✨│ai-help

[copper pebble]

cloud/faster

RVC v2

[hearty jasper]

Yup, but idk why I cant generate, says error after converting

[copper pebble]

how much was ur audio long

[hearty jasper]

Under 10 seconds

[copper pebble]

and did u tried duplicate the space??

[copper pebble]

hmm how many times did u tried?

[hearty jasper]

2 times for 10 second MP3, tried 1 time for a 2 minute mp3

Maybe I'll try next time

[static gyro]

heloo could you help me fix this?

[copper pebble]

you should use waw not mp3

[copper pebble]

what r u using bro??

seeems pretty old

[subtle rover]

anyone have the huggingspace link to make a cover?

training is finally done! thank u poopbert and everyone else for helping!!!

[static gyro]

https://colab.research.google.com/drive/1UUUVl4eVRp9WoN3aEMpwSs2EdlP0ji6C#scrollTo=w4wXvoez9Rce
i train from code

[copper pebble]

access denied

use rvc disconnected btw

its the new colab

[north halo]

what makes the voices sound bad? I messed around with the settings to make this one voice sound good but when I use it on a discord call it just sounds terrible.

[rare goblet]

Ayo? @north halo level 2 !!!

[subtle rover]

link to make a cover?

pls help

[static gyro]

https://colab.research.google.com/drive/1PorBmm4bUjY1-dk4f7XKrO7Mx51wbqYU

[copper pebble]

https://discord.com/channels/1159260121998827560/1167844377046040769

[subtle rover]

thank u!!

[static gyro]

what is that?

[fickle mason]

Okay so, When you run Dependencies, it automatically runs 'Download Pre-trained models'. It doesnt mention NOT downloading it in the RVC disconnected doc thing @copper pebble

[copper pebble]

the new traiing colab

[copper pebble]

ur right i got it wrong sorry

im helping to much people at once

while traiing

[copper pebble]

did u have the same error ??

[static gyro]

disconnect prevention looks like step one

what do i do?

[copper pebble]

https://discord.com/channels/1159260121998827560/1160926856548200531 heres a guide bro

[sturdy echo]

hi how can i take song from youtube and put my voice on it

[static gyro]

so now im in a loop.

[copper pebble]

https://discord.com/channels/1159260121998827560/1167844377046040769

[static gyro]

says to come here

[copper pebble]

lol

sure rvc disconnected is useful just for small dataset

[subtle rover]

should i share the google drive link of the training file

[static gyro]

so what would you suggest?

[rare goblet]

Ayo? @static gyro level 4 !!!

[static gyro]

skill issue

[copper pebble]

yh u need the download link of your model

drive,hf,mediafire.etc.

[subtle rover]

ok

[copper pebble]

using that

[static gyro]

because my datasets are as large as they need to be, some of them are 30 minutes so i can get it under 200 epoch

[copper pebble]

cuz what were u using has some banned codes for free users of colab

[copper pebble]

maybe u can do that

try

yh

[static gyro]

how to do this?

[subtle rover]

thx

[copper pebble]

1st u need to run "Depencies"

in the colab

and wait till he says "Connect drive"

[subtle rover]

whch one should i pick lol

i have no idea what any of these are

[fickle mason]

tehy need to up ur pay

I think i figured it out

i didnt change expierment_name to my model name

[copper pebble]

i was thinking about that

[copper pebble]

ahahhah

[copper pebble]

the ones that u named

[subtle rover]

none r there

these r random names

im confused

[copper pebble]

u sure that u downloaded the model good??

[subtle rover]

i added the gogle drive link of the folder

google

[copper pebble]

that are others one

[static gyro]

unable to load

what do i do?

[copper pebble]

just need the zipname

like "shit.zip"

[subtle rover]

am i just supposed to link the .pth file?

[rare goblet]

Ayo? @subtle rover level 6 !!!

[copper pebble]

refresh

[subtle rover]

i did

still random names

[copper pebble]

send me the link of your model pls

[copper pebble]

do you have the .pth and the .index in a zipped folder??

[subtle rover]

not zipped

jsut a regular folder

[minor zenith]

uhm, is RVC supposed to auto disconnect itself? if so is there a way i can disable that? i can't make anything when it disconnects in the middle of it

[subtle rover]

the one rvc exported

lemme zip it and upload

[copper pebble]

yh please

To everyone im about to get off in like 20 min

[copper pebble]

what r u using

[minor zenith]

RVC v2

[static gyro]

it disconnected after epoch 17

[copper pebble]

wth?

why

[barren swift]

that's the method, what are you running it on
local installation or google colab or something else

[copper pebble]

what did it say

[copper pebble]

THANKS

[minor zenith]

uhh google colab i think.

[static gyro]

still going but it disconected the gpu

no error message

[barren swift]

google colab disconnects you if you go afk for too long, but worry not for I have a solution

[copper pebble]

i think its supposed to be like that

[rare goblet]

Ayo? @copper pebble level 24 !!!

[copper pebble]

STFU

[copper pebble]

what colb r u using

[copper pebble]

its called "disconnected" after all

its like a stealth mode against google

[minor zenith]

uhhh...like the tunnel? (I'm sorry i'm confused)

[rare goblet]

Ayo? @minor zenith level 1 !!!

[barren swift]

what is the website called

that you are using google colab on

[copper pebble]

can u send me the link of it pls?

[minor zenith]

https://bf9d3a5fadfa1dc892.gradio.live this one ig?

[toxic scroll]

how to listen voice received with neuro

[copper pebble]

thats old

you can use this https://huggingface.co/spaces/TheStinger/Ilaria_RVC

[minor zenith]

oh- there's a new one? i haven't been keeping up with this stuff for a bit

[barren swift]

either that one or use RVC disconnected

https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS#scrollTo=Yj-_npuA_HRB

[copper pebble]

thats for training

[toxic scroll]

how to listen voice received with neuro?

[barren swift]

they didn't specify what they wanted to use it for

[copper pebble]

yh but this was used for covers

[minor zenith]

oh i'm not making any. I don't think i could unless i use Elevenlabs to make custom voices

[barren swift]

i think you could do it, google colab puts all the work on google's servers and not your computer

as long as you follow the instructions on the website it's not that hard

i've been training a model lately and as soon as I figure out how huggingface works ima get that model maker role

[copper pebble]

https://discord.com/channels/1159260121998827560/1167844377046040769 heres a simple guide for ilaria rvc (covers)

https://discord.com/channels/1159260121998827560/1160926856548200531 heres a guide for rvc disconnected (training (making models))

About to get off so dm me or ping me just for serious problems

bye

[quick flame]

yo can anyone give me the link to that google docs or whatever it was where you can feed the huggingface link and an mp3 file

no clue how to explain it further

[minor zenith]

iIaria's RVC keeps giving me errors..

[digital falcon]

can someone tell me the new version?

[rare goblet]

Ayo? @digital falcon level 1 !!!

[digital falcon]

of vc

??????

anyone

[wispy burrow]

Realtime/W-Okada?

[digital falcon]

what

[toxic scroll]

how to listen voice received with neuro?

[rare goblet]

Ayo? @toxic scroll level 1 !!!

[digital falcon]

No like i need the version like ex: "v.1.0.0.3"

[minor zenith]

is there any other versions of RVC that i can make covers with? Ilaria's RVC keeps giving me errors every time i try to convert

[wispy burrow]

Since you refer to it as VC, I assume you want a realtime Voice Changer..
Realtime-VC (RVC)? Or W-Okada?

[fickle mason]

im suing whoever said 0% epoch is normal

@copper pebble I forgot to say thanks for all the help < 3

[copper pebble]

no problem mate

[minor zenith]

if there isnt any at all then i'm just gonna go back to the google colab cause this isn't working at all for me

[wispy burrow]

AICoverGen maybe?

[digital falcon]

i need help on this problem

anyone?

[wispy burrow]

I knew it...That's W-Okada.
#🔍│help-w-okada

[static gyro]

im tryingExtracting model files:
In Google Drive, locate "rvcDisconnected* -> "ModelName" folder.

i see this what do i do?

it says it should be here but its not. what do i do?

[minor zenith]

okay, so...im a bit confused. i'm trying to upload the audio file in AICoverGen but nothing gets uploaded

[static gyro]

you still here to help?

[static gyro]

Please is anyone here?

[dense dirge]

someone here has the link to rvc disconected colab?

[near flame]

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

[quartz heart]

where

@violet heron

[static gyro]

can some one plese tell me how to extract model?

[violet heron]

You do it after

When you inference

Like make covers

Same exact thing with kits

You can change the pitch for the train

[static gyro]

can you help me?

[violet heron]

Try finding it on colab itself

Files button on the left

[static gyro]

ok.. where? thank you for helping btw

[violet heron]

Mangio RVC fork I think!??

[static gyro]

ahh. so where is the auto save to google drve?

[violet heron]

Ok

Oh your saving?

[static gyro]

yeah suposed to

[dense dirge]

that dsnt work for me

[rare goblet]

Ayo? @dense dirge level 4 !!!

[static gyro]

so it should go to my google drive right?

[dense dirge]

the rvc disconected xd

[violet heron]

I would wait for it to finish training

[violet heron]

You can using the cell below

[static gyro]

i did already once before

[violet heron]

Or you can directly download

[static gyro]

cell below what?

[violet heron]

Weights folder on there

3 dots

Download

[static gyro]

what cell do i press for saving?

you said cell below

[static gyro]

no function called "download"

[violet heron]

You can only select 1 file

[dense dirge]

help pls

[static gyro]

ok got it

[rare goblet]

Ayo? @static gyro level 5 !!!

[static gyro]

so this tutorial is wrong got it, please share

this one i mean

[autumn knoll]

Hello! What are the differences between RVC v2 & w-okada, since the addition of real-time vc gui to RVC v2? Other than RVC also allows model training and UVR 5? Does w-okada has real-time conversion features that RVC v2 doesn't have?

[dense dirge]

how i can fix this

other times i have trained

but now this appears

it says colab has a gpu limit

what is it

and when can i refresh it

[static gyro]

your gpu usage might have expired for the day. wait 12 hours

[worthy grove]

D:\RVC1006Nvidia>D:\RVC1006Nvidia\runtime\python.exe EasierGUI.py --pycmd D:\RVC1006Nvidia\runtime\python.exe --port 7897 Traceback (most recent call last): File "D:\RVC1006Nvidia\EasierGUI.py", line 22, in <module> from i18n import I18nAuto ImportError: cannot import name 'I18nAuto' from 'i18n' (unknown location)

[rare goblet]

Ayo? @worthy grove level 1 !!!

[worthy grove]

The actual module contains said class

So I am not sure what is up

[dense dirge]

okk ty

[brittle wing]

why do my voices have a lisp

i dont speak with a lisp

[odd flower]

How can we tts using rvc models?

[rare goblet]

Ayo? @odd flower level 1 !!!

[fickle mason]

i put BOTH .pth and index in a .zip folder right

[fickle mason]

yuh

[odd flower]

?

[fickle mason]

go to #guides it shows tts

#1159513888199540817

[odd flower]

Tnx

[static gyro]

how do i get index?

[minor zenith]

so is there a way to make the ai scream? like if you scream into the mic like you got scared or something is there no way that the ai can replicate that?

to try and explain what i mean:

[rare goblet]

Ayo? @minor zenith level 2 !!!

[near flame]

colab files > Mangio-RVC-Fork > logs > ModelName

[static gyro]

where is it?

[near flame]

make sure to run index training first

[snow sequoia]

hej

so my thing does not make a sound

like when i do start it does not work

[static gyro]

where is that?

[near flame]

it's the last cell

after export model

[static gyro]

thank you

[brittle wing]

yo guys

when im using large voice files ( larger than 20 seconds smth ) i keep getting an error about gradio saving to temp file smth

using mangio-rvc-fork

man i even went as far as fucking modifying the code

[static gyro]

i pressed it but nothing happened

[brittle wing]

it used to give me an error saying nontype has no object "tobytes"

now its saying dtype

[tame mica]

it should be on your google drive

[static gyro]

could you guide me?

oh i see it

[prisma sluice]

It doesnt work for me, the audio respons i get is just my normal mic

[brittle wing]

anyone knows how to fix this? i had this problem for days now

it only happens on large files

apparently the "data" is nontype

but the thing is im pretty sure the data is supposed to be the wav song or smth, just the data of the audio, so it shouldnt be blank, why is it returning blank?

[tame mica]

is it that hard to split those 20+ seconds into 5 seconds

[brittle wing]

does anybody know how do i get a non-web version of rvc v2?

[trail iris]

#1159949278270193734

[brittle wing]

can you explane please?

[copper pebble]

back mfs

[tranquil cliff]

Hey guys

[copper pebble]

@fickle mason did u complete that model??

[languid schooner]

I am having a bit of troube with being able to hear my self with the ai voice

and when I can sometimes hear myself its very choppy and lagy

any sugestions?

[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

[slim geyser]

Need to train it on screaming, but it'll make the rest of the model sound more like it's yelling since rvc kind of equalizes all of the training data.

[hushed quartz]

my friend has an issue where using realtime rvc works just fine asides from some popping noises and we changed some settings to get it to stop temporarily, but inevitably it came back.
would that indicate the device is insufficient to run it? I have a very similar device to the one the person is running realtime rvc and can do it on there fine so i'm confsued as to what the issue might be

[rare goblet]

Ayo? @hushed quartz level 1 !!!

[fickle mason]

The one time it worked, I didn’t know where the .pth file was. So when I went to weights RUNTIME DISCONNECTED I LOST EVERYTHING. 10th times a charm, at this point nothing phases me

[copper pebble]

bruhhhhh

[fickle mason]

Yuh

as u can tell I have good luck

[tame mica]

uh huh

[fossil garnet]

How do I fix the voices just being default and not working

Like none of the default voices are working now they’re all at zero and 0.1

This happened when I tried, adding a new model voice

[ornate tartan]

hi guys how can i delete a sound

[rare goblet]

Ayo? @brittle wing level 1 !!!

[sudden halo]

What softwares do we need for using models?

[polar hound]

entire #1159513888199540817 section for u to read. easygui, mangio, applio, etc. pick one and any of those guides will tell you everything

[rare goblet]

Ayo? @polar hound level 1 !!!

[sudden halo]

Thank you!!!!

[rare goblet]

Ayo? @sudden halo level 1 !!!

[lament barn]

I want to model my own voice, I have a text reading voice recording close to 10 minutes. a voice recording in which I only speak without any clean music. since it is a Turkish voice recording, I had problems with some modeling sites. is there any site where I can make a model in Turkish, or does anyone know a website or application where I can make a 5000 epoch model?

[rare goblet]

Ayo? @lament barn level 1 !!!

[violet heron]

Your model would overtrain if it’s 5000 epochs

[teal surge]

i am trying to get the program to work using the google collab could someone help getting it to work?

if you can help send me a DM

[molten pecan]

Are you trying to make a model?

Or just convert an audio?

[teal surge]

i dont know how to open the program lol

[molten pecan]

Ah

But what do you want to do with collab?

[errant shadow]

Let me ask if rvc v3 is available yet

[rare goblet]

Ayo? @errant shadow level 1 !!!

Ayo? @teal surge level 1 !!!

[carmine hearth]

Not yet

[errant shadow]

Thank

[teal surge]

all the tutorials look different to what i see

[errant shadow]

/level

[rare goblet]

Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote

[carmine hearth]

https://discord.com/channels/1159260121998827560/1159290666019471421

[teal surge]

never mind ill try later

[devout notch]

so im running the trainer on collab and its running tensorboard but this is what i get

am i accidentally using an older version of the collab?

[oblique venture]

Colab cant run tensorboard unfortunately, tho you can get the event file and then run the tensorboard on your computer from it

[vocal cipher]

does rvc v2 on huggingface work on android

[pseudo warren]

hi, someone here speak spanish? I need help with the AI of Brendon Urie TwT

[glad zealot]

if you're training it using colab yes, but if you want to do it directly on your phone, no

[hybrid spire]

So I'm just using kits ai website to make voice models, but the free version just allows up to 10 min of audio to use. Supposedly if I made the audio length shorter by speeding it up (cutting the length in half in audacity), would that help with adding more audio in making voice models, or does that not help at all?

[glad zealot]

the longer your dataset the better, most of the time.. but you cant really download models you made on kits ai so wont really matter if you wanna use it for anything other than cover

[hybrid spire]

That's what I mean though, bc of that 10min limit. So what I'm thinking is speeding up audio and packing more audio into that 10min length. Would it matter either way?

[oblique ocean]

Can someone help me?

I did everything correctly but it keeps giving me an error and it keeps loading infinitely

[brittle wing]

i used a dataset longer then 10 mins and it worked for free

[full spindle]

Hello, a question, I already trained my model and I have the "modelo.index"... but I still don't know how to make the file "TADC_modelo.pth"

[hybrid spire]

How exactly? Doesn't the kits ai website prevent that, unless you get a premium account?

[rare goblet]

Ayo? @hybrid spire level 1 !!!

[brittle wing]

Im sorry but I have no idea

[wooden orchid]

why i cant open the app

[devout summit]

Hi, does it really work?

[daring verge]

lemme hear it first

[green birch]

Sounds good imm send it in the am

[umbral egret]

what if my model gonna overtrain

its bad?

[neon valve]

Help

Why can't I upload picture here?

[golden lake]

guys how can i make a cover with me own voice?

[ember sable]

can anyone help me? all voice model rvc it not working for me

i use the newest ver

[rare goblet]

Ayo? @ember sable level 1 !!!

[ember sable]

should i just use .14 one?

[restive kiln]

what's the diff between rvc and #🔍│help-w-okada

[mystic ferry]

Yes

[neon valve]

Is there any way to reduce "CUDA out of memory problem?"

[rare goblet]

Ayo? @neon valve level 1 !!!

[tulip cloak]

just use colab

that error occurs when your gpu is not sufficient

[neon valve]

Runtime disconnected

I really want everything done locally

And I really don't want to keep pestering you for favour

[glossy coyote]

-rvc

[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

[pine mango]

yo guys thisgood so far?

im going for 600 epochs

[proper shale]

thats uh

literally just 600 steps

way too early to say anything

[pine mango]

holy fuck i just realized my dataset is incorrect and i cant proceed futher

:')

[copper pebble]

what gpu do u have??

[orchid glade]

Can I create more data sets using my own model to make it better or is that dumb

[ember sable]

do i need to download rvc to use rvc voice model on voice changer?

[orchid glade]

do I cut my data set into 5 or 10 second clips

will it be better if i do that ?

[rare goblet]

Ayo? @orchid glade level 1 !!!

[orchid glade]

do I need the file to be wav for it to be good ?

[neon valve]

Sorry for the late reply

[neon valve]

You sound like an European monarch brewing up the most inbred heir possible

[neon valve]

Also will reducing number of CPU processes used help?

[pine mango]

yo guys anyone got a dataset maker?

like to convery my vocals into a dataset file

[carmine hearth]

I'm not sure what you mean by looking for a dataset maker... Erasing instrumentals with UVR solves 90% of dataset creation.

https://docs.aihub.wtf/guide-to-create-a-model/dataset-creation

[ember sable]

?

[paper wave]

???

[pine mango]

so i had this one program that would turn my clean .wav voice clips into a dataset zip folder but turns out it bugged out recently and is not working

so im tryna find another one that can do the same for the vocals i have

[proper shale]

Have you connected the Drive to the right Google account?

[pine mango]

hello mr mod

[paper wave]

yeah

[proper shale]

hi mr cat48

[pine mango]

:D

[paper wave]

bye i gotta sleep

it's 11 pm

[proper shale]

ok then

[pine mango]

goodnight

[proper shale]

yea gn

[pine mango]

yo dude can you help me w/ this?

if you free that is

[proper shale]

uh, honestly idk a dataset maker

or zipper

or whatever it is 😭

[pine mango]

ah alr no problem

[pine mango]

💀

[proper shale]

i mean

you don't need that

[winter seal]

training a model while its still connecting

[pine mango]

i kinda do lol im making my model

[proper shale]

unless you're working with huge amounts of audios

zipping manually should work just fine

[pine mango]

yo what

[proper shale]

...yeah

[pine mango]

alr bro wait a sec

[proper shale]

even if i zip 1,2k ogg files it works

...real story...

[pine mango]

i got these 2 min vocals (above file is a combination of them all)

so you saying if i just zip this folder its gonna work?

[proper shale]

yea

just use the combination tbh

[proper shale]

no need for them to be separate

[pine mango]

yeah im gonna do that

[proper shale]

ok nice

should work just fine then

[pine mango]

yo nice

alr im zip it rq

[proper shale]

ighty

[rare goblet]

Ayo? @pine mango level 6 !!!

[pine mango]

lets see if this work

s

[copper pebble]

How much VRAM do u have?

[pine mango]

did not work ;-;

[proper shale]

what colab/fork r u using

[pine mango]

https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw#scrollTo=Sb5fzhzEXK8X

[pine mango]

.

[proper shale]

just use RVC Disconnected

theres a guide for it in #1159513888199540817 , link is in #1159949278270193734

[pine mango]

-colab

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[pine mango]

this one easier?

The one i was using had a yt tutorial

[proper shale]

RVC Disconnected is just a bit more reliable

[pine mango]

oh

[proper shale]

A bit harder tho, but the guides will help

[pine mango]

does lemme see if there are any tutorials about this one

[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

How to Make AI Covers using Ilaria RVC

Link: Documentation

Create a model with RVC disconnected (colab)

Link: Google Docs

Credits: Nick088, Angetyde and Kanav

[bleak dust]

should I use crepe or rmvpe for singing? I have an rtx 2060

[proper shale]

is ur dataset fully denoised

[bleak dust]

you mean the voice? I am not sure, I just started with this and I got a model from here

how do I check?

[proper shale]

oh just use rvmpe

if some parts suck u can just use crepe

sorry, i thought you were training a voice model

[bleak dust]

ohhh, thanks

[rare goblet]

Ayo? @bleak dust level 1 !!!

[proper shale]

np ;)

[proper shale]

ur nvidia one

[proper shale]

Keep at 2 always

[rare goblet]

Ayo? @neon valve level 2 !!!

[neon valve]

Oh

I've always use 11

thank

[copper pebble]

yh

[proper shale]

np ;)

[orchid glade]

do I cut my data set into 5 or 10 second clips
will it be better if i do that ?
do I need the file to be wav for it to be good ?

[hot violet]

as my experience,it should be better with cutted audio and wav format 32 type float is good setting

[thorny grail]

did voice models get deleted

[copper pebble]

#1159289700490694666

[thorny grail]

It doesn't have jordan peterson anymore lol

[copper pebble]

give me a sec

[thorny grail]

okk

[copper pebble]

https://www.weights.gg/download?fileUrl=https://models.weights.gg/clm72rcrh0ltgcctc412gxmgb.zip

here u go

[thorny grail]

Thanks. How'd you find it though

[rare goblet]

Ayo? @thorny grail level 1 !!!

[thorny grail]

Also, which collab is the best one? Are they all similar or are there really poor quality ones?

@copper pebble

[copper pebble]

covers (inference) or training?'

[copper pebble]

weights.gg

[thorny grail]

no training, just covers

online

[copper pebble]

a lot of model of the old server didnt get upload but they still are here

[thorny grail]

this is really helpful

[copper pebble]

you can also do /search in #🔍│find-models

[thorny grail]

ok

[copper pebble]

u can use this

give me a momnt

[thorny grail]

sure

[copper pebble]

https://discord.com/channels/1159260121998827560/1167844377046040769

easy tut

easy platform

[neon valve]

Is VRAM and display memory the same thing?

[thorny grail]

ok thanks!

[neon valve]

If so I got 3.8G

[copper pebble]

mhh i think its not enogh

a guy had like 5.2 and still want enough

[neon valve]

How can I get around such low VRAM

[acoustic shale]

Time to make him sing barbie girl

[copper pebble]

dms are opens if u need help

[neon valve]

Is there any way?

[copper pebble]

real

[acoustic shale]

Or maybe some fucking disney song

Oh I know the one

[copper pebble]

im not that good with local processing so

[copper pebble]

caramelldansen

[acoustic shale]

Let it go

Wanna hear him belt

[thorny grail]

Thanks

[astral jungle]

Yo

[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

[copper pebble]

yo

[odd flower]

is there a way to directly use the RVC models for TTS(text to speech) apart from the method mention in the #1159513888199540817 section?

[copper pebble]

actually no, if u want to use them u need to convert

[odd flower]

ok

[slender ravine]

when i try opening the RVC folder it lags my pc

[frail stratus]

How can I use ai voice here?

[proper shale]

#1159513888199540817

[molten pecan]

Do you have a good pc?

[spring badge]

where do I upload the voice i downloaded to Ilaria RVC

[copper pebble]

you just need the download link of it

and copy it in "download models" in ilaria rvc

[spring badge]

it keeps saying this italian shit

[copper pebble]

u keep putting the same wrong shit link

send me the link blud

[spring badge]

https://huggingface.co/sail-rvc/HomerSimpson2333333/tree/main

[molten pecan]

[copper pebble]

install the pth and the index and make a zip folder with them

then upload the zipped folder as a model om huggingface

its seems like the bro that made that model shared index and pth in two different folders so u have to do this shit

or u can just find another model of him

[copper pebble]

Julia

[molten pecan]

Rickie

[copper pebble]

[molten pecan]

skull emoji

[copper pebble]

real

[proud fox]

no

[stark grove]

doing a model

but it can't detected my GPU

[vale birch]

which one is that

[cobalt oyster]

Is there anything like txt to speech rvc not eleven labs

[vale cypress]

Hi

[rare goblet]

Ayo? @vale cypress level 1 !!!

[vale cypress]

Is there a working colab or gradio for training?

[west swallow]

Voice seems to be cutting out a ton when I'm trying this. I do have an AMD GPU and I can see this software doesn't support that because I can only use my cpu and that may be the issue.

I would show my settings but I don't have permissions to send images

[sand iris]

which software ?

[vale cypress]

I really need one...

[west swallow]

Not sure if i'm providing the right info here but here is the exact version with the file name for it MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.10b

[sand iris]

Oh yeah, you got the nvidia version

[west swallow]

is there an amd version

[sand iris]

yes

[west swallow]

oh sick okay

[sand iris]

the "DirectML" one

[west swallow]

Okay so I have the AMD version now and there is gpu0, gpu1, gpu2 how do I know what to select

[sand iris]

0

[west swallow]

okay

[sand iris]

pinned messages in #🔍│help-w-okada should have you covered

[west swallow]

ok cool

[ripe river]

hii, im trying to download the voice model for Serj Tankian and it says that the file doesnt exist???, What should i do??

[sharp island]

Wtf does “skip load preprocessed dataset files” and “import model” cells if not resuming” mean

[copper pebble]

https://huggingface.co/spaces/TheStinger/Ilaria_TTS

[copper pebble]

why asking this if ur not resuming??

[sharp island]

Resuming?

[copper pebble]

https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS?authuser=0#scrollTo=Yj-_npuA_HRB

[copper pebble]

r u trying to resume??

[sharp island]

I have no idea what means

[copper pebble]

ME TOO AHHAHA

[rare goblet]

Ayo? @copper pebble level 25 !!!

[copper pebble]

well i know what it means

just resume ur training

but i dunno if that works on disconnected

[copper pebble]

restart runtime

[stark grove]

alr thanks

[sharp island]

Like, u stop and then go back

[stark grove]

ok

[dense wagon]

what does this mean

i keep getting this

when doing feature exraction

should i continue to train ???

or restart the runtime google colab

[stark grove]

oh uh

[molten pecan]

Qué programa estás usando?

[dense wagon]

means u used up all the free gpu you were given

gotta pay or wait like a day or 2 to use

again

[stark grove]

... GOD DAMNIT

[dense wagon]

rip 😭

[stark grove]

welll there is a way

use another browser

[dense wagon]

another way is u can make an alt account

ian think using another browser will work

[stark grove]

hear me out i'm using chrome and it works

[dense wagon]

wait you already switched to another browser ??? or

[stark grove]

yeah

[dense wagon]

bruh how

it shouldnt even work

or maybe

idk

[stark grove]

and yet it works

[dense wagon]

smh and ive been paying but alr 😂

[stark grove]

damn man sorry to hear that

but basically if you lose gpu in a browser

like opera gx

move to another browser

and try again

[dense wagon]

too lazy imma juss keep paying

[rare goblet]

Ayo? @dense wagon level 9 !!!

[stark grove]

but... is free

[rare goblet]

Ayo? @stark grove level 4 !!!

[dense wagon]

🗿

lazy

[copper pebble]

use another account

@stark grove r u italian??

[stark grove]

1. well yes tee hee
2. yeah i'm Italian

just found out i was using another account

oops sorry

[copper pebble]

aaaaaaaaaa

e dillo prima no haha

[stark grove]

lol

[copper pebble]

#1159291287430778933

vieni qua magari ti aiutiamo meglio

[stark grove]

ho già risolta grazie :D

[copper pebble]

okok nel caso pingami o vieni in quella chat :))

[stark grove]

okie

[shy zealot]

**Method 1 🔎: ** The search accuracy is 100% 💕

Yeat V2

Upload the <t:1694426508>
Server uploaded: AI HUB (RIP)
Channel uploaded: AI HUB (RIP) Post
Profile Owner: x13xbeatzzxx

Author

x13xBeatzZxx

Discord User

x13xbeatzzxx

Internal ID

8454

Language

EN 🇺🇸

Type

RVC V2

Size

51 MB

Extractor

N/A

Epoch

0

Steps

0

Powered by ModelAI.net, Made with Love ❤️

[daring cradle]

I have a problem with the voice changer, all voices I download run on 20 000+ ping. Whereas those already included run normally around 300 ping. Does anyone have a solution for this ?

[half cove]

#🔍│help-w-okada

[daring cradle]

my bad

[earnest glen]

how do i make ai cover sounds

like i need a program or visit a webpage

[mystic ferry]

What is this error im getting?

Mangio rvc fork, using mangio-crepe, trying to make a sound clip using cho'gath model https://www.weights.gg/models/clm731spw1f1jcctcg1kob01e

[molten pecan]

You can do both ways

-rvc

[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

[mystic ferry]

no man istg i've seen ur nickname in scp secret labaratory before

full ss from start of logs

first I do a test run and it works and whenever I try to load this model or mangio-crepe, unsure which causes the problem, this occurs

nvm something is wrong with the index

when I run it without the index file it's fine

[copper pebble]

@slender plinth @low shard per caso puoi vedere se riesci ad aiutarlo??

sto per staccare che sono ucciso

[low shard]

done

[copper pebble]

grazie buonanotte

[low shard]

buonanotte

[molten pecan]

@weak heron a

[weak heron]

ok so i cant get the files and stuff but I have all the music and recordings for the ai cover

[molten pecan]

Files for what?

[weak heron]

the ai voice

im trying to make a ai cover of peppa pig singing I miss you by blink-182

[molten pecan]

Okay gotcha

What do you want to use to convert the vocals?

[weak heron]

what do you mean by convert them?

if you are talking about just having the vocals with no music I already have them

[molten pecan]

Make the vocals from I Miss You sung by Peppa

How do you want to do that?

[weak heron]

im not really sure on how to but I think there was a peppa voice in this server

[molten pecan]

Oh okay

[shy zealot]

**Method 1 🔎: ** The search accuracy is 100% 💕

Peppa Pig

Upload the <t:1694426534>
Server uploaded: AI HUB (RIP)
Channel uploaded: AI HUB (RIP) Post
Profile Owner: tamarind2_s

Author

tamarind2-s

Discord User

tamarind2_s

Internal ID

8689

Language

EN 🇺🇸

Type

N/A

Size

0 MB

Extractor

N/A

Epoch

0

Steps

0

Powered by ModelAI.net, Made with Love ❤️

[molten pecan]

There are various ways and some of them don't require downloading the model locally

Wanna use the method I use?

[edgy tangle]

why mangio-crepe pitch extraction its too slow?

[weak heron]

sure

[paper wave]

it can't find the file, that's the issue (help guys)

[weak heron]

do you need the vocals for the song I have them if you need them

[molten pecan]

Woops

Not that one

https://huggingface.co/spaces/r3gm/Aesthetic_RVC_Inference_HF

This one

[molten pecan]

Yeah of course the acapella is needed

Okay so, make an account in huggingface first

[weak heron]

ok

k now wut?

[void pond]

can I train a voice with ogg files or do I have to change them to wav or mp3?

[proper shale]

You can

Even if you have 1200 files

[paper wave]

is it the more audio files the better the ai will turn out to be

[proper shale]

no

it's about how long your dataset is and the quality

actually, more files will make ur performance worse

[void pond]

oh, this is good to know, I downloaded a voicepack from a character from a game which happens to be like, 200 files of around 2 seconds each

I guess I shouldn't do that next time

[proper shale]

ehhh i think u should be fine

performance will be kinda slower tho

[void pond]

Its my first time trying to train and yeah, it does feel really slow, thought it might just be my gpu

[proper shale]

oh

what gpu?

[void pond]

rtx 3050

I love it but its far from the best

[deft finch]

How do I make an Ai cover?

[void pond]

https://www.youtube.com/watch?v=hB7zFyP99CY

[rare goblet]

Ayo? @void pond level 1 !!!

[void pond]

with a few expections, this vid helped me a lot

[deft finch]

Is there any that are not so complicated?

[void pond]

Probably not, this is the only way as far as I know. But, then again, I've started today, I did make a few covers with pre-trained models and now I'm trying to make my own

Its not really complicated, you just need to find a model, download the programs, and not get file paths mixed up

you just need to pay attention and it goes mostly smooth

I think my training thing got stuck at Epoch 2

but it is still using 100% of my gpu

[honest pond]

better use a colab

if is going to slow

[edgy tangle]

help! Im training locally after a few months and when i was doing the pitch extraction...

It takes up to 1000 seconds

That has never taken this long in the times I created voices

thread:3, f0ing, Hop-Length:64: 29%|██████████████ | 2/7 [29:43<1:14:19, 891.82s/it]

Normally when I do this it takes about 3 minutes in total.

[rare goblet]

Ayo? @edgy tangle level 1 !!!

[edgy tangle]

I have an RTX 3050 4gb and a Core i5 12th

[void pond]

Google collab?

[honest pond]

yes

[void pond]

it is still stuck at epock 2 I think

[honest pond]

use more cpu procesors

[honest pond]

can be

[void pond]

All right, I'll try to wait for now and see what I can do, worst case I'll use collab next time

thanks!

[edgy tangle]

i should use all?

[honest pond]

How many do you have?

[edgy tangle]

i have 12

[honest pond]

use 9 otherwise your pc will crash a little bit.

[edgy tangle]

Gradio uses 8 processors by default

[honest pond]

try 12 then

although if you already had it there, you might be using the processor for something else.

[edgy tangle]

k

[brittle wing]

(ill send the full image if needed) im training an rvc right now and it's stopping after the first epoch. What can I do?

Id love to use a google collab but google is blocking ai now

[edgy tangle]

thread:3, f0ing, Hop-Length:64: 20%|██████████ | 1/5 [00:00<00:00, 1008.97it/s]

its not changing

[void pond]

Ah really? I was about to ask how to do that, I am still stuck at epoch 2

[brittle wing]

Yeah you wont even get to an epoch because of googles rate limiters :(

[void pond]

I have asked this already but, is it normal to be stuck at only 2? I thought these would go over to like 500 epochs so I think this thing will never fininsh

[brittle wing]

I have no idea! I have mine set to 60 epochs and i havent even gotten to the second one LOL

[void pond]

like damn its like this for an hour already

[honest pond]

if the first epoch lasted one minute there is definitely a problem if the second one takes more than 1h

[edgy tangle]

im going to die

finally

idk why my cpu is on 2.72 ghz

it can reach 4.40 ghz

[glad zealot]

Is it rvc disconnected?

[eternal turtle]

yo web or training on computer?

[shy zealot]

**Method 1 🔎: ** The search accuracy is 100% 💕

Bruno Powroznik

Upload the <t:1694426488>
Server uploaded: AI HUB (RIP)
Channel uploaded: AI HUB (RIP) Post
Profile Owner: analogspiderweb

Author

analogspiderweb

Discord User

analogspiderweb

Internal ID

8275

Language

EN 🇺🇸

Type

N/A

Size

162 MB

Extractor

N/A

Epoch

0

Steps

0

Powered by ModelAI.net, Made with Love ❤️

[edgy tangle]

i think im using an old version of rvc

[rare goblet]

Ayo? @edgy tangle level 2 !!!

[edgy tangle]

I'm going to install everything again, maybe it will help a little

[void pond]

Yeah, its still stuck at 2

[rare goblet]

Ayo? @void pond level 2 !!!

[void pond]

NO NOT THAT TWO

[honest pond]

magnio?

[tough void]

(using ilaria rvc) how can i delete one of the models i downloaded?

[rare goblet]

Ayo? @tough void level 1 !!!

[edgy tangle]

idk

just

beta

[honest pond]

try to use appilo

[tough void]

i put a link to the wrong model and named it ariana grande, but now when i try to put the correct link with the same name it says error

[honest pond]

it seems to me that it is the only one they support.

[edgy tangle]

i just downloaded rvc from Ai hub guides

there is not crepe

for pitch extraction

wow

this version is a lot faster

[void pond]

got me stuck on epoch 2

[edgy tangle]

not for me

[void pond]

damn

[edgy tangle]

rmvpe is better than crepe/mangio-crepe?

ooh

this rvc doesn't have crepe so

im going to use rmvpe_gpu

for infer i should use crepe or rmvpe?

[void pond]

let me ask this just to be sure once more

can I use ogg for training?

[brittle wing]

help..

wait actually i got that to work

[rare goblet]

Ayo? @brittle wing level 2 !!!

[brittle wing]

i cant figure this out

because i sure as hell did run feature extraction

and it ran successfully

[brave eagle]

Hi everyone! I'm having a very bizzare problem. I'm trying to train a model locally and the webui constantly errors out every step of the way. Nevertheless, process is still running fine in the console. I tried this with the regular RVC and the Mangio fork.

[main patio]

How do I get a huggingface link for a model from weights.gg?

[brittle wing]

Please somebody help me, my feature extraction went through just fine.

[daring verge]

lemme see logs of feature extraction

[brittle wing]

hold on just a second im trying to rerun it in microsoft edge instead

[brittle wing]

ill be honest i think i fucked it up extra

[molten pecan]

Ultimate vocal remover 5

[brittle wing]

highly recommend this one

[proven pecan]

Anyone have an RVC that does't sound terrible lol?

[brittle wing]

wdym

[proven pecan]

I can't find any for audiobooks that doesn't sound bad

[brittle wing]

like.. Any? it might be ur settings itself

[proven pecan]

what setting would that be?

[brittle wing]

like the pitch and the f0 method

[proven pecan]

not it

It adds weird "z" sound to certain letters

stuff like that

[brittle wing]

yeaah it makes it crunchy like that

maybe the ones youre using dont have enough training?

[proven pecan]

20 minutes not enough for training?

[molten pecan]

f0 method?

What's that for?

[glad zealot]

That's feature extraction iirc

[languid wigeon]

someone help me?

[molten pecan]

Wait what?

I don't understand your question sorry

[molten pecan]

What's the matter?

Mhm

Yeah

Oh okay

Okay

I recommend using UVR5

Ultimate vocal remover 5

Select the deecho algorithm and the windows size to 320

If you don't have a good computer use the huggingface space of uvr5

[brittle wing]

how do I download AIs from applio?

[molten pecan]

By the way for the cleanest results it's best if you first remove the reverb and then the harmonies

[glad zealot]

No idea.. maybe harmony first then reverb?

[molten pecan]

Also for removing harmonies use hp5 karaoke, and for backup vocals use hp6

[molten pecan]

Nope. You first remove the reverb and then the other backing vocals

[brittle wing]

or kits ai, whichever

[rare goblet]

Ayo? @brittle wing level 3 !!!

[molten pecan]

Idk how to use applio, sorry

[brittle wing]

lolol its okay

[molten pecan]

Wanna use the app that I use?

[brittle wing]

well I want a specific model thats on it

[molten pecan]

It's super simple to use

[brittle wing]

[molten pecan]

Have you checked on weights.gg?

Or on #1159289700490694666

[brittle wing]

yes :P(

:(*

[molten pecan]

Oh okay

So the model you want is only available on kits.ai?

[brittle wing]

yes

[molten pecan]

I see

Yeah it's a shame when that happens 😿

[brittle wing]

I tried making my own

[molten pecan]

Do you know the name of the person who did it?

I might be able to find it on huggingface

[brittle wing]

Yes, I tried searching for them but they havent given any info

oh that might work

sebong thingz

[molten pecan]

Tell me their nickname and ai model

[brittle wing]

https://app.kits.ai/convert/shared/jeonghan-Vlr2v-CcpU
Jeonghan

[molten pecan]

Gotcha

I'll search for them on hugginface

By the way

Do you dry your hair with a hair drier?

Or do you not?

[brittle wing]

. . . .

Towel dry

[molten pecan]

I thought so

[brittle wing]

?!?!?

[molten pecan]

I asked that same question to any autistic person I meet here

[brittle wing]

theyre bad for ur hair if you do it everyday 🤞

UR DSAIUGNFSJDGHSFD

[molten pecan]

For my own research :3

[brittle wing]

LMFAOOO

they also burn me

[molten pecan]

Grace your carrd website thing is so sweet and so cool 😭

[brittle wing]

IS IT? thanku!

[molten pecan]

I got distracted browsing it 😭

[brittle wing]

that makes me happy actually

[molten pecan]

You did such a good job congrats!!

You should ^-^ I love it when people rant like that

[brittle wing]

i go crazy when it comes to talking ab my interests

I set up the "verkwan gallery" page with 75% of the photos being already in my downloads 😭

[molten pecan]

Oh my god I haven't explored that page yet

I was wondering what it was

[brittle wing]

HEHESHF i like to abuse the elements as much as i can

[molten pecan]

Woah and I suppose your phone is full of pics of kpop idols, right?

[brittle wing]

bc theres an element limit on carrd free

[brittle wing]

yes..

though those pics are on my computer

[molten pecan]

Oh I see

That's admirable to me😭

It takes time and dedication

[brittle wing]

im passionate about it to say the least

[molten pecan]

Oh my god wait what's that for?

[brittle wing]

well some of it is for a kpop bot i use

and I submit images for it

but other ones are just personal fun collections

[molten pecan]

Ohhhhh really?

[brittle wing]

bc i like groups that are really irrelevant sometimes

and half of the stuff is lost media

[molten pecan]

Oh my god this is supposed to be the channel for rvc stuff 😭

[brittle wing]

I JUST REALIZED THAT TOO

LOLOLO

OOPS

[molten pecan]

😭😭😭 it's fine

[brittle wing]

COUGH COUGH.. AI!!! AI!!

anyways .. .

I friended u hold on i can ramble to you in dms if you want hehehe

[molten pecan]

Grace you are so fun to talk to I wish I had a friend like you

Good idea then

I'll first search the model

_<

[brittle wing]

[molten pecan]

@brittle wing I found TWO!!

https://huggingface.co/models?search=Jeonghan

[brittle wing]

👀

[molten pecan]

Sorry for taking so long 😭

I was reading your page

[brittle wing]

OOH the first one I used, let me try the second one

ITS OK HEHE

[molten pecan]

And I know this sounds weird but

You have the best DNI section ever

[brittle wing]

HFJSDGJHSDGJHDS

I think thats good then

[molten pecan]

I'm sooo glad to see someone who thinks like I do

Anyways

[brittle wing]

[molten pecan]

Yeah so, what are you going to use for inference?

[brittle wing]

like song recreation?

sorry im new to the terms

[molten pecan]

Oh that's fine

[brittle wing]

because if so - catallena by orange caramel

[molten pecan]

Pffft 😭

Why the emote omg😭

[brittle wing]

ldllsdfl

[molten pecan]

Inference is converting an audio so an AI model sings it

If that makes sense

[brittle wing]

Ohhhh yes yes

okay so yes - catallena by orange caramel

[molten pecan]

Okay

But are you going to use kits.ai?

Applio?

The huggingface spaces?

Gradio?

Google collab?

[brittle wing]

I've used the kits.ai one just once, and now i downloaded the one you found so im using a rvc gui program

[molten pecan]

Ohh okay

[brittle wing]

results of the kitsai one 😁 this is the model i wanted to download

[molten pecan]

The one I found? Wdym?

[brittle wing]

Oh the one you sent me, one of the two

[molten pecan]

Oh okay

[brittle wing]

its so awkward without the music omg

[molten pecan]

My question was, what app are you going to use to make the inference?

Not the model itself

[brittle wing]

Ooooh, RVC GUI

I like that one, its straight forward

[rare goblet]

Ayo? @brittle wing level 4 !!!

[molten pecan]

Oh gotcha

Can you send me the link to it?

I forgot what GUI was 😭

[brittle wing]

yesss one sec

[molten pecan]

Yes

[brittle wing]

https://github.com/Tiger14n/RVC-GUI/releases/

[molten pecan]

Interesting

Gosh I should have gone to bed an hour ago but I want to help you so badly 😭

Can we talk in dms now?

[glad zealot]

Proceeds to use the rvc

Should be the one I have on kaggle right?

@karmic oxide ya still have the link where you got yours from? Imma try if I can port it to kaggle

[wary crystal]

Question, Crepe and Mangio Crepe what are the differences, and which one is beter?

[shy zealot]

**Method 3 🔎: ** The search accuracy is 40% 🥺

Carmen

Upload the <t:1689625112>
Server uploaded: AI Hispano
Channel uploaded: AI Hispano Post
Profile Owner: elmewtwom

Author

ElMewtwo-M

Discord User

elmewtwom

Internal ID

1409

Language

EN 🇺🇸

Type

RVC V2

Size

0 MB

Extractor

Harvest

Epoch

200

Steps

0

Powered by ModelAI.net, Made with Love ❤️

**Method 3 🔎: ** The search accuracy is 40% 🥺

Carmen

Upload the <t:1689625112>
Server uploaded: AI Hispano
Channel uploaded: AI Hispano Post
Profile Owner: elmewtwom

Author

ElMewtwo-M

Discord User

elmewtwom

Internal ID

1409

Language

EN 🇺🇸

Type

RVC V2

Size

0 MB

Extractor

Harvest

Epoch

200

Steps

0

Powered by ModelAI.net, Made with Love ❤️

[coral pebble]

hello can someone help me please, does anyone know what I can do to make my model sound less robotic? I've trained it to 350 epochs

[rare goblet]

Ayo? @coral pebble level 1 !!!

[eternal turtle]

increase dataset?

if it's robotic, then it's probably not isolated enough

[devout notch]

the models i make are too large in size

is there a way to make em smaller?

[karmic kelp]

how can i create my own voice?

[split rose]

why does it give me an error FileNotFoundError: [WinError 3] The system cannot find the path specified: mi-test/3_feature256 when i try to train with "pitch guidance" set to false?

[dawn meteor]

how do i train my own ai voice locally with rvc gui

[mystic ferry]

Did you extract features?

[mystic ferry]

I don't know any other apps, but I use Mangio-RVC fork and if you need help figuring things out on it I could help

Ill send the huggingface link one sec

https://huggingface.co/MangioRVC/Mangio-RVC-Huggingface/tree/main

Okay I have school rn so ill check back in 50 min

[split rose]

ah missed this step, ty

[brittle wing]

yess

[frozen quest]

Ahh ok

Is 3 mins dataset enough or need a longer dataset

[waxen cliff]

if i had a dataset a little over an hour long, how many epochs should i do

[proper shale]

A longer dataset would be better

10+ mins at least

[fossil glen]

-rvc

[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

[fossil glen]

-colab

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[proper shale]

TensorBoard will tell you that.

There is not a specific number that you should use because all training sessions are different because of dataset size, quality and batch_size. Therefore, a model could be very undertrained at, let's say, 200 epochs, but for someone else with a smaller batch_size, 100 epochs could already overtrain.

The best way to prevent undertraining or overtraining is by checking TensorBoard. loss/g and loss/d will help you determine whether your model needs more training or not.

I hope this clears things up for you.

[mystic ferry]

I mean ofc a longer one would be better if available, but I did make one with I'd say around 2mins dataset and it turned out okay (definitely not the best but not too bad)

[mystic ferry]

Hey, I had this question before, but noone responded, do you maybe know how to open up tensorboard? Whenever I open the .bat or .exe file (forgot which it is) it just opens and instantly closes

Im using local Mangio RVC fork

[coral galleon]

Hey guys this is my first time training a model fopr an upcoming project could someone help me as I keep getting this error inside of command prompt:

runtime\python.exe trainset_preprocess_pipeline_print.py C:\Users\Username\OneDrive\Desktop\RVC-beta0717\Houston Jones 40000 11 C:\Users\Username\OneDrive\Desktop\RVC-beta0717/logs/Houston Jones False
Traceback (most recent call last):
File "C:\Users\Username\OneDrive\Desktop\RVC-beta0717\trainset_preprocess_pipeline_print.py", line 8, in <module>
sr = int(sys.argv[2])
ValueError: invalid literal for int() with base 10: '-'

Can someone please help?

[mystic ferry]

Did you run the preprocess?

[coral galleon]

I'm using the RVC web bat

but ive just discovered I can use the google version