✨│ai-help

[limpid silo]

15

im running it again and im on epoch 60

yet no pth file

i mean i see some pth files under logs

but i assume they arent it

huh

i will try those

uhh i am not sure what those are but i am just following the guide from the server

https://aihubdocs.github.io/en/rvc/local/mainline/

i think its just called RVC web-UI.. im not sure if I am even talking about the same thing

[glad zealot]

The actual model files will be in assets/weights

Kewk just need for it to finish training ig

Training all the epochs set

Or continue training using the same settings but different epoch count to something near the current epoch count on the console to get the model

Ye not sure how to know what epoch it was at

Ohh

Check logs/train log

train.log

It should show there what the epoch count was

[limpid silo]

i did not

[limpid silo]

i can try this

currently at 102 i will stop and restart with 105

[glad zealot]

Yup

You can resume

Just make sure all of your settings except for the epoch count is the same

[limpid silo]

well my problem is that there is no stop option

hence i cant resume

but once i restart it picks up from where i left on so

[glad zealot]

Yup

It will auto stop when it reaches the epoch you set it too

[limpid silo]

by looking at the log msg INFO:kurokumoFinal:Loaded checkpoint './logs\kurokumoFinal\D_6200.pth' (epoch 100)

that pth file might be a breakpoints..? idk

[glad zealot]

Wut

[limpid silo]

dw about it i was just reading through the cmd

[glad zealot]

That just means the last epoch saved/logged was 100 I think

Before you closed it

[limpid silo]

yeah i was wondering if D_6200.pth was the file that i can use

[glad zealot]

Nope

[limpid silo]

:L

[glad zealot]

The only usable file there are those in assets/weights

Everything else is only used for training

[limpid silo]

kk just 4 more epoch until it will be over

iwlyk if i see anything under weights

[glad zealot]

Yup if it reaches the last epoch it should save it in assets/weights

It should say something like "training complete, you can check the train.log" or something

[limpid silo]

lfg

[glad zealot]

Noice

You can close it now since it's saved

[brave garnet]

Voicechanger Settings (Okada)

⠀

Settings for Nvidia GPUs

F0 Det.: rmvpe (suggested for all series)

RTX 40-series: 80-96 chunk | +16384 extra
RTX 30-series: 96-112 chunk | +16384 extra
RTX 20-series: 112-128 chunk | +16384 extra
GTX 16-series: 128-192 chunk | +8192 extra
GTX 10-series: 128-192 chunk | +8192 extra

Advanced Settings

Protocol : Sio or Rest
Crossfade: 4096 start 0.2 end 0.8
Trancate: 300
Silencefront: Off
Protect: 0.5
RVC Quality: Low

⠀

Made with love by Ilaria 💖

[woeful tendon]

does anyone know how to fix web server looping

[rancid comet]

how to fix warning intaizling voice data?

[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

[spark kraken]

does it look like it's overtraining?

[rare goblet]

Ayo? @spark kraken level 1 !!!

[proud elbow]

what epoch number in 70k steps? anyway still you gotta test several checkpoints in RVC inference

[spark kraken]

900 epochs and 38 minutes dataset batch size 8

[candid sail]

I can't find my added_ .index file. But I have all the checkpoint .pth in weights folder. What should I do?

[rare goblet]

Ayo? @candid sail level 1 !!!

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[odd shale]

You forgot to press the "train Index" button

[odd shale]

I think it's overtraining.

[frosty shadow]

can anyone tell me how do i make my own voice rvc model
any youtube tutorial or guide?

[odd shale]

We got a plenty of guides you can follow.

Also, you wanna make a model of who? Yourself or a character?

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[molten pecan]

Yo creo que es mejor compartir los docs que usar ese comando

Es mucho más conveniente

Acá hice una guía acerca de hacer modelos de voz

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

[odd shale]

Aunque tampoco está mal resumirle a las personas textualmente el como hacer modelos.

Digo yo pues, lo que haría un Helper.

Así sea por amabilidad.

No simplemente estamparles un link a todas sus preguntas.

Aprende eso Julia.

[molten pecan]

Pero si mis docs ya está resumido y organizado

Y ese comando da 8 links diferentes. Un principiante se perdería en dónde empezar

[odd shale]

Yo sé, pero igual Julia.

[molten pecan]

Yo y el resto de los helpers enviamos ese link dando algo de contexto

[odd shale]

De hecho.. Eso es lo que traté de decir.

[molten pecan]

Solo digo, ten cuidado con enviar guías obsoletas

[odd shale]

Yo no tengo problemas con tu guia, tú eres fabulosa en todo lo que haces Julia.

Solo digo, tampoco es para responder con un solo link a todo lo que pregunte un usuario random nuevo.

La tortuga te lo dice.

🐢

No la lastimes..

[molten pecan]

No me lastimé, solo digo que no es conveniente envíar guías obsoletas

Y la mayoría del tiempo envíamos un link dando algo de contexto

https://aihubdocs.github.io/en/ here's everything you need to know

Además yo la verdad creo que al usuario le da igual si le das bastante contexto

[odd shale]

Yo también hago eso Julia.

[molten pecan]

Ellos vienen a hacer sus modelos y ya

[odd shale]

Almenos soy yo que me gusta socializar y explicar 😭

[odd shale]

Umm.. Yo había dicho "no lastimes a la tortuga" en broma, osea no lo dije en serio.

Ya que, bueno, me gustan las tortugas.

🐢

Pero si, usaré tus guias nuevas.

[quartz aurora]

hello

i am using mangio rvc to train voices,but it tells me i have no nvidia gpu,i have an amd gpu

what should i do?

[finite galleon]

Currently there are no rvc forks that supports amd gpu i am pretty sure

[sand iris]

Yeah they all dropped support at the same time lol

time to praise the🐧

[finite galleon]

Still wonder why

[sand iris]

DirectML is a piece of shit

It's like, a workaround at best

[heady hedge]

Wow, that's loud. Do it again

[sand iris]

?

[odd shale]

?

[tacit grail]

How do i fix this?

talking about at 1:22 and 1:47

[wintry palm]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[odd shale]

It can be fixed if you remove the backing vocals from the sample you used to infer.

That acoustic version of Lean Wit Me has backing vocals, so you must remove them.

[brittle wing]

am i better off running RVC online if i have a nvidia gtx 960

im having a problem with a model i want to run and according to the creator it works for them (they are running RVC locally)

anytime i try to download the model on any huggingface gui, theres an error

i even tried one of the OG colabs that end up disconnecting after 5 mins and other models work

[rare goblet]

Ayo? @brittle wing level 3 !!!

[brittle wing]

but not that one

(not my model I've just uploaded it to huggingface for personal use)

the og google drive link for the model

[brittle wing]

and here is the error i get from the colab

which may give a better idea of the issue

im essentially just familiar with RVC from the golden age of RVC before google sniped it

[tacit grail]

How do I remove backing vocals?

[odd shale]

You must use UVR.

More specific the Karaoke 2 or BVE models.

[limpid silo]

so i just successfully trained and tried my own model. but I found out that the model's voice sounds like shit during certain pitch. will training with different sample on top of my model possible fix the problem? can it make it worse? I might try to train my model more with the samples with the pitch that sounds weird rn

[tacit grail]

Is there a specific model to remove only the backing vocals? I am using uvr but it's separating all vocals from the song and putting all vocals together in one file

[odd shale]

Do you have the latest version of UVR?

You must download the BVE model in the VR arch model download tab

[tacit grail]

Do I use it on the song first or do I use it after I separate the vocals entirely from the song?

Or would that just add too much artifacting?

[tacit grail]

Ok, I'm using the BVE VR model on the unedited and vocal only version of the song and it completely ruins it

[fair ravine]

hey gentleman I don't find the "MODELS" folder to put the newest in after I installed MMVCServerSIO (edit: I created it but I don't know if is good)

and every time I say something to the model it sounds after the phrase like an echo or something else

I seen that I have to write in this part of the server

[small ivy]

When i try selecting my micro in the AI voice changer it says a “ Permission denied by system “ someone know how to fix it?

[short pecan]

attempting to redub/convert audio to sound like the selected voice that I made my own model of. However, it doesnt work and keeps barking out ERROR. Heres the cmd output to help troubleshoot. what do I need to do in order to fix?
user interface is MangioRVC btw.

heres the settings currently used

[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

[odd shale]

You downloaded the model?

[left dirge]

I got a problem, each time I download a voice model and use it, for some reason the lag it has its enormous, yet, any of the models that come with the voice changer by default work just fine

[odd shale]

@viscid moss Help him plz.

[short pecan]

thanks bot this tooooooooooootaly helps -_-

[serene wigeon]

?collab

[rare goblet]

Ayo? @serene wigeon level 6 !!!

[serene wigeon]

how can i bring up list of collab links

[viscid moss]

hmmmm idk how to fix that

[odd shale]

Me neither

[odd shale]

-colab

[azure marsh]

Suggestions for @serene wigeon

☁️ Google Colabs

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

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[serene wigeon]

dash, dang it

[odd shale]

Which are your PC specs?

[left dirge]

Ryzen 5 3600

32 Gb of memory

Amd Radeon RX 6800

[rare goblet]

Ayo? @left dirge level 1 !!!

[left dirge]

Should be enough to run the voice changer

[odd shale]

Yeah, it should.

-realtime

[azure marsh]

This interaction has expired, use the command /guides realtime if you wish to see it again.

[short pecan]

@helpers

[rare goblet]

Ayo? @short pecan level 5 !!!

[odd shale]

You can't do that XD

[queen meadow]

why i cant send pic

[short pecan]

it worked one time but eh oh well

[queen meadow]

[VCClient] wait web server...10 http://127.0.0.1:18888/
[VCClient] wait web server...20 http://127.0.0.1:18888/
[VCClient] wait web server...30 http://127.0.0.1:18888/
[VCClient] wait web server...40 http://127.0.0.1:18888/
[VCClient] wait web server...50 http://127.0.0.1:18888/
[VCClient] wait web server...60 http://127.0.0.1:18888/
[VCClient] wait web server...70 http://127.0.0.1:18888/
[VCClient] wait web server...80 http://127.0.0.1:18888/
[VCClient] wait web server...90 http://127.0.0.1:18888/
[VCClient] wait web server...100 http://127.0.0.1:18888/
[VCClient] wait web server...110 http://127.0.0.1:18888/
[VCClient] wait web server...120 http://127.0.0.1:18888/

[short pecan]

just tryina get attention to what i got going on here

[queen meadow]

i get this in the cmd

[left dirge]

Alright this helps. Thanks, this should fix the issue as I never exported it to onnx

[tacit grail]

Yes I did. I used the model and tried with the actual song and the song with just the vocals and it majorly messes it up. I tried using the karaoke model and I can't really tell a difference from that and the entirety of the vocals

[queen meadow]

anyone knows solution?

[odd shale]

Do you still have the original file of that acoustic cover of Lean Wit Me?

[tacit grail]

Yes

[odd shale]

Try to isolate it with any vocal model in UVR

But before that, enable the BVE model in Vocal Splitter options

[left dirge]

Alright I did this, completely fixed the issue, thanks again

[tacit grail]

So desperate the vocals from the song then separate the back vocals from that result?

[odd shale]

First, select your preferred model for isolation like MDX23C or Voc FT

Then right click and click on "advanced MDX NET settings"

[short pecan]

anyone have a solution to this? Id like to get started learning how to redub audio with RVC

[odd shale]

Then click on Vocal Splitter options, and put the settings like this:

[round crest]

does it exist some good and quick wait to filter audio files to get better quality running appolio to generate the vocals

[odd shale]

Sadly i don't know how to solve that.

[odd shale]

Nope.

[round crest]

ok thx

[odd shale]

You're welcome buddy

[quartz aurora]

what should i do when this appears

[odd shale]

Send full screenshot of the GUI

[left dirge]

Also I kind of forgot, how do I make it so I don't hear myself when using the voice changer? So I can use it without hearing a playback to myself

[short pecan]

its because your outputting it to your headphones or speakers

in order to effectively use the voice changer, use a virtual audio cable that takes the output of the voicechanger, and loops it back into a new input for where-ever your going to apply the voice to

https://vb-audio.com/Cable/

[tacit grail]

Uh ok, I'm in vocal split options. Now what?

[short pecan]

simply put, install this, then afterward put the cable "input" for your voice changers output, and the cable "output" for the input of the app you want to have hear the result, like discord or a video game.

[left dirge]

I'm already doing that

I have a virtual cable to discord

But it still loops back to me

[tacit grail]

Oh wait nvm hang on let me try it now

[short pecan]

may i see the settings for the input and output for your voice changer and your discord? screenshots would be nice.

[left dirge]

I can be heard on Discord with the custom voice without an issue, but it loops back to me

[deep bear]

why i cant upload model in google collab (im trying this 3-4 times) my internet is good

[short pecan]

https://tenor.com/view/the-rock-hmm-think-thinks-thinking-gif-12573421

Is there something that you have that would play it back like steam or something?

hmm.....

try looking at this perhaps?

[left dirge]

There shouldn't. I don't even have steam open up

I had to turn it down in the volume mixer to not hear myself

Also. Its not the default input and output either

But I keep hearing myself, the sound comes from the voice changer app

[rare goblet]

Ayo? @left dirge level 2 !!!

[left dirge]

Not from somewhere else

[rare goblet]

Ayo? @tacit grail level 4 !!!

[short pecan]

strange thing is, my VB is called line and then microphone for some reason, yet it works fine, no issues

i just assumed that its the same as input and output for what its called for others

maybe not?

[tacit grail]

I separated the vocals first then màde it separate only the main vocals and that's what the result is #✨│ai-help message

[odd shale]

Umm, these are the backing vocals.

Try again but select Lead Vocals

It should give you the backing and lead stems

[tacit grail]

It says "main vocals only" is checked though

[odd shale]

Try with Lead Vocals only

[short pecan]

@left dirge check the properties of the VB input and the output?

ex:

@left dirge is this enabled? i turned it on and applied and i can hear myself

[left dirge]

Huh, I'll check that

But it should not be

because I can't hear myself when I turn down the volume on the voice changer app

It is not

Properties of both are working correctly too

[tacit grail]

Lead vocals only result :/

[short pecan]

still though its an interesting clue

that its specifically coming from the Voice Changer Client Window and nothing else

this is all of my stuff here for the window

[left dirge]

This is all of my stuff here

[short pecan]

wait

the output under ServerIO Analyzer

change that to your headphones?

[tacit grail]

#✨│ai-help message I even tried it with and without the deverb setting and it's still only getting the back vocals not the main vocals

[short pecan]

cause mines on the headphones, never done anything with ServerIO nor do i know what it does, but maybe thats the culprit?

@left dirge

[prime rapids]

can someone pls help me every time when i try to train my model and i set the epoch number and i click "train model" it doesnt train. There are no epoches getting processed in the command bar.

[odd shale]

Just in any case, delete that sample.

[short pecan]

tryina help him while I hope for someone to help me with my woes

[odd shale]

Posting copyrighted samples of songs is forbidden.

[prime rapids]

can u pls help me

[tacit grail]

Oh wasn't aware, my bad. Was just trying to show what was happening

[odd shale]

I'll say that the problem you got is pretty weird, maybe you're not setting something properly.

I've never had problems with BVE on UVR

[odd shale]

Send screenshot

[prime rapids]

i tried but i dont have pic perms

[rare goblet]

Ayo? @prime rapids level 1 !!!

[prime rapids]

can i send it in dms

[odd shale]

Do it.

Habla en español.

Delete this sample too

[prime rapids]

oh it works now

i reinstalled it and it didnt help

[odd shale]

There are no pretrains selected.

🐢

Also, which is your dataset length?

You shouldn't put it on 1

Also, if your dataset is below 10 mins, check the "Cache all training sets" checkmark

[prime rapids]

oki ty

[prime rapids]

its below

10 minutes

[odd shale]

Check that then

[prime rapids]

ok ty

[odd shale]

And only check or "Save only the latest ckpt" or "save a small model at each checkpoint"

Not both or else your D and G files will bug out

[prime rapids]

ok tysm

[odd shale]

Being honest i don't even know how to solve that problem

[left dirge]

Could be the problem, but honestly, thanks for the help but welp. I can just turn down the volume in the sound mixer, I won't worry more about it

[azure marsh]

RVC v2 Tutorial for Paperspace

💡 Click on the button below to open the guide.

Credits: LollenApe

[short pecan]

np, tried my best

[tacit grail]

:/ well that sucks. It seems I can't progress unless I figure out how to fix these types of issues. If anyone can help me further then please @ me and I'll reply whenever I can. About to clock in for work in 30 minutes

[prime rapids]

im so confused

[odd shale]

Umm... It seems you commited the mistake of downloading the inference-only version of mangio fork

Download the Train-Infer one.

[prime rapids]

is it the one that opens the command thing and downloads the file through it

i used that one

oh nvm

oki ty sorry for being slow

[odd shale]

Download the full portable version of Mangio Fork.

[prime rapids]

ok ill try

[rare goblet]

Ayo? @prime rapids level 2 !!!

[left dirge]

Also, does someone know how counter strike handles sound devices? Cause there is no option to set up a sound device

[sand iris]

It should have the option to pick an input audio device now from what I heard

was added on the latest big update

[left dirge]

Really? I'll check again then. I did update the game

Yeah, alright then nevermind, Counter updated that, no more problems

My audio sounds chunky as fuck on counter but welp

[odd shale]

Another user had the same issue as you

[sand iris]

almost always a matter of putting CHUNK to a higher value to allow it to process in time

[gloomy maple]

with what do ppl make ai covers now

its was a google doc

then ilara

whats the easiest way?

[odd shale]

Ilaria RVC.

[gloomy maple]

where can i use that?

i had a file downloaded but that one doesnt work anymore

[odd shale]

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

Use this one that got an temp fix by Poopmaster

[gloomy maple]

is there like a how to use it somewhere

[odd shale]

Just execute the first cell

Wait till it finishes installing their things

Then execute the second one

And click on the Public URL that will get generated.

[gloomy maple]

alr tysm

[half cove]

main one should be workin

[gloomy maple]

how long does that normally take? its still downloading

[odd shale]

2 mins

[gloomy maple]

oh its done now alr

[dense sapphire]

does someone know why when i run the go-web.bat it opens, and then shows: C:\Users\user1111\Desktop\RVC1006Nvidia>pause
Press any key to continue . . .
and doesnt do anything after that

[candid dust]

need help with the elevenlabs plugin tts i cant find it in the drop down anywhere

i restarted applio after installing it

[dawn nebula]

did u download the righjt version btw

[viral cobalt]

What's the best RVC for training in terms of speed? [Ryzen 9 7500, RTX 4090]

[onyx lodge]

Hey maybe a dumb question but are there any specific settings/algorithm to use to make a model do better screaming vocals? Or will that just depend on the model itself?

[serene wigeon]

has anyone noticed some models take more than six minutes to upload to applio

[tawny oriole]

Hey, I made a zip folder, and how do I make ai covers with that?
(I wanna use it for my private)

[lean owl]

uh what is this?

[hot lagoon]

by "lead and then add em back together"
What do you mean by that?

[half cove]

just put em at the same time in like audacity

[lean owl]

what does this mean?

it happened all of a sudden

and i cant start the rvc disconnected colab

[timid flower]

How can i use rvc but without a gpu?

[odd shale]

You can use RVC without GPU.

But it will be only for inference and if you got a good amount of RAM

[fresh tendon]

why does the voice changer use so much cpu

[fresh tendon]

like how come fortnite only uses 4% of my cpu but the voice changer uses 70% 💀

[odd shale]

IDK

[fresh tendon]

is it a bitcoin miner

or smth like that

[odd shale]

XDDDDDDDDDD

[fresh tendon]

how is that a dumb question

[rare goblet]

Ayo? @fresh tendon level 1 !!!

[fresh tendon]

i mean it literally makes my pc loud asf

for a voice changer

[sand iris]

those figures seem exessive

[fresh tendon]

Nope

[sand iris]

only 4% for fortnite ?

hmmmm~2

[rare goblet]

Ayo? @sand iris level 72 !!!

[sand iris]

show your voice changer settings anyways

[fresh tendon]

it was 4 b4

but alr

same thing

[fresh tendon]

oh that makes more sense

cause of the chunks thing right

[sand iris]

depends

[fresh tendon]

70% still crazy tho

[sand iris]

kinda ? but that's not most of the CPU load

[fresh tendon]

wym

[sand iris]

the other settings are what affect CPU load the most

send a screenshot of your settings here pls

just to see if they're fine

[fresh tendon]

alr

cause the vc is so fun to use but that just kills it frfr

what do i change

its only when i turn it on so it must be the settings

[rare goblet]

Ayo? @fresh tendon level 2 !!!

[fresh tendon]

u think it could be the extra?

or crepe (idk what that is)

[sand iris]

extra is hella intensive on CPU

[fresh tendon]

ic the problem

[sand iris]

crepe is also really taxing to run, use rmvpe instead

• I don't see what the index slider is at, but if it's at non-zero it's also really CPU-heavy

[fresh tendon]

16k is fine but once i put anything above it jumps to 60%

[fresh tendon]

wym

index slider

[fresh tendon]

wont that ruin the voice

[sand iris]

no

if anything rmvpe is both faster and faster

[fresh tendon]

w

its kinda robotic with low extra

ngl i got scared when i saw that 70% but now its chill

w mans for helping

fixed it

[hushed lance]

How do I turn a zip folder into an ai cover link to use?

@molten pecan

[molten pecan]

https://aihubdocs.github.io/en/essentials/uploading-models-to-hugging-face/

[hushed lance]

tyyyyyyyyyy

[spare shale]

@viscid moss

[~] Converting voice using RVC...
gin_channels: 256 self.spk_embed_dim: 109
<All keys matched successfully>
Traceback (most recent call last):
File "/content/FIX/src/main.py", line 291, in song_cover_pipeline
voice_change(voice_model, main_vocals_dereverb_path, ai_vocals_path, pitch_change, f0_method, index_rate, filter_radius, rms_mix_rate, protect, crepe_hop_length, is_webui)
File "/content/FIX/src/main.py", line 201, in voice_change
rvc_infer(rvc_index_path, index_rate, vocals_path, output_path, pitch_change, f0_method, cpt, version, net_g, filter_radius, tgt_sr, rms_mix_rate, protect, crepe_hop_length, vc, hubert_model)
File "/content/FIX/src/rvc.py", line 150, in rvc_infer
audio_opt = vc.pipeline(hubert_model, net_g, 0, audio, input_path, times, pitch_change, f0_method, index_path, index_rate, if_f0, filter_radius, tgt_sr, 0, rms_mix_rate, version, protect, crepe_hop_length)
File "/content/FIX/src/vc_infer_pipeline.py", line 571, in pipeline
self.vc(
File "/content/FIX/src/vc_infer_pipeline.py", line 424, in vc
npy = np.sum(big_npy[ix] * np.expand_dims(weight, axis=2), axis=1)
IndexError: index -1 is out of bounds for axis 0 with size 0

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/content/FIX/src/main.py", line 346, in <module>
cover_path = song_cover_pipeline(args.song_input, rvc_dirname, args.pitch_change, args.keep_files,
File "/content/FIX/src/main.py", line 316, in song_cover_pipeline
raise_exception(str(e), is_webui)
File "/content/FIX/src/main.py", line 85, in raise_exception
raise Exception(error_msg)
Exception: index -1 is out of bounds for axis 0 with size 0
1

@~@?

[viscid moss]

Bad index. Trainded instead of ADDED

Which model?

[spare shale]

This is my Private model 😅

[viscid moss]

a

[viscid moss]

can u show me the files? U can blur the names

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[spare shale]

1 Minute

2 Files

[viscid moss]

That's the incorrect index

u need the ADDED one

[spare shale]

AAA xD

Yes I remembered

-1 Model

And I can’t use it without an Index?

[viscid moss]

Hmmmm I think not

I don't remember

[half cove]

u can use models without index

[spare shale]

?

[viscid moss]

But on CoverGen i'm not sure

[half cove]

is it different?

i dont see why it would be

[spare shale]

I know that this can be in EasyGUI

[hushed lance]

i need help

@molten pecan

[glad zealot]

Hard code? Idk I just make it work

[hushed lance]

so i just downloaded the local rvc thing but now what do i do after i go to go-web?

[molten pecan]

https://aihubdocs.github.io/en/

Here's everything you need

[hushed lance]

ik

i dont get it

[half cove]

[hushed lance]

i need your help

@molten pecan pl

[molten pecan]

Please don't beg for help and offer context

[hushed lance]

what????????

i have to pay to ask for help?????????

[molten pecan]

I said beg, not pay

[hushed lance]

im not begging

im asking for help....

[molten pecan]

Just tell me what you need help with

[hushed lance]

why cant you just help..

can i ss

[molten pecan]

Yes

[hushed lance]

kk

@molten pecan

what do i do?
after i clcik go web

it keeps showing that

[molten pecan]

Uhm wow okay

[hushed lance]

?

what

[molten pecan]

Seems like it's opening with the notes app by default

[hushed lance]

how do i make it open

[molten pecan]

That's not supposed to happen

[hushed lance]

ughhhhhhhhhhhhhh

im gonna cry
i just want to make songs

[molten pecan]

Google "how to reset open by default file extension"

[hushed lance]

😭

ok

[molten pecan]

It's okay if you are frustrated, just remember to not take it into this server

[hushed lance]

ok

[molten pecan]

We are human, don't get mad at us for things out of our control

[hushed lance]

ok

will you stay

[molten pecan]

I'm busy rn. @half cove can help you

[hushed lance]

@half cove help pls

[half cove]

It's 1am 😭

What u need help with

[hushed lance]

@half cove

it keeps doing this

when i click go web on rvc

[half cove]

Im not on pc rn

[hushed lance]

omgggg

[rare goblet]

Ayo? @hushed lance level 3 !!!

[hushed lance]

nobody can help

cries

cries 10x

[runic meteor]

anyone have a link to the colab for making covers

[glad zealot]

You can use it on Chromebook?

wut?

[hushed lance]

help meeeeeeee

[glad zealot]

Not sure how or if it even works on Chromebook locally

[hushed lance]

i miss the google share one

that was so easy

this is so complicated

[glad zealot]

Which one are you using?

The name of it

[hushed lance]

the file?

[glad zealot]

The folder ig

[hushed lance]

kk

@glad zealot

[glad zealot]

Hmm I think that's mainline

[runic meteor]

Do you have the link to the colab to make covers

[rare goblet]

Ayo? @runic meteor level 2 !!!

[glad zealot]

There's no public google version link for it atm

[runic meteor]

Oh

[glad zealot]

But rvc v2 disconnected should be working

[carmine hearth]

colabs #🤖│bots message

[hushed lance]

this is crazy

i just wanna make ai songs lmfaoooooooo

[glad zealot]

Ah you can just use covergen then or Ilariarvc

[hushed lance]

whats that

girl i havent used ai since august lol

@glad zealot

[glad zealot]

You can go to #🤖│bots and do -colab

Look for IlariaRvc or AICoverGen

[hushed lance]

ok

which one is better?

??

[glad zealot]

AICoverGen will try to automatically make a full cover with the background music included

[hushed lance]

i have isolated vocals already

[glad zealot]

Ilaria Rvc is only for vocals, must be cleaned

[hushed lance]

kk

[glad zealot]

Yup Ilariarvc is better then

[hushed lance]

kk

tyyyyyyyyyyyyyyyyyyyyy

ily

[glad zealot]

Np

[hushed lance]

also can yall unban my main

LMAOOOOOOOOOOOOOOOOO

@glad zealot

[glad zealot]

Ask admins for that :D

[hushed lance]

i think it got banned cause i was play arguing with someone

ok

@glad zealot

hey

um

why wont a huggingace link work when i keep putting it in for the voice model

is there a certain way i have to put it in?

@molten pecan

@half cove

hello...............

..

[lean owl]

is there a way to train models on huggingface

[odd shale]

No, there's no way.

You must use local, Colab or Paperspace.

[lean owl]

oof

wait whats paperspace

ive never heard of that actually

oh wait its paid

[half cove]

Send the link

[glad zealot]

Kaggle and saturncloud

[odd shale]

Saturncloud?

[glad zealot]

Ye

Basically kaggle but worse imo XD

[odd shale]

Why exactly?

XDD

[glad zealot]

Max 10gb storage, just 1 gpu

Cant use tunnels

:>

[odd shale]

I hope that page gets shut down.

[glad zealot]

Lmao

Well that's on free

[odd shale]

Nobody is crazy enough to use it.

[glad zealot]

At least you can use cli on it

Ssh I mean

[brittle wing]

Does anyone know how to make an RVC voicemodel? I have a bunch of voiceclips but I don't know how to make myself a model.

[molten pecan]

https://aihubdocs.github.io/en/ here's everything you need

[earnest marten]

There are very suspicious spike drops on both g and d at different points

[rare goblet]

Ayo? @earnest marten level 2 !!!

[earnest marten]

According to the guide, it could be mode collapse, here's the same graphs with 0.8 smoothness

[dawn nebula]

zoom out

[earnest marten]

What do you mean by zoom out?

Already had ignore outliers turned off

So it's basically showing everything in the graph

[silent sun]

Can someone help. Me

I can't open the app

I'm on mac os Catalina

It bounce and then gives me a error

[limpid spindle]

I do not even know where to begin to start on this stuff

[frigid pelican]

Well what are you trying to do?

[limpid spindle]

Well, I'm trying to use the John Ward voice from unholy trinity for fun and I have the voice file downloaded from this server but that's it

[frigid pelican]

Dude I had the same problem you have Right now! I solved it though and I can help

[limpid spindle]

Thank you, I'd love that

I don't get a single thing about this and I am not at all tech savy

[frigid pelican]

Soooo, you have both the pth file and the index file in your downloads right?

[limpid spindle]

No idea at all, I can direct you to the file that's in this server

[frigid pelican]

Yeah let's do that.

[limpid spindle]

Okay

John Ward (FAITH: The Unholy Trinity), rmpve
https://huggingface.co/MrCrimson/JohnWard/resolve/main/JohnWard.zip

Author: _mr_crimson (@willow shell)
Created: <t:1690503658>
Type: RVC
Algorithm: Rmvpe
Epochs: 300

https://cdn.discordapp.com/attachments/1134279298749042718/1134279298992316527/John_Ward.webp

This is the file I downloaded

I clicked the https file

[frigid pelican]

Ok so instead of clicking on the file have you tried copying the link?

[limpid spindle]

I just clicked the https and downloaded it, should I copy it?

[rare goblet]

Ayo? @limpid spindle level 2 !!!

[frigid pelican]

Don't click the link, if your on a phone, tap and hold on the link and copy the link

[limpid spindle]

Okay, I'm on pc so I dunno if it makes it any easier

[frigid pelican]

Thats ok, double click on the link then

Then copy

[rare goblet]

Ayo? @frigid pelican level 3 !!!

[limpid spindle]

Okay and what do I do with the link

[frigid pelican]

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

This is where we insert the voice into the song vocals or whatever

[limpid spindle]

Okay, so it'll do it that if I talk, it'll cover over my voice live?

[frigid pelican]

Not like that, this part is a bit tricky but don't worry its very easy when you know what to press

[limpid spindle]

Okay \

How can I make it to where it covers my voice as I talk?

[frigid pelican]

This voice converter is not made for live talking, its made for recording of voices or song vocal conversion

[limpid spindle]

Ohhhhh, I'm looking for live talking

[frigid pelican]

Oooooooooo

There's a site I know

But it ain't free

[limpid spindle]

Fuck man, that sucks

How much is it?

[frigid pelican]

Uhh 5to 300 dollars

Its called voice ai

[limpid spindle]

Oh hell nah, I'm a broke ass highschool student lol

Oh yeah I know about that I think

[frigid pelican]

You get like 30 seconds for free but it sounds like someone dying

[limpid spindle]

Damn

[frigid pelican]

Well that it, unless you wanna change song vocals I can't help ya

[limpid spindle]

Damn, well, maybe, I really do appreciate your help a ton

[frigid pelican]

No problem

[wet niche]

hey can someone help?

everytime i try to use anything that is marked as "rvc" it brings this error

Looks like there's a bit of a problem.
unknwon message

If you clear the information being managed by this app, it may be recoverable.

Initialize

Reload without initialize

Error
unhandledrejection
no error stack
TypeError: Cannot read properties of null (reading 'modelSlots')
TypeError: Cannot read properties of null (reading 'modelSlots')
at i (http://127.0.0.1:18888/index.js:2:1277059)

[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

[dawn nebula]

what software are u using

[trim isle]

can anyone lmk what kind of audio clips are good for testing voices? i'm refraining from using my own voice bc I have a very strong accent that messes up most models

[rare goblet]

Ayo? @brittle wing level 2 !!!

[hot lagoon]

Do you know where is the BVE.onx ?
My trying to replace the kar 2 on ai cover gen with bve

[molten pecan]

Try doing this

When doing inference, lower the search feature ratio to 0

[trim isle]

i've tried a few diff values including 0 and got no improvement on that front

[molten pecan]

@dawn nebula wait you can change the vocal isolation models in AICoverGen?

If so, you def need to change them in the prezip

It used outdated ones iirc

[molten pecan]

Dang. Then idk what to do

[trim isle]

i just wanna know where to go for like clean samples just for testing purposes

i've used songs with UVR but when i try like normal speech it's sloppy at best

[molten pecan]

Seems like you didn't train well enough

Did you use TensorBoard?

[cerulean hedge]

man when i try to put this link in ilaria rvc (web)

https://huggingface.co/Evoled/Bruucid_ZPZL/resolve/main/Lucid_PZPL.zip?download=true

it does not work

does anyone know

oh wait

it started working

wtf

[proper shale]

you do need to duplicate the space yk

[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

[finite galleon]

Vocft, kar2, and the foxjoy dereverb used on aicovergen are mdx architecture based, bve is vr architecture based so it isnt as simple as changing the models files

[odd shale]

Also, BVE is meant to be used in chain with an isolation model via Vocal Splitter Options

First isolation, then backing/lead separation

So, for now BVE is unusable in colab

[viral cobalt]

I asked this earlier with no luck, so sorry for spamming but: What's the best RVC for training in terms of speed? [Ryzen 9 7500, RTX 4090]

[rare goblet]

Ayo? @viral cobalt level 1 !!!

[odd shale]

Try with Mangio Fork or Applio 3.0.7

A new version of Applio got released

[bitter comet]

anyone know why when creating data sets and seperating the voice in uvr5 why there is some sort of static noise when removing the background noise?

[mortal stirrup]

How to open mangio hub i forgot actually

[odd shale]

Some vocal models from UVR tend to produce noise.

I suggesr you to use Melband Roformer and MDX23C on MVSEP

[proper shale]

open go-web.bat

[mortal stirrup]

Thanksss

[proper shale]

yw :)

[odd shale]

Mangio Hub?

[mortal stirrup]

bro i want to train a model

[rare goblet]

Ayo? @mortal stirrup level 1 !!!

[mortal stirrup]

how can i ?

[tropic flare]

@odd shale hello

[rare goblet]

Ayo? @tropic flare level 1 !!!

[tropic flare]

please

[odd shale]

🐢

[tropic flare]

I want a link to download the rvc program

[tropic flare]

@odd shale

I want a link to download the rvc program
Modern version
@odd shale@proper shale@proper shale@molten pecan

[odd shale]

Don't mass ping staffers, please.

[tropic flare]

im sorry

[odd shale]

You can download Mangio Fork ig.

[molten pecan]

please avoid pinging too much

i'll give you the link

1 sec

[tropic flare]

sorry

[tropic flare]

okay thx

[tropic flare]

i want download the rvc program

[molten pecan]

https://aihubdocs.github.io/en/rvc/local/mainline/

there's the link + usage guide

good luck

[tropic flare]

this rvc program?

[tropic flare]

I did not understand anything

I want to download rvc
just

[molten pecan]

the installation tutorial is in the page

please read thoroughly

[tropic flare]

This is the modern and developed version of rvc?

[barren forge]

Are there any Turkish voices on this server?

[molten pecan]

it's the original version of RVC

it's up-to-date

[tropic flare]

https://huggingface.co/RavenCutie21/OkadaVersion15Backups/resolve/main/MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.15.zip?download=true

this

why?

[molten pecan]

that's real-time voice changer

[tropic flare]

yess

[rare goblet]

Ayo? @tropic flare level 2 !!!

[tropic flare]

i want a

voice changer

[barren forge]

julia you can help me?

Are there any Turkish voices on this server?

[molten pecan]

then ask in #🔍│help-w-okada

[tropic flare]

here pl;s

[molten pecan]

there surely is. search in #1175430844685484042

or in https://weights.gg

[mortal stirrup]

anyway i can troll people with real time voice changing in valorant

[tropic flare]

#🔍│help-w-okada

[barren forge]

thanks buddy

[tropic flare]

@molten pecan@brittle wing

[molten pecan]

i don't know how to use the voice changer

this is not the channel for voice changer

please ask in #🔍│help-w-okada

[half cove]

raven deleted all her stuff i think

[molten pecan]

you are being irresponsible. don't make me mute you

[tropic flare]

why?

[half cove]

i dont know

[tropic flare]

How do I download it then?

voice changer program

[half cove]

just use the main download

[tropic flare]

how

Is there a difference between them?

[half cove]

nvidia download

[half cove]

i think ravens uses a old version thats more stable or somthin

[tropic flare]

what this

[half cove]

nvidia voicechanger download

[tropic flare]

this voice changer program The modern and improved version is better?

@half cove

[mortal stirrup]

does anyone have arijit singh model?

[half cove]

if you have nvidia gpu then download that

[tropic flare]

yes

[tropic flare]

https://huggingface.co/RavenCutie21/OkadaVersion15Backups/resolve/main/MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.15.zip?download=true
Was this version better or not?

[brittle wing]

how do I use D_(some numbers).pth models? there's no checkpoint models in my weights folder

[half cove]

it was more stable according to raven

[tropic flare]

I also want to Virtual Audio Cables Modern and developed version

[tropic flare]

Has it been permanently deleted?

and how raven

[half cove]

someone prob has it saved

[half cove]

raven is a ai hub user that left

[half cove]

use VAC by Muzychenko

[tropic flare]

what this

[half cove]

virtual audio cable

[tropic flare]

anyone?

[sacred sand]

I have the same question

[half cove]

i recommend the first one

[tropic flare]

What is the difference between them?

[tropic flare]

I want to use it on Discord and FiveM
@half cove

[half cove]

both work

ive just seen less problems with the first one

[tropic flare]

If I download first

@half coveWHAT THIS?

Is this version better than the second option? Healthy?

[half cove]

the download

[tropic flare]

OKAY

This is the best version of Virtual Audio Cables@half cove??

[half cove]

i told you to use VAC by Muzychenko....

both work

[proud matrix]

every time i try to create a cover using rmvpe it just says error every time

anyone know why?

[tropic flare]

this best version yes? @half cove

[half cove]

as i said before VAC by Muzychenko seems to have less problems but both will work

[molten pecan]

bnuuy

in which version of RVC?

[proud matrix]

im not 100% sure on the version I just used 1 click install on pinokio for it

[proper shale]

Next time don't ping every staff you've seen on the help channels, please.

[sand iris]

I dodged the mass pinging

epic win

[modest isle]

do i have to download python for the program to run?

[rare goblet]

Ayo? @modest isle level 1 !!!

[lapis geyser]

downloaded amd version of RVC and there's no go-web, is there no go-web for the amd version?

[sand iris]

Yeah they removed it randomly

[lapis geyser]

so there's no way now to create voice models for amd gpus?

[sand iris]

You never could create models with AMD GPUs on windows

you could only do voice conversion with existing models

[zinc cloud]

yo im using llaria huggingface and when i press convert it just says error

help plz

[lapis geyser]

Aight, thanks ❤️

[sand iris]

it's unfortunate

take a look at online options I guess

[zinc cloud]

plz

[lapis geyser]

I did check the one mentioned in aihubdocs, says you have to babysit the whole process which could take hours.

[rare goblet]

Ayo? @lapis geyser level 1 !!!

[lapis geyser]

I'm not that invested lol

[zinc cloud]

@sand iris

help

bro

[odd shale]

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

Use this colab.

[rigid dirge]

hey rencently moved to Applio where in the folder does the .index file go and the "voice model" section

[rare goblet]

Ayo? @rigid dirge level 1 !!!

[rigid dirge]

i mean here lol

[bitter comet]

hi i just looked into spek to see the khz and its at 15 i heard some people say multiply it by two which target sample should i choose 40k 48k or 32k

[glacial slate]

hello im new to this kinda thing so i downloaded rvc gui but it says that it doesnt support amd cards. is there any way that i can make my gpu work?

[rigid dirge]

no made it myself

basically this and so far this is all i found related to my model i made

(in the files)

the index apeared but the actual file didnt 😭

istg

your ment to train the model first right

or the index?

or does it not matter

?

do you not

💀

ok good

thats waht i did

let me get the python script

one sec

it just says this

[rare goblet]

Ayo? @rigid dirge level 2 !!!

[rigid dirge]

then this

ima try one think and if it dosnt work il come back

[quartz aurora]

does someone know what is the best version of wav2lip at the moment?

[rigid dirge]

wait

that works again

?

istg i remeber google blocking all that shit

😭

[quartz aurora]

the wav2lip? or what

[rigid dirge]

idk i remember a while back in the old ai hub server a bit before it got deleted rvc or the online google collab thing got wiped didnt it

thats why it became mostly local

[primal light]

Is it better to split up dialog for the dataset or to keep the entire dialog

[rare goblet]

Ayo? @primal light level 1 !!!

[odd shale]

Split it

[primal light]

👍

[rigid dirge]

for rvc is there a tutorial online

actually dosnt matter

im just stupid 😭

its fine i think i get it

its just like the old collab

sorry i havnt used colab since most of it got deleted 😭

google wiped that shit of the face of the earth istg

thanks tho

istg boykissers >>>

https://tenor.com/view/boykisser-meme-gif-6685457771534397088

[azure marsh]

Suggestions for @rigid dirge

☁️ Google Colabs

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

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[rigid dirge]

FUCK

no gifs

:<

ah well anyway femboys and boykissers> 😭

[proud elbow]

that should be fine for batch size 2-4, or try the original RVC which may be abit more lightweight

[rigid dirge]

ye il try collab thanks 😭

thank you pookie

[proud elbow]

still feasible for 5-10 mins dataset, also no time limit like in colab

[latent sand]

how to dowload rcv guı

[rigid dirge]

should i split them up btw

or just one big wav file

[proud elbow]

doesn't matter, as RVC preprocess automatically splits and trim silences

[rigid dirge]

oh really

thats actually really cool lol

thanks

[proud elbow]

literally recommended spec esp for batch size 8

[rigid dirge]

should i keep this at default btw of just does it depend

[rare goblet]

Ayo? @rigid dirge level 3 !!!

[low shard]

those are different type of pretrains, check out more about them here #📰│dev-updates message, #📰│dev-updates message. See which matches more your need

[bitter comet]

so i have like 15-20 minutes of data set i just trained 10 epocs and the voice sounds pretty good already its still kind of glitchy but that's it if i want to improve the quality will i have to train longer like more epocs or will i have to add more minutes to the data set?

[half cove]

if your dataset is over 10 min use original if its under use 0v2 and rin_3 is for clean datasets but i havent tested much

[proper shale]

yeah train for longer, 10 epochs is too soon to say that model is bad or good

[frozen sandal]

the applio rvc on google colab is extremely slow today for some reason, taking like 20 minutes to download a file, what is happening?

[odd shale]

RVC GUI?

[bitter comet]

alright thanks i already noticed at epoc 15 that the quality sounds pretty insane somehow the glitching occured cause i was still locally training the model so rn i don't have issues with the voice but the accent is the one that lacks currently and i believe training more will help that part out 🙂

[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

[spiral hamlet]

Does anybody know how to solve this??

[proper shale]

screwed up on dataset formatting

[rigid dirge]

hey i have the pth file for what i need you can see it and im using applio but where do i put it for it to apear in "voice model"

[proper shale]

pth in weights, but that one pth wont work bcuz its for training

[rigid dirge]

oh ok there was two pth tho let me find the other one

should i use this one?

[proper shale]

yeah this

[rigid dirge]

ight thanks that helps

[boreal axle]

Ik this sounds like a dumb issue, but I have everything setup right in rvc client but when i try enable the voice changer, I dont hear anyrthing, and the passthru option does make sound come through my cable, but its just my actual voice.

just jnoticed

for some reason my "vol" is always at 0

when everythings setup right

yep

when can I send screenshots? I dont have perms here

Can I DM you all of my settings?

[rare goblet]

Ayo? @boreal axle level 1 !!!

[boreal axle]

Alright

oh ye i just was able to

lol

all my settings

idk why it isnt picking up my voice at all

i did notice that it shows my volume changing once i enable passthru

but ofcourse that isnt my voice changer voice, just me

What chunk should I test with?

volume is still always 0

its just not picking up my voice which is weird

[unreal geyser]

you don’t have a monitor

your voice is coming out of the cable

[boreal axle]

im using voicemod to hear myself because im using a soundboard with this

[unreal geyser]

oh

[boreal axle]

either way its still not picking up my mic

like i said this is always stuck at 0

alr

ye

im using voicemod for a soundboard

still stuck at 0

[unreal geyser]

can you hear yourself when passthru is enabled

[rare goblet]

Ayo? @unreal geyser level 1 !!!

[boreal axle]

yes

which is weird

but as soon as its off my vol is set to 0

wait i dont even hear myself with passthru on server anymore

[unreal geyser]

weird

[boreal axle]

uhhh

this doesnt seem right

💀

[rare goblet]

Ayo? @wraith kayak level 4 !!!

[boreal axle]

uhhhhhhhhhh

should i try reinstall rvc

already tried

[unreal geyser]

yea

reinstall

[boreal axle]

na im gona try a reinstall first

seems faster than just debugging something that might not even be fixable

[half cove]

delete the pretrain and model_dir folders and restart okada

[boreal axle]

im just reinstalling the whole thing

bro its the first thing i thought of

😭

well no not really

its just'

easier

than

debugging

[half cove]

ok

[sand iris]

I mean we've got pretty much all issues troubleshooted already

[proper shale]

unstable training happens when you have a big dataset and small batch size

id cut down to 10-20 mins and train on batch 8

[half cove]

also make sure its clean

also

that doesnt seem to be finished training yet

[dawn nebula]

pick the best ones in ur dataset and limit to 10 minutes

yeah

[half cove]

you mean a one of the new pretrains?

[proper shale]

eh, not really

[half cove]

id use ov2 if your dataset is shorter than 10min

if your talking about this please use it 😭

[odd shale]

You did make sure to truncate silences?

[sharp zodiac]

does anyone know how to use the realtime ai voice?

[odd shale]

-realtime

[azure marsh]

This interaction has expired, use the command /guides realtime if you wish to see it again.

[odd shale]

There you go

[sharp zodiac]

thanks

[odd shale]

You did make sure to also delete background music?

Or background noise?

[half cove]

id use mdx23c

for removing music

and if anything remains use melband roformer

[odd shale]

I suggest you to use MDX23C and Melband Roformer

[proper shale]

just keep em to a minimum, like 0.125 is ok

[hexed cave]

Hello, am trying to merge Hu Tao's ai voice model with the cupid (twin version) vocals. But the whispers and background noises make it sound weird

[full marsh]

-colab

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[hexed cave]

its very breath-y, and its not smooth

like the whispers make it sound like gibberish

[sharp zodiac]

i cant figure out how to download it

[hexed cave]

if anyone has tips on polishing ai covers then let me know

[rare goblet]

Ayo? @hexed cave level 1 !!!

[odd shale]

You must obviously isolate the lead vocals first

[hexed cave]

yeye

I converted the Hu Tao model with the cupid vocals

but it came out breathy and not smooth

bc of the whispers n background voices

this si what I mean

mhm

I got the vocals from the album version of the song

alright

https://www.youtube.com/watch?v=dc3eD6C6Gbk

this is what she sonds like

for the voice model

just using this as reference

alrighty

god damn

I can't find a clean audio where it doesn't include the background vocals

[odd shale]

use MDX23C

[sharp zodiac]

my rvc realtime isnt working, is there a fix for it

[rare goblet]

Ayo? @sharp zodiac level 1 !!!

[hexed cave]

I tried ballin' its not bad

I found a cover

where they only sing the main lurics and not the background.

https://www.youtube.com/watch?v=BKsa4alYTGU&t=0s

[hexed cave]

ouh-?

oh-

im using a site to help me

[rare goblet]

Ayo? @hexed cave level 2 !!!

[hexed cave]

heard about uvr before

I just hope