✨│ai-help

[flint vapor]

I know that he/she tried also on colab but can’t listen same

[sand iris]

huhcat

[brittle wing]

should i reset my mac ? 😂

[sand iris]

Nah not worth it x)

sourced this from the python package's website... sounds like it should've fixed everything but it somehow didn't

[forest sluice]

So it pinpoints to either:

• The software is broken (not very likely)
• Wrong mic selection
• Permission shenanigans

[lucid crow]

no.

[sand iris]

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/issues/1718
user here had the same issue

[lucid crow]

it is correct

[brittle wing]

hmm

[lucid crow]

can anyone help meeeeeeeeeeee

[sand iris]

A slow machine will not be able to run voice changer

no way to avoid that

...you can run voice-changer on google servers though

[lucid crow]

doesnt work

[sand iris]

this doesn't ? #🔍│help-w-okada message

[brittle wing]

so now what? bye-bye?

[sand iris]

Trying to see if I can find anything

any signs of this failing or did it run properly btw ?

[brittle wing]

properly run

[sand iris]

fuck

[manic sapphire]

Can i use this with potato laptop

[brittle wing]

pls see dm

[sand iris]

Depends on how potato it is

[manic sapphire]

I hope i can use it

[sick sluice]

is there an rvc program better than this one

than the japanese one

[sand iris]

Which one are you using?

They're pretty much all the same nowadays, just some have some things added for convenience

[inland coral]

Hi,I have a sound recording of 1.42 minutes, I made 300 epochs and it turned out to be a very different sound, like a monster voice :) can you make the sound file for me or help me?

[halcyon oxide]

can anyone help me use ilaryia rvc?I dont know how to use different ai voices

[brittle wing]

my friend is facing this issue in collab because of this he is unable to use the colab voice changer can somebody help

[rare goblet]

Ayo? @brittle wing level 2 !!!

[sand iris]

which link is he using?

for colab

[brittle wing]

the github link

[sand iris]

oh wrong one then

[sand iris]

#🔍│help-w-okada message

Up to date one is there

[jaunty parcel]

Does anyone know thats going on with Google colab? None of my models will load for step 2
It says JSONDecodeError

[half cove]

Don't use the old kalomaze easygui colab just use ila rvc (inference only) or the new easygui colab

[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

[lament eagle]

Is there any UVR colab

[jaunty parcel]

I'm currently using easygui colab and haven't had any issues with downloading the models when I did it last week. So I'm wondering if anyone is having issues at the moment

[olive crow]

@proper shale

[proper shale]

you should delete the /content/dataset/ folder path part, then do preprocessing and feature extraction again

ig that's it

[olive crow]

do i just delete it and leave it on blank

@proper shale

[proper shale]

leave the path part blank ye

[olive crow]

ok

still does the same thing

@proper shale

[proper shale]

huh

probably just an easy GUI moment then, idk

wait wait you're using a t4

what

are you on Colab

[olive crow]

yeah google colab

[proper shale]

yeah if you don't have colab pro easyGUI's Gradio part will disconnect the colab runtime

[olive crow]

is there any alternetivs?

[proper shale]

id recommend using RVC Disconnected;

https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS

if you want the guides: https://aihubdocs.github.io/en

[half cove]

Send me the link of it

Your prob using an old version

[jaunty parcel]

https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=F6zbsuNs6xZF

[half cove]

thats a old version

[jaunty parcel]

do you happen to have the link to the new one?

[rare goblet]

Ayo? @jaunty parcel level 1 !!!

[half cove]

-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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[jaunty parcel]

ty ❤️

[inland coral]

How can I download stable diffusion?

[brittle wing]

Is this good for training?

Like, it doesn't give an error, or it disconnects.

[odd shale]

I don't know, sorry.

There should be SD collabs.

[inland coral]

So, if I send a dataset, can anyone make it into an audio file for me? I need it so much I can't do it :(

please help

[flint vapor]

Is not already an audio file?

[inland coral]

In the sense that I have a sound, it is a file that I can use in rvc to clone it.

[flint vapor]

If you have a clean dataset you can already train

[inland coral]

Even though I did a lot it didn't work

[flint vapor]

which format is the file audio

[inland coral]

waw

[rare goblet]

Ayo? @inland coral level 2 !!!

[flint vapor]

If you have any request for a model you can do in #1159289738314919936

Wav is good

[inland coral]

thanks

[flint vapor]

You are welcome 😄

[night relic]

Hello

[flint vapor]

Hi!

[night relic]

I want a google collab notebook for upscaling and enhancing videos to crispy 4k quality using ESRGAN. It should have realesrgan-x4plus-anime and 4X_DigitalFrames_2.5 models to upscale my videos

[flint vapor]

This is rvc help channel so unfortunately i don't know how to help well about upscale video, you can ask on #🔍│help-ai-art because they are more expert about upscaling, and ai images and video

[night relic]

Ok I'll ask on that

[rare goblet]

Ayo? @night relic level 1 !!!

[flint vapor]

Ok 😄

[night relic]

Ok thanks

[half cove]

#🔍│help-ai-art

oh shit kakao already sent it

[flint vapor]

[scarlet inlet]

i need help , my volume to change my voice is always on 0 , is it normal?

" vol: 0 "

" but : 0ms "

" res : 0ms "

[flint vapor]

Okada right?

[scarlet inlet]

hmmmmm , idk what's okada but i have " real time voice changer "

[flint vapor]

Ok yes is okada

[scarlet inlet]

oh okayy

[flint vapor]

Are you sure you put your default microphone in input options?

[scarlet inlet]

yea

[brittle wing]

hey guys, can someone tell i we can train models on mac m1 using mangio rvc fork?

[flint vapor]

Have you click start? Can i see your okada settings?

[scarlet inlet]

yess !!*

of coursez

[rare goblet]

Ayo? @scarlet inlet level 1 !!!

[scarlet inlet]

but how

[brittle wing]

.

[flint vapor]

Just a screen

[scarlet inlet]

i can show you

okay

and paste it here?

[flint vapor]

Yes 😄

[flint vapor]

https://docs.google.com/document/d/1KKKE7hoyGXMw-Lg0JWx16R8xz3OfxADjwEYJTqzDO1k/edit#heading=h.8vqd8m4fh76q look for mac os help

[flint vapor]

On f0 detector put rmvpe

Wait, which gpu do you have?

[scarlet inlet]

euhmm where?

[scarlet inlet]

how ? in my system info?

[scarlet inlet]

i found it

its okay

[flint vapor]

You can see in task manager

[scarlet inlet]

but my gpu???

[scarlet inlet]

okay

[flint vapor]

On performance

[flint vapor]

Ok you have amd

[half cove]

click on gpu0

[flint vapor]

You downloaded the amd version of okada?

[half cove]

yep

that is the directml version

[flint vapor]

Yes true

[flint vapor]

from here

[scarlet inlet]

euhmmm i think no

[half cove]

if it has gpu dropdown its nvidia if it has the cpu gpu0 gpu1 etc its directml

[flint vapor]

Click on gpu0 so we see the name of gpu 😄

[flint vapor]

No no you have sorry, i didn't see under

[half cove]

also your chunk is suuuper high

[half cove]

and extra

[scarlet inlet]

i choose what number for chunk?

[half cove]

can you check the name of your gpu

doesnt show on the image

gets cut off

[hybrid coral]

Should I be using MVSEP to isolate background noise and artifacts for the future AI models I'll make from now on? (I'm a rvc ai model maker who makes ai models in mobile)

[half cove]

mvsep is good

[scarlet inlet]

gpu = graphic card???

[half cove]

yes

[flint vapor]

Yes

[scarlet inlet]

oh my bad

[scarlet inlet]

my graphic card is : AMD R9 290X

[rare goblet]

Ayo? @scarlet inlet level 2 !!!

[flint vapor]

How much of vram have?

4 or 8?

[scarlet inlet]

VRAM
4096 MB
GDDR5 1250 MHz

^^

[flint vapor]

Is a normal gpu, is not so good for okada but should work a bit, strange that you don't have volume

You can try fix your chunk

[flint vapor]

on f0 det put rmvpe onnx

[scarlet inlet]

yea

[scarlet inlet]

okay

ty

[gilded wigeon]

plz find brawlstars fangs model link😭😭

[flint vapor]

You are welcome 😄

[flint vapor]

This is not request channel

you can find on #🔍│find-models

[scarlet inlet]

i put chunk on 112 , its correct?

[flint vapor]

or request on https://discord.com/channels/1159260121998827560/1159289738314919936

[flint vapor]

You can try

[gilded wigeon]

how can find fang's model link

[rare goblet]

Ayo? @gilded wigeon level 1 !!!

[flint vapor]

https://aihubdocs.github.io/en/essentials/voice-models/ these are the website guides where you can find models

if the model don't exist you can request

for free or payment

https://www.weights.gg/it try search here

[flint vapor]

https://www.weights.gg/it/models/clpusn0is002r4nu69mf68ggz this is the model you searching

[gilded wigeon]

thanks bro👍👍

[flint vapor]

You are welcome

[scarlet inlet]

i tryed all , there have bugs and lags and not respond randomly.

[flint vapor]

You tried to reboot pc?

Sometimes the first installation can have some bugs

[scarlet inlet]

yea i rebooted it for installation of CABLE output to try something

[flint vapor]

I check if there is a common issue about this error wait

[scarlet inlet]

okay thxxxx

[flint vapor]

I checked here https://docs.google.com/document/d/e/2PACX-1vQIwJ3MVidhgEaXwWFl0xpVonVOVfneaNVADd7-NMWFgPIsfWWhG8NNqzQMsXDIOGlBIfxscoIm2_6I/pub but i don't find the error you have

You have check if your mic is working fine

?

[scarlet inlet]

euhmmm

[flint vapor]

On output put cable input

[scarlet inlet]

my mic???

[flint vapor]

Yes

[scarlet inlet]

yeaaa is working in discord

my mic

[scarlet inlet]

i need to on output cable input you say?

[flint vapor]

Yes

[scarlet inlet]

okay wait

[flint vapor]

Also on serverIo analyzer put cable input

[scarlet inlet]

i choose the server option , and after i'm gonna to choose output cable input ?

windows direct sound ou wasapi?

[flint vapor]

Wasapi no

[scarlet inlet]

okay

[flint vapor]

MME or windows directsound should be fine

You can try all if you see is not working

on monitor put your speaker if you want hear yourself

if you don't want hear yourself is ok

[scarlet inlet]

okay

[rare goblet]

Ayo? @scarlet inlet level 3 !!!

[flint vapor]

But should work also on client option the volume

[flint vapor]

You put so much low chunk

[scarlet inlet]

i tried all chunk numbers possible

[flint vapor]

increase a bit

[flint vapor]

put audio on client

[scarlet inlet]

okay

[flint vapor]

and serverIO Analyzer put cablee output

[gilded wigeon]

how can download model link?

[flint vapor]

I send you a guide

https://aihubdocs.github.io/en/essentials/voice-models/

[rare goblet]

Ayo? @flint vapor level 25 !!!

[scarlet inlet]

i said " hello " i wait 1 minute and i heared it but i tried again there no have sound

[scarlet inlet]

i did it

[flint vapor]

Ok from here you need to fix chunk and extra

[scarlet inlet]

okay

[flint vapor]

You need to play with them and see who is working better

[scarlet inlet]

okay i gonna try that

[flint vapor]

On tensorboard there isn't precise number of epochs

[brittle wing]

Anyone has a mangio-rv-fork colab guide? i wanna use colab to train models for FREE

[flint vapor]

https://aihubdocs.github.io/en/rvc/cloud/rvc-disconnected/

[brittle wing]

thanks

for training, is mangio-crepe good or rmvpe

[half cove]

rmvpe

[flint vapor]

Rmvpe

[brittle wing]

ok

[gritty tide]

hihi, I was following a tutorial video on this and I when I launch it I'm only able to use tiny_crepe and another rmvpe_onnx, I have a NVIDIA 4090 and I was wondering if I downloaded the wrong version as all the other ones are N/A, could somebody direct me to the correct version pls?

[molten pecan]

is that okada/voice changer?

if so, ask in #🔍│help-w-okada

[gritty tide]

ouh, my bad, sorry!

[molten pecan]

np =3

[formal pecan]

Users of the paid version of Google Colab do not need to reinstall every time?

[brittle wing]

how do i save my models and settings in colab

[formal pecan]

How do I fix these errors?

[brittle wing]

is it possible to remove like background noises in it T_T since i can hear the fan in it

[flint vapor]

Yes

Oh you deleted message , however yes when you have amd you need onnx

@brittle wing

[brittle wing]

i just installed okada

i have a really good pc but im on amd so

the voice is awful maybe u can help me ?

[rare goblet]

Ayo? @brittle wing level 1 !!!

[brittle wing]

im using an onnx voice ai

[flint vapor]

On f0 det put rmvpe onnx

and on audio use client not server

[brittle wing]

okay

chunk and extra are good ?

[flint vapor]

Depends from gpu

what gpu you have?

[brittle wing]

wtf the ms

[flint vapor]

is the latency

[brittle wing]

amd 5700 xt

so its really nice

[flint vapor]

You have 4gb of vram?

[brittle wing]

is it detected ?

[brittle wing]

hmm idk

[flint vapor]

It should work so so for okada, but should work

When the voice is awful is most because of chunk and extra

[brittle wing]

https://www.gigabyte.com/fr/Graphics-Card/GV-R57XTGAMING-OC-8GD-rev-10#kf

no its 8

[flint vapor]

Ok so is more better

[flint vapor]

Mostly of these problems is about these

[brittle wing]

the problem is the latency

[flint vapor]

Yes so is about chunk

[brittle wing]

res latency

[flint vapor]

You need to play with chunk and extra settings

and see what settings can fix a bit the voice

[flint vapor]

try also to enable some noises options like echo and sup2

[brittle wing]

maybe my gpu is not detected ???

[flint vapor]

No don' t worry the gpu is detected

[brittle wing]

do i stay on gpu0

[flint vapor]

because here is red

gpu 0 should be your amd 5700 xt

you can check on task manager also

[brittle wing]

when im using the basic voice its really fine

[flint vapor]

You converted the custom voice in onnx?

[rare goblet]

Ayo? @brittle wing level 2 !!!

[brittle wing]

well see

[flint vapor]

Ok wait that generate onnx, sorry but i never did a onnx conversion before because i have nvidia version

For any errors you can see if you encounter one of these problems here: https://docs.google.com/document/d/e/2PACX-1vQIwJ3MVidhgEaXwWFl0xpVonVOVfneaNVADd7-NMWFgPIsfWWhG8NNqzQMsXDIOGlBIfxscoIm2_6I/pub

is a guide with the most commons errors

[brittle wing]

is it good ,

[brittle wing]

did it

its really better

[flint vapor]

Yes but don't worry the serverio analyzer don't mean a lot

[flint vapor]

as output put cable input

[brittle wing]

index always to 1 ?

[brittle wing]

didn't installed it is it needed ?

[flint vapor]

For talk on discord and other apps yes

[flint vapor]

0.5 is good

[brittle wing]

do u use these ?

echo sup or anything ?

what is this ?

[flint vapor]

This is ok don't worry

[flint vapor]

try to enable echo

and sup2

[brittle wing]

okay I think im good

ty a lot

do u know other good place to dl voice ?

[flint vapor]

Weights.gg

[brittle wing]

its just that amd is really bad cause u always need to convert the voice etc

[flint vapor]

Or also here in the server

i recommend weights.gg

[brittle wing]

🙏

[flint vapor]

because there are models also before october 5

in october 5 we had another server but got deleted by discord

[flint vapor]

*before of

so inside weights.gg you can find also old models

[craggy jackal]

is google colabs broken?

[brittle wing]

its better if there is more epochs no ?

[flint vapor]

No don't worry

Not always if a model have more epoch than other is good

[craggy jackal]

is google colabs broken?

[rare goblet]

Ayo? @craggy jackal level 1 !!!

[brittle wing]

ok ill install vc cable

[flint vapor]

Ok 😄

[flint vapor]

Hi! Idk the situations about colab

but you can try different colabs if you encounter some error

-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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[latent kettle]

like what ?

[flint vapor]

You trying to get the model maker role?

If you have uploaded your model you need to wait that a QC will review your model and approve you the role

[latent kettle]

I'm just reading it .. but yes .. for future

[flint vapor]

Oh ok 😄

[latent kettle]

I mean.. image is missing

[flint vapor]

Maybe is some bug

When someone approve you you will receive a message in the #1159514067187277865

Rejected if you will be rejected

[latent kettle]

is it hard to get a model maker?

[flint vapor]

and accepted if you passed the check

[flint vapor]

No, only model master is more hard

[flint vapor]

But i think that now a bit yes

many models qc and type of tests are changed

so you need to do a good model for obtain the role

[latent kettle]

such as

what kind of test?

[half cove]

you just need one good model

[flint vapor]

Don't worry not a test with question

they do a test if the model sounds good

after you uploaded

and if all is ok you will obtain the role 😄

[half cove]

https://aihubdocs.github.io/en/extra/model-maker-role/

[flint vapor]

Like samples and other things

[half cove]

should have all the information

[flint vapor]

Here is the guide

[flint vapor]

@latent kettle

[tidal sky]

i cant hear the models anyone can help?

[flint vapor]

In what sense?

Where you can't hear?

[latent kettle]

what kind of test will qc do .. like how it check quality

[tidal sky]

like i dont hear the audio example of the models

[flint vapor]

They have some samples and they check if your model sounds good

[flint vapor]

on voice models or where?

[tidal sky]

yea here on the discord

[flint vapor]

You can't hear the sample provided by the model maker?

[tidal sky]

yes

[half cove]

listening to your audio and checking it using a spectrogram

[flint vapor]

It should be a discord problem

Because i can hear samples now

If you can try to download them

[tidal sky]

but i can hear everything on discord beside that

[flint vapor]

Try if you can download the sample

[tidal sky]

i did it but still cant hear them

[flint vapor]

Which model you want to listen sample?

[tidal sky]

im just checking if theres any models that i can use

[rare goblet]

Ayo? @tidal sky level 1 !!!

[tidal sky]

dont have a spesific one

[flint vapor]

You tried to refresh discord?

[tidal sky]

yes i did

[flint vapor]

Very strange, you are on mobile or pc?

[tidal sky]

pc

[flint vapor]

browser version or client?

[tidal sky]

client

[flint vapor]

Try if on browser version you can hear

[tidal sky]

okay gimme a min

[flint vapor]

If you need a specific sample and you can't listen tell me and i will download for you and send you

[tidal sky]

okay im on browser and it worked ty

[flint vapor]

You welcome 😄

[serene hull]

wtf is this? help me

[proper shale]

it's good

[latent kettle]

how can I become a model master?

[brittle wing]

which is the best and quickest epoch to make a singing model?

[flint vapor]

You need to have done at least 10 good models

and superate a quiz of 7 questions

[latent kettle]

ohh.. what kind of questions

[latent kettle]

where ? in #🔍│find-models ? or in #1159514067187277865 ?

[flint vapor]

Also in weights.gg

in get model maker is for obtain the model maker role

[brittle wing]

..

[flint vapor]

and in find model is for search models

[flint vapor]

There is not a precise one

[dusk glen]

what is epoch?

[latent kettle]

basically.. in weights.gg.. I have to create a profile and have to upload best quality models in it ?

[flint vapor]

Yes but is connected with your discord account

[latent kettle]

if you read a book for 1 time.. that is just like 1 epoch

[flint vapor]

Is an amount of time of a model like a train cycle that learn from the dataset you give

[latent kettle]

after that .. I have to send any application or something else?

[dusk glen]

hmm, anyone know how to fix annoying sound when using rvc

[rare goblet]

Ayo? @dusk glen level 1 !!!

[flint vapor]

Yes, but before you need model maker role

[flint vapor]

you have to do /jointeam

and apply for model master

[flint vapor]

also 10 hq models and reply to all the question in the quiz

[latent kettle]

okay .. I got it . thanks a lot .. I'll try if i could get it .. not for now.. but in future maybe 😅

[flint vapor]

You welcome 😄

[latent kettle]

one more question.. can I Train models on Mac studio m2 Max 2023

[flint vapor]

Yes i think

which version of rvc you using?

Mainline,Mangio or applio?

Don't worry there aren't stupid question

I send you a guide

https://aihubdocs.github.io/en/essentials/how-to-make-voice-models/

is all written here 😄

You can find all info on how make a model

Yes 😄

Depends from where you use rvc, and on which gpu

the time for make a model

*train

[latent kettle]

I don't use applio and maingo rvc.. and I don't know what is mainline

[flint vapor]

Depends from every gpu, are you using on colab?

[rare goblet]

Ayo? @brittle wing level 1 !!!

[flint vapor]

Are the versions of rvc

local versions

With a 3060 of 12gb and a dataset of 15 minutes to me require like 1 hour

No don't worry is not hard to set up 😄

You can find all info in the guide, also on which local version choose

[latent kettle]

which is best any easy to install/use on Mac studio m2 max 2023 ? can I Train on it.. if yes how .. any guidance

[flint vapor]

Idk well about videos, most info are written on the guide

[brittle wing]

Are there any acurate bpm detector websites?

[flint vapor]

Personally i use mangio and i know that mangio is also on mac, but idk if also applio is on mac

[brittle wing]

Are there any acurate bpm detector websites?

[sand iris]

don't use mangio on mac

[latent kettle]

how can I install it on Mac.. any guide

[sand iris]

pls

[flint vapor]

Yes wait i will send

[sand iris]

there isn't a whole lot of mac users on mac

[flint vapor]

No i don't have on mac haha

i have on windows

[sand iris]

if you need anything I can help, but the process is kind of a pain

[latent kettle]

if you device have sensors. then you can

[sand iris]

because the rvc devs wrote stupid shit, and apple doesn't handle things in a convenient way

[flint vapor]

I never tried on mac😭

[sand iris]

(turns out mangio's mac scripts are even more stupid ! yay !)

[flint vapor]

You just need a good clean dataset

around 15-20 minutes

also 10 sometimes

[brittle wing]

How?

[flint vapor]

dataset is the file audio you use for train your model

Yes it must be most clear at possible

[latent kettle]

like if you have samsung s10 it have a dedicated sensor on its backside.. you can use it with samsung health

[brittle wing]

i was talking about the bpm of a song

[latent kettle]

ayo.... 😆😆

[flint vapor]

You welcome 😄

[latent kettle]

yes. vocalremover.org

[flint vapor]

for clean dataset and remove vocals i recommend Ultimate Vocal Remover

[latent kettle]

he/she wants to check bpm of a song.. (online)

[flint vapor]

Oh ok ok 😄

my bad

[latent kettle]

what about RVC on mac

[flint vapor]

@sand iris says that he/she not recommend to use mangio on mac, i can send you the guide same if you want to read

[sand iris]

RVC on mac should be something along the lines of :
downloading the RVC code
running run.sh
....

but a whole lot of shit can go wrong

[flint vapor]

https://docs.google.com/document/d/1KKKE7hoyGXMw-Lg0JWx16R8xz3OfxADjwEYJTqzDO1k/edit#heading=h.8vqd8m4fh76q this is the guide but is full of bug the installation

[latent kettle]

see my reply above.. I thought he/she was talking about blood pressure 🤣🤣.

[sand iris]

if python >= 3.11 is installed, it borks
if there's a space in any folder's name, it borks
if brew isn't installed, it borks
if xcode isn't setup to compile c++, it borks
it borks randomly sometimes
it can't train models no matter what

[brittle wing]

Can anyone try to remaster this song? https://www.youtube.com/watch?v=9eroflaT9c4&ab_channel=LasVariedadesdeAndrés dm me for a remaster of the song

[flint vapor]

If you have any request you have to ask in request-model

[latent kettle]

so .. I think it is useless to use it on mac

[flint vapor]

You can use on colab instead

[sand iris]

if you want to convert voices from existing models, it's useful

but in all other cases... yeah

[brittle wing]

Im trying to remaster this song

[rare goblet]

Ayo? @brittle wing level 15 !!!

[latent kettle]

because of runtime disconnected and limited access i don't Want to use it

[flint vapor]

Yes but this channel is support for rvc

[flint vapor]

There is also paperspace but you need a trial

[brittle wing]

Any more ai services that can upscale the song??

[latent kettle]

paperspace .. Idk about this

[flint vapor]

Idk, rvc is not about upscale songs

[flint vapor]

Yes is a cloud like colab but for use you need a trial

is written on the guide

[latent kettle]

is it better than colab?

[flint vapor]

Gpu should be bit better

but as i said is on trial

Is not free

[latent kettle]

🤕

[flint vapor]

[half cove]

could use kaggle ig

[latent kettle]

BTW.. thanks for your support... 😊

[flint vapor]

You are welcome 😄

[latent kettle]

what is this ?

[brittle wing]

DONT INSTALL RVC ON MAC

I have myself experience... today only

[sand iris]

kek

[flint vapor]

Yes but calm down😭

[half cove]

https://www.kaggle.com/

[brittle wing]

i cant

[half cove]

need phone number

[sand iris]

He's been through the 7 stages of grief

[flint vapor]

Haha

[molten pecan]

Man we must make a Kaggle guide one day

Hopefully Hina updates the stuff constantly

[flint vapor]

He saw the hell

[half cove]

i should test kaggle i havent yet

[flint vapor]

What happened?

[brittle wing]

@sand iris knows

[brittle wing]

error after error

at the end, i asked google gemini. all started to work fine

then i wasnt able to train models.......

rest is history

[haughty flame]

I think this question belongs here:

Should I stop training after certain mode collapses, or should I just trust that it's not overtrained?

At this point, I've already had three mode collapses, but the g/total line still doesn't seem to be going up in the graph:

INFO:JacobCollier:Train Epoch: 160 [0%]
INFO:JacobCollier:[10176, 9.803199879555537e-05]
INFO:JacobCollier:loss_disc=3.324, loss_gen=4.239, loss_fm=15.370,loss_mel=24.370, loss_kl=1.109

This are the graphics:

[haughty schooner]

i get this error constantly 😦
Text length too long (200 < 4811), using segments: lanz.wav
but the trainingfile looks good, 61 sentences and the sample got sliced into 61 samples.
so i tried to generate configuration for training, which seemed to work as well. but when i start training, i get this one:
utf-8' codec can't decode byte 0x81 in position 38: invalid start byte
im using gradio and try to train a TTS model

[brittle wing]

[spiral osprey]

Can anyone help me resolve this?

[rare thorn]

-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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[rare goblet]

Ayo? @rare thorn level 2 !!!

[odd shale]

Share full screenshot

[brittle wing]

can i find a tutorial on how to use the RVC hugginface version

[spiral osprey]

i don't have.

[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

[brazen leaf]

Hello everyone

i am facing an error when i click on convert button

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 72.00 MiB (GPU 0; 2.00 GiB total capacity; 808.67 MiB already allocated; 0 bytes free; 1.14 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

can you guide me how i can solve this issue ?

https://github.com/Mangio621/Mangio-RVC-Fork/releases

that is the link where i download the RVC

[sand iris]

GPUs with 2GB of vram are hell to work with

you need to make sure as little of it is being used by other programs, else it doesn't work because it's running out of space

[brazen leaf]

i am trying to train any model, i am just try to clone simple voice

[rare goblet]

Ayo? @brazen leaf level 1 !!!

[sand iris]

oh training will not work

needs at the very least 6GB of VRAM

[brazen leaf]

i have 16GB ram

[sand iris]

GPU RAM not system

[brazen leaf]

oh got you Sir

how i can solve this issue please ?

I have installed fresh copy of Mangio-RVC-v23.7.0_INFER_TRAIN.bat in windows 10 and my system specification is Intel Core i7 5500U CPU @ 2.40GHz 2.40 GHz . and 16GB RAM and 64bit windows 10 and NVIDIA GPU 940M

[sand iris]

close as many applications as possible

and hope it will work

[brazen leaf]

now this time i did not using any other application. only CMD and Chrome Browser opened.

[brittle wing]

?

[haughty schooner]

Text length too long (200 < 4811), using segments: lanz.wav. i get this error constantly 😦
but the trainingfile looks good, 61 sentences and the sample got sliced into 61 samples.
so i tried to generate configuration for training, which seemed to work as well. but when i start training, i get this one:
utf-8' codec can't decode byte 0x81 in position 38: invalid start byte
im using gradio and try to train a TTS model

[brittle wing]

this is my first time training with this colab, last time i trained was before google nerfed the old og colab 😭

[spiral osprey]

pls someone help me.

[brazen leaf]

This is my system GPU information

[warm dune]

im facing the same error

[brazen leaf]

if i did not select any model just trying to change the pitch i got this error

NameError: name 'cpt' is not defined

[low shard]

i think its the old easygui colab, you can use the newer, which is https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw#scrollTo=G0MEhFM19Vq6, or look at other ways on how to inference (make ai covers) online in our docs: https://aihubdocs.github.io/en

[brittle wing]

also im using RVC disconnected

[spiral osprey]

and in relation to the rvc v2 disconnect which one will replace it?

[brazen leaf]

this is the screenshot

[low shard]

can you show also the input you putted and the name of the .zip that contains the dataset?

[low shard]

rvcdisconnected still works fine for training

[rare goblet]

Ayo? @brittle wing level 4 !!!

[warm dune]

where do we input the voice model?

[spiral osprey]

It doesn't work for me.

[brittle wing]

im re-running it rn

just to see if it works this time

[spiral osprey]

I just want to do a cover there, I don't want to train his voice here!

[low shard]

check out the docs, there are different colabs for making ai covers which explain where to input it

[low shard]

yea i gave you the links, in the new easygui colab you can both inference (make ai cover) and train, if you wanna make an ai cover scroll down, also it doesnt have ui

there are also other colabs (such as ilaria rvc which i personally use) for making ai covers with UI (interface)

[brittle wing]

yeah same error

re-ran it

idk whats going on lmfao

its found the dataset clearly

you can see each file there

[low shard]

did you upload the SiouxsieSoux.zip inside of the rvcdisconnected folder in your google drive?

[brittle wing]

yes

[spiral osprey]

This easy colab has an interface, I want to do vocals with an interface.

[brittle wing]

use ilaria

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

this one works fine for me

has a gui and everything

[brittle wing]

"No such file or directory: '/content/temp_dataset/D:/SiouxsieSioux' -> '/content/dataset/SiouxsieSioux'"

but its right there

[low shard]

heres also the guide for ilaria rvc https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/, also the new easygui colab doesnt have ui

[spiral osprey]

and the rvc v2 discconnected?

[low shard]

just use ilaria rvc

[low shard]

it doesnt have ui and its only for training

[brittle wing]

rvc disconnected is for training only

[spiral osprey]

Will it be off the air?

[low shard]

wdym off the air?

[spiral osprey]

yeah

[brittle wing]

[spiral osprey]

I want to train it too but rvc v2 disconnected is down. I can't train him there.

[haughty schooner]

Text length too long (200 < 4811), using segments: lanz.wav. i get this error constantly 😦
but the trainingfile looks good, 61 sentences and the sample got sliced into 61 samples.
so i tried to generate configuration for training, which seemed to work as well. but when i start training, i get this one:
utf-8' codec can't decode byte 0x81 in position 38: invalid start byte.
im using gradio and try to train a TTS model does anyone has some experience with this?

[low shard]

rvcdisconnected isnt down wdym

[low shard]

all of them are wavs right?

[brittle wing]

yeah

[low shard]

click on runtime, disconnect and delete runtime, and restart everything it prolly bugged

[low shard]

what you trying to do ?

[brittle wing]

#✨│ai-help message
i did that here

but ill give it one more go

[low shard]

yea but before you just re ran it, without deleting the runtime i guess, or did you?

[brittle wing]

i deleted the runtime

im very familiar with how colab works

i just havent used this colab specifically

[warm dune]

Thank you for this one

[spiral osprey]

I can't use it and even other people can't use it either.

[brittle wing]

are you having the same issue as me

[spiral osprey]

yeah.

[brittle wing]

huh.

[spiral osprey]

For you to see, I've been saying for a long time that he's been inactive. What I meant to say is that it's not working. they just don't understand me lol.

[haughty schooner]

i want to train a tts-model

[rare goblet]

Ayo? @haughty schooner level 3 !!!

[haughty schooner]

https://www.youtube.com/watch?v=p31Ax_A5VKA&t=890s this version specificaly.

"2024-02-25 19:39:56 | INFO | httpx | HTTP Request: POST http://127.0.0.1:7860/api/predict "HTTP/1.1 200 OK"
2024-02-25 19:39:56 | INFO | httpx | HTTP Request: POST http://127.0.0.1:7860/reset "HTTP/1.1 200 OK"
2024-02-25 19:39:56 | INFO | httpx | HTTP Request: POST http://127.0.0.1:7860/api/predict "HTTP/1.1 200 OK"
2024-02-25 19:39:56 | INFO | httpx | HTTP Request: POST http://127.0.0.1:7860/reset "HTTP/1.1 200 OK"
2024-02-25 19:41:28 | INFO | httpx | HTTP Request: POST http://127.0.0.1:7860/api/predict "HTTP/1.1 200 OK"
2024-02-25 19:41:28 | INFO | httpx | HTTP Request: POST http://127.0.0.1:7860/reset "HTTP/1.1 200 OK" this is something i also see at every command i give

[brittle wing]

alright well fuck it

disconnected doesnt work

ill try easygui

wait so i have to upload the full unzipped folder

or does it have to be one long continuous audio file

im confused

[low shard]

That looks like a rvc fork gui with tortoise tts but never seen that thing before, sorry but i don't know

[haughty schooner]

its exactly that, a rvc fork with tortoise tts 😄 its actually the only one i know. i produce tons of music, but discovered this AI-stuff quite recently. may i ask what do you use?

[low shard]

Usually i just generate a tts audio with edge tts with a premade model that is the same gender and language of the rvc model im gonna use later, and then use ilaria rvc (a google colab way to use rvc as i can't locally) to use the audio as an input, download and usenthe rvc model on it

[haughty schooner]

you need a special kind of graphics card for that, right?

gpu

[molten pecan]

found the problem. the dataset must not have spaces or special characters

[viral sand]

i want to download ai voice changer, theres many versions of MMVCS like every single one have diffrent size (798MB or 3GB), what is the difference between them? which one should i better download? @molten pecan

[molten pecan]

ah idk about the voice changer

I don't use it

ask in #🔍│help-w-okada

[brittle wing]

ah

well nvm im using easygui and training it there

[molten pecan]

you might get errors too if you don't rename the dataset

[brittle wing]

oh

well its training fine atm?

30 epochs atm

[half cove]

@low shard who else had the problem?

[molten pecan]

uhh what?

[molten pecan]

EasyGUI has been working correctly lately to my knowledge

[brittle wing]

yeah

[half cove]

he was using disconnected before

[brittle wing]

you said i might still get errors

yeah

now using easygui

and its working fine

[molten pecan]

btw remember that, if you want the best results, do not input a semi arbitrary total epochs amount

[molten pecan]

seems like easygui can handle datasets with spaces/special characters

didn't know that

[brittle wing]

for a 15 minute dataset what should i set it to then

[fluid latch]

i think its because your dataset folder was located in a folder named "D:" which is weird first of all and yea probably no likey symbols

[molten pecan]

https://aihubdocs.github.io/en/rvc-resources/epochs-overtraining--tensorboard/ here's everything you need to know

=3

[brittle wing]

cheers

oh well i know that nvm

[low shard]

I did it online in google colab , yes for rvc you need a GPU

[brittle wing]

wait

if i were to set it to like 800 epochs

but it begins overtraining at lets say, 500 epochs

would it make a difference if i just set it to 500 epochs from the beginning

or will i get the exact same result

[molten pecan]

nope. here's the thing

[low shard]

#✨│ai-help message

[molten pecan]

if you scroll up you'll see a value named "save frequency" or sum like that

[brittle wing]

ah well its fine then, i tend to just put something higher than i expect then find out where it starts overtraining

[half cove]

screenshot

[molten pecan]

that determines at how many epochs the model will be saved. so if you leave it at like 10, the model will get saved after the epoch 10, 20, 30 etc

[brittle wing]

usually checking by checking the time of overtraining on loss/g/total

and then going through logs and finding the epoch at that time

[molten pecan]

you should look after the step, not epoch

example

if the overtraining point was at the Step 600, get the checkpoint that was closest to before the overtraining point

the checkpoints will be saved with this format

modelname_epoch_step

example: kalomaze_e60_s120.pth

[brittle wing]

i know that

i mean like

it tells you what time it was saved

so lets say it starts overtraining at 4:15

ill go through the logs, find 4:15

and use that checkpoint

[molten pecan]

hm I see

[brittle wing]

that was one of the methods i was told a while back

[molten pecan]

basing off the time might get you lost

imo searching for the Step is more convenient

but if it works, go for it ig

[flint vapor]

https://tenor.com/view/goofy-run-chat-cat-hello-chat-gif-1980506103906038006

Oh here gif don't work

[brittle wing]

yeah i havent trained in ages, had to relearn some stuff

[rare goblet]

Ayo? @brittle wing level 5 !!!

[brittle wing]

i trained before google nerfed the og colab

the golden era 😔

[solar geode]

helloo, I am using easy gui but the files are not saved on my drive and I cannot resume the training

[proud elk]

Yo was wondering if anyone here could feed my curiosity on voice training. I have been using the mangio RVC web and its worked perfect, voice sounds quite natural. I was curious though, are there other training methods/programs out there that could potentially use some different algorithms to achieve different results? Purely curious.

[rare goblet]

Ayo? @proud elk level 2 !!!

[brittle wing]

so with Ilaria RVC we can only use models via huggingface?

[molten pecan]

The model link must be from either Google Drive or Hugging Face

[valid hamlet]

how can i find a voice model?

[half cove]

#1175430844685484042 or weights.gg

[molten pecan]

#1175430844685484042 or https://weights.gg/

[molten pecan]

Grill yourself poop

[valid hamlet]

can i find a really good voices?

[molten pecan]

I won

[molten pecan]

Yeah sure

[valid hamlet]

lemme see

can u send me a girl one?

[molten pecan]

If you want guaranteed quality, you can buy one from our model masters #1191429836321849435

[molten pecan]

Ilaria and MrM0dz have some great ones

Check their profile on #1191429836321849435

[valid hamlet]

lemme see sec

[molten pecan]

You didn't send the link

Embed failure

Therefore

Nah, I won

[thin stump]

Mines had the lowest dip at exactly 370 epochs at 15.88k steps since the saving interval was set to 10. Should I use the 15480k - 360 epochs, or 15910k - 370 epoch?

[molten pecan]

15910

[thin stump]

Thank you

[molten pecan]

yw

https://cdn.discordapp.com/emojis/1166312934096523265.gif?size=64

[neon mulch]

can some1 help me

#1211414024093302784 message

[odd shale]

Please be more specific.

[odd shale]

Please, use the #1159289738314919936 channel if you wanna request a model.

[rapid niche]

Does anyone know why the epochs will save to the logs folder on colab but not to my google drive?

[jovial galleon]

yo i accidentaly closed my gradio window while training how can i restore it

[dawn nebula]

if you didnt close your console it will continue training

but if you need to restart, just put the same exp name

and settings

then continue

[jovial galleon]

yeah i just wanna train feature index after

but ill try that

thank you

[dawn nebula]

yeah

all g

[rough cove]

I can't find where the echo is coming from or how to disable it

[brittle wing]

what echo

[rough cove]

after I installed the virtual audio cable I started getting a repeat of my mic

[odd shale]

Sup2 option.

Tick that checkmark on the W Okada gui.

[rough cove]

still there

[brittle wing]

when using easygui, to continue training do i run everything again except the index training and feature extraction?

[rough cove]

my dumbass, nvm solved it

[rare goblet]

Ayo? @rough cove level 1 !!!

[rough cove]

I swear I had my output to line1 instead of my headset

[brittle wing]

oh and ive got these in my bin randomly, am i safe to just delete these

[odd shale]

Yeah... For this reason you must NOT keep ticked both "save an checkpoint each epoch" and "save only latest checkpoint"

[brittle wing]

ah

so im okay to delete them then

[brittle wing]

also am i correct in saying this

[odd shale]

Or else the D and G files won't have it's proper numbers corresponding to each .pth file

[brittle wing]

ah

in that case ill turn that off when i continue training

[odd shale]

It's safe only if you finished training your model

[odd shale]

No, pay attention.

You must only keep ticked only one of these both options on the Gradio

[brittle wing]

im using easygui

theres no option for that

[odd shale]

F

[lime star]

Can someone please explain to me why this Google Collab no long works? It was my favorite one 🥺 https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7

[brittle wing]

ah

the OG

[void wedge]

ive been having problems with that too

it wont let me download modelsa

[brittle wing]

idk

but honestly, use ilaria

[tacit grail]

Im not sure why but im getting an error when trying to convert with Applio. I have the input audio and the index file selected but shbortly after pressing convert it says error

[void wedge]

where can i find that

[brittle wing]

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

[brittle wing]

?

[vapid zodiac]

i tried to use this one and when i try to download the model it says error

[left cosmos]

guys nobody is active in the w-okada anybody can help me with W-okada voice changer rn?

[rare goblet]

Ayo? @left cosmos level 1 !!!

[lime star]

Awesome thanks 🙂

[brittle wing]

a few workarounds that have worked for me:

1. Download the model, unzip it, zip it back up, upload it to ur google drive and then copy that link and use it
2. If its a huggingface link you can also try putting "?download=true" on the end of the link

[vapid zodiac]

its a huggingface link and even with the ?download=true it Still says error

[rancid mortar]

what is s thresh

[brittle wing]

damn

try the first thing then idk

[left cosmos]

BROO

someone help with w-okada

[brittle wing]

am i correct in saying this? i want to continue training but dont want to fuck anything up

[lean monolith]

1. Choose the model. what do i put here?

it didn`t as the model like here https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing

ask*

im speaking about Ilaria rvc

[rare goblet]

Ayo? @lean monolith level 1 !!!

[lean monolith]

nvm i was being stupid

[void wedge]

ok

so im trying ilara

i do not know how to

add models

in the fucking slightestr

and it wont let me do shit\

[rare goblet]

Ayo? @void wedge level 1 !!!

[void wedge]

HOW DO I DO ANYTHING

[lean monolith]

I KNOW

also thank you vap for saving my day, gd

[void wedge]

nvm i figured it out

[cinder wyvern]

keep getting this error when trying to import some huggingface models into the rvc colab, its never done this before

JSONDecodeError Traceback (most recent call last)
<ipython-input-6-d1791a7398f3> 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)

[half cove]

use ila rvc for inference

-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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[spare shale]

https://colab.research.google.com/github/poiqazwsx/RVC-Notebooks/blob/main/Ilaria_RVC_Tempfix.ipynb

@half cove Sorry for the ping!
How long will it work?

[half cove]

i think normal should be fixed but im not 100% sure

[spare shale]

I mean, How Long Will He Disconnect? 😅*

[half cove]

?

[spare shale]

How long does it take for colab to shut down Ilaria?

@~@

Sorry, I have a fever and my head hurts, I can’t write

[half cove]

like how long after opening it will it disconnect?

[spare shale]

Yes!

[half cove]

like 3h

[spare shale]

I mean it

[spare shale]

Thanks: 3

And sorry for my message

[half cove]

why sorry

[spare shale]

Because I didn't write my message very correctly

[half cove]

its fine

[spare shale]

милота!

[cinder wyvern]

i get an error when downloading the models on ilaria too

is everything just broken rn or something

[half cove]

give me your model link

download on ila rvc should be workin

[cinder wyvern]

https://huggingface.co/Geovenox/ATEEZ/resolve/main/hongjoong 1.zip?download=true

[thin stump]

what?

hugginface btw

[half cove]

that should work

does it error while downloading or infering?

[cinder wyvern]

downloading

[finite galleon]

Give us the full error that is shown on the colab

[brittle wing]

retraining on a seperate account

using easygui

[half cove]

does it show up on the inference tab tho?

[cinder wyvern]

no when i click refresh theres still nothing in the dropdown

[half cove]

fuck

i cant be bothered testing it rn

i have too many things

that im doin

[dim tree]

My application seems bugged. It won

[rare goblet]

Ayo? @dim tree level 1 !!!

[dim tree]

t show me my stuff </3

[spare shale]

#🔍│help-w-okada

[dim tree]

My apologies.

[lucid crow]

hello i cant hear myself and i cant see my gpu

can someone help

@spare shale

@spare shale

[spare shale]

@lucid crow Sorry but I can only help if it's colab

No need for ping!!

[lucid crow]

ok dm me.

[spare shale]

T-T

np*

[lucid crow]

sent u a dm

[spare shale]

Please stop pinging me

[odd shale]

It's really necessary to sent DMs to staffers for help?

You can just tell your issue here.

[odd shale]

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

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

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

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

__**BE PRODUCTIVE.**__ 🤝

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

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

[odd shale]

People doesn't even know how to explain RVC issues.

[abstract kestrel]

https://youtu.be/RSY9yEQS_B0?si=I5uRHaI1d8aOKmI9 What is the font of this thumbnail?

[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

[brittle wing]

Can anyone answer this question, is there a way for me to train a model and increase the epochs? I have a 250 epochs model, and I wanted to increase it.

[rancid mortar]

what does indux do

[frigid cypress]

can someone help me? i downloaded a voice from #1175430844685484042 to use in the mangio rvc web, put it in the weights folder, and refreshed the voice list however nothing is showing up. can someone help?

[spare shale]

What?

[half cove]

So you have the training files?

[half cove]

Accents

[half cove]

Screenshot

Your folder

[brittle wing]

I have the .pth and added

[frigid cypress]

cant send attachments

[half cove]

Pth in weights index in logs

[half cove]

Talk a bit and level up

[frigid cypress]

bru

[half cove]

Wait wrong person

You can't train with just those

[hot lagoon]

anyong know a way to dedecrypt swf files?

[gritty radish]

i trained my model and got the .pth file but idk where to get the index file from

[proud elbow]

de-decrypt means to re-encrypt it?
there should be some softwares to extract assets from .swf you can search, but many of the swfs especially made in 2000's era have 32 kbps mp3 compressed audio

[hot lagoon]

my trying to find one that dose extract the swf files but my not sure what to use to get those extract

[dawn nebula]

1. .pth files are not these

2. you have to train feature index to get .index files

[brittle wing]

Can someone help me to find more voice lines of this guy? (It's the ''announcer'' of Forza Horizon 5 Latin American Spanish dubbed)

[solemn dagger]

Hey guys, I'm trying to use the RIN_E custom pretrain but everytime I try to train I keep getting the following error: "IndexError: list of index out of range"

This was after I generated the index, and then it proceeds to generate the index again.

Each time I try to train it gives the same error and generates the index again.

I'm using Appolio V3.0.7 and I have the pretrain in the ApplioV3.0.7\rvc\pretraineds\pretraineds_custom

Is that where they're supposed to be? Any help is much appreciated.

[rare goblet]

Ayo? @solemn dagger level 1 !!!

[solemn dagger]

Model folders are full, and the paths are correct. Am I supposed to keep "pretrained" and "Custom Pretrained" on at the same time?

Because thats what I am doing rn.

Also heres what the error I was talking about

Yeah. This is only my second time making a model though, so idk if that helps

[dark spear]

I´m sorry to bother, but does anyone else besides me noticed that the Ilaria RVC google collab makes the models sound quite different from the clasic link that stopped working? They sound weird, lower quality...

[solemn dagger]

So thats what the 40k meant when I downloaded it... lol. Thank you Light. If you want anything from the fornite store hmu.

Yeah my dataset is in 48k. I'll export it as 40k and try again.

I thought you meant that the sample rate of my dataset matched what was selected in Applio. I should have read closer

[dark spear]

I thought the same, but I have used my models plenty of times before and I know how they sound and how to make them sound good, and with this link... Idk, maybe I´ll just wait and hope the other one gets fixed.

Are there any other alternative to use RVC v2 models in the cloud?

[tacit grail]

How do you fix voice cracks from AI?

what about readding vocal effects after voice conversion? sorry, im pretty sure i asked this already but i forgot what yall said

like reverb, echo, autotune, etc that an original audio from a song has that isnt the instrumentals

yea they have to be removeed before voice inference but what about after? How do I add them back with that characters sounding reverb and other effects... oh wait i remember now...

i just gotta learn how to mix vocals

what if its pretty clear already?

[twin heron]

Trying to upload a model from Huggingface but I keep getting this error: JSONDecodeError: Expecting value: line 1 column 1 (char 0)

This is the colab link I'm using: https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7

[feral saffron]

Me too

[tacit grail]

There should be a way on Applio to make it redo a specific part of a song without having to cut that part of the song and rerun it with that file then add it back

[twin heron]

What's wrong with this colab is it stupid?

[tacit grail]

Which Pitch Extraction Algorithm is the best on Applio? Is it a "it depends" sotuation?

So default, got it

[serene wigeon]

depends how its coded, it might trigger that error message inappropriately

or maybe it looks at the magic bytes of the file instead of the file extension

[twin heron]

Either way it doesn't seem like I can find a single properly working colab cloud

[serene wigeon]

applio has been doing well for me, but i understand the frustration that they are hit or miss

[twin heron]

Someone needs to update that one thread in #1159513888199540817 cause I notice most of those are misses

[static oar]

is there like a suggested gpu for it?

the program

[ruby wedge]

why do i keep geting a JSONDecodeError

[zenith crystal]

https://media.discordapp.net/attachments/1007057623717261326/1211553667757514782/image.png?ex=65ee9e47&is=65dc2947&hm=5bd954a057935cc7b16181501206e3279cc2a641bf0b9e8e021eeabc7f171ab4&=&format=webp&quality=lossless

[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

[royal jacinth]

Does the newest version of applio support AMD gpus? Because the old version used to allow me to select my amd gpu, but the new version does let me select my GPU. Is there a way to select it, or is it already selected, and if so, how can I see if it's using it and not the cpu?

[zenith crystal]

is there smth wrong with the google collab

i used it like a month ago but now im having problems

[ruby wedge]

yea smth like that

[zenith crystal]

how do i fix it

[ruby wedge]

idk

i have that too brother

[fossil kindle]

I'm havin trouble with that too 😓 guess I'll be waiting for help with ya

[zenith crystal]

😭

[rare goblet]

Ayo? @zenith crystal level 1 !!!

[fossil kindle]

https://tenor.com/view/persona-3-persona-3-reload-persona-reload-sees-gif-979312620180196338

[fossil kindle]

@zenith crystal @ruby wedge I might be making some progress I'll keep you updated

[zenith crystal]

bet

[serene hull]

is problem solved?

nah

[rare goblet]

Ayo? @serene hull level 2 !!!

[spare robin]

Anyone know if it's possible to change your default slider settings in RVC Applio for your Model Interface on startup? like the audio envolope, voice filtering, search feature ratio, etc.

[rare goblet]

Ayo? @spare robin level 2 !!!

[serene hull]

have somebody offline version of this? may be its should be working

[glad zealot]

Yes but you would have to edit the code yourself to change it

[spare robin]

When you say code do you mean modifying parameters in the bat file or json file or in python?

apologizes for my ignorance.

[fossil kindle]

@zenith crystal https://youtu.be/0nTfLpqfN1Q?si=EPdbGHjQyCRVaZdO

[rare goblet]

Ayo? @fossil kindle level 1 !!!

[glad zealot]

In the main python file

[spare robin]

gotcha

[indigo olive]

yoh

who can help me?

[gloomy palm]

hey guys i need help with where to place my downloaded models for rvc

[sand iris]

that's a very old link

-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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[sand iris]

RVC Disconnected and easyGUI are both for model training