✨│ai-help

[hollow yew]

I did

I did

Did the whole process

Signed in my account

Duplicate space or whatever

So uhh

Still didn't work

[half cove]

ok show me what it said when you downloaded the model

[hollow yew]

I'll do it tomorrow

Probably my wifi at this point

:((

Too tired

[half cove]

rip

[wary cedar]

@round fog i think i figured it out. im not 100% on it, im doing 750 epochs

[plucky locust]

whitebeard is not a registered model! Use /list to list all models, or use /register to register a new model.

Community Models

goku (Goku (English Dub))
luffy (Luffy (English Dub))
mickey (Mickey Mouse)
peter (Peter Griffin)
plankton (Plankton)
spongebob (Spongebob)
squidward (Squidward)
waluigi (Waluigi)
... and 200+ more

Check out all of the public community models at https://www.word.band/voice_models

You can use any model using the model id (bolded) in the /cover command. For example:
/cover model:goku youtube:https://www.youtube.com/watch?v=s9FCCiEs84k

To see your private models and register new ones, please start a DM with me!

[round fog]

hey my bad i didn’t have signal

[rare goblet]

Ayo? @round fog level 35 !!!

[wary cedar]

your chillin brother

[round fog]

if you could show ur setting i double check everything

[wary cedar]

i watched some yt vid, and got a base understanding but yeah

[round fog]

to make sure you don’t waste like

4 hours

[round fog]

okay sounds good

[wary cedar]

how we lookin @round fog

[green parrot]

my processing status is stuck here:

INFO:torch.nn.parallel.distributed:Reducer buckets have been rebuilt in this iteration.
INFO:IziMusic:====> Epoch: 1 [2023-12-25 16:57:37] | (0:01:45.971008)
INFO:IziMusic:Train Epoch: 2 [80%]
INFO:IziMusic:[200, 9.99875e-05]
INFO:IziMusic:loss_disc=3.764, loss_gen=3.726, loss_fm=19.185,loss_mel=23.460, loss_kl=2.518
INFO:IziMusic:====> Epoch: 2 [2023-12-25 16:59:15] | (0:01:37.413507)
anyone know how to fix this?

[sand iris]

there's no error there

[green parrot]

ok, should I just wait? its stuck since about 20 mins

[proud elbow]

why u chose harvest instead of rmvpe_gpu? (and I suggest to learn more from #1159513888199540817 instead of mostly outdated YT videos)

[wary cedar]

idk some guy told me to 😭

is it gonna be worse?

[proper shale]

not necessarily

but rvmpe would be better

[proud elbow]

better and faster extraction, cause it's state-of-the-art now

[wary cedar]

also me personally i find it easier to learn from visuals rather than reading something i dont understand :/

[wary cedar]

ah i see

sorry im being stupid

anyway its almost done

only 250 epocs left

started abt an hr or so ago

[stark geyser]

Hi all ! I have the program rvc-tts-AINetSD (edge). I downloaded a couple of voices, put them in a folder, the model is detected but the voice is not, what should I do with this?

[round fog]

would be better to change a few things but it’s too late now lol

[wary cedar]

what would you have recommended and how do i test the model

[rare goblet]

Ayo? @wary cedar level 4 !!!

[wary cedar]

its done

uh @round fog

i got two pth files

[tribal cosmos]

should i collapse feature or not? i want the best results

[regal meteor]

this is probably a silly question but is there a way to convert multiple files

like not at once, just like a queue or smthing

[fluid latch]

theres a batch mode somewhere it hink

[glad zealot]

yes

[fluid latch]

i never used it myself tho

[regal meteor]

which one is that? im currently using jus the regular rvc gui

[glad zealot]

same

[glad zealot]

mainline

[regal meteor]

is there a link to it somewhere?

[violet heron]

-local

[azure marsh]

Suggestions for @regal meteor

All Working Local Forks 🖥️

• Applio-RVC-Fork, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub

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, -colabs, -overtrain, -realtime, -rvc, -help

[violet heron]

Ok

[wary cedar]

did allat

idk what else to do now

[violet heron]

I’m the weights folder

[wary cedar]

?

[violet heron]

Wait nvm

Ok

To use your model go to inference

[wary cedar]

yuh

now what?

[violet heron]

Refresh and select your model

[wary cedar]

done

[tribal cosmos]

should i collapse feature or not? i want the best results

[violet heron]

Then in the box upload the audio file you want to convert

[wary cedar]

i wanna convert text not audio

i wanna test its txt to speech

[violet heron]

Use something like Elevenlabs or EdgeTTS and convert that

[wary cedar]

grabbed a file with just voicelines

and tried converting

shit just errored on me

[sand iris]

spaces in filename

[wary cedar]

huh

thats just how the filename is.

[sand iris]

Yeah it craps itself when there's any spaces in folder/file names

[wary cedar]

alr one sec

[violet heron]

Rename to have no spaces

[sand iris]

Rename to something without, and it'll most likely work

[wary cedar]

done converting rn

forgot how big the vid actually is.

this is gonna take years

lemme get a shorter clip

one sec

[tribal cosmos]

should i collapse feature or not? i want the best results

[wary cedar]

How do i use eleven labs for this shi

[willow arch]

i cant download tensorborad and this thing wont run without it

[wary cedar]

also how do i make my folder into the pth file cuz rn all i see when i go into the folder is a d_23333333.pth and a g_23333333.pth and a bunch of random stuffs

[wary cedar]

How do i upload my own voice to it.

thats what i was asking.

[violet heron]

You generate it, download it, then convert with RVC

[willow arch]

wheneveri click continue it closes and nothin happens

[wary cedar]

what.

brother

i just trained it in rvc

...

[violet heron]

I’m saying

You select a random model on Elevenlabs, you generate

You go into RVC

You convert the audio you got off Elevenlabs

[brittle wing]

is "Faster training" Such a thing?

[wary cedar]

bro what. why are you explaining this so complicated.

[violet heron]

It’s called better GPU

[brittle wing]

What is the website???

[wary cedar]

elevenlabs

[violet heron]

Im saying on Elevenlabs generate it, download it, then on RVC you convert that

[wary cedar]

oh. thats so aids 😭 nah but lets say i wanted to use the model for like instant voice changer

how would i do that

[violet heron]

You would download Okada which has real time conversion

[wary cedar]

but how do i upload my model to it. because rn theres two pth files

[violet heron]

These two?

These aren’t the model files

[wary cedar]

yuhg

then what is

...

[violet heron]

It’s whatever you named the model .pth

[wary cedar]

That doesent show up

[violet heron]

Located in the weights folder

[wary cedar]

its just a folder

theres no weights folder.

nvm found it

[proud elbow]

assets\weights probably for the latest version

[pliant remnant]

anyone can help me rn?

[wary cedar]

sorry for my autism

[violet heron]

If I know

[willow arch]

for me, i have applio and it says no tensorboard and whenever i continue, it doesnt work

[wary cedar]

https://cdn.discordapp.com/emojis/925780626517426247.gif?size=48&name=Disco_Cat&quality=lossless

[violet heron]

Vencord moment

[wary cedar]

on jah

[violet heron]

What file are you opening

[willow arch]

go-applio

[violet heron]

Tensor is not installed then

[willow arch]

how do i download that

[pliant remnant]

i swear, every time i need help the thing fixes lol

[wary cedar]

wheres the indexes for it?

[rare goblet]

Ayo? @wary cedar level 5 !!!

[violet heron]

In the logs folder

[wary cedar]

checked there theres no index for my file name

are indexes even worth using

[violet heron]

The index simulates accent

[wary cedar]

so wheres mine then... im so confused

[violet heron]

Your supposed to have it if you clicked on “Train Feature Index”

[wary cedar]

wheres there

that

oh fuck

i forgor

welp

time do to that shit again

fuck

[violet heron]

You don’t have to retrain the model

[wary cedar]

just the index?

[pliant remnant]

how to make that there is completely no tearing or breathing in vocals?

[violet heron]

Only make the index using the same dataset/file name

[violet heron]

Yeah if you already have the pth

[wary cedar]

alr bet

thats why it sounded so fucked

[wary cedar]

same ammount of epochs and keep settings the same?

oh

wtf

that was way too fast

or is that normal

i clicked it and it legit did it instantly

[wary cedar]

its not in the logs folder still

but its in the The-Finals folder i made

i got an error

audio path is fine tho

alr fixed it somehow

[violet heron]

Index only takes a few seconds

[wary cedar]

why index not used

[violet heron]

But is it in logs?

[wary cedar]

no man

i already said that

its directly in the folder for the thing

[violet heron]

The added index

That’s your model index

[wary cedar]

so then where tf is the actual index 😭

i trained it but bro just dint show up

[wary cedar]

yeah. i used that

[violet heron]

The added index is your index

[wary cedar]

but it says index not used

oh

im a fucking idiot

im actually slow

i just realized

why

im so fucking stupid

still sounds like shit tho iwl

like listen to this https://www.youtube.com/watch?v=kZ87wiHps9s&t=1s

the nthis

[proud elbow]

AI VOICE that path has a space, no wonder it doesn't work

[wary cedar]

oh.

im even more slow

ill try it with allat

[wary cedar]

sounds the same tbh

https://www.youtube.com/watch?v=kZ87wiHps9s&t=1s <-- tryna make it sound like this

[radiant pendant]

Does anyone know, is there some kind of epoch calculator?

[wary cedar]

like wtf

[pliant remnant]

sounds like markiplier lol

[wary cedar]

LITERALLY

like wtf

its so fucked

[pliant remnant]

wdym

[wary cedar]

im tryna get it to sound like the yt vid above

https://www.youtube.com/watch?v=kZ87wiHps9s&t=1s

this one

[pliant remnant]

i saw it

[wary cedar]

yuh

it sounds nothing like the vid

[pliant remnant]

i mean in video it has more energy or something

idk how to describe it

[wary cedar]

sounds so shit no emotion bro sounds dead and robotic

[pliant remnant]

yeah yeah

[wary cedar]

yuh

[pliant remnant]

in video its more energetic

[wary cedar]

so how do i make it have more emotion

cuz that games commentary legit uses ai

for the voices

[pliant remnant]

idk, maybe they did something else

[rare goblet]

Ayo? @pliant remnant level 2 !!!

[pliant remnant]

bro, every time i replay this i wait for him to say "my name is Markiplier!"

[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

[covert sedge]

help i was training my model and this popped out of nowhere

now i'm afraid my 2 hours training process will disappear

[latent kettle]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[olive heron]

I need help

[covert sedge]

anyone? 😭 idk what to do..

[brittle wing]

How i fix my mic repet what i hear

[queen schooner]

idk if this is the channel for it but it works fine, the delay is there but i just dont hear any model voices that i put?

[covert sedge]

so is there any day limit for RVC Disconnected? if yes how long it takes to use it again

[violet heron]

4 hours iirc

You can always use a alt

[covert sedge]

yeah but is it a limit depending on hours you're on colab site?

[violet heron]

Did you even read my message

[covert sedge]

kk

[uneven scroll]

how do i access past "checkpoints" of a model's training? I am using the colab rvc disconnected

and also, are these "checkpoints" complete models on their own??

[violet heron]

Yeah, but you also need a index for a “complete” model

[uneven scroll]

wait if it doesnt save the index at the checkpoint how would i get it

[wary cedar]

what would be the best settings for this

[violet heron]

Train feature index iirc

[wary cedar]

48k sample rate im assuming?

[violet heron]

Or train index

[wary cedar]

i got more clips brother, gotta retrain this bitch

[uneven scroll]

tysm

[violet heron]

Depends on your audio file for sample rate

[wary cedar]

what ya mean...

😭

its a simple mp3 file

taken directly from a game pretty much

just obs and yoinked from the game

should i use pitch guidance or not

[violet heron]

I would set your sample rate to as close as the mp3/wav as possible

So I would use the number 40k for this model

I usually don’t touch pitch guidance

[uneven scroll]

so I could just load it into the (import model from drive to notebook) thing and then do index training using that?? sorry this is my last question i appreciate u

[wary cedar]

so have it off or leave it on

[violet heron]

Load dataset and same name/settings then train the index

A full model is a pth and index

[violet heron]

What it was set to

[wary cedar]

on

[rare goblet]

Ayo? @wary cedar level 6 !!!

[violet heron]

Ok then

[half cove]

dont you check where the cutoff is instead then double it

[wary cedar]

alri, what abt for feature extraction and training

i did 750 epochs and it wasent even close to overtraining before

[violet heron]

Oh forgot that

Hi Poopmaster

[half cove]

hai

[wary cedar]

what should i do for training, i did save frequency 50, epochs 750 but ima increase to 1250 and batch size (idk what it does so i left it)

[uneven scroll]

what does overtraining sound like

[wary cedar]

not sure, but some fella told me to use some app beginning w t forgor the name to check for overtraining

[uneven scroll]

o ok

[half cove]

tensorboard

[wary cedar]

yes

[half cove]

would sound kinda robotic

[brittle wing]

heelp, i have delay

like 3-4 seconds

i have una rx 6600 xt

i want to play valorant with w okada

[misty topaz]

i need help, the bot wordband bot tells me the thread link isnt rcv 2, but it is

[flat nacelle]

what is a good duration for voice samples and how many epochs should i have to get the best quality model? also do i have to split my audio sample into many different files for it to work best?

[rare goblet]

Ayo? @flat nacelle level 2 !!!

[flat nacelle]

thanks, one last question, say my training set is 30 minutes, how many epochs should i do?

would too many epochs just make the quality worse?

alright thank you, I appreciate it

[calm crown]

Does anyone has an applio colab link that works?

[flat nacelle]

sorry if i sound stupid but whats a mel graph

[brittle wing]

Hello, I have a model, A link to use it Google colab give me

[half cove]

you want a link for inference on rvc?

so like making covers with already made models

but online

[brittle wing]

Yes

[brittle wing]

Yes

[half cove]

you can try ila rvc

https://huggingface.co/spaces/TheStinger/Ilaria_RVC make sure you duplicate it

here is the guide https://docs.google.com/document/d/186_a8d32iYxr1UAEm1LAFQvLu2w3huVIg-KTlTHshg8/edit

[wary cedar]

yo yo yo

[half cove]

yea

[wary cedar]

gimme a good like. tts program

i have a custom model i made

and ive been using elevenlabs tts then using rvc afterwards using the audio file from elevenlabs

i need sum i could do directly

can be local idm. id honestly prefer local

[half cove]

hmm

try bark

[wary cedar]

huh

[half cove]

it should have emotions and stuff like that

[wary cedar]

this ting?

[half cove]

yes

[wary cedar]

btw do yk a good thing for writing python code

seems like ima need it

[half cove]

hm?

[wary cedar]

like a program

where i could write up python code

ik thonny exists cuz. well i used to use it in school

[brittle wing]

thank you so much

[wary cedar]

uh.

wait

im being slow

[rare goblet]

Ayo? @thorny musk level 2 !!!

[wary cedar]

uh

[half cove]

if its too hard to install you can try the colab or spaces

it should have it

[wary cedar]

huh?

quetion is

how do i upload my own model into it

[sand iris]

[wary cedar]

BRO IDK WHAT IM DOING 😭

[sand iris]

I see that

[wary cedar]

moderator laughs at newbie

[sand iris]

ya ain't supposed to use pip from python

[wary cedar]

im just tryna make a model man. i have it done i just wanna test my tts

[wary cedar]

... 😭

listen do yk any simple tts programs?

where i could use my own model

local would be prefered

[sand iris]

edge-tts supposedly

doesn't quite run locally, but it should do

[wary cedar]

i just see python and im instantly killing myself

[sand iris]

lmao fair

[wary cedar]

https://tenor.com/view/low-int-intelligence-specimen-cat-meme-gif-25294795

embed fail

low intelligence specimen = me

[sand iris]

if you have python installed already, you should be able to open cmd via windows search

[wary cedar]

yeah. ima try applio apperantly it has it

[sand iris]

and then pip install edge-tts (or the command you tried earlier for bark)

[thorny musk]

someone using Applio RVC?

[wary cedar]

is it any good?

taking 5 decades to install fr

[thorny musk]

idk

[wary cedar]

FOR FUCK

[thorny musk]

i try install it because w-okada don't work with my radeon gpu

and w-okada using my cpu not gpu

i hear about applio have amd support

[thorny musk]

i try from installer exe

[wary cedar]

FFFFFFFFFFFFFUUUUUUUUUUUUUUUUUUUUUCK

[sand iris]

ya right

No python install on this system apparently

[wary cedar]

there is

[sand iris]

pip3 maybe ?

Or it came without pip which could also happen

[flat nacelle]

is it normal for one epoch to take 6 minutes?

[thorny musk]

maybe try in terminal? (windows terminal) ?

not cmd

[wary cedar]

shit gpu

[wary cedar]

will try

[thorny musk]

and run as admin

[wary cedar]

well fuck

[gloomy maple]

i downloaded a juice wrld model with a huggingface link, if i upload it to huggingface and then record something it should give me ai juice wrld voice right or not? (im new 2 this shit im sorry)

[rare goblet]

Ayo? @gloomy maple level 1 !!!

[sand iris]

what does python -m pip output ?

[thorny musk]

run terminal as administrator

maybe help

[wary cedar]

i did.

[sand iris]

Yeah I know what it is. When you installed python you didn't put "add python to PATH"

[wary cedar]

idk man

lemme reinstall python

[sand iris]

It's a sneaky little box you have to tick on the installer

[wary cedar]

i swear. if i just spent an entire fuckin day

making a model

and it sounds like ass

ima kms

[sand iris]

[thorny musk]

what gpu

[wary cedar]

ok pip should be installed by now

well once this finishes

[magic cloud]

how do u get rid of the robotic sounds??

[wary cedar]

by training your model more.

[magic cloud]

how

[wary cedar]

adding more epochs

[magic cloud]

its not even my model

[half cove]

1. he could be using another persons model
2. overtraining exists

[wary cedar]

yuh overtrainign was gonan be my next guess

tbh

[wary cedar]

simply dont use that model :))

[magic cloud]

but others are worse, they sound even more robotic

[half cove]

i mean it could be more reasons like f0 method used

[wary cedar]

find one that doesent.

[half cove]

or tweaking with the settings

[wary cedar]

if all of them are robotic

then your settings most likely

are fucked

obviously i am slow

[magic cloud]

how to adjust my settings

[wary cedar]

so i wouldent know anyway

[magic cloud]

i already followed a lot of vids

[half cove]

what are you using

local or cloud

[magic cloud]

local

[half cove]

ok

mangio mainline or applio

[magic cloud]

what 😭

[wary cedar]

NOTHING FUCKING CHANGED

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAH

[half cove]

what does ur folder say

[wary cedar]

reinstalled python n everything

[half cove]

did you do add to path on the installer

[magic cloud]

MMVCServerSIO

[rare goblet]

Ayo? @magic cloud level 1 !!!

[wary cedar]

Y E S.

[half cove]

oh thats voice changer

[half cove]

[magic cloud]

yea

[half cove]

ask in #🔍│help-w-okada

[wary cedar]

reionstalling

what do i click

cuz

seems what i clicked dint fucking help

[half cove]

add python to enviroment variables

[wary cedar]

reinstalling rq

[rare goblet]

Ayo? @wary cedar level 7 !!!

[wary cedar]

one minuteo

[hollow moon]

Why is my Cable output not working anyone else got this problem?

[plucky bloom]

um, hello? I'm sorry, i don't know how to create AI voice models- can someone tell me a place to start, pls?

[wary cedar]

yo @round fog you alive my brother? any way to just use tts with my model. im so tired spent an entire day tryna get this bitch to work and still cant find a way

[topaz acorn]

is that to execute the program? bc i also just installed python and it didnt start the web interface, so i prob didnt install what i needed? bc i installed it as quick setup not as custom

[hollow moon]

My normal mic makes the ai sound completely fine but when i turn onn the cable output and it just sounds like a demon

[wary cedar]

ima kms.

[wary cedar]

im gonna kms man. this shit dont work

[plucky bloom]

uuuuuh

[rare goblet]

Ayo? @plucky bloom level 1 !!!

[half cove]

you need python to run rvc

[plucky bloom]

can i ask what rvc is?

[topaz acorn]

ye i installed python but it doesnt work :c

C:\WINDOWS\system32>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897
Das System kann den angegebenen Pfad nicht finden.

C:\WINDOWS\system32>pause
Drücken Sie eine beliebige Taste . . .

(cant find path etc)

[wary cedar]

finally got it now what do i do

[half cove]

did you allow environment variables?

[topaz acorn]

ugh i wish id remember what the exact short form was lul but its basically the thing that makes ai voices possible or its the category at least that ai voices fall into, here its basically about the training thingy where u can make rvc models for ur ai voices :3 hope i explained it right haha bc im also not rly the big brain of ai voice stuff ^^

[rare goblet]

Ayo? @topaz acorn level 1 !!!

[topaz acorn]

ill start the setup again and check if i can do that, bc i didnt check any additional things i guess

[plucky bloom]

no probs! i really appreciate it! im trying to do my first ai voice but i simply dont understand where to start-

[topaz acorn]

there are sum gud vids, id say this one might help u out :3
https://www.youtube.com/watch?v=ixB9oalT3cQ

[plucky bloom]

thank you!

[topaz acorn]

its what im using rn ^^

[topaz acorn]

dw, hope it helps <3

[topaz acorn]

what should i check? only the enviroment variables?

[half cove]

didnt it have a few others already checked

[round fog]

use something like eleven labs for text to speech

[topaz acorn]

nu, this is how i got it rn, i uninstalled it and this is basically the reinstall thing

[round fog]

and then take that audio and just convert it

with rvc

[wary cedar]

mf i did that

i wanna just test it directly

i dont like how its sounding too much

its close

its just not there yet

[round fog]

idk what u want from me lol

[half cove]

do the first 4

[round fog]

that’s just how u do it

:p

[wary cedar]

im sorry 😭

id rather do like uh local tts

if possible

[topaz acorn]

it auto checked precompile standard library, should i do that too or uncheck it?

[half cove]

i dont think that should affect anything

just leave it

[topaz acorn]

kk

[coarse harbor]

I seem to be having problems cause the program no longer opens for me. It just opens "Localhost:7897" in browser and that's it.

Idk what is causing it cause it was working fine before

"The program", I mean Mangio and Applio

Both do the same thing

Mangio didn't used to and I don't know why it is now.

[half cove]

send screenshot

[topaz acorn]

oki nicu ty <3 i think it works now :3 ill just try to do it now and might ask stuff again later, btw how many epochs should i do for 15 min? or is there maybe a thumb rule that i could use? like 1h = 400 epochs or something and just calculate it up or down?

[half cove]

you should use tensorboard to check but that might be hard for new users so id just test different epochs

you can try 400

[topaz acorn]

oh what is tensorboard? :o

[topaz acorn]

kk

[trim slate]

Everything should be working normally but input won’t pick up voice

[coarse harbor]

Of what exactly? The browser is just blank and this is what the cmd looks like https://cdn.discordapp.com/attachments/652641690670071838/1188991476139565066/thfghj.png?ex=659c899a&is=658a149a&hm=6615602cc9db6979f12cc7da6838a9c97c7b040659d6a5c06ed939dcc373bf78&

[rare goblet]

Ayo? @coarse harbor level 1 !!!

[coarse harbor]

Oh

can I not post links?

(I know it wasn't letting me upload it directly)

[half cove]

its just white?

[coarse harbor]

Yes

[half cove]

what does the cmd say

[coarse harbor]

Set fp16_run to false in 32k.json
Set fp16_run to false in 40k.json
Set fp16_run to false in 48k.json
Use Language: en_US
Running on local URL:  http://0.0.0.0:7897

To create a public link, set `share=True` in `launch()`.
"```

[half cove]

it should be working

[coarse harbor]

I haven't changed any settings or anything so Idk why it's acting up

[half cove]

i guess try a diff browser or somthin

[coarse harbor]

Weird, it loaded in chrome. I always was doing it in Firefox and wasn't having problems before.

Hmm

[half cove]

it could have updated and broke somthin

[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

[coarse harbor]

Hmm, possibly. Thanks for the help.

[topaz acorn]

for now im just testing it with 8min audio to go trough the steps but it didnt do anything :c it was loading but i dont get a done message at the right side as i should :c what could be the problem?

[rare goblet]

Ayo? @topaz acorn level 2 !!!

[topaz acorn]

ill restart it i guess, maybe that helps >.< or running it as admin or smtbg

huh? :c

or maybe if that helps, heres the rest

hope anyone has time for this :<

[candid lichen]

@sleek nova so how exactly do you improve the dataset?

Sorry I'm still kinda new to all of this

Or is there a online tutorial I can follow

[sleek nova]

It is your initial audio files you use. Mostly looking for clean vocals with no background noise. Also minimal amounts of silence.

The guides tab has alot of info regarding it all

[candid lichen]

Okay well the model I made of myself was pretty clean because I recorded my own voice, so if that's good then the epochs will make it better in some degree right

#1159513888199540817

[thorny musk]

applio HOW MUCHH

fasteeer

[candid lichen]

Cause I'm training with colab

Is that a good source to train with

[sleek nova]

I have only used local so havent read too much into the colab

[candid lichen]

Dang alright guess I'll have to figure it out, I didn't see anything in the guides with colab training

[sleek nova]

What I've seen in my models is the larger the dataset the less epochs were needed.

[candid lichen]

My model being like 800+MB

[sleek nova]

Can always test the models at different save intervals and compare them to your liking

[wide parcel]

man i swear every time i update applio i get more problems
ImportError: cannot import name 'weight_norm' from 'torch.nn.utils.parametrizations' (F:\applio\new\Applio-RVC-Fork\env\lib\site-packages\torch\nn\utils\parametrizations.py)

anyone else had this problem or know a fix?

[candid lichen]

Save intervals?

[sleek nova]

At least for local, you can specify it to create a useable model save at different intervals, like every 20 epochs or so

[half cove]

for colab

[candid lichen]

So that's like a test for how the quality will sound?

Mine only has this

There's more but that's what your referring to right

[sleek nova]

The checkmark in the image is what is being referred to.

Saying to save a model

[warped garnet]

Hey, guys!

[candid lichen]

Okay maybe I should try to figure it out somehow or get a guide cause I understand the gist of what you're getting at

My colab link is kinda basic from giftmasters

[thorny musk]

does anyone have applio tutorial?

[candid lichen]

#1159513888199540817

Maybe

[sleek nova]

There is a guide in there labeled All-in-one English Documentation. It mentions colab

And applio

[wary cedar]

someone tell me a way to just use txt to speech with my model. i dont wanna use elevenlabs txt to speech then use the file to apply my model. I wanna just straight up use my model and use it as a txt to speech. Id be happy to run it locally

[half cove]

just use anything my dude

like

edge tts

[wary cedar]

idk how to.,

[half cove]

you dont have to do it local

use colab or spaces

[wary cedar]

id prefer to do it local

cuz then i dont have to upload my model

or make an account

or wtv else the fuckin sites want from me

id rather do it locally

[half cove]

try this rqhttps://huggingface.co/spaces/TheStinger/Ilaria_TTS

[wary cedar]

need to make an account

to upload my model

and i aint making an account

[half cove]

what?

for the tts

no

[wary cedar]

if i wanna use my custom model

yes

i already tried that

unless i missed sum

[half cove]

incognito to show no login

[wary cedar]

and how exactly am i meant to use my own model.

without uploading it.

[half cove]

you dont use rvc models with tts

you get the audio from tts

[wary cedar]

... your joking.

FUCKK

[half cove]

then

[wary cedar]

so your telling me

i wasted

all my fucking day

making some fucking rvc model.

[half cove]

what tf was ur ass trying

[wary cedar]

TO MAKE A TTS MODEL

fuck

ima kms

its over

[half cove]

ok so

[wary cedar]

https://cdn.discordapp.com/emojis/819629246435164212.webp?size=48&name=dead~1&quality=lossless

[half cove]

the tts uses a premade voice to say somthing

then

you use

rvc to make it the voice you want

[wary cedar]

yeah

ive had poor results using that

so

[half cove]

[wary cedar]

can i make a tts model

somehow

i want

[half cove]

i mean you could use bark to train a tts model without needing to convert it using rvc

[warped garnet]

https://tenor.com/view/ai-generated-ai-ai-art-jarvis-artwork-gif-27298200

[wary cedar]

can i train a model locally

...

[half cove]

if you have a good gpu

[wary cedar]

or can i convert a rvc model to a tts model

[half cove]

no

[wary cedar]

fak

[half cove]

rvc is only audio to audio

[wary cedar]

any way to do this shi locally?

[half cove]

applio should have the tts preinstalled

[wary cedar]

see

i cant run that

cuz it says im missing torch

yet i cant get torch cuz it gives me some random error

reinstalling applio

so it has a tts trainer? @half cove

[half cove]

not training tts

using tts then converting using rvc

[wary cedar]

then i dont need that shit 😭

thats not my issue

my issue is it comes out funky

i want it to just be tts

i want to make a tts model

straight tts nothing else

no other processes involved

[thorny musk]

\

why applio doesnt detect my amd GPU

[half cove]

what gpu do you have

[thorny musk]

amd radeon rx6950XT

[half cove]

did you select amd in the installer?

[thorny musk]

there was no option to select GPU anywhere in the installer

[half cove]

What are you using

[thorny musk]

I thought that applio would change the voice live, straight to the virtual cable

[thorny musk]

Applio

[rare goblet]

Ayo? @thorny musk level 3 !!!

[thorny musk]

instalkę

[half cove]

Oh yea

[thorny musk]

installer

from git

[half cove]

I swear it gives an install for amd

Last time I tried

[half cove]

Your trying to use voice changer?

[thorny musk]

nothing to choose gpu when instal

[thorny musk]

No

because No gpu

[half cove]

I think your talking about realtime rvc

I'd use wokada for that

https://rentry.co/VoiceChangerGuide

[thorny musk]

Ye like in wokada

[thorny musk]

I have wokada but wokada didnt see my gpu and i must use CPU

Ryzen 7 5700X

On wokada my % gpu is 0%

[half cove]

Ask in #🔍│help-w-okada then lil bro

[thorny musk]

when cpu go into 50%-70%

tomorrow

ty

byebye

[half cove]

Bai

[modern surge]

If I have an AI but only its index and its .pth, can I finish training it with another dataset or similar?

[heady fable]

How do I make the voice sound better?

[violet heron]

That’s how it’s supposed to be

A .pth and a index

[fluid topaz]

get a better dataset, or clean it better

[wide parcel]

shout out poopmaster

[wide parcel]

what’s with the auto reacts 😭

[spring turtle]

any reason why whenever i run a conversion while using focusrite 2i2 as input for my dt990s after i click convert it makes all audio incoming inaudible and glitchy and stuttery, any fix for this?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[low lynx]

Hi when i use the “RVC V2 Hugging version” i download the model and name it and it shows “succeed” but when i try to select the model in “choose your model” it doesnt let me choose anything.. just to write. If i do write the model’s name i gave it then it disappears after i click on smth else.

Can someone please help?

[acoustic forge]

Using easiergui index not used is always present even when using an index

[harsh scarab]

You're placed in Cover Generator Queue. Current position: 1

[slate pawn]

can anyone help in this index file that't not working?

[proper pivot]

Help

I just trained my model

And there's no pth

[bitter scarab]

yo what batch size is responsible for? at first i thought the more the better, but that's not the case it seems

[proper pivot]

Nvm

[rare goblet]

Ayo? @proper pivot level 1 !!!

[proper pivot]

Btw what's the difference between opt path and input path

How to use it

Thanks

[proper pivot]

I found it in weights

[proper pivot]

Help

[dense moon]

Does anyone here use a mac?

I’m so close to making my own model but for some reason its not letting me to train feature index or even train the model, ive done all previous steps and they were all fine, except that apparently it doesn’t detect my gpu.. even tho the feature extraction worked just fine

[proper shale]

well I don't believe you can train on Mac, even with m1 chips

[proud elbow]

I don't think there's decent support for training with mac GPU, the alternative is to train in colab cloud (RVC disconnected)

[proper shale]

it's a performance setting, uses more VRAM to train faster. However, it can affect the quality of the model in general, if the value is too high. It's a thing I recommend changing and trying out different values

[dense moon]

Ive tried colab but i keep getting disconnected from its runtime or smth 😭

[proper shale]

Probably using a restricted one then

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[proper shale]

Specifically RVC Disconnected works quite well for model training

[dense moon]

alright i’ll give it a try.. lyk if any error shows up

[rare goblet]

Ayo? @dense moon level 1 !!!

[dense moon]

In previous ones, the feature extraction didn’t work or the message thing wasn’t even popping up

[proper shale]

oh

btw there's a guide

i strongly recommend reading it

[dense moon]

I have watched like 5 tutorials in over a day trying to make mine

all run into errors.. just bcs im using a mac

[proper shale]

https://docs.google.com/document/d/1XuxQYiqEhYrdYeCZRRLrmV_ciMKo0bV-jTCGHu_-5Cc/edit?usp=drivesdk

this will help with the colab

[dense moon]

Alright2 ty

[proper shale]

yw :) good luck on the model!!!

[dense moon]

“Cache all training sets” is if my audio is under 10 mins right?

[proper shale]

don't think that's needed

[dense moon]

alrite

its running 🏃

i did 250 epochs

[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

[dense moon]

finished it! I got the the index and pth file into a zipped folder, now where can i test it? The rest of the steps is just about uploading it to the discord xd

[thorny musk]

i downloaded directml version

[marble owl]

Hello i have a question why my RVC web window will show error after maybe 3 or 4 times converted a file, it works after restarting but any idea why is my gpu full maybe?

[brittle wing]

ehh, what happen?

[candid dagger]

I want to sing a song a 50 cent, but the result is not as fresh as expected. I use crepe method but output is kind of glitchy

[proud elbow]

make sure the isolated vocal steam is clean, then use rmvpe

[molten pecan]

crepe has its pros but it's more prone to artifacting. When that happens switch to RMVPE

And yes make sure the vocals are clean and you use a WAV file

[warm shoal]

hello

[hearty pawn]

Hey @proper shale
We're trying to run exported ONNX voice models purely on web browsers, you can try the base LLVC deployment here: https://ort-onnx-poc.vercel.app/

I was wondering if it's possible to run the exported RVC models standalone (with minimal Python/pytorch) and if there are any scripts out there? as LLVC is almost impossible to train on custom voices and we're trying to plug in community trained ONNX RVC models instead, the closest i could find is https://github.com/isletennos/MMVC_Trainer/blob/main/onnx_export.py but it's a bit unclear how it would work on raw wav/mp3 files (as I'm not familiar with neural net architectures of these models)

would appreciate help with this, we also intend to open-source the resulting code so that in-game voice-to-voice in voice chats is able to supported by even indie games, would appreciate from anyone else here too

[stray acorn]

is there a way to add lyrics to a midi file with voice synthesiser only?

[proper pivot]

Help

[astral jungle]

Why not use Hina Mod AI Cover Gen

[proper pivot]

I can't

Always Error while using gradio

Help

[rare goblet]

Ayo? @proper pivot level 2 !!!

[finite galleon]

Did you run the first cell

[brittle wing]

yeah, nobody cares

[edgy hollow]

whats the most realistic female model? asking for a friend

[dense moon]

May i know where is a gui i can use to test my voice model?

[tame mica]

"asking for a friend"

[covert sedge]

is it possible that my covers sounds a bit off cuz i'm using RVC GUI?

[tame mica]

no
likely the model, pitch extractor or acapella

[covert sedge]

idk i've seen some messages here saying they GUI is outdated and not recomended anymore

i'll try with Original RVC then

[tame mica]

gui shouldn't matter iirc

[safe spear]

hey, i want to train a model on a 17 min audio's dataset, how many epochs do you recommend me to train it ??

[covert sedge]

how to run Original RVC

[fossil glen]

-local

[azure marsh]

Suggestions for @covert sedge

All Working Local Forks 🖥️

• Applio-RVC-Fork, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub

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, -colabs, -overtrain, -realtime, -rvc, -help

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[molten pecan]

-egirl

[azure marsh]

Suggestion for @edgy hollow

Searching for the perfect RVC e-girl voice model?

Look no further – we've got you covered! Our RVC e-girl voicemodel is in high demand, and we've streamlined the process to make it effortlessly accessible for you.

[molten pecan]

https://rentry.org/ilarvc_inf_guide use this website instead

[molten pecan]

staff approved

[fossil glen]

how can use rvcv3?

[molten pecan]

RVCv3 models are compatible with RVCv2

don't worry

@violet heron @half cove

[fossil glen]

i mean how to train

use rvcv3

[deft minnow]

@molten pecan acepta mi solicitacion plsss

[molten pecan]

@silver anvil is a beta tester

[molten pecan]

necesitas ayuda con RVC?

[deft minnow]

necesito hablar contigo

[molten pecan]

no hablo por MD

[deft minnow]

no, en verdad no, necesito de un argentino

[molten pecan]

para qué xd

[fossil glen]

its official or from ai hub engineer?

[deft minnow]

necesito hacer una compra en argentina, yo puedo pagarte para hacerla para mi

de verdad

[molten pecan]

yeach check their roles

[half cove]

don actually confuse people pls @molten pecan

[deft minnow]

estoy buscando argentinos por toda parte

[silver anvil]

what?

i made that

[molten pecan]

😭

[silver anvil]

what u need

[molten pecan]

lo siento pero yo ya no confío en hacer transferencias. suceden muchas estafas por Discord

[half cove]

ohh

[molten pecan]

tendré que decirte que no, por mi propia seguridad

[half cove]

its the egirl thing

[deft minnow]

comprendo

[rare goblet]

Ayo? @deft minnow level 1 !!!

[deft minnow]

perdon por ser invasivo

[silver anvil]

use gift cards

or paypal

[molten pecan]

no pasa nada. suerte con lo tuyo

[silver anvil]

u can also use steam money gifts

[molten pecan]

I have no clue how any of that works 😭

hey Squish can I DM you?

[silver anvil]

ye

[fossil glen]

is there a rvc v3 google colab version?

[brave garnet]

@silver anvil

Link utili a cura di Ilaria

⠀

• Link alle guide, programmi e consigli utili

Link

⠀

Made with love by Ilaria 💖

[silver anvil]

rvc v3 is not real

[molten pecan]

wait wtf did I touch

[inner hazel]

hey, so when i try to train with rvc, when i do the feature extraction for the pitch, it just stops at the 2nd sample

any idea?

im using colab pro+

[fossil glen]

just modified ver? its still based on rvcv2?

[silver anvil]

no, its not real

its not a thing

[violet heron]

Why are you paying for colab pro

You don’t need it

[inner hazel]

im working with my pc

[inner hazel]

so i have too keep all resources

[half cove]

i see alot of people who buy colab +

[fossil glen]

[violet heron]

But why did you pay for colab pro

[inner hazel]

and i use it for other stuff too

[half cove]

for some people its just more simple

[inner hazel]

^

[half cove]

than somthing like paperspace

or others

[inner hazel]

im used to it

[half cove]

you dont need to setup

[inner hazel]

so yeah thats the point im at now

[rare goblet]

Ayo? @inner hazel level 1 !!!

[violet heron]

It ain’t that hard to learn the other free colabs

[violet heron]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[violet heron]

Not even RVMPE

[inner hazel]

and it is getting stuck there

[half cove]

@inner hazel use the easygui

[inner hazel]

i only used so vits until now

[violet heron]

Your using a outdated colab

[violet heron]

Use “RVC Disconnected” for training a model

[covert sedge]

i didn't like Original RVC, now i'm trying to do cover on Huggingface Spaces version

hope it turn out good

[half cove]

do this

[half cove]

with paid u can use gui

[inner hazel]

okok let me look into that colab

[proven hill]

MERRY CHRISTMAS AI HUB!

Celebrate Christmas with the highest quality girl voices for the voice changer or regular rvc v2!

Soft Speaking Female: https://ko-fi.com/s/9c8a5b192e

American E-Girl: https://ko-fi.com/s/e21ec90d1b

Cute & Calm Girl: https://ko-fi.com/s/c8cf0c2129

And you will find many others here, FOR FREE!! https://ko-fi.com/ilariaowo/shop

Have a great Xmas Eve! :LC_WhiteCat5:

[inner hazel]

do the samples on easygui have to be i a zip or just lying around?

[molten pecan]

I heard paperspace is better

I have respect for paperspace

[half cove]

i never said better

i said more simple

[molten pecan]

colab is really cluttered imo 😭

[half cove]

colab+ has ui

[molten pecan]

@copper pebble oh yeah I do still use svc!

how did you know?

^^

[molten pecan]

NO WAY

aw man 😭

[molten pecan]

you always send this sticker 😭

[half cove]

#🔍│help-w-okada

[inner hazel]

doesnt look that good

on train index

[vagrant marsh]

The MS just increases and the voice quality doesn't get better. How can I fix it?

this is the realtime rvc

[proper shale]

using cpu

with crepe too

and 1 chunk

[vagrant marsh]

just 1 chunk works?

[proper shale]

...no

you need like 300 if u want something (maybe) usable

[vagrant marsh]

300 chunks?

I can't use my GPU bc it is GT 610

very incompatible

[proper shale]

yeah?

[vagrant marsh]

Same problem

[proper shale]

no gpu in the world can do 1 chunk

[proper shale]

try using dio as ur f0

[vagrant marsh]

okay

[proper shale]

tho i believe cpu will still be meh sadly

you kinda need a GPU for rvc stuff sadly

[vagrant marsh]

and the same problem stills

the ms of the res still bugged

[vagrant marsh]

oh, my...

and there is no free colab to do this

is there?

[proper shale]

...there is, check the pins in the #🔍│help-w-okada channel

[vagrant marsh]

okay

thx

[proven hill]

just to no clutter the channel with the wrong issues, next time use #🔍│help-w-okada, thanks for the collaboration!
(you can continue here for now, just to not lose the topic <3)

[vagrant marsh]

👍

[spring turtle]

any reason why whenever i run a conversion while using focusrite 2i2 as input for my dt990s after i click convert it makes all audio incoming inaudible and glitchy and stuttery, any fix for this?

[hidden prawn]

What program is that???

[proper shale]

W-Okada

[proper shale]

No clue

[lone sundial]

why is it saying this?

[proper shale]

Do you have some sort of log/Colab output we can see

[lone sundial]

one sec

here

[proper shale]

...this doesn't help

are u running locally?

[lone sundial]

no public

[lone sundial]

sorry im a newbie

[proper shale]

...so you're probably running on a restricted Colab

[proper shale]

it's ok, dw ;)

[lone sundial]

uh what does that mean😭

[proper shale]

so in September, some colabs got restricted cuz they had this GUI

[proven hill]

use Hugging Face

especially Ilaria RVC

[lone sundial]

ohhhh

[proven hill]

slower than colab but at least works

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[proper shale]

and basically they disconnect immediately

[proper shale]

Ilaria RVC the best on

[lone sundial]

ohh

[lone sundial]

ok thank you

[proven hill]

(not paid to say this)

[proper shale]

there's a guide for it, strongly recommend

reading it

[lone sundial]

where is it?

[proper shale]

yeah no! definitely not

k wheres the promotion

[lone sundial]

LOL

[proper shale]

https://rentry.org/ilarvc_inf_guide

[lone sundial]

ok thank you

[rare goblet]

Ayo? @lone sundial level 2 !!!

[proper shale]

yw :) good luck on the cover

[lone sundial]

thx🥹

[floral heron]

how do you get the applio uvr5 to work? The models it points to don’t exist

[proven hill]

use uvr5 locally, works better

[inner hazel]

on easygui i always get this error NameError Traceback (most recent call last)
<ipython-input-8-59a7acf51464> in train_index(exp_dir1, version19)
31 big_npy = (
---> 32 MiniBatchKMeans(
33 n_clusters=10000,

NameError: name 'MiniBatchKMeans' is not defined

During handling of the above exception, another exception occurred:

NameError Traceback (most recent call last)
1 frames
<ipython-input-8-59a7acf51464> in train_index(exp_dir1, version19)
41 )
42 except:
---> 43 info = traceback.format_exc()
44 logger.info(info)
45 infos.append(info)

NameError: name 'traceback' is not defined

any idea?

it happens on "train index"

[rare goblet]

Ayo? @inner hazel level 2 !!!

[inner hazel]

im using this easygui colab

"NameError: name 'MiniBatchKMeans' is not defined"; "NameError: name 'traceback' is not defined"

[proper shale]

i recommend just using rvc Disconnected

[shy zealot]

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

Ghost Mountain

Upload the <t:1686389063>
Server uploaded: AI Hispano
Channel uploaded: AI Hispano Post
Profile Owner: 8cta

Author

octa

Discord User

8cta

Internal ID

273

Language

ES 🇪🇸

Type

RVC V2

Size

0 MB

Extractor

PM

Epoch

500

Steps

0

Powered by ModelAI.net, Made with Love ❤️

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

Ghost from Call Of Duty: Warzone 2.0

Upload the <t:1690495063>
Server uploaded: AI HUB (RIP)
Channel uploaded: AI HUB (RIP) Post
Profile Owner: dreadvr

Author

.exodus

Discord User

dreadvr

Internal ID

2176

Language

FR 🇫🇷

Type

RVC V2

Size

113 MB

Extractor

N/A

Epoch

160

Steps

0

Powered by ModelAI.net, Made with Love ❤️

[half cove]

@brittle wing#🔍│find-models

[modern surge]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[thin sparrow]

for mac m1 which batch size did you recomend me?

[simple hornet]

Naruto I will end you in broly's voice

[violet heron]

[inner hazel]

btw, is 1700 samples too much ?

like

1 epoch = 2mins

[covert sedge]

idk if it helps

[inner hazel]

not really, i just started training

but i guess i have too much samples

[covert sedge]

what's your total samples length then

[inner hazel]

about 4h