested and my mic is still working

[proven hill]

@low shard

[brazen ore]

Does training wear down gpu?

[abstract ember]

hello, im trying to make ai cover of tokai teio singing through patches of violet, but at the chorus she sounds like she ran out of breath, how to tune?

[lusty pelican]

How do it use the json and index file of the voice model i downloaded?

[pastel oak]

Send screenshot of the program

[knotty moth]

she might lack of stamina or need some endurance recovery skill you could get from some support card

• make sure the input voice has been isolated & denoised properly
• if the model struggles at that pitch level, try tuning pitch shift or try another better model

[hard badge]

Which voice conversion software should I use if I want to preserve emotion, significantly alter the original voice, and interfere with it in real-time on an average model using a Tesla T4 GPU? I only intend to use it for talking, not singing..

[knotty moth]

if it keeps overheating and doesnt have proper cooling, airflow or thermal paste

esp laptop gpus tend to overheat easily

[knotty moth]

pth & index file are used

index file isnt mandatory but helps reproducing accent

[solemn fulcrum]

Do you know where I can find this same server but it's from Brazil?

[hard badge]

↑

[knotty moth]

continuing your discussion, here is guide & download for the deiteris fork

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

[maiden wren]

Try getting deiteris one maybe it will fix the issue

Its only 3 gigs and you can create a thread in #1192011222023950368 if you have furthermore issues

Yeah you need to use deiteris one

That versuon didnt work for me too

So i used deiteris

Np

[hard badge]

. ↑

[knotty moth]

sorry but we dont support the og version, only deiteris fork

the fork one is argued to have better quality, which could solve ur problem

[maiden wren]

I reccomend the nvidia one which is 3gb

Yes hold on

I reccomend going to site and finding this link and click on it to download deiteris fork for nvidia gpu

@untold socket

After download Extract the zip and open file named MMVCServerSIO.exe

np

np

[hard badge]

.↑
. |

[maiden wren]

np

[knotty moth]

dml is for AMD/intel gpu, and cuda is for Nvidia gpu

[maiden wren]

yeah

just click this link since you use rtx 3060 this link is good it will automatically start download

brb soon

back

[hard badge]

Which voice conversion software should I use if I want to preserve emotion, significantly alter the original voice, and interfere with it in real-time on an average model using a Tesla T4 GPU? I only intend to use it for talking, not singing..

[maiden wren]

I dont know how tesla t4 gpu's work but i heard its from nvidia right?

[hard badge]

Yes

It's the weakest turing generation data center gpu

[maiden wren]

oh,

so you want voice changer is that right

[hard badge]

Yes

[maiden wren]

i don't know if it can preserve emotion after the original voice but we have 1 that can change voices though

[zealous basin]

I’m Guys I am Alex – a creative wizard with 15+ years of experience in design, development, animation, and more. 🧙‍♂️✨
These days, I’m deep in the world of AI tools and TikTok content creation, helping people grow and monetize their profiles. 📱💸

Whether you’re just starting out or looking to scale fast,
I’m here to share what works and support your journey.

💬 DM me anytime – I’m always happy to help!

[hard badge]

What is it?

[sturdy patio]

yo any good alternatives for w okada? or any good forks?

[maiden wren]

its a voice changer named deiteris fork

[hard badge]

Thanks

[maiden wren]

the site for it is in https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/ you can find download link there i reccomend downloading one for nvidia

[hard badge]

Can I just git clone?

[lusty pelican]

is deiteris the same as the rvc client?

[maiden wren]

this is the best one

[sturdy patio]

ty

[hard badge]

The best option?

[maiden wren]

yeah atleast thats what i find best

[sturdy patio]

wait how do i extract it?

it's a .002 and if i change it .zip it doesnt work at all

am i missing summin

[maiden wren]

do you have nvidia gpu?

or amd

@sturdy patio

[sturdy patio]

nvm i dont like to read

its ok boss

im gonna read the tutorial

i dont wanna piss yall off

thanks boss

[maiden wren]

are you struggling to find download link?

[sturdy patio]

nope i got it

i just wasnt reading

[maiden wren]

nice

[sturdy patio]

my bad lol

[maiden wren]

its okay

[low shard]

Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

are you looking to do e girl trolling/catfishing?

[low shard]

are you looking to do e girl trolling/catfishing?

[hoary magnet]

i need help

When i open my start_http file it doesnt load or do anything the first time i did it it worked.

[low shard]

start_http is apart of the old original wokada, you deffo followed an old youtube video tutorial

[low shard]

Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[hoary magnet]

C:\Users\pible\Downloads\Ai Voice Changer\MMVCServerSIO>MMVCServerSIO.exe -p 18888 --https false --content_vec_500 pretrain/checkpoint_best_legacy_500.pt --content_vec_500_onnx pretrain/content_vec_500.onnx --content_vec_500_onnx_on true --hubert_base pretrain/hubert_base.pt --hubert_base_jp pretrain/rinna_hubert_base_jp.pt --hubert_soft pretrain/hubert/hubert-soft-0d54a1f4.pt --nsf_hifigan pretrain/nsf_hifigan/model --crepe_onnx_full pretrain/crepe_onnx_full.onnx --crepe_onnx_tiny pretrain/crepe_onnx_tiny.onnx --rmvpe pretrain/rmvpe.pt --model_dir model_dir --samples samples.json

this is the link

https://www.youtube.com/watch?v=SxdnGxicJOg&lc=UgweM0-AOj85q_BkZgB4AaABAg

wait it worked now

sorry i was impatient lol

[low shard]

that one is outdated asf, it uses an over year old version of original wokada, and vb audio cable creates issues on windows
delete the zip, folder and uninstall from windows app settings

also, i saw its about ai girl voice, are you trying to do e girl trolling/catfishing?

[hoary magnet]

no

[low shard]

the one you're using is OUTDATED, don't use it

[hoary magnet]

im just learning to download the voice changer

[hoary magnet]

send me one

[low shard]

what's your pc pgu, operating system, and what do you want to do?

[hoary magnet]

3060 nvidia

gpu

[low shard]

like:

• ai covers
• tts
• e girl trolling/catfishing
• roleplay with realtime voice changer

[hoary magnet]

i just want to have fun with the voice changer i am nvidia gpu 3060 and my operating system is windows

i wanna do some random voices with the thing lol

[low shard]

windows 11, right?
also im guessing voices like spongebob or villagers lol

[hoary magnet]

yeah windows 11 pro and i might do those

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

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

• Vonovox

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

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even 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]

you can either use wokada deiteris fork or vonovox

they are both the best, the only difference in your case since you got nvidia and are on windows, is the User Interface and that vonovox got paid voice effects like low quality mic

[sturdy patio]

fuckkk

anyone know why the voice model i cloned sounds a bit static and a bit off? It can pick up some stuff and replaces words with other words like how are ya -> brocolli

[low shard]

Hello, to understand more your issue and know how to help, Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[maiden wren]

make chunk higher

set extra to 2.7

try 1000

okay,

idk if he was catfishin or not

he didnt specifiy

ohh

sorry didnt notice

[rugged turtle]

hi ai

where is ai

ai wake up

[low shard]

😭

[low shard]

Hello, do you need any help?

[crisp kettle]

when I run on kaggle it occasionally stutters (I followed the guide and put everything higher/lower than recommended just to make sure it was not too heavy on the gpu). Besides, on the draft session info, everything seems good except the cpu is at 100-120% (I made sure to use gpu to run the vc). Is something wrong or is it supposed to be a bit stuttering like that?

[low shard]

Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[simple ore]

voice changer on kaggle would have some latency, that's internet for ya

[low shard]

well yes that's true because of the internet connection,
but let's check if they are using the updated one and also if they meant latency or other issues like cutoff since they said stuttering

[crisp kettle]

• 3060 rtx but on kaggle I chose T4 x2
• windows 11 latest
• find out why it stutters randomly
• the most recommended one in the guides channel
• I turned it off but I will boot it up if you still want it

[low shard]

3060 rtx
if you got an rtx 3060, you can do it locally, is there a reason why you're using cloud?

[crisp kettle]

everything is perfect when I use the local version so nothing to say

[crisp kettle]

ah yeah, sometimes I want to play gpu intense games

[low shard]

could you please tell the games name, lower the graphics to the lowest 1080p 60fps, then show a screenshot of your local wokada deiteris fork settings ?

just to check if maybe you could do it locally

[crisp kettle]

I have been using it locally since the begining. But I also want to sometimes play heavy games

[crisp kettle]

something like the finals, control, black myth wukong, etc... Those are pretty intense even on low

[crisp kettle]

I tried with the finals, it became unusable when in game

[low shard]

ohh, so you already tried? even tried asio/wasapi with disabled jit compilation?

but yeah you're right they are pretty intense, i was just checking if maybe you could try but if you don't want to, you can just share a screenshot of the kaggle settings and check if there's anything you can adjust

[crisp kettle]

everything execpt asio/wasapi as the noise was interfering too much (my mic is not that bad)

[low shard]

i see, maybe let's check if there's any kaggle settings i could help you adjust?

[crisp kettle]

not a pro but I know some tricks and played around the setting. Either using cloud or run on a different device lol

[crisp kettle]

about asio/wasapi?

[low shard]

i meant other things like disabled jit compilation and the extra, since you said you have noise issues with wasapi/asio and you can't fix that unless you use a 3rd party tool noise reduction

[crisp kettle]

I also did use a different machine so I can play heavy games (it worked). But the other machine spec is kinda meh so I got bit greedy (used to the performance of 3060

[crisp kettle]

tried them all. Shouldn't I keep extra at 2.7s? I want to keep it at the best quality possible so I don't really want to reduce it

[low shard]

alright nvm then, i mean you could keep extra a bit lower for less delay, but i understand you want the best quality
either you can try sacrificing a bit the quality using a lower extra and fcpe as f0, or this might be related to connection issues, since on cloud there's more delay

[crisp kettle]

oh I still want to ask you about kaggle. When one is using it, should the CPU be at 100-120% constantly? Is that normal?

followed everything and pretty sure that everything is set to the GPU

Is that the reason? If not then maybe like the other guy said, the internet introduces some stuttering

[hard badge]

Is 24 GB of storage enough for https://github.com/deiteris/voice-changer?

[simple ore]

the voice changer is ~5-6GB

• models

[brisk trail]

Hello lovely peoples! I'm a beginner model maker and wanted to know what sample rate does in applio voice training

[knotty moth]

other than the RTX 3060 rig, what is that spec?

[hard badge]

But for the full thing with dependencies too.

[simple ore]

about 5-6GB

[minor badger]

mine keeps saying "trial"

any fix

[low shard]

you installed the vac trial, not lite, it might be you didn't follow the right guide

what's your pc gpu & operating system? what do you want to do?

[minor badger]

fixed it thanks

[low shard]

Do you may want me to check your settings? You can send a screenshot and I can help

[minor badger]

nty

[crisp kettle]

nvidia 1650 4gb (gaming laptop)

[knotty moth]

it should work as bare minimum unless you need lower delay for competitive gaming

[winged garnet]

hi, which rvc is good now? i just need voice quality and not performance. (rtx 5000)

and how many epochs should i train the model for? my audio file has more than 300 pre-cut wav files. their total duration is 1 hour 17 minutes

[queen adder]

thats a very powerful Geforce

[low shard]

Please Elaborate:

• your PC GPU (like yes rtx 50 serie, but which?)
• your operating system
• what you want to do
• what tutorial link are you using (if any)
• a screenshot of the program (if any)

also, be aware that RVC means Retrieval-based-Voice-Conversion, not realtime voice changer

[low shard]

there's no right amount, the only way is using the tensorboard

[winged garnet]

rtx 5060ti
win 11
i want play game with my friend and share model
no tutorial
no program

[low shard]

i want play game with my friend and share model
so, use realtime voice changer right? not train models? because those are 2 different programs and things

[winged garnet]

yes. i have trained my model but i dont know what software to use for real time changer. i have trained with applio

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

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

• Vonovox

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

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even 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]

you can either use wokada deiteris fork or vonovox in your case, the major difference is the user interface, and that vonovox got paid voice effects like low quality mic (like bitcrushing), but you can check their pros and cons if you want

[pale gyro]

Why is this?:

w-okada voice changer ERROR: Could not find a version that satisfies the requirement faiss-gpu (from versions: none)

[winged garnet]

ok tysm

[low shard]

Please Elaborate:

• your PC GPU (like yes rtx 50 serie, but which?)
• your operating system
• what you want to do
• what tutorial link are you using (if any)
• a screenshot of the program (if any)

[low shard]

please let me know!

[pale gyro]

It's google CoLab

[low shard]

could you please elaborate everything i asked?

google colab is just a cloud computing service for people with a bad pc, there's thousands of notebooks

[pale gyro]

Sure:

• My PC iGPU: Intel Skylate Intel HD 520
• Windows11
• Run back w-okada VCC since I need it for a job and last time I used it was months ago
• I used an old collab link but I lost it so I just googled a new one (Also used github repo for storage instead of google drive)
• For screenshot It's not present right now since I closed the tab and only copied the error. But I will bring it within few minutes.

Also we can't even attach screenshots. But here: https://i.imgur.com/PciCTTN.png

[pale gyro]

You also want the link of the colab project ?

[low shard]

yes, it would help

[pale gyro]

It's this one: https://colab.research.google.com/github/w-okada/voice-changer/blob/master/Realtime_Voice_Changer_on_Colab.ipynb

[low shard]

thanks for the info, even tho i dont really understand what you mean by for a job, do you need to:

• tts
• ai covers
• e girl trolling/catfishing
• roleplay in games
• roleplay in vc
• etc?
  Asking since there's different programs for each task

[pale gyro]

Need it for voice acting

[low shard]

you sure you need it realtime, and not on pre-recorded audios?

[pale gyro]

I can either do pre recorded as I used to do, or real time. But ofc I prefer pre recorded.

[low shard]

then you're using the wrong program

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

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

[pale gyro]

What else is available then?

[low shard]

Wokada is only for realtime, you'd need Applio instead

[low shard]

also that is an outdated version of original wokada lol

[crisp kettle]

yeah it work great, just sometimes I leave it at the office so I was wondering if cloud is a good option...

[low shard]

Applio colab, read up the guide and let me know!

[pale gyro]

Well I still would need a vm to test it on like google colab.

[low shard]

it runs on the cloud computing service google colab, it's the same service cloud, just different program

[pale gyro]

welp...

I used the colab version lol

[low shard]

Yeah, use the applio colab version

[viral mason]

I'm confused on what they're trying to do

Do they want okada or to train

[pale gyro]

No train things

[queen adder]

wait that sounds similar to stranger things

O_O

[pale gyro]

Actually after downloading pretrained models somehow dosn't show under the models list

[pale gyro]

well I didn't watch it so idk.

[queen adder]

i mean the title

and the text

train things = stranger things

it sounds similar

[pale gyro]

oh wew

Also that pack wouldnt even work on Skylate GPU's

or any intel gpu

it's only for AMD as I see

[earnest forge]

What really is Sample rate when regarding training with v2?

[pale gyro]

Ok finally I found the original w-okada colab version

[viral mason]

Oh god the original okada

[earnest forge]

gotta admit I'd only recently figured this; but the og 'Kada always gave me this unexplainable autotune when I talked. Deiteris fixed that 🙏

[spare eagle]

-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

[pale gyro]

Nvm it also failed

Connecting to the vm now seems impossible it's stuck there.

[knotty oyster]

GPU: NVIDIA GeForce RTX 3060

64 bit

Just want to troll in game with it

https://www.youtube.com/watch?v=SxdnGxicJOg&lc=UgzGz88-UBa6zY_UGn54AaABAg

[queen adder]

very powerful aaah

[viral mason]

Pretty much only Nvidia users (for now) get to use the good ai voice changers

[knotty oyster]

what?

[queen adder]

i mean our geforce its too powerful

[knotty oyster]

ah so i cant use it cause i have a 3060?

[viral mason]

Vonovox might get availability for amd users at some point, the creator even said he's working on getting a card for amd

[queen adder]

no

i mean

its very powerful and thats good

very very good

perfect

🗿

[knotty oyster]

ah i see, ty!

[west rampart]

is there a one step way to generate audio using tts and an rvc model

I was looking to use the kokoro tts model

[low shard]

don't use original wokada, you shouldn't even use wokada for pre-recorded audios

use the applio colab, read it

[low shard]

the original wokada colab is broken, and it's not the right program for you, use the applio colab

[low shard]

that video is outdated asf, dont use youtube for realtime voice changers, it uses an over year old version of original wokada, and vb audio cable creates issues on windows
delete the folder, zip and uninstall from windows app settings

Just want to troll in game with it
are you looking to do e girl trolling/catfish?

[west rampart]

ive legit edited the kokoro local tts gui code to play the sound to my output device instead of just leaving it as an output file then routing it through to the voice changer

which now that I think about it seems extremely inefficient

[low shard]

no it's not related to that

[pale gyro]

It dosn't even work '-'

[low shard]

the applio colab works, could you elaborate the issue?

[pale gyro]

Downloading pre trained models wont show them on list. And even using it's custom path will give runtime error

[low shard]

pre-trained models are only used as a base for training models, are you looking to train models?

[pale gyro]

oh bru..

Are all models rvcs btw? So I use #1175430844685484042

[low shard]

😭 if you only want to use models on pre-recorded audios, dont touch any pre-trains

[low shard]

not all, but most yes, just be sure they have the RVC tag

[pale gyro]

duh only 10 results

[low shard]

there are over 10k rvc models, try restarting your discord

might be there's a lot of post forums, you could also try using https://weights.com/models

[junior pulsar]

Where can I download the okada voice changer?

I have a amd gpu on windows

[pale gyro]

from web

github

[junior pulsar]

Can you send me it?

[low shard]

which amd gpu?
which windows version?
and what do you want to do?:

• ai covers
• tts
• e girl trolling/catfishing
• roleplay in vc
• roleplay in games

[low shard]

any updates?

[pale gyro]

?

[low shard]

@pale gyro please don't suggest that one, it's outdated

it's original wokada

[pale gyro]

It was from the original github repo. But the english one?

[low shard]

yeah, it's outdated, wokada deiteris fork (modified version) b2332 is the one with the best performance at the moment

[pale gyro]

they should update links on the md file lol

[low shard]

the original one isn't used anymore

[pale gyro]

I see

[low shard]

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

Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[worn silo]

Hello guys

[pale gyro]

oh wow the whole #1175430844685484042 didnt help me lol

Time for gogoling

[low shard]

!give-media-perms 1h @stoic frigate

[low shard]

https://weights.com/models

start the program and used just like today before and the previous days
what program? there could be thousands of AI programs

[worn silo]

guys i want to switch from gpt api to my own local model and lower the latency time and the accuracy should be similar to gpt, so i need to finetune a model. which model should i choose for this? like should it be smaller like distilBERT,Tinyllama or some few 8b parm models?

[low shard]

oh you're using original wokada, it's not suggested anymore, you deffo used a old youtube tutorial

delete the zip, folder, and uninstall vb audio cable from windows app settings

vb audio cable also creates issues on windows

[pale gyro]

Well bruh Applio dosn't even work

How am I supposed to show the models under Voice Models menu ?

[low shard]

send the tutorial link you're using, i feel like you're using the Novision AI Girl voice tutorial

[pale gyro]

I even set them into: ..\ApplioV3.2.9\rvc\models\kenki

[low shard]

you shouldn't use that, it got less performance and quality, it's outdated and as you see its more bugged

[viral mason]

https://cdn.discordapp.com/attachments/1194442192111013888/1316496021039349880/attachment.gif

[pale gyro]

Bro ?!

https://i.imgur.com/TUWzKE0.png

Wait I can post pics wew

[low shard]

because you need to download the models first

[pale gyro]

And this?

[low shard]

did you do as in https://docs.aihub.gg/rvc/cloud/applio-colab/#1-upload-voice-model ?

[pale gyro]

I didn't use it on colab since it dosn't work with Drive

Oh I see. It does install it on /log wow

I installed them into the rvc/models path

Uhh... Isn't the model supposed to use the GPU instead of the CPU?

[low shard]

what? you need to use colab, your gpu is old

you can't do it locally

[pale gyro]

Not old but not supported (Intel)

This is why we wanted VCC instead

[low shard]

you shouldn't use local, you can just use the Applio colab, original wokada and applio both have colab versions

[pale gyro]

Had to use rmvpe algo for it.

[spare eagle]

-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

[idle elbow]

so wwhen i talk with okada it doesnt work in my game

i have to tab out to hear it

[low shard]

Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[idle elbow]

come to the vc

@low shard

[chilly hound]

I need help with voice changer on colab

[low shard]

I can't VC, it's better you elaborate what I asked

[low shard]

Please Elaborate, those are crucial infos:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[rare glen]

how to add my own voice or i cant?

[low shard]

you'd need to train your own rvc model of your own voice

i hope you didn't come here from yt tuts, they all old

[rare glen]

um

whats the new method

[low shard]

for what?
Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

depends on which thing you want to do, there's new programs for different things

[rare glen]

to make ai covers

of my own voice

[low shard]

yeah don't follow the rvc mangio or rvc gui tutorial you see on youtube, please answer the infos i asked u

[rare glen]

uh idk the gpu
windows
ai cover
tutorial outdated

and idk which program

[low shard]

uh idk the gpu
that's the most crucial thing for ai

You can check your pc gpu on Windows via:

ctrl+shift+esc (task manager) -> Performance tab -> GPU

also I'm guessing windows 11

[rare glen]

yeah

[rare glen]

um ok but why does gpu matter

[low shard]

why does gpu matter
yes, A LOT, AI is intensive and complex, this one runs on your hardware, it isn't a 1 click company product like chatgpt that runs on a remote good pc cloud server, it's open source and driven by the community

[rare glen]

oh so it's one of those

ok wait i have to sleep can i do it tomorrow instead

[low shard]

alr, let me know

[wind ivy]

Hey, how do I run this: https://github.com/deepbeepmeep/Wan2GP, on my laptop which has these specs:

1. CPU: Intel CORE i7 12th Gen
2. Memory: 16GB
3. GPU: NVIDIA GeForce RTX 3050 Ti Laptop GPU (4GB VRAM)
4. OS: Windows 11

[upper narwhal]

Hello everyone, I have a question: What is the best program for making AI covers and which is the best for creating a voice model? I'm unsure which one to use, and I want to do everything locally.

It doesn't have to be the best, just the one that people usually use the most.

[low shard]

what's your pc gpu and operating system?

[low shard]

i don't think you can run it

[upper narwhal]

CPU: Ryzen 5 7600 6-Core Processor 3.80 GHz
GPU: Nvidia GeForce RTX 5060 8GB
Memory: 16GB 6000mhz
OS: Windows 11

[low shard]

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

• Applio: The most suggested RVC Fork with some extra features like TTS, kinda faster and simpler but same quality tho
• Mainline: The original RVC

[wind ivy]

Really, even though that is already the super optimized version?

I really have to upgrade my current setup

[low shard]

well i didn't try it, but in the readme it says that there are models that go as low as 6gb vram, and you got 4gb vram

[analog fiber]

https://www.youtube.com/shorts/Y2tzig9aty8
what type of AI is used here?

[low shard]

8gb vram is the bare minimum nowdays, you got half of that

[low shard]

@upper narwhal also let me know for any issues

I think it would be better to use Applio

not sure about rtx 50 serie support on mainline

[upper narwhal]

No problem, I'll test both and see which one gives me the best results. Thanks for the help!

[low shard]

which one gives me the best results
they should give the same results, the major difference is just Applio is easier and a bit more optimized, the quality is the same, because the original rvc devs left it to rot in 2023

[wind ivy]

Fortunately, my country, Malaysia, already released a laptop with an RTX 5090 GPU that has 24GB of VRAM, but it’s priced at a whopping MYR29,049, which is equal to $6,886.91

But it’s too expensive for me

My current setup costed around, ~MYR5,699

[simple ore]

not without a bunch of code changes

[low shard]

shouldn't it be possible with just updating pytorch?

[analog obsidian]

no

[simple ore]

you can bump pytorch, but then you need to go and change all torch.load() calls

[viral mason]

what are these accounts that just put images

are they stupid

I ain't clicking that shit

[low shard]

mrbeast scams 😭

[low shard]

what if they use it from source? is the main branch updated for that atleast?

[viral mason]

do you know what they even do? is it just an image or does it contain like a virus orr

[low shard]

no virus, u just open it, there's an image saying smt like "mr beast is doing twitter giveaway go there"

[viral mason]

most pointless scam accounts ever 😭

[simple ore]

dont think so?

[knotty moth]

having their image perms disabled is somehow safer

[wind ivy]

You know? I created an easy Python package manager for Windows only, that helps manage your Python environment, installs packages, search PyPi, remove packages, etc…
For full details, go here: https://github.com/ROCKYANDTHEPAWPATROL20/PENVCreator/releases

Of course, the release’s description is generated by ChatGPT

[simple ore]

(and the rest of it? 🙂

[wind ivy]

Are you mentioning me? If yes, then the code itself and Markdown all generated by AI

I have recently updated the software: https://github.com/ROCKYANDTHEPAWPATROL20/PENVCreator/releases/tag/penvcreator2.0

[river lintel]

hii i wanted to use the w okada voice changer to troll my friends but its stuterring and has a big delay why is that so?

ryzen 7 5800x
rx 6650 xt
windwos 11
32gb ram

[low shard]

troll my friends
are you looking to do e girl trolling/catfishing?

[river lintel]

not e girl just generally use a differnet voice i am a girl haha

like a troll voice

[low shard]

share the tutorial link u used

[viral mason]

first ever normal person to join for the voice changer

mark this in history

[river lintel]

i just installed it with the github informations and by searching the chat

i uninstall and reinstalled it again and it worked for like a couple minutes but now it doesnt again

also there’s a bit of static/noise in my microphone even when i dont talk. can that cause probelms?

[quasi condor]

@hollow sparrow

[hollow sparrow]

yo im here

[quasi condor]

so ur saying its using your cpu instead of ur gpu?

[hollow sparrow]

yea, i selected gpu 0 dont know what that is but i checked task manager and all the powers going to my cpu

[quasi condor]

u gotta change the cpu on okada to ur gpu

[hollow sparrow]

wdym

[quasi condor]

is okada open?

[hollow sparrow]

this thing?

yea

[quasi condor]

send a ss of okada

[hollow sparrow]

where dms

[quasi condor]

u can send it here

[hollow sparrow]

okadas like the actual voice changer right, sorry if im slow just new

[quasi condor]

yea

[hollow sparrow]

alr give mea sec computers laggy

what like this?

interface?

[quasi condor]

oh no ur using the old one

this one is outdated

[hollow sparrow]

where can i get a new one

[quasi condor]

type -realtime on bot commands and click the first one

[hollow sparrow]

here

now what

[quasi condor]

its a guide

[hollow sparrow]

oh alr

[quasi condor]

also what's your pc gpu?

[hollow sparrow]

rx 580

[quasi condor]

click this one

download amd on windows

[hollow sparrow]

hm?

[quasi condor]

@low shard i needs help im at outside rn

[errant ginkgo]

Does anyone have a google colab link for ai cover?

[sage lagoon]

what causes my mic to spike? like there's just occasionally a word that's super sharp whenever i use an ai voice. or is this just a 'me' issue?

[simple ore]

look at loss_avg_50 charts instead

[hollow nova]

whats the best app to merge voice models? Just been using wokada fork to merge. I also assume it’s not possible to merge models with different sample rates right?

[rigid bobcat]

hello everyone I just downloaded and started to use W-Okada Voice Changer with the install version of vcclient_win_std_2.0.78-beta.zip ... I own rx9070xt and was hoping to get faster realtime output voice changing. But results were so chopped and unlisteneable. I wonder if its about the version I installed or the Chunk/Extra settings I have... For example only avaliable range to have real time voice is usually Chunk value of 72000 [1.5 sec] with Extra 144000 [3 Sec]

[short oracle]

I am using 5080 laptop version, windows 11, i want to reduce delay betwen me speaking and the ai voice changer outputing it

[low shard]

Huh

[low shard]

Can u share the link? There's different programs

[low shard]

Do not trust yt tuts, delete the zip folder and uninstall vb audio cable from windows app settings

What's your PC GPU, os and what do you want to do?

[low shard]

Did you first check if your PC GPU is enough?

[quasi condor]

i said that an hour ago o- o

[errant ginkgo]

I use phone,not pc

[quasi condor]

u gotta use pc to use that

[low shard]

Wokada deiteris fork, and no not different sample rates

[quasi condor]

oh its colab its a bit difficult to use tht on mobile

[low shard]

I'd suggest weights.com In your case since it's mobile friendly

[quasi condor]

weights - a bit buggy on mobile :p

[low shard]

Huh? Works fine for me

[quasi condor]

its buggy for me cuz my phone is literally 3 years old

[low shard]

You encountered bugs? Would be better to report them in discord.gg/weights

[low shard]

What phone do you have?

[quasi condor]

i thought u knew o - o

[low shard]

I forgot tbh

😭

[quasi condor]

breh

[low shard]

You're using outdated original wokada and prolly vb audio cable,
Delete the zip folder and uninstall vb audio cable from windows app settings

What do you want to do?:

• ai covers
• TTS
• e girl trolling/catfishsing
• roleplay in games
• roleplay in vc

[quasi condor]

tho i should update my phone

its telling me to update to android 15

should i?

[low shard]

What do you want to do?:

• ai covers
• TTS
• roleplay in games
• roleplay in vc

What tutorial link are you using? Share a screenshot of your settings

[low shard]

If ur phone is old and always lags every second, maybe it's just better to change it

[quasi condor]

eh screw it im risking it

[golden walrus]

sir, can i ask if you know how to make the voice less chipmunk ?

[low shard]

wdym?

[golden walrus]

i mean

somehow my pitch and formant make me sound like 12 yo kid

or even less

[quasi condor]

are u using okada or smt else?

[golden walrus]

Vonovox

[quasi condor]

o

[golden walrus]

i mean, it could be my data

fault

[quasi condor]

well i don't use vonovox @low shard u can do this 😄

[golden walrus]

i thought it's just pitch and formant stuffs

[rigid bobcat]

roleplay in games and vc

[winged garnet]

hey, i played a game with realtime voice changer with my friend and it was quite interesting, now i want to try livestreaming with rvc is it possible? for female voice my sister is a vtuber so i can totally ask her to record for about 2-3 hours or can i get her voice on stream to make a quality model? i also asked grok AI and grok said i need both my original voice and my sister's voice as "base" and "target" is it necessary? to train a quality model how long does it take to train data. i will prioritize voice quality and not care about how long it takes to train the model or if it is harmful to the pc. my current pc is i5 14400f, rtx 5060ti 16gb, 32gb ddr5 6000. with my configuration can i train a high quality model? and for my needs which rvc should i use to train the model and which rvc to change voice in real time?

[low shard]

delete everything

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

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

• Vonovox

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

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even 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]

1st link wokada deiteris fork, read it

[winged garnet]

reply me pls

[pseudo steppe]

is this still working?

[low shard]

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

• Applio: The most suggested RVC Fork with some extra features like TTS, kinda faster and simpler but same quality tho
• Mainline: The original RVC

[low shard]

yes, but you said got an rtx 4060

[pseudo steppe]

umm i need for my friend

[low shard]

what pc gpu does your friend have?

[pseudo steppe]

rx 580

[low shard]

that's bare minimum, is he here in the server? what does he want to do exactly?

[winged garnet]

also asked grok AI and grok said i need both my original voice and my sister's voice as "base" and "target" is it necessary? Or do I just need to train my sister's voice? For best quality, how much data do I need to prepare?

[pseudo steppe]

for discord only

[low shard]

no, you just need to train the voice, chatbots don't know about RVC, use the docs https://docs.aihub,gg

[low shard]

should work, id suggest him to join here if he got any issues

[pseudo steppe]

the issue is he can't open that application

[low shard]

tell him to join here, would be better he elaborates why he can't open it

[winged garnet]

How much data do I need to get the best possible voice quality for livestreaming?

[pseudo steppe]

@drifting ether

[drifting ether]

I'm not on my pc for now

[pseudo steppe]

alright

[short oracle]

Cant rn, and want it to vc with friends and use female voice (trans)

[low shard]

Ohh, that's great!

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

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

• Vonovox

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

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even 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]

be sure you're using either wokada deiteris fork or vonovox with vac lite, and not original wokada with vb audio cable

[glossy obsidian]

Hello, I establish a small hospital, I wanna making healthcare software using AI. so identify illness with voice. is it possible?

[short oracle]

already doing that

using wokada fork

[low shard]

b2332?

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

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

it would be better u also show a screenshot of ur wokada deiteris fork later

[short oracle]

uhh i am on trip rn will be back in likee

4 to t days

u good at that time?

[low shard]

i mean sure u can @ me

enjoy the trip

[paper bloom]

hey i have a question when i use that owakada voice changer thing sometimes i can hear other ppl thru my headset and it changes there voice too

it only happens sometimes

i have to restart my pc to get rid of it or am i doing something wrong?

[pastel oak]

It means your microphone is picking up your headphones volume, so reduce your headphones volume, move mic further away, move the slider for noise gating under the F0 option further to the right

[low shard]

that's an over year old of original wokada, it's outdated asf

don't follow yt tuts, don't use vb audio cable it creates issues on windows such as crackling

delete the zip, folder and uninstall vb audio cable from windows app settings

please elaborate:

• ur pc gpu
• operating system
• what you want to do?:
  • AI Covers
  • TTS
  • E girl trolling/Catfishing
  • Roleplay in call
  • Roleplay in games

[naive horizon]

Hi, sorry to barge in, I'll ask again:

GPU: 3070 Laptop
OS: Win10
What I'm trying to do: I'm trying to get into streaming as a hobby after work. I just dont wish my voice to be recognized, while not making it obvious I am using a voice changer. Not robotic, or too high pitched, just another human. That's the bare mininmum - preferably I have a character I can play as the voice. I am male and it'll also be a younger male.

[iron venture]

can someone help me get the latest version i have an nvidia gpu and windows the only versions i see is a year old

[low shard]

will help in https://discord.com/channels/1159260121998827560/1399115654220742676

[low shard]

dont follow yt tuts they are all old, tell your pc gpu, OS and
are you looking to do e girl trolling/catfishing?

[iron venture]

tiktoks i have a 1070 and a i7

[low shard]

tiktoks
like e girl trolling/catfishing catching pred tiktoks?
and what's your OS?

[iron venture]

no just singing i have windows 10

[low shard]

just singing? then you're using the wrong program

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

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

[iron venture]

i was gonna use it for normal talking too for background talking in videos

[low shard]

i think what you need is inference (use models) on pre-recorded audios

not realtime voice changing

RVC doesnt mean realtime voice changer

[iron venture]

i know that i need to figure out how to use both

[low shard]

you don't need to use both, it's just a waste of time, wokada isn't the right program for what you need to do

[iron venture]

ill figure it out myself

voice models tab

i guess look on yt for working ones

[low shard]

are you looking to do e girl trolling/catfishing?

[low shard]

yt are all outdated, never trust yt for RVC nor Wokada, AI moves at sonic speed

[iron venture]

im talking about suggested models from the voice models tab

[low shard]

Your Nvidia GPU is good enough to do inference (use models) locally (on ur pc), not the best to train (make models) even if still possible

You can:

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

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

[low shard]

I gave you help for what you need, or do you actually need to use AI for something else as e girl trolling? I think you're confusing programs

[iron venture]

no i dont need it for e girl trolling im literally a girl

why do you keep saying that

[viral mason]

YouTube uses outdated stuff if you wanna use buggy broken stuff go ahead

[viral mason]

A lot of creepy losers with no life come here to download the voice changer for that purpose

It's way too common

[iron venture]

oh wow

[low shard]

then you can read the guides I sent you, it will help you understand how to use the program

[normal drift]

Hello, Im trying to use weights' AI voice model in RVC AI cover maker, but it seems like there's no way to copy the direct DL link ? I can't right click the DL button. I wanna paste it into the url text field to download instead of having to download the model locally to my laptop then upload it again

[low shard]

I wanna paste it into the url text field
you can't do that, the only way is you have to download it on your laptop then upload it yourself

[normal drift]

Oh, but iirc there used to be such a feature, why was it removed ?

I used to be able to copy the direct DL url of model from there

[iron venture]

how u fix bad ms

[low shard]

nope, there was never for weights.com, there is (and still is) only for HuggingFace.com (a general AI platform), those are 2 different sites

[low shard]

show a screenshot of your program

!give-media-perms 1h @iron venture

and also tell what u doing, like playing marvel rivals with it

[normal drift]

Back then it was called weights.gg I think, but whats the reason for not allowing a direct DL link ?

[low shard]

Back then it was called weights.gg I think
nope, I'm there before even weights.gg was ever a thing, it was never possible, you had to always do the manual download, and you can't do anything about it, they do have things like login so other sites don't just scrape them iirc

[normal drift]

Oh, that makes sense, thank you

[low shard]

are you using it in a game? you said you wanted to sing for tiktok live or smt ?

which program are you using it with, that matters a lot

[normal drift]

Btw, is there any AI now that can do remix or mashup of input audio(s) ?

[iron venture]

im in a discord call with my friends its still high ms

[low shard]

No need to ping me, I saw your message above that you deleted it, and no I'm not aware, i don't think there is one

[low shard]

maybe try like suno, not sure

[low shard]

set f0 to rmvpe without onnx

extra to 1.5

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

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

[normal drift]

Btw, when using RVC AI cover with input song of male voice but model is female singer, am I expected to change something in setting like pitch or just leave them at default ?

[low shard]

dont expect low delay, your gpu is ancient

[low shard]

try playing with the pitch

[iron venture]

yea true

[normal drift]

Thank you, btw in AI art the strength of the model is controlled by CFG I think ? So in AI audio, how can I strengthen the effect of the singer's model to influence & make the output sound more like them singing ? Is this possible ?

[low shard]

those are 2 completely different AIs lol

[low shard]

lmk

[normal drift]

🥲 Sorry, I don't understand the technicality behind these AI stuffs

But is there a way to influence the strength of the AI voice model ?

[low shard]

nope in RVC, the only thing similar to what you said is just the index ratio, where you can control the influence of the trained accent

[normal drift]

Oh, the accent, I think I wanna give that a try aswell to see if it gets the effect I want

Is it this option ?

[low shard]

yes

[iron venture]

lol idek atp it was better a couple hours ago

[low shard]

you sure that's what you want to do? you were talking about tiktok tts before #✨│ai-help message, this isn't the right program for you, why you using this one?

[iron venture]

my friends using it while talking to me im gonna use the same thing

works fine for them

i fixed it

[low shard]

what? that's different from what you asked

[normal drift]

What are your personally favorite/ recommended AI voice models ? I might find precious hidden gems I never knew existed before this way

[viral mason]

just check the voice models tab

[iron venture]

is it bad i wanna just do whatever i want with the client? i dont have a set thing im just trying to have fun with it?

[viral mason]

it was even in his name

[ornate comet]

16.0 GB (15.4 GB usable) - AMD Ryzen 5 PRO 4650G with Radeon Graphics 3.70 GHz working here ?

[cold cave]

@low shard Do you know how to save storage from training. When I train models from applio noui, it fills up my storage automatically

[low shard]

working for what? what do you want to do?:

• ai covers
• tts
• e girl trolling/catfishing
• making rvc models
• use rvc on pre-recorded audios

[low shard]

clean up storage from your drive
be sure to train with save only latest on
leave autobackup cooldown to like 15
Enable cleanup if this is your first time training a model and you're not resuming

[cold cave]

I hope it works

[knotty moth]

enable the "Save only latest checkpoint" option

[cold cave]

I already did

It works

[knotty moth]

I have no idea since I'm using kaggle but ig it doesn't overwrite G & D files in Drive storage

so you let it store in the colab internal storage

[cold cave]

What about fast training in applio no ui?

[knotty moth]

[naive arrow]

Yo is there anything like this live voice changer but it changes from like English to a different language

[tepid moss]

This one's a tricky question. Since it involves the math part of AI.

Everytime I study the math formulas that make these functions possible I can never fully comprehend how, lets say writing a prompt is the same as a math function that is shown in a paper...

Or how an image diffusion model turns noise into an image and is also explained by a math function... You can say that I did not grasped all the pieces of this part.

Would any of you guys could give me a better example on how to understand these?

[placid holly]

did anyone ever train on f0: FCPE for W-Okada?
I heard FCPE is lightweight for Okada
but i don't know the Quality From FCPE train to FCPE Okada, not yet tried that stuff

[viral mason]

who would train on fcpe when rmvpe exists

[sly heart]

i already did the all steps but why my mics doesnt turn on

[vestal bear]

what is the best girl model for vietnamese voice

[hallow thistle]

For a Vietnamese voice model, try find one in #1175430844685484042.

[late badge]

Can someone tell me what the best most accurate ai juice wrld voice model is? Or is there any such thing as premium models that people sell that are alot better?

[low shard]

Can someone tell me what the best most accurate ai juice wrld voice model is?

There are thousands of rvc models, the only way is to listen to the demo and try it yourself

Or is there any such thing as premium models that people sell that are alot better?

Nope, paid models are not allowed

[low shard]

Are you trying to do e girl trolling/catfishsing?

[low shard]

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

Please Elaborate:

• your PC GPU
• your operating system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

Maybe try https://github.com/Sharrnah/whispering-ui

[quaint minnow]

Hey, so I would like to convert a recording I made into a different voice for a hard techno song I'm currently making, I have a gtx 1660 super and I'm using Windows 10, what program should I use to convert my voice using another voice model?

[low shard]

Your Nvidia GPU is good enough to do inference (use models) locally (on ur pc), not the best to train (make models) even if still possible

You can:

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

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

Let me know for any issues!

[quaint minnow]

Thanks! I'll give those a try

do you know if there are any "neutral" rap voice models that are available?

I need a rap voice but not of any famous artist

if that's even a thing

[low shard]

Ehhh, RVC AI model need to be trained on a specific voice, you could try searching for less popular ones, or try merging/fusing models in applio (but honestly not many people use that feature)

[quaint minnow]

I'll try that

finding a less popular one that is

[broken urchin]

hey

i was wondering can i run w okada deiteris fork on GTX 1070 gpu

i currently have the RTX 3050, will the GTX 1070 have much worse performance?

[hallow thistle]

You can run W-Okada with GeForce GTX 1070. However, GeForce GTX 1070 would have worse performance if compared to GeForce RTX 3050.

[broken urchin]

and why is that? the power of the gpus is very similar

the gtx card is a bit stronger in power too

[hallow thistle]

That's an expected question. While some claimed that GTX 1070 would outperform RTX 3050 over gaming at certain situations. The thing is some reported that some GTX 10/16 GPUs couldn't run W-Okada with a graphic demanding game at the same time without lowering the settings.

[simple ore]

1070

3050

3050 is 1.5x faster in fp32 and 100x faster in fp16

[broken urchin]

oh so thats why

makes sense

[simple ore]

3050 has 80 tensor cores, 1070 has none

tensor cores is what gives it a boost in AI

[viscid osprey]

what voice changer should i use for a 7800 XT?

i've been using w-okada fork for a while now

[hallow thistle]

Use this one. https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#download-amd-intel-and-cpu-on-windows

[viscid osprey]

it kinda sucks? or it just doesn't really give me what i want, the breaths are broken

thanks!

[hallow thistle]

There's a better W-Okada. The old original one is broken.

[viscid osprey]

i think i have this one

[low shard]

you got wokada deiteris fork b2332? (specifically this version since it's the latest)

and vac lite?

you can show a screenshot, so i can also check your settings/advanced settings

[viscid osprey]

i'll boot it up rq

yep

i was looking for something new

[low shard]

show a screenshot of your whole settings & advanced settings

[low shard]

there's nothing new for non-nvidia users (There is Vonovox only for Nvidia users), also i feel likke this might be related to your settings or models

[viscid osprey]

eh
might be the models, i'm just looking for realistic ones

[low shard]

your normal settings are fine, but you can adjust the advanced ones

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

• Advanced Settings -> Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Advanced Settings -> Crossfade Lenght: Controls how smoothly the AI stitches different processed parts "chunks" of your voice back together. 0.1 for fastest voice, 0.15 for improved quality but increases delay by ~50 ms
• Reduce the delay on Windows via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

also, try other models, and remember that RVC models got limitations on non speech sounds, such as laughing and screaming

[viscid osprey]

hmm, i'll try that

[pastel lantern]

I am going to try and create visual art (image or video) with the help of prompt and AI
is StableDifussion good enough for it?
I have integrated graphics so it takes time but with StableDiffussion I can generate pack of 20 HD images in 40-50 minutes (some get's blurry or unfocused)
help me out with this one

[low shard]

I have integrated graphics
you shouldn't even do it locally then

https://github.com/rupeshs/fastsdcpu there's this if you want, but it isn't the best quality

i would rather use cloud or get a better pc in your case

[pastel lantern]

hmm I use laptop maybe in a year or two wil get pc

[pastel lantern]

I'll look into cloud then

[viscid osprey]

what models could you recommend that handle non-speech better?

[low shard]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#models-to-try , you can also search for models, some might have been trained on things like laughing, but as I said, it's an RVC limitation, it's been like this for 2 years and since the devs left it to rot, it will probably never do an actual realistic laugh since i'm guessing you're looking for that

[viscid osprey]

and i can only use RVC on w-okada?

[low shard]

There's no AI Voice Model that surpassed RVC in quality yet, and yes

there are other STS AI Models like So-VITS-SVC, Beatrice 2.0, but those are lower quality and old, no one really uses them

[hallow thistle]

RVC voice models can be used on W-Okada.

While you could use Beatrice voice model in older W-Okada versions, it was later removed in fork W-Okada, leaving only RVC.

[rancid pine]

What should I do if I've reached the daily quota for ZeroGPU? (Ilaria RVC)

Do people not train models locally anymore?

[hallow thistle]

Try another different online RVC option like on Kaggle or Google Colab.

[low shard]

Either wait, pay for huggingface pro, or use something else

[low shard]

Ofc they do, Ilaria RVC Zero isn't even for training models

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

[proven hill]

use it locally

[hallow thistle]

Also, some people here train voice model locally on their PC with a GPU. They not all quit training voice model locally entirely.

[rancid pine]

It was a separate question, I am aware it's not for training models.

I don't have my PC specs with me right now, it'll be out of my reach for a while, but I use Windows

[rancid pine]

Link to the guide? Is it preferable to do so on cloud or local, or does it depend?

[hallow thistle]

A specific Ilaria RVC fork on Hugging Face can't be used to train a voice model, but the mainline one with full functions of similar name does.

[hallow thistle]

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[rancid pine]

I still have to navigate through the categories

[hallow thistle]

These are options where to run RVC online.

[low shard]

there's info in our docs, but it would be better you check your PC GPU because it's crucial for AI training first

[rancid pine]

Is it preferable to train a model on cloud over general cases, then?

[low shard]

it's suggested to train a model on cloud only if your pc gpu is not good lol, overall cloud got limited free time

there's no difference in the program code, it's the same quality, cloud is used only for GPU Poor users

[rancid pine]

I see

Thanks

I'm pretty certain my PC can handle the load because I've trained a few models locally before

[low shard]

if you're sure about that, then you can just use Applio from our ai hub docs locally

[hallow thistle]

It is preferred to train a voice model online when you don't have a good GPU (lower than RTX 20 GPU), but online options are usually limited in GPU usage time. RVC on online and local all function exactly the same; the difference is which RVC fork you use.

[rancid pine]

I'm trying AICoverMaker Kaggle again, and it gives me this prompt: Installing requirements ERROR: Could not open requirements file: [Errno 2] No such file or directory: 'requirements.txt'

[weak tartan]

-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

[low shard]

@viscid moss could u check this?

[viscid moss]

Hmmm

Lemme check

[simple ore]

not cloned the repo before running install?

that's the only reason for reqs missing

[rancid pine]

No, already cloned

[viscid moss]

Are u sure that u are running the notebook from this link?
https://www.kaggle.com/code/eddycrack864/rvc-ai-cover-maker-ui

• running the latest version notebook (v6)

[rancid pine]

Yes, I am using the correct link. How do I check if I'm running the latest version Notebook?

[viscid moss]

Before u press Copy and Edit, choose version 6 of 6

Working on my test

[rancid pine]

Hmm, that is what I chose before

[viscid moss]

If u have persistence enable, disable it. Maybe something gets corrupt

Try again with a clean env

[rancid pine]

This requires internet on, doesn't it?

[viscid moss]

yep

[low shard]

yeah, which means you need to verify your phone number

[rancid pine]

Tried to, but it's a lost cause. Apparently, it couldn't verify my number and I would have to go through the hassle of identity verification. Thanks for the help nonetheless

[low shard]

Contact kaggle support about it https://www.kaggle.com/contact

or, just use local since you said you got a good enough pc

[viral mason]

Ew why is there no persistence

[viscid moss]

cuz it's for testing only xd

[viral mason]

That's a guaranteed loss of all progress as soon as you leave the space 💔

[hallow thistle]

I select CPU mode at first to test a notebook on Google Colab and Kaggle.

[lethal iron]

why does my comfyui zluda take hours to make its first photo it has been on 57% for almost an hour now

[arctic bronze]

Hi everyone, sorry for the text, I'm using a translator, I'm asking for help, I need a person to make a voice model of my friend, his birthday is on August 14th, I was trying to make a voice model and calculated that 150 epochs will be processed after 20 hours.
In general, who wouldn't find it difficult to help me?

[hardy schooner]

Hi, have anyone here worked with ssd_mobilenet_v2_fpnlite_640x640_coco17_tpu-8? I have trained the model and tested it in local env - everything works. After that I converted it to tensorFlow.js format and pushed it to browser (Blazor). The code itself seems to work but the output I'm getting is nonsense. Is there anyone that is willing to help me with this?

[lethal iron]

CUDA error: CUBLAS_STATUS_INTERNAL_ERROR when calling cublasSgemm( handle, opa, opb, m, n, k, &alpha, a, lda, b, ldb, &beta, c, ldc) can someone help me with this error

[viral mason]

you could use Applio and create the voice model yourself, it's completely free and easy to learn

[simple ore]

dependson what you're trying to do

[muted needle]

Seeking Native English Speaker for Exciting Opportunities!
I am a AI developer.
Are you a native English speaker with excellent communication skills? I'm looking for a native English speaker collaborated me!
I need only language skill not anything.
If you want to collaborate with me we can discuss just now.

[simple ore]

what's the scam?

[pastel orbit]

i wanted to upload image here but nvm

[worthy valve]

I HAVE A GPU Nvidia RTX

Win 10
AI Covers
I want to see if the program is updated or not. I had it a long time ago, but I formatted my PC.

[cedar sky]

Guys, who knows what is the best offline audio transcribing model for mobile devices?

[low shard]

which?

[low shard]

you said realtime voice changer first, what do you actually want to do?

[worthy valve]

3060

[worthy valve]

I want to see if the program is updated or not. I had it a long time ago, but I formatted my PC.

[low shard]

yeah there are different programs

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

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

depends on what you want to do

[worthy valve]

Roleplay

[low shard]

so, just to clarify, you want to do roleplay like spongebob in games? because you have been changing the reason previously to ai covers

[worthy valve]

like spongebob

[worthy valve]

Or a character from a TV series

[worthy valve]

The desired goal is to see how far the AI progresses, but I enjoy role-playing with friends or other games like VR C.

[low shard]

alright

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Wokada Deiteris Fork

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

• Vonovox

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

• Original Wokada

Not suggested nor maintained, older versions in youtube tuts are even 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]

you can use either wokada deiteris fork or vonovox

they have similar performance and quality, u can check more about their pros and cons in their guides

[worthy valve]

I have tried RTX 3060 12GB which one is best for my card

I don't know why when I use this there is a clipping in the sound and the sound is not good

[low shard]

that's an over year old version of original wokada