✨│ai-help

[viscid moss]

U re using an outdated colab probably

Use the new ones:

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

More commands: -audio, -local, -overtrain, -realtime, -rvc, -help

[rustic locust]

now what

im done with the training

[thin stump]

Did you remember your step count?

[rustic locust]

nope

how do i find it

I think it's 3k

[thin stump]

Download that in your weights folder and test that with your index

[rustic locust]

which one?

there's 4

[thin stump]

Did you stop the training? Download the one that says e_3000

or the one closest to it

[rustic locust]

okay i downloaded it

[rustic locust]

what now?

[thin stump]

Zip it in a folder with the added index. That's your voice model https://docs.aihub.wtf/essentials/how-to-make-ai-cover/

[rustic locust]

okay

do I run local URL or public URL?

I'm on this screen

[thin stump]

Never used it before so I can't help

[rustic locust]

I thought you made one yesterday

[thin stump]

tryreplay.io is what I use to make ai covers

[warm notch]

does anyone know the optimal amount of epochs to use for 10 minutes of audio for my model

[rustic locust]

okay I'm getting replay

[thin stump]

The TB will tell you. Past the red circle is overtraining https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/#--monitoring

[half cove]

its always good to test different checkpoints

[warm notch]

I use RVC1006Nvidia Do you know where I can check it on this?

Cause Ive just been using 100 epochs for everything

[lavish minnow]

Can you do him

[rustic locust]

It sure is taking long to download

[thin stump]

I never used local so I can't help

[thin stump]

If the recent version doesn't work you can downgrade to the 2/10 version. That's what i'm using

[thin stump]

I mean 2/10/2023 Version 3.5.3

[rare goblet]

Ayo? @lavish minnow level 2 !!!

[lavish minnow]

https://tenor.com/view/lets-play-league-lets-play-league-of-legends-league-of-legends-league-lets-play-gif-24554643

v

v

v

v

v

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[proper shale]

the guide explains it, doesn't it?

[rugged stirrup]

Would it be a good idea to use 32k even though this is probably needing 40k

[proper shale]

has a harsh line on 16khz

so yeah 32k

[rugged stirrup]

Alright

[finite galleon]

Normally with adobe enhance you want to stay at 32k

[rugged stirrup]

Yeah this dataset was likely put through it

I didnt make it

[finite galleon]

Adobe always has that wierd line at 16k

[rugged stirrup]

Adobe enhance messes with the quality alot or atleast the speech

[proper shale]

yeah that has been enhanced b4

[finite galleon]

It makes it sound like no breaths and just flat

I dont know what the term is

[rugged stirrup]

It makes the speech sound hollow

[finite galleon]

I use it as a last resort if any overlapping sfx or random noise couldnt be removed by any uvr models

[rugged stirrup]

We need better SFX removal models for bird sounds and whooshes

[rustic locust]

I have my model and mp3 in replay. what now?

[narrow maple]

uh hello
im new here
i need help cuz im trying to get rvc working over a google colab
and im getting an error

[proud elbow]

I'd never try adobe enhance if there's RX 10

[proper shale]

what's the error?

[narrow maple]

"JSONDecodeError"
god i wish i could send a screenshot-

but i respect that you cant send images here

[narrow maple]

idk what i am doing wrong, but everytime i try to load step 2 i get this error

[proper shale]

yeah thats an outdated colab

[narrow maple]

ops

[rare goblet]

Ayo? @narrow maple level 1 !!!

[proper shale]

https://docs.aihub.wtf/ has some links and tutorials

[narrow maple]

thank you

[broken lagoon]

what do i do

[narrow maple]

thats the same issue i had

[proper shale]

outdated colab, check https://docs.aihub.wtf/ for more recent stuff

and tutorials too

[narrow maple]

again, very much thank you

[proper shale]

yw :)

[warm notch]

So I tested out 500 epochs and im about 300 in. I set it to save every 100 epochs and I have a D and G file. It has both options on here do I need to put both paths in order for it to recognize the presave? Also If I stop the training I could just start it back up with this option at 300?

[proper shale]

1. Keep those pretrains as is, those are just the base pretrains and you don't need to use the D/G files in the logs folder.

2. You can just resume training with the same settings, it'll work unless something messed up

[warm notch]

oh ok. I might have ran into an issue. The model has the same model name as another one I made since this one is just updated with higher epochs. Will it overwrite my other one? Or will it cause an issue

should I just delete the old .pth file

[proper shale]

Maybe move those files out of logs and weights

To prevent that type of stuff

But in case you wanna delete that might work too

[proud elbow]

move out or rename the folder name in logs

[warm notch]

it says the file is open in another program

I might just delete it and hope for the best

ok so whats happening is the saves are going into the old folder along with the old presave .pth files.

in logs

[proper shale]

oh yeah

predictable tbf

you can delete the old ones then

[warm notch]

how do I know which ones are the old ones and which ones are the new ones

[proud elbow]

"Date modified"

[warm notch]

wait they're all modified today lol. I think its automatically deleting the old ones

I might be fine

[proud elbow]

if ur not sure, start over the training with the same configurations as last resort

[rustic locust]

can I get the link to the collab again

[rare goblet]

Ayo? @rustic locust level 4 !!!

[proper shale]

which one

[proud elbow]

-colab

[azure marsh]

Suggestions for @rustic locust

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

More commands: -audio, -local, -overtrain, -realtime, -rvc, -help

[warm notch]

Okay so I think I figured out the issue. I used to use an old RVC and I transferred over the .pth files without all the logs. So the reason the logs were all modified today is because its the only one. But the .pth file in my weights is still the same name as the one being trained. So in theory couldn't I just delete the old .pth file in weights?

[rare goblet]

Ayo? @warm notch level 4 !!!

[proper shale]

yeah

if it isn't interfering with logs dw about it

[warm notch]

ok yea I just deleted it. Thanks guys appreciate it

[proper shale]

yw! :)

[half cove]

Bandit plus

[rugged stirrup]

I used that it removes most sfx but not all

[narrow maple]

im having a 1033 error in middle of downloading..

[half cove]

Best we have rn

[rustic locust]

bruh

I cant use it

I thought it was free

[half cove]

Internet issue I'm guessing

[half cove]

It is free

And open source

[rustic locust]

Well it's saying i have to pay because I can't use my GPU or something

[proper shale]

oh yeah your GPU quota might be up

[rustic locust]

how to fix it

[proper shale]

you'd have to wait for 24 hours

[rustic locust]

oh

[proud elbow]

free gpu for one session daily

[half cove]

Or use an alt

[proper shale]

give us that error log

[narrow maple]

like this?
i dont really understand what you mean with log
do you mean the link?

[proper shale]

are you using cloudflare url?

oh wait that might just be how the old ila rvc is after not being fixed

Try this one:

https://colab.research.google.com/drive/1mHKTGH5e3SAyDSBss1KtiYRbDdQzwSMs

[narrow maple]

thanks, i was just using the linked one in the earlier page

[hoary acorn]

Someone help me in Mx 5 vc pls

[narrow maple]

idk if im doing right here-

[jade adder]

yo how do i use the ai voice

voice ai says i need to play credits

[jade adder]

nvm

i got it

[vague niche]

is there a online resource for rvc tts?

i tried to get it running on mangio but i couldnt find it

[proud elbow]

it's not TTS, it's voice conversion, but u can do convert any TTS output with an RVC model

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[white mulch]

wait so what should be in it 😭

[pure sandal]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[steady nova]

"can't find zip file" while zip file is there

[rare goblet]

Ayo? @steady nova level 1 !!!

[brittle wing]

wait how do you do it

[echo pecan]

-help

[azure marsh]

Command has expired

[rare goblet]

Ayo? @echo pecan level 1 !!!

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[umbral raven]

hey

does it normally take long for the client to load?

its been on the cmd window for a while now

[low shard]

That looks like a wokada realtime voice changer problem, pls ask in #1192011222023950368 or #🔍│help-w-okada as its not rvc

[inland zephyr]

Hey guys by any chance i can use this to talk on call on my phone ?

[umbral raven]

not really

[inland zephyr]

hmm i see

[proper shale]

yeah no

it wouldn't run on phone sadly

[inland zephyr]

some weird things i can do to create a chain or interlink between both pc and mobile

?

[unique hull]

i am opning for the first time and yeah i'm still waiting

[halcyon rain]

I was trying to download the RVC on this page, but when I did, I got a message "Open file with..." and I selected Google. Then I couldn't unzip it anymore, and when I open the file, all it does is download itself again. I don't know if I'm explaining myself correctly

[proper shale]

yeah no, you'd have to use PC only

[proper shale]

are you on phone?

[halcyon rain]

No, I'm on PC

[proper shale]

and what fork r u trying to download

[halcyon rain]

Sorry, what was a fork?

Sorry for the unknowing, practically new here 😅

[proper shale]

I'm just asking the specific rvc program you were trying to download

[proper shale]

It's alright :p

[sacred urchin]

Hi, I have problem with launch this
error:
D:\RVC1006Nvidia>runtime\python.exe gui_v1.py
'runtime\python.exe' is not recognized as an internal or external command,
operable program or batch file.

D:\RVC1006Nvidia>pause
Press any key to continue . . .

[proper shale]

do any of the folders before "RVC1006Nvidia" have spaces in them? this could cause that

[fluid topaz]

open go-realtime-gui.bat

[halcyon rain]

The one on the huggingface page that is linked in the Installing & Opening section, in the Mainline section of the document you passed before

[rare goblet]

Ayo? @halcyon rain level 1 !!!

[halcyon rain]

This document

[sacred urchin]

Program files etc?

[sacred urchin]

I know mate

[proper shale]

the folder names specifically, if they have spaces, remove em

[proper shale]

go realtime is for realtime*

if that's what you're going for...

[sacred urchin]

wtf

[proper shale]

Try to install 7zip, find the file and open with that

[unique hull]

oh no nvm

[rare goblet]

Ayo? @unique hull level 1 !!!

[turbid pagoda]

Is anyone on here a kits AI user. And run into an issue while trying to train and it just keeps telling you to try again because servers are busy?? I've been hitting the train button for like an hour. I'm paying 9 dollars a month for it to not work lol

[rare goblet]

Ayo? @turbid pagoda level 5 !!!

[fluid topaz]

no idea

[sacred urchin]

Still not working

[rare goblet]

Ayo? @sacred urchin level 1 !!!

[proper shale]

yeah I'm sorry but we really don't recommend kits here... colab pro would probably be a better idea tbh

[fluid topaz]

seems like you are missing the runtime folder, either re extract the .zip or redownload it, looks corrupted

[sacred urchin]

do you know what his name is?

[fluid topaz]

name?
RVC1006Nvidia.zip

[sacred urchin]

runtime folder I mean

[fluid topaz]

its runtime

check if its exist there, inside your rvc folder

[sacred urchin]

I have

[fluid topaz]

what about python.exe inside that runtime folder

[coral garnet]

Guys why doesnt my VB-cable activate on discord?

I cant upload files?

[fluid topaz]

set your default input into vbcable output

[coral garnet]

@fluid topaz

add me

gonna dm you

[fluid topaz]

aight

[sacred urchin]

Haven't, only python39.zip, poython39.pth, python.cat, pyexpat.pyd

• directories and files

[fluid topaz]

i think your antivirus deleted it somehow

[sacred urchin]

It's have sense, I see that folder in packed file and have python.exe

[fluid topaz]

yeah copy the one in packed to your runtime folder

after that try starting the gui

[sacred urchin]

I will

[sacred urchin]

working, ty

[unique hull]

nevermind, my files have spaces in them so it didn't work

[brittle wing]

got those 2 errors any ideas how to fix them?

[brittle wing]

can anyone dm me or ping me the latest rvc link of google collab

[rare goblet]

Ayo? @brittle wing level 1 !!!

[steady nova]

how do I create download link to it to connect it with ai cover song app?

I need download link of it like on #1175430844685484042

[proven hill]

voicechanger voice tutorial NOW NOW NOW!!!!!!!!!!!!

[unique hull]

hey guys
if i trained an Ai model, how exactly do i use it for text-to-speech ?

[glad berry]

i have the voice changer but it doesnt change my voice

[hushed grail]

can someone help

cant run it bc its stuck in 'initializing'

is it a wifi issue or am i doing smth wrong?

[brittle wing]

is there a tutorial on how to use the applio colab?

[pastel oak]

wokada?

[sinful nexus]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

More commands: -audio, -local, -overtrain, -realtime, -rvc, -help

[brittle wing]

yo do you know how to get credits for free in voice ai?

[low shard]

We aren't affiliated with voice.ai, we use rvc and wokada which are all free and open source instead

[brittle wing]

Guys is there a caddicarus model somewhere cuz i can't find one in the voice channel.

[scenic creek]

NameError: name '_private' is not defined

on last step

[molten iris]

hi I would like to know why rvc use my processor and not my gpu even when I change the settings (I have 6700xt and download the amd version)

[hushed grail]

yes

[rare goblet]

Ayo? @hushed grail level 1 !!!

[hushed grail]

(mb its the other one)

[brittle wing]

Does anyone know?

[pastel oak]

so wokada or not

if yes delete model_dir and pretrain folders

[brittle wing]

ahh okay

[rare goblet]

Ayo? @brittle wing level 1 !!!

[hushed grail]

it is

mb

[pastel oak]

do this and launch again

[scenic creek]

''so-vits-svc 4.0'' topic in help rvc

help plz

[brittle wing]

where do i download rvc

nvm found it

so,how do i make models in it?

[pastel oak]

ah that one is not for making models

try this https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/

[low shard]

check our docs https://docs.aihub.wtf

[scenic creek]

where i can find normal working rvc. i just get many troubles with them

[rare goblet]

Ayo? @scenic creek level 1 !!!

[low shard]

did you follow a yt tut or our docs ?

check our docs https://docs.aihub.wtf

[fluid matrix]

who told you to try and train on a mac 😭

use a colab instead

https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/

this one for example

you should only run inference on a mac

i did

check the link

and follow the guide

you can resume it

just read the thing in full before you do anything, all you need to know is there

[tame mica]

can u not

[frank flare]

hey where can i download rvc?

[tame mica]

-rvc

[azure marsh]

Suggestion for @frank flare

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[tame mica]

https://docs.aihub.wtf/

[frank flare]

thank you

[brittle wing]

Guys is there a caddicarus model somewhere cuz i can't find one in the voice channel.

[brittle wing]

There is none i both sadly.

[rare goblet]

Ayo? @brittle wing level 1 !!!

[fluid matrix]

if cant find it in both places, then it probably doesnt exist yet

https://www.weights.gg/models/clm730deb1btycctc3cmjr2th

but i just found it.....

you didnt search for it lol

[brittle wing]

Oh wait my bad.

[fluid matrix]

wait

shit

that model is zipped horribly and old

it might not work great

but it has the index and pth, so just try it out

[brittle wing]

Ok,if it doesn’t work is there any other ones?

[fluid matrix]

nah it should work dont worry

also no thats the only one unfortunately

[brittle wing]

Oh ok.

[nova trench]

Hello! Sorry, it's me again. I've tried RVC for the first time now, but it seems that I'm getting this error whenever I try converting. What do I do?

[wispy burrow]

What error is displayed in the cmd?

[drifting notch]

Anyone know how to reduce clipping and stuff ?

[cinder compass]

yo anyone know a google colab that works for the models? the other ones i’ve used say error

[narrow maple]

that means the colab is outdated
i had the issue yesterday

hold on im linking you the one im using rn..

[proud elbow]

use 32-bit float WAV while processing the audio. if you have to convert to PCM WAV or FLAC, normalize below 0 dB first.

[narrow maple]

https://colab.research.google.com/drive/16LkwvFZeudTpUOsE_6bMjOq2qkxFo8Hr?usp=sharing#scrollTo=9yQA9h6N2YV7

[rare goblet]

Ayo? @narrow maple level 2 !!!

[narrow maple]

@cinder compass

[vapid gust]

what pretrain model do i use?
i have a 1.20min dataset very clean

[narrow maple]

i wanted to ask how are sites called that remove instrumental?
im asking cuz normally i get my acapellas from youtube, but rn theres no acapella on youtube of the song

[brittle wing]

Rvc gui says running in system python what do i do?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[low shard]

rvc gui? are you watching an old yt tut?

[narrow maple]

i was yesterday ngl-

like i landed on the outdated google colab cuz of a youtube tut

[brittle wing]

What do i use that is easy to set up then?

[unique hull]

hey, how can i stop training my model ?

[low shard]

check our docs instead https://docs.aihub.wtf

yt tuts are old

[brittle wing]

Oh ok.

[brittle wing]

do I need a graphics card to use these softwares ? or is it possible to use them without it ?
since I don't have a desktop neither a nividia nor amd graphics cards

[rare goblet]

Ayo? @brittle wing level 1 !!!

[drifting notch]

dummy terms pls

i dont know wth that is lmao

[elfin bay]

Stuck in this part, it won't detect my gpu, it's built in with my computer btw

[proper shale]

if you don't have a dedicated nvidia gpu you can't train locally

sorry

[elfin bay]

Dam

Fat L

[proper shale]

that's just how PyTorch and stuff works so

¯_(ツ)_/¯

u can use colab tho

check https://docs.aihub.wtf/ :)

[elfin bay]

I have been using collab, it's just the limit sucks lmao

But ig I gotta work with what I have💀

[proper shale]

Yeah

[elfin bay]

Do you know how long that limit is?

[proper shale]

4 hours

[elfin bay]

Dam

Hey thank tho!

Thanks*

[proper shale]

yw :)

[ionic roost]

Yo, is there a software I can use to test what notes my model is hitting?

Whatever saved audio I have, I want to directly compare the notes and frequency to the real deal voice

[soft stratus]

@proper shale
Elevenlabs doesn't want to make voice

[proper shale]

you need to sign up, as the notification said

[soft stratus]

Ok

[brittle wing]

WHTA THE HELL IS THAT?!?

[queen snow]

uhm

I don’t know how to do this

Itches head

[brittle wing]

@thin stump

[honest lantern]

when i try to use rvc collab . i get "ERR_NGROK_6024" any help ? sorry for bad english .

[thin stump]

That's a collapse because the dataset has been damaged

[brittle wing]

Litsa the dancer told me to set the target sample rate to 32k and now I don't understand a thing

[thin stump]

What did you use for the batch size?

[brittle wing]

I Denoised the dataset

[brittle wing]

What is batch size...

[thin stump]

Could you dm me your dataset I could check it right now

[brittle wing]

Of course yes.

[mystic pumice]

does anyone have the repository of EasierGUI?

im getting this very same error but i found how to fix it

[rare goblet]

Ayo? @mystic pumice level 1 !!!

[mystic pumice]

i jsut need an old version

the scrip needs a deffined fuction that just isnt there

if i had to guess it was deleted by mistake during an update

because the other function is still there

okay i found it on another one of their projects

ha!

fixed it!

glad i learned basic python 3 years ago lol

can i just put files here?

[winged rampart]

yoo so im trying to make an ai cover and I have problem with step 2 of the "Easy GUI"(Google Colab) because it turns red when I try to put the voice model to the url:

can someone help

[proper shale]

this colab is outdated, check https://docs.aihub.wtf/ for newer colabs that work

[winged rampart]

thanks

is there any guide for it or?

[proper shale]

the docs have guides for each colab they list... so yeah

[frigid finch]

I'm using the Ilaria RVC and have restarted the colab space like 4 times now but still getting the red error message, model port & audio file are all loaded and working but it still won't generate the output. is there anything I'm missing?

I've tried a couple different voice models as well but no dice

[brittle wing]

The models won't export bru

[rare goblet]

Ayo? @brittle wing level 1 !!!

[frigid finch]

You're not supposed to have the model pth in the bit where it says model??

I probably just used the wrong word

[brittle wing]

Nah when i click export voice model it doesn't go nowhere

[proper shale]

what are you using?

[brittle wing]

How do i fix this?

[proper shale]

check cmd

command line

[sonic vortex]

ghvuihvgh

[brittle wing]

how?

[proper shale]

oh yeah ur on colab

[brittle wing]

mhm

[proper shale]

just check the colab page then, see what the output says

[brittle wing]

the only error i see

[proper shale]

i mean... is w-okada working?

if it is u shouldn't mind the "frequent errors occur" stuff

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[hexed plinth]

Does anyone know where the folder location (rvc) is located?I want to uninstall the app

[coral osprey]

this happens to me too lol

[rustic locust]

I need help again

It's not letting me download the _e30 wieght

[rare goblet]

Ayo? @sand island level 1 !!!

[rustic locust]

it acts like it's downloading and then doesn't

and the _e45 one won't download either

[sand island]

may have gotten my last ask for help comedically taken down since i wanted to link I got the local download link for windows off of the dev updates, but was asking if anybody was suffering the same issue that I'm suffering with how in Ilaria RVC after you use the assistant to download more pretrains the entire training tab breaks and the training buttons disappear and you can't type in the model name section too

[marble forge]

clicking train feature index does nothing

[rustic locust]

NVM i fixed the download issue. I had downloads for that site blocked...

[brittle wing]

Rvc 2

[crystal agate]

can i have some help w the voice changer client it's not recording anything n the pitch for my voice will like fluctuate

[odd wind]

i need help pls!

my voice changer sounds croaky

not clear

they dont sound bad when i play the video

but when i speak with it it sounds very robotic

[haughty moss]

Model architecture not defined faceing issue

[rare goblet]

Ayo? @haughty moss level 4 !!!

[crystal agate]

.

[rare goblet]

Ayo? @crystal agate level 1 !!!

[willow shell]

My question is not RVC related but I cant use the usermodel command

It simply says "the application didnt respond"

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[spare idol]

getting this error

---

ModuleNotFoundError Traceback (most recent call last)
<ipython-input-4-1af50e2c3de6> in <cell line: 4>()
2 #@markdown Link the URL path to the model (Mega, Drive, etc.) and start the code
3
----> 4 from mega import Mega
5 import os
6 import shutil

ModuleNotFoundError: No module named 'mega'

---

NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.

To view examples of installing some common dependencies, click the
"Open Examples" button below.

[cerulean phoenix]

Hello, everybody I have new project for fine tune Typhoon LLM model
Let me know if you have any experience in fine tune LLM

[marble forge]

hello?!?!

[nova trench]

Hello! I hope it isn't a bother, but may I ask for recommended settings for this? ;; I'm having a hard time experimenting since each coversion takes about 20 minutes, and I'm not quite sure where to start

[rare goblet]

Ayo? @brittle wing level 3 !!!

[nova trench]

Thank you so much!! I'll give this a spin o7

[oak lintel]

@west shuttle could you give me your settings for that model that you made? I've been trying to figure it out

[cobalt trout]

im new on making covers

[balmy trout]

what is the best model to separate stems

like best best, the slowest

[exotic valley]

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
chex 0.1.86 requires numpy>=1.24.1, but you have numpy 1.23.5 which is incompatible.
pandas-stubs 2.0.3.230814 requires numpy>=1.25.0; python_version >= "3.9", but you have numpy 1.23.5 which is incompatible.
torchtext 0.17.1 requires torch==2.2.1, but you have torch 2.0.1 which is incompatible.
torchvision 0.17.1+cu121 requires torch==2.2.1, but you have torch 2.0.1 which is incompatible.

I am getting this error while installing dependencies on colab

[north wasp]

Guys I am kinda dumbfounded. Can I ask here for support of voice ai aswell?

[balmy trout]

whats the difference between these

[trim halo]

What was the website that was modeled after politicians?

[thin stump]

Model makers use BS Roformer now. There should be a beta version on it on UVR except pic related is an mvsep exclusive

[rare goblet]

Ayo? @thin stump level 22 !!!

[trim halo]

I think it was called something like weights.gg.

[upper furnace]

I'll explain, I downloaded a local version of Gptsovits but when I try to convert the text I wrote into an audio file it tells me this, what do I do?

[balmy trout]

isee

does it separate stems?

[atomic zephyr]

fuck what does this mean

[thin stump]

It separates the vocal and the instrumentals. But if you're asking for separate parts (drums, bass etc.) , that's more of a band lab thing

[atomic zephyr]

oop wait, nvm its alright

[upper furnace]

?

[thin stump]

Use Eemposter's no webui version. Look around for the link

[upper furnace]

Oh, do you need another file to get it running?

[finite galleon]

When did you download that?

I checked t2s_model.py on the github and there wasnt a line 507

[upper furnace]

Yesterday

[upper furnace]

Why?

[finite galleon]

Making sure it wasnt outdated

Did you input the audio reference?

[upper furnace]

a

[upper furnace]

yes

(I'm downloading the better version from AIHub.)

[finite galleon]

Ss me your webui

[upper furnace]

Ah, I deleted it to download a better one😅

[finite galleon]

Which?

[upper furnace]

Now, with the recent version, what do I do?

@finite galleon

[latent kettle]

is there any one colab notebook in which i can use UVR [for both isolation of vocals and deEcoreverb]

[upper furnace]

Help

[mortal cipher]

yo so im tryna make custom AI carti songs over my beat how would I go of doing that
do i need to record my own vocals then put carti on it?

[rare goblet]

Ayo? @mortal cipher level 1 !!!

[mortal cipher]

hell ye

[upper furnace]

@finite galleon

[lament furnace]

Would 32-bit float be better or use 64?

[proud elbow]

32-bit is more commonly used by most audio softwares and has enough fidelity

[lament furnace]

So using the default PCM format is not recommended?

[proud elbow]

peaks above 0 db will be clipped, unlike the float one

also has much more dynamic range

16-bit is -96 dB noise floor
24-bit is -144 dB noise floor

32-bit float is far more range

[lament furnace]

What about the audio files we downloaded, don't they have their own wav type by then?

Wouldn't processing them into 32 bit be bad or

is that not how it works?

[proud elbow]

for example youtube download?
I used yt-dlp with converted format flac
the downloaded flac will be 24-bit pcm but typically enough as long it doesnt have much clipped samples, and then u can convert it to 32-bit float wav for further processing

[lament furnace]

I should use yt-dlp instead if so

Thank you for your inputt

[finite galleon]

Lemme see your full webui

[rare goblet]

Ayo? @finite galleon level 37 !!!

[finite galleon]

Move your cmd out of the ss

Oh wait nvm

[finite galleon]

Try testing the pretrained gpt and sovits

[atomic mesa]

where do I put the voice models that I downloaded

[lean owl]

anyone else getting this issue on duplicated spaces?

mainly on private spaces

[fluid sail]

soo im trying to put a voice zip file into voice.ai and its too big, even tho im pretty sure its meant for it, any ideas

its a valorant sage file and its not working cus its too big for voice.ai

pls help

[rare goblet]

Ayo? @fluid sail level 1 !!!

[fluid sail]

not you arcane

you cant help

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[mortal cipher]

im tryna make custom AI carti songs over my beat how would I go of doing that
do i need to record my own vocals then put carti on it?

someone gon see it eventually😭

[brazen elm]

Coming from Futurepedia and could not figure out how to make a new RVC. Please help me by making a Sole (Anticon Rapper) voice model.

[sinful sedge]

is there a rvc model for african rebel voice?

like deep voice

[mortal cipher]

aight thanks

[odd shale]

Nope. Ig you can make one yourself or request it on #1159289738314919936 channel

[odd shale]

We got guides you could use.

[odd shale]

Of course you need to record your own vocals.

[mortal cipher]

my fault i was high as hell😂

[odd shale]

Altho if you want a good Carti model, you could make it yourself using studio sessions of him that are leaked on YT

[mortal cipher]

aight thanks

[odd shale]

Umm, nop. Instead you would have to check the Tensorboard g/total graph.

Because you can never know how much epochs are enough for a model.

Also use 8 on batch size.

[odd shale]

We got guides you can use.

https://docs.aihub.wtf/

Like this one

[odd shale]

The epoch is how many times your dataset is trained i think.

Ig you should check the weights folder.

[rare goblet]

Ayo? @earnest jasper level 7 !!!

[earnest jasper]

what is this error

training doesn't starts

just loads and completes the 'training' column

tried on different alt accounts too

same problem

[onyx rampart]

anyone got any advice on how to sort out vocal impurities in song? because it can be very mixed with outputs

this is the separated vocals for quality reference

oh it's mutiple voices

that would explain

done just figured it out, but I have no idea how to fix that still

actually I'm surprised it was able to be so consistent with multiple layers of vocals

I'm just figuring out how to seperate backing vocals

[cyan patio]

In #info, under the "working forks", the first one in the local section, Applio, the link doesn't work, and hasn't for ages

[onyx rampart]

oh

[cyan patio]

Dont know where to report this, the section said to go here

[onyx rampart]

https://github.com/IAHispano/Applio

This?

I did find it on github by the looks of it

[cyan patio]

https://github.com/IAHispano/Applio-RVC-Fork/releases/tag/v2.0.0

That is the broken link

[onyx rampart]

Yeah it's the same guy, he just renamed his repo

[cyan patio]

All I'm doing its telling you so you can fix the link, trying to help #1192011222023950368

[onyx rampart]

I can't fix the link dude

I'm just some random member

¯_(ツ)_/¯

[cyan patio]

No worries m8

[glad zealot]

what pretrain are you using? people have said ov2 seems buggy atm if you are using it

[west shuttle]

hi, i used v2, 48k, RMVPE, 1 hop lenght (i'm using applio 3.09 for this one, new versions doesn't allow this and set auto to 128), then the standard settings. Actually that doesn't make a big difference in the final model, the most important part is still the dataset

[oak lintel]

Does it depends on my mic mostly cuz it's a good mic but the sounds and delays are pretty bad

Delayed, buzzing and low sounds

It's like messy in general

[west shuttle]

hahaha mine too

That's why i use all the filters in okada

[oak lintel]

Does it work fine for you?

[odd shale]

It's buggy because the current version is a test version

[west shuttle]

Not as good as i would like, because my voice sounds bad. I want to buy new mic but there is no money and all my models are for free haha

[earnest jasper]

it's OV2 only

😭

i'll wait for it to get fix 😅

[oak lintel]

I have a pretty good mic(hyperx quadcast) can i have your settings and try it out?

[odd shale]

OV2 won't get fixed till a new polished version gets released.

[west shuttle]

I have a 5 years old 20$ mic, there are no settings, just the filters okada has haha

[onyx rampart]

I have heard applio twice now, how come it's popular?

[glad zealot]

can you set it to 48k on the top cell?

idk

[onyx rampart]

I have been using mangio rvc

Does it yield better results than mangio?

Or is it just for training models only??

[trail violet]

how do i remove static noises or make it less AI

[rare goblet]

Ayo? @trail violet level 1 !!!

[oak lintel]

😭 even the creator itself

I want to make it more clear, no buzzing and stuff

[west shuttle]

Haha but i'm happy, because if my models sound that good with a cheap mic, then they can sound even better with a good one... Or that's what i want to believe

[west shuttle]

Mmm... I don't really know, Nvidia broadcast maybe? (I don't like it, but it's the last resource)

[fluid sail]

soo im trying to put a voice zip file into voice.ai and its too big, even tho im pretty sure its meant for it, any ideas
its a valorant sage file and its not working cus its too big for voice.ai
pls help

[low shard]

maybe you could check https://docs.aihub.wtf/rvc/resources/vocal-isolation/

[oak lintel]

I'm using the VB cable, hyperx (a really good mic) yet it still sounds pretty bad

[low shard]

ty for reporting that

[onyx rampart]

this has sent me to a weird looking webpage?

NEVERMIND IT DIDN'T LOAD RIGHT, ignore me

[low shard]

lmao dw

[earnest jasper]

my audio's 32k

idk what you are asking for 😭

[west shuttle]

I use voicemeeter as virtual mic and i recommend Maono Pd200X or the Fiffine K688, are cheaper and (for me) way better (Maono wins against Fiffine too)

[glad zealot]

ye just change the thing cell above to 48k

cause weord

¯_(ツ)_/¯

[low shard]

also fixed

[frozen abyss]

someone that knows how to open the goggle colab thing? for the ai covers youtube tut will be good

[dire shadow]

hi i have this error, how can i fix this? "is not an accepted origin. (further occurrences of this error will be logged with level INFO)"

[oak lintel]

Can you record a clip of you using it 😭 cuz my mic is good but it sounds so ass

[glad zealot]

what are you using

[dire shadow]

what you mean?

[rare goblet]

Ayo? @dire shadow level 1 !!!

[glad zealot]

what program

[dire shadow]

i am using colab

[glad zealot]

yes what is it?

[dire shadow]

Real time voice changer

[glad zealot]

what link are you using?

[dire shadow]

this https://colab.research.google.com/?

[glad zealot]

no the link to the voice changer

[earnest jasper]

how to do that 😅

[west shuttle]

The Maono? I don't have it, i tested it with a friend sorry. I u mean the voicemeeter, it's the same as vb cable

[glad zealot]

u know what lemme send ya something in dm

[dire shadow]

https://colab.research.google.com/github/w-okada/voice-changer/blob/master/Realtime_Voice_Changer_on_Colab.ipynb#scrollTo=lLWQuUd7WW9U

[dire shadow]

this?

[glad zealot]

https://colab.research.google.com/github/hinabl/voice-changer-colab/blob/master/Hina_Modified_Realtime_Voice_Changer_on_Colab.ipynb

use this

[dire shadow]

thats the same?

[oak lintel]

😭 sounds really bad though

[rare goblet]

Ayo? @oak lintel level 3 !!!

[glad zealot]

yes you can ignore the upload model part tho since thats broken

[dire shadow]

okay thank you so much but i cant do it now i will try it later

[brittle wing]

Heya guys, is there a way where I can know how much epochs is recommended when training the ai model?

[glad zealot]

there no real way to know what around 200-300 epochs is a good spot on most datasets

[merry fossil]

hey..
im new and wanted to ask if i can choose any of this to install?
https://discord.com/channels/1159260121998827560/1159620272316502096

[brittle wing]

Do the number of epochs depend on how long or clean the dataset is?

[plush compass]

With 4060TI 8 RAM, there will be no local problems, right?

[glad zealot]

longer means it trains longer so it should only need less epochs
cleaner dataset means it doesnt need to train as much but i still recomed it being 200-300

[brittle wing]

Okay. Thanks for the help!

[proper shale]

It varies a lot, but it does seem to increase depending on how long the dataset is

Honestly the best method is just checking the graphs, don't attach yourself too much to epoch values.

[eager siren]

is this right?

[proper shale]

ig just put the number of the gpu to 0, then you should be good to go :)

[eager siren]

I don't know how to use pretrain in local rvc, how do I do it? which file to use?

[proper shale]

ig dont use custom pretrains

just use the normal ones

they should be in the "pretrained" folder, just copy their respective paths

[brittle wing]

i cant read jap so

[rare goblet]

Ayo? @brittle wing level 2 !!!

[brittle wing]

how to install RVC

[proper shale]

https://docs.aihub.wtf

or if u want w-okada: https://rentry.co/VoiceChangerGuide

[brittle wing]

this one issnt specifi ill try bot tho

[proper shale]

if u want the realtime one then check the latter

[brittle wing]

i see okay

but wouldnt the none real time one be better?

[proper shale]

it depends on what u want tbf

but yes, technically, a non realtime one would be better in quality bcuz it doesnt need to be rushed xd

[brittle wing]

yeah i only have a rx 580 gpu

and a ryzen 9 7850x3d

[proper shale]

...oh

well uh

bad news for you

amd and ai don't go well together

[brittle wing]

cant afford nvidea

[proper shale]

i mean... you have a ryzen 9

[brittle wing]

i bought it like 9 months ago

with a motherboard

then got fired

[proper shale]

oof

[brittle wing]

so i never got to finish it

until i moved to my old house

and i just put it in my old pc

all i had to buy was ram

and reuse the rest of the parts

[proper shale]

shiit

thats tough man

either way, sadly RVC uses stuff that only NVIDIA GPUs support.... and no conversion for AMD GPUs exist (unless you want to use linux and ROCm)

[dawn nebula]

i think for now use colab

[proper shale]

^

nvm apparently rocm doesnt work well on 5xx cards

yeah colab is ur only option

[proud elbow]

non realtime inference using cpu is slow but fine tho

[odd wind]

can i use text to speech instead of voice???

[rare goblet]

Ayo? @odd wind level 1 !!!

[wispy burrow]

This was the very first time I was able to train with a Batch Size of 4.

After testing out Applio 3.10, I can't open Ilaria-RVC anymore...
It worked before but now I get this error:
ImportError: cannot import name 'pretrain_helper' from 'tools' (C:\Users\User\Documents\AI\Ilaria_RVC_Beta_1\runtime\lib\site-packages\tools_init_.py)

[eager siren]

???

[proper shale]

it identified ur gpu though?

huh

[hushed geyser]

hi, i need help fix Index training error on RVC v2 Disconnected. /content/Mangio-RVC-Fork/logs/experiment_gura/3_feature768

Exception Traceback (most recent call last)
<ipython-input-24-6c93ec206f26> in <cell line: 30>()
29 listdir_res = list(os.listdir(feature_dir))
30 if len(listdir_res) == 0:
---> 31 raise Exception("No features exist for this model yet. Did you run Feature Extraction?")
32
33 try:

Exception: No features exist for this model yet. Did you run Feature Extraction?

help me fix this 😢

[proper shale]

i think you screwed up on formatting the dataset

make sure that:

1. dataset folder name is the same as the experiment_name in the Set Training Variables step.

2. you're not compressing the files as .rar. this causes issues.

3. in case you did all of those, it could probably be an audio issue, try a re-export in Audacity as a .wav with all audios you used.

[wispy burrow]

You can't train your Feature Index if you didn't do step 2b: Extract Feature

[proper shale]

and yeah that too, but it's most likely what i've pointed out

[hushed geyser]

okay. thanks

[eager siren]

When training, it does not identify the GPU

[proper shale]

but then later on it gives you the gpu number

just make sure you've set the gpu number to 0

[wispy burrow]

But it did when you launched Applio?

[eager siren]

yeah

[rare goblet]

Ayo? @eager siren level 3 !!!

[eager siren]

however, in version 3.0.8 still.. updated version 3.1.1 does not work here

😭

I'll try again here

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[tidal cairn]

can someone link me with newest illaria rvc collab for inference?

[proper shale]

https://colab.research.google.com/drive/1mHKTGH5e3SAyDSBss1KtiYRbDdQzwSMs

this is beta tho so

[tidal cairn]

whats's the most recent collab for inferences, this one seems like it's for something else @proper shale

[proper shale]

this is for inference too

just has a training module

[tidal cairn]

okay thank you so much

[proper shale]

yw

:)

[odd shale]

Hello MJ.

I missed you.

[proper shale]

hi leo

[odd shale]

[proper shale]

tutl

[odd shale]

Tootling and eating lettuce

While making models.

[proper shale]

based

[odd shale]

Ur based too

[proper shale]

thx :)

[odd shale]

🐢 ❤️

[atomic zephyr]

what does this mean exactly??
like, what IS the console? the webpage?

[proper shale]

this is for local stuff

[dry oak]

do i habe to buy the vm sub for custom modules

[proper shale]

?

[atomic zephyr]

yeah im doing it locally bcs why not

[proper shale]

so the command line is the console

[atomic zephyr]

..,,
and is that, the webpage??...,,

[dry oak]

on voice mod when im trying to upload this servers modules

[proper shale]

aka command prompt

aka this

[atomic zephyr]

oh real as fuck

well i closed that shit, so i guess it's not working

[proper shale]

dont think voicemod supports those models

[atomic zephyr]

does this mean tensor is done downloading

[rare goblet]

Ayo? @atomic zephyr level 2 !!!

[atomic zephyr]

i guess so

[proper shale]

yeah

[serene bluff]

anyone know why my voice takes so long to load

[atomic zephyr]

i pressed ctrl c and nothing happened

[serene bluff]

when i press start and speak in takes like 30 sec to talk

[atomic zephyr]

then i pressed it three times and it asked if i wanted to terminate the thing
said no / n, and it just closed on its onw

:(

[proper shale]

well that's the thing u press to quit so

[proper shale]

do you have an amd gpu?

[serene bluff]

yep

is that a problem?

[proper shale]

do you know if you're using the right version for ur gpu?

the amd version has four buttons to select the gpu

[serene bluff]

im think i am can i dm u

[proper shale]

sure

[atomic zephyr]

where the run selection at :(

[brittle wing]

how do i use amd gpu

[proper shale]

on what

[brittle wing]

on rvc

[proper shale]

can u give a full screenshot of tb real quick

[atomic zephyr]

yah i got you

[proper shale]

try the refresh button real quick

[proper shale]

hm, local forks don't work with amd gpus, afaik

[atomic zephyr]

like, the the orange one or the normal refresh page

[brittle wing]

fuck

is there any way to use it?

[proper shale]

the orange one yea

[brittle wing]

or the one that support it

[proper shale]

see, problem is

nothing supports amd

[mighty mirage]

what does RVC V2 mean? is that something I select here?

[brittle wing]

mf

[proper shale]

RVC is the model architecture

[mighty mirage]

Oh so do I just select RVC still?

[proper shale]

yeah just keep as rvc

[mighty mirage]

I see, Thank you!

[proper shale]

yw

[brittle wing]

gotta spend million to generate 1 ai cover fr

[proper shale]

colab:

https://tenor.com/view/transitions-kinemaster-black-guy-suit-tiktok-gif-25279479

[brittle wing]

i gotta wait along time and it can shut down randomly

[proper shale]

actually you could still run locally tbf

it would just take longer to make the ai cover

because cpu

[brittle wing]

true

[proper shale]

outdated colabs do that yeah cuz webui restrictions

[atomic zephyr]

its not doing anything..,,

[proper shale]

so idk what u prefer tbf

[proper shale]

are you training yet?

i think it only shows info when training

[atomic zephyr]

im not sure lol, im just trying to figure this out for the first time

[proper shale]

yeah you need to train first then

and then itll show u the graphs

[atomic zephyr]

do i train it on this

[proper shale]

yup!

[atomic zephyr]

cool

now i just have to start it back up oh my god

[atomic zephyr]

can i like, dm you for help about this LMAO

[proper shale]

I'd rather keep this here, cuz like... people can search for their issues later yk

and maybe find this part useful

who knows

[atomic zephyr]

yeah

well then

what EXACTLY do i do with this..,,,,

[proper shale]

just insert the path where the dataset folder is on

u can right click and copy path on windows iirc

[atomic zephyr]

ok, jesus, i am STUPID
is it this?

[proper shale]

no, it's the folder that has the audios u want to train a model from

it can be outside of the rvc folder tho

[atomic zephyr]

.,, oh shit where the fuck do i put the audio

..,,

[rare goblet]

Ayo? @atomic zephyr level 3 !!!

[atomic zephyr]

thank you arcane

[wispy burrow]

Regular RVC does not have a specific location for your dataset.
Simply create a folder that contains your audio files.
The folder can be anywhere.
You can have it outside the RVCNvidia1006 folder.
Then copy the path of the folder with your voice clips.
And paste it in that field.

[atomic zephyr]

can i just name it anything

like "base_audios" or somethiong

[wispy burrow]

You can.

[atomic zephyr]

awesome

and the audio file is the audio that consists of like, the voice/thing that I want to make the model of, right?

[wispy burrow]

Yes.

Just make sure you are using a normal folder and not a zip.

[atomic zephyr]

yeppp

[atomic zephyr]

so, just like this? then what do i do for that singer ID thing?

[wispy burrow]

Do you have RVC and that folder on OneDrive? I think that could cause some issues.
Leave speaker/singer ID on 0.

@atomic zephyr

[rotund cairn]

i got a ? why does my rvc voices like make weird sound after i talk

[turbid pagoda]

Is there an easy GUI like the in the past where the path and index files populate after training and I can just download them like a regular output. I've tried the rvc and applio but I can't find the index files and the path files wouldn't download to my drive after a read the directions. I need the simplest one there is where I don't have to go digging for the output and just download them.

[wispy burrow]

Also, if you don't know this yet, there is a really great guide.
Here.
https://docs.aihub.wtf/rvc/local/mainline/#4-select-dataset

[rotund cairn]

i got a ? why does my rvc voices like make weird sound after i talk

[rare goblet]

Ayo? @rotund cairn level 1 !!!

[atomic zephyr]

oh thank you, also, no it’s not really on the one drive thing, it just kind of defaults to that

[turbid pagoda]

I'm talking like an easy GUI where I can download the index and path files to my computer itself not just the drive

[analog fern]

step 2 aint downloading the model

JSONDecodeError Traceback (most recent call last)

<ipython-input-32-4979bd4e0ff3> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:

5 frames

/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

[tame atlas]

My discord crashes when I turn on ai voice
Does anyone know how to solve this problem?

[dawn nebula]

what software u using

[dawn nebula]

is the model broken

[analog fern]

it turns out the model probably is broken

[rare goblet]

Ayo? @analog fern level 1 !!!

[analog fern]

according to other users on here

i will try ilaria rvc

[tame atlas]

Realtime Voice Changer Client v.1.5.3.18a onnxgpu-cuda

[rare goblet]

Ayo? @tame atlas level 1 !!!

[dawn nebula]

weird error

[tame atlas]

there is no error, discord just restarts when i turn on ai voice

[dawn nebula]

discord has hardware acceleration

probs turn that off

[analog fern]

works guys!!

[tame atlas]

tried it, didn't help

[dawn nebula]

great to hear that

[dawn nebula]

idk then i have limited knowledge on okada

[bold mantle]

anyone here

i need help with smth

i got a voice and downloaded it some guy said there would be a voice cable or smth thing i dont see any

also does voice ai work for it bcuz its free and im broke as hell

[proper shale]

we aint helping you to catfish ppl

[bold mantle]

ok

[proper shale]

but just know the info's out there

[bold mantle]

ok

where

[odd shale]

Just use public models posted on voice-models channel bruh.

We ain't going to help you to catfish or scam

Also why you want to catfish someone?

[bold mantle]

bcuzzzzzzz

hes my friend

[rare gale]

help me please. When loading the voice model, the couple writes an error immediately.. the screenshot cannot be sent here..0(
Here he writes in red: # File exists, proceed with creation of creds and client
JSONDecodeError: Expecting value: line 1 column 1 (char 0)

[idle ermine]

On the Gui step 2 whenever I paste any link i’ve tried many models it’s give me a error? help

[idle ermine]

Sorry, can you elaborate please thank you.

[rare goblet]

Ayo? @idle ermine level 1 !!!

[idle ermine]

@rare gale

[thin stump]

• you could post images now

[proper shale]

Using outdated colab, check https://docs.aihub.wtf/ for newer ones and guides too

[proper shale]

same thing @idle ermine

[idle ermine]

Rn im trying a diffrent Gui

I belive I was in outdated one

[eager siren]

3 hours and I didn't even reach 100 epoch....it's not worth doing it in a local RVC 🫠

gtx 1660 super doesn't seem to be doing well

[rare gale]

I did it through Google easy GUI..

[low shard]

yea and its outdated

use the ones in our docs instead

it doesnt matter if the colab you found its by google or not, its outdated lol

[wispy burrow]

Not surprised. It's a GTX...
Out of curiosity, how long did each Epoch take?

[young plume]

mine keeps crashing its weird

[wispy burrow]

We are going to need a bit more info.
What are you using?
What were you trying to do?
What was the error?
...etc.

[fading drum]

so if i accidentally stopped training that was already done, and clicking "train feature index" DOESNT make an index, how do i GET the index

[eager siren]

3 minutes?

[fading drum]

i was training a voice, then the voice finished training and then i accidentally restarted my pc before i could finish the training, what do i need to change and set in order to train the index file?

[fading drum]

yeah thats normal for me too

[fading drum]

you can reduce the time by keeping the cmd window in the foreground

or changing the pythons priority and affinity in task manager

[vast fractal]

Does a GeForce rtx 3070 work for rmvpe training, because every time I try training a voice on rmvpe my computer over heats, and it all becomes laggy and unusable

[fading drum]

thats what happens

its doing alot of calculations, you need high vram

[wispy burrow]

Man, that's still faster than mine...
I usually sit between 4 minutes to 4 minutes 20 seconds...
But I have GTX 1060 6GB...

[vast fractal]

But it overheats and shuts down

[fading drum]

how many gb of vram do you have?

[vast fractal]

Like 14

[fading drum]

what cooling do you have? i have like 2 extra fans, 1 pointing to my gpu, and one pointing away

[vast fractal]

6 fans on top and on the side and liquid cooling

[fading drum]

huh

[lament estuary]

whats the voice everyone uses?

[fading drum]

maybe do this

[fading drum]

literally any voice at all

[lament estuary]

like the e girl

[fading drum]

thats the magic of rvc

[lament estuary]

catfish voice

[fading drum]

look up rvc e girl voice

[lament estuary]

where?

[fading drum]

on weights.gg

[lament estuary]

ok

[amber zephyr]

what are these new pretrained models for?

the last 2

[wispy burrow]

Use the exact same Model Name and RVC version as your previous training session.
Then Train/generate Index.

[fading drum]

also make sure to turn on the "cache all training sets to GPU memory"

[fading drum]

welll, im guessing italia is for an italian sounding voice

[fading drum]

what app are you using?

[amber zephyr]

its a website rvc disconnected. i was getting ready to train some models before i stumbled upon this

[thin stump]

the last one is for russian

[amber zephyr]

ok and what is rin again i forgot?

[fading drum]

the last one is for russian apparently

[fading drum]

do i need to use the pre-trained base model g/d path thing?

[thin stump]

Sensitive to noise and needs to be a clean studio recording. Ov2 wants clean audio too except it's only for short datasets (6 mins or less)

[vast fractal]

What rvc fork do you recommend

[rare goblet]

Ayo? @vast fractal level 2 !!!

[amber zephyr]

ok

[atomic zephyr]

i thought i just needed it to be an audio file...,, not a xml source file or whatever..,,, :((

[wispy burrow]

No. Thatshouldn't matter for your INDEX.

[atomic zephyr]

i wish i can just ASK someone to make the model FOR ME but i dont wanna PAY so i have to do it MYSELFFF

[wispy burrow]

I'm not following...

[fading drum]

ummm is it an nvidia?

[atomic zephyr]

im also not getting that "copy as path" option...,,

[rare goblet]

Ayo? @fading drum level 4 !!!

[fading drum]

you can just go to the navbar in explorer, copy that then add (filename) and add quotes on either end of the whole link

[atomic zephyr]

the link with the guide had this as an example, so im thinking "aw it needs to be this, huh"
but i dunno lols

[atomic zephyr]

can you provide an example

i am very stupid

[fading drum]

so like "C:Music\Dookie\fart.mp3"

[fading drum]

folder, folder, folder, item

[atomic zephyr]

LMAOO okk

[fading drum]

welcome to the club

[fading drum]

do you have 3rd party explorer plugins? like winrar, or winzip? peazip? or a file converter

[atomic zephyr]

i have winrar, yet

[fading drum]

because they usually break some windows explorer things

[atomic zephyr]

yes*

also heree

like this

question marl

mark*

[fading drum]

maybe look up a support ticket or smth online

[atomic zephyr]

for what...,,

[fading drum]

yeah, that should work, i personally wouldnt use onedrive because it will have to download to your pc anyway

[fading drum]

for the "no copy as path" thing

[atomic zephyr]

yeah but it kind of just defaults to that so I have nothing else to use

[wispy burrow]

remove the ""
And add the file extension at the end.

[fading drum]

the "" is important, and it doesnt need the extension

so then it will use all the audio files in the folder

[atomic zephyr]

i was about to say lmao

[atomic zephyr]

theres only one wav too

that's the exact file name too

[fading drum]

oohh wait, is tha to the specific item? i meant only to the folder

[wispy burrow]

I though that was the audio file and not a folder.

[atomic zephyr]

ohh

[fading drum]

but if youre doing the file you need the file extension

[fading drum]

nah mb man

[atomic zephyr]

yeah the folder is "base_audios" and the wav is "HOUR LONG THANG"

ok lemme take that out then

there we go

i think

lets continue

[fading drum]

yeahp

[wispy burrow]

Should work.

[rare goblet]

Ayo? @wispy burrow level 22 !!!

[atomic zephyr]

no fucking way.

what's that video again
"im gonna break my monitour - im really gonna do it"

[fading drum]

whats your gpu?

[atomic zephyr]

i genuinely have no idea
i got this PC from my brother, so i have NO IDEA about ANY of the specs.

[rare goblet]

Ayo? @atomic zephyr level 4 !!!

[fading drum]

ok. go to device manager, and open up display adapters

[fading drum]

are there 2 gpus?

[atomic zephyr]

i..,, i have no idea

lamo

[fading drum]

the the device manager?

[atomic zephyr]

this?

[fading drum]

bigger ss pls

[fading drum]

intel gpus are not supported

[atomic zephyr]

ending my life oh my god

[fading drum]

hanng on

is that ALL thats there

[wispy burrow]

That is a GPU, but is there an NVIDIA as well?

[atomic zephyr]

yes lmao

[atomic zephyr]

uhhh

[fading drum]

are you able to open your pc

is it a laptop?

[atomic zephyr]

how do i check for that..,,

[atomic zephyr]

like pop a side of the case or smth