✨│ai-help

[worthy jasper]

it wont work on discord ?

[velvet eagle]

bro idk how to activate

[rare goblet]

Ayo? @velvet eagle level 1 !!!

[velvet eagle]

i extract it

how?

[bright meteor]

Hello!

[velvet eagle]

hi

[worthy jasper]

create a folder on ur desktop

[velvet eagle]

alr

[worthy jasper]

called ai mods or whatever

[proper shale]

yeah but u gotta mess with those

[bright meteor]

I just wanna know, what is the minimum MacOS version for Wokada?

[proper shale]

lol

[lethal sky]

which settings

[proper shale]

tune

[worthy jasper]

extract it into tha folder

[proper shale]

https://rentry.co/VoiceChangerGuide yall

[proper shale]

also dont make folders with spaces in their names

breaks shit

[proper shale]

u should use Ventura at least

[velvet eagle]

done now?

[proper shale]

@velvet eagle

[velvet eagle]

thx

[proper shale]

yw :)

[bright meteor]

alright thanks

[velvet eagle]

idk but do i extract all those zips that sayyyyys?

[violet heron]

Yes

[molten pecan]

😭

[velvet eagle]

ik how bruh

alr

[haughty pulsar]

[primal vector]

I tried posting a model of Isabelle from "Isabelle ruins everything" and it got deleted for "poor quality." don't know why cause the voice demos I used sounded fine and the model was 500 epochs. Don't know if the bot just glitched or I did something wrong.

[half cove]

It says who to contact for any questions

[primal vector]

got it

[velvet eagle]

does anyone have a yt tut

[proper shale]

They're all outdated really

So it wouldn't really help

[haughty pulsar]

I'd say the guides are pretty straightforward without videos

[proper shale]

^

[haughty pulsar]

It even has images

[snow finch]

i search this ai model https://www.youtube.com/watch?v=7HgWUobYjHU help pls

[molten pecan]

https://rvc-docs.github.io/Docs/ here's a tutorial

Four methods to search voice models

[velvet eagle]

nvm got it

[snow finch]

but it's not available anywhere except on app.kits. But there's probably a fee.

[violet heron]

Check weights.gg

[molten pecan]

[snow finch]

People on this site are just asking to make this model

[rare goblet]

Ayo? @snow finch level 1 !!!

[violet heron]

No that’s where the models are hosted

You click the 3 dots and click download model

[snow finch]

It's not there, I checked everywhere, but youtube has a lot of ai covers, maybe it's private but I don't think so.

[violet heron]

What’s the model name?

[snow finch]

Yung Trappa

[violet heron]

Do they go under any other name?

[snow finch]

No

[violet heron]

Like real name???

Also the person you requested the model of might have requested a takedown

That might be why

You can always request for a model in #1159289738314919936

[snow finch]

I'll give it a try, thanks.

[hexed falcon]

somebody has a youtube like

montanablack

eligella

trymacs?

fibi

[violet heron]

Huh

Are you looking for voice models of those people?

[hexed falcon]

voice model

yes

sir

[violet heron]

Check #1175430844685484042 or the website weights.gg

[hexed falcon]

thanks

[hexed falcon]

can you help me search pls

[rare goblet]

Ayo? @hexed falcon level 1 !!!

[violet heron]

Why

[hexed falcon]

pls

i ned Montanablack

[violet heron]

You just click the search icon and type it in

[hexed falcon]

Trymacs

Can you give it a try

[violet heron]

You can

[half cove]

The file was too large

[rotund prawn]

How the hell
45s

Should I do 5 epochs?

[proper shale]

just use the latest version

lol

[rotund prawn]

lol

[modest drum]

oh my GOD

it ran out

im gonna scream

i dont like this game anymore

[haughty pulsar]

[frail plover]

I'm using RVC v2 Disconnected to basically clone my own voice for a school project. I'm following the "Create a model with RVC disconnected" guide that i found here on discord. I am using one audio file of 1 hour 30 minutes of good quality with no background noise or reverb. The guide told me i should set the total epochs based on dataset length, but i just put it at 500 since i saw other voice models were generally also around 400 / 500 epochs. I've let it run for like 2 hours 35 minutes, and its only at epoch 80 so far, with it doing one epoch ever 2 minutes roughly. So doing the whole 500 epochs is like 16 hours? Is that normal? should i go do the whole 500 epochs? and if i want to stop it, test it, and then maybe continue training based on the results how do i do that? right now it automatically saves every 30 epochs or so, i assume when it saves it also saves to my rvcLogs.zip file on google Drive, although im not sure

[molten pecan]

https://rvc-docs.github.io/Docs/ check the training guide

Everything you need to know is in there

Epochs, overtraining, re-training, etc

[frail plover]

thanks!

[molten pecan]

yw! Any suggestions/questions you lmk

[frail plover]

will do 🙂

[dire panther]

hi

i need realy help

real time voice changer client is safe i dowload but it isnt works

[spare shale]

Have you tried to Run it through Administrator?

[dire panther]

no i uninstal

[rare goblet]

Ayo? @dire panther level 1 !!!

[dire panther]

is it safe

im wory about have a virus in it

@spare shale

[spare shale]

Nah

[dire panther]

W-OKADA

named

[spare shale]

Seriously Virus?

[dire panther]

idk

[spare shale]

Now W-OKA I use a lot of people, I don’t think it exists

• You can Check on sites (Which I don't know)

[haughty pulsar]

W-okada is not a virus

[dire panther]

thanks lot

[haughty pulsar]

All the code is open source

It's public

[dire panther]

oh

[spare shale]

Yes W-OKA It's Open Source

I really don't use it 😅

[dire panther]

Thank you so much i relaxed

[spare shale]

I don't use it because I don't need it 2. It won't work on my laptop

[spare shale]

https://media.discordapp.net/attachments/1167948353833816154/1194044119706767500/gif_make_mikus_18.gif?ex=65aeeb3e&is=659c763e&hm=ab3f85fa1a1e7c023fb7d5569a94bbe1cdbabff67495f9f4aec517f0e1a236c2&

[placid storm]

can someone tell me how to install applio v2
?

[modern epoch]

anyone know the directory to put models?

[proper shale]

pth in weights, index in logs

[modern epoch]

i dont have a folder named weights

[proper shale]

maybe its inside assets?

[modern epoch]

just making sure version 1.5.3.17b is the newst right

or should i use the google colab ver

[proper shale]

oh nvm ur using W-Okada

https://rentry.co/VoiceChangerGuide

explains everything you need to know

and model_dir is the folder with ur models btw

[brittle wing]

Uhm I hate to ask but can someone send me a YouTube video or something on how to use the AI voices and how to set it up!

?

[modern epoch]

why what are you using

[proper shale]

nah i thought u were using some local rvc fork

[proper shale]

https://rvc-docs.github.io/Docs/

[brittle wing]

THANK YOU! ^^

[rare goblet]

Ayo? @brittle wing level 1 !!!

[proper shale]

You're welcome :)

[brittle wing]

oh uhm do i have to download a rvc?

to make an ai cover?

[violet heron]

You can use it via a HuggingFace space or google colab

-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-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

[violet heron]

https://rentry.org/ilarvc_inf_guide here’s the guide for Ilaria RVC

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb here’s the applio colab too

[brittle wing]

ahhh i dont think i can do this without a tutorial, is there like a video i can watch so i can do this step by step?

this is so confusing

[snow hazel]

I'm trying to make ai covers on mac but it says uvr wont open because apple can't check for malicious software.

[violet heron]

https://rvc-docs.github.io

[violet heron]

Not a Mac user but there should be able to open it anyways

[snow hazel]

Nope

[brittle wing]

site not found

[snow hazel]

anyone know what to do?

[molten pecan]

https://rvc-docs.github.io/Docs/ my website is much more convenient

Copy it to your clipboard

[snow hazel]

I'm not getting information

I did the one that said intel

[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 graphics card do u have

[snow hazel]

Intel HD Graphics 4000 1536 MG@dawn nebula

[dawn nebula]

use colab

do u need training or inference

[snow hazel]

No I just want to make Ai covers

Also How do I do colab?

[molten pecan]

https://rvc-docs.github.io/Docs/

Here's everything you need to know

[snow hazel]

What about mangio

[molten pecan]

Yes. Be sure to read before asking obvious questions pls

[snow hazel]

I just don't know what to do about UVR

[half cove]

https://rvc-docs.github.io/Docs/vocal-isolation--datasets/uvr5--mvsep/

[molten pecan]

Why do they keep doing this

[half cove]

yes

[molten pecan]

They just ask instead of taking 3 seconds to use the search tool

[half cove]

its ai hub users

[molten pecan]

[half cove]

you cant expect much

[molten pecan]

😭

[half cove]

[snow hazel]

what about for mac?

[half cove]

uvr on mac?

just use mvsep

its part of the guide

[snow hazel]

It said on the webaite some installers foe mac for UVR

[half cove]

you could do that

then ig

idk about uvr on mac

i guess it should work actually

[spare shale]

I need Help T-T

[viral mason]

disconnect will not work

what did I do to deserve this

[spare shale]

How can I describe this Picture?! What kind of Pose is this?!

[viral mason]

looks like he's trying to show you something cool

[half cove]

make sure your dataset name and experiment name are the same

[viral mason]

is that a new update?

[spare shale]

M.. What do call this pose?!

[half cove]

no?

[viral mason]

it's never been like that before tho and worked fine

[viral mason]

this is what I'm trying to train

[snow hazel]

I just don't know what one to do.

[rare goblet]

Ayo? @snow hazel level 8 !!!

[storm kernel]

I don't find the link/guide to install RVC locally with an AMD gpu can someone help me please?

[half cove]

it says on the guide tho?

[viral mason]

I think I found a trend?, 0 second clips are untrainable

[snow hazel]

It just says that it's avalible on mac.

[molten pecan]

#🔍│help-w-okada

[molten pecan]

https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/RVC1006AMD_Intel.7z

https://rvc-docs.github.io/Docs/

[molten pecan]

Here's a complete guide for RVC if you want

[storm kernel]

thank you

[snow hazel]

There is 3 intalls for mac. I don't know which one will work.

[brittle wing]

idk why i hear an echo but it's like whispers

[storm kernel]

Is there a specific python version I have to install?

[rare goblet]

Ayo? @storm kernel level 1 !!!

[brittle wing]

how do i solve this?

[molten pecan]

Just latest one, to my knowledge

[molten pecan]

In voice changer?

[storm kernel]

I have a python error when I start the go-web file

[brittle wing]

yes

[molten pecan]

#🔍│help-w-okada

[molten pecan]

Odd. Are you executing as admin?

[storm kernel]

no

when I start it as admin it says path not found

[brittle wing]

ty

[rare goblet]

Ayo? @brittle wing level 2 !!!

[molten pecan]

https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/RVC-beta.7z use this one instead

Idk why the AMD one is broken

I keep forgetting that

[storm kernel]

I'll try

[haughty pulsar]

No?

For nvidia should be https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/RVC1006Nvidia.7z

[storm kernel]

I am asking for AMD

and it worked Julia thank you

[haughty pulsar]

https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/RVC1006AMD_Intel.7z

[storm kernel]

this version does not work she said

[haughty pulsar]

What

Now I'm confused

[storm kernel]

I have a python error when I start this go-web file

[haughty pulsar]

What did you unzip using?

And in what folder

[storm kernel]

winrar

[storm kernel]

It's in a separate folder I did not move anything

[haughty pulsar]

I just use 7zip

Maybe they broke it or something

[prisma crag]

where do i find this part?

[hot lagoon]

Well need more training or nah?

[prisma crag]

my models arent loading

colab version

[violet heron]

Click upload

[prisma crag]

yeah i did that

when it finishes

it just

doesnt appear

[supple wasp]

how do you use the voice models

i am so confused

i need help

:(((((((

[onyx shale]

Use RVC-GUI

Then import the zip

[supple wasp]

rvcgui???

[onyx shale]

Yeah you can download it if you Google it

[supple wasp]

ill try thank you

[onyx shale]

Np

[violet heron]

For what?

[violet heron]

It’s outdated

Don’t use

[onyx shale]

No it's still good

[violet heron]

Doesn’t even have RVMPE

And you can’t even train

[onyx shale]

The average user doesn't need that stuff

[violet heron]

Also lacks some other options

[violet heron]

RVMPE is the best pitch extraction for covers

[prisma crag]

sooooo how do i use load in the models

[violet heron]

It will sound better compared to harvest, crepe, etc

[supple wasp]

just for like voice trolling on roblox

[rare goblet]

Ayo? @supple wasp level 1 !!!

[violet heron]

You would use W-Okada

[supple wasp]

and vr chat and stuff

[violet heron]

-realtime

[azure marsh]

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

[violet heron]

Read the first one

[onyx shale]

But how do I MAKE a voice model? All I know is that I need a ZIP of the person's voice files, which I have. So now what?

[rare goblet]

Ayo? @onyx shale level 1 !!!

[violet heron]

Don’t use RVC-GUI

You can’t make voice models with it

[onyx shale]

So how do I make the model?

[violet heron]

What’s your GPU?

[onyx shale]

Like computer GPU?

[supple wasp]

i am still extremely confused

[violet heron]

Yes

[violet heron]

https://rentry.co/VoiceChangerGuide

Guide for it

[onyx shale]

Nvidia RTX

[violet heron]

Tells you everything you need to do

[violet heron]

Yeah but which one

[onyx shale]

4080

[violet heron]

Ok you can locally train

-local

[azure marsh]

All Working Local Forks 🖥️

• 🍏 Applio, 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
• GPT-SoVITS, by RVC-Boss GitHub

Need some help? 🤔

You can find more info on the "❔┋guides" channel. If you can't find your answer, feel free to ask for help in "help-rvc". Credits to Faze Masta and Antasma for compiling these links.

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

[violet heron]

Download Original RVC, if you need TTS use applio instead

[onyx shale]

How do I make it though?

[violet heron]

-rvc

[azure marsh]

Suggestion for @onyx shale

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)

[violet heron]

https://rentry.co/TrainingVoiceModels

[onyx shale]

Can you just use your words? It's hard to understand

This stuff is hard to understand

[violet heron]

You go to the train tab

You input the dataset location

You set epochs to 500

You set it to rvmpe

You set batch size to like 6

You click train model

Then you click train feature index

[snow hazel]

What do I do?

On mac btw

[violet heron]

Before this you process your dataset

[onyx shale]

The train tab where?

[violet heron]

There is different tabs on RVC

[prisma crag]

can someone help me out my uploaded models arent appearing

[rare goblet]

Ayo? @prisma crag level 2 !!!

[violet heron]

“Inference” “Train” etc

[onyx shale]

I thought you hated RVC

[violet heron]

RVC-GUI

[onyx shale]

Ohhhhh

[violet heron]

which is a very outdated fork of RVC that uses a weird ui

[onyx shale]

I see 🍴

So where's the original?

[violet heron]

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/releases

Download the “For Nvidia GPU users:” one

[scenic gull]

is this usable?

[dawn nebula]

is ur soothing 0.99+

[dawn nebula]

keep training

[scenic gull]

this is if i set it to 0.987 smoothing

[violet heron]

Extract everything and drop this file in the folder

And run it

[scenic gull]

and ignore outliers

[violet heron]

They stopped giving it automatically for some reason

[snow hazel]

Why does ny mac not like UVR?

[violet heron]

#1192011222023950368

[dawn nebula]

keep training

[dawn nebula]

use mvsep.com

if u register, the queues will be a lot quicker

[scenic gull]

thanks

[snow hazel]

I've used it before and I don't likw the quality it gives out.

[dawn nebula]

u probs pick the wrong models

[hot lagoon]

keep going?

[storm kernel]

I get this error message while downloading the model in RVC colab, but the link is set to "anyone with the link" and it's not private can someone help me please ?

[rare goblet]

Ayo? @storm kernel level 2 !!!

[storm kernel]

It was working before but it's not recently

[scenic gull]

i accidentally did train model before train feature index what do i do

[digital osprey]

help me, please

how to make own voice model?

[violet heron]

-rvc

[azure marsh]

Suggestion for @digital osprey

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)

[scenic gull]

i accidentally did train model before train feature index what do i do

[rare goblet]

Ayo? @scenic gull level 3 !!!

[violet heron]

Your fine

[scenic gull]

ty

-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)

[scenic gull]

-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-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

[molten pecan]

https://rvc-docs.github.io/Docs/

[timid valve]

would you recommend using pitch guidance when training models?

the data set is literally just 30 minutes of someone talking

so no singing or anything

[molten pecan]

Yeah it's always recommended

[timid valve]

alrighty, thanks 👍

[molten pecan]

yw!

[wise wave]

What the link to dolwend it

the voice changer

[molten pecan]

-okada

-realtime

[azure marsh]

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

[molten pecan]

Maybe it's in the link

[dawn nebula]

bro i think there pitch guidance option should be eliminated it should be automatically on

[molten pecan]

Or the model architecture one 😭

[dawn nebula]

yeah defs

then only rmvpe and rmvpe_gpu

[timid valve]

is the latter one more optimised for gpus?

[scenic gull]

local training is suddenly taking a lot longer, what could cause this?

did i just leave the pc on for too long?

[molten pecan]

Yeah it uses more GPU power

Therefore a faster process

It's available for training only

[timid valve]

👍

[rare goblet]

Ayo? @timid valve level 1 !!!

[timid valve]

thank you

one last question

what batch size per GPU do you recommend? running on a RTX 3070 with 8GB of VRAM

[molten pecan]

You use it according to your VRAM, yes

8 in this case

Except for short datasets (around 2 minutes)

In those cases use 4

[timid valve]

you are being too helpful, thank you very much

[molten pecan]

yw! Feel free to ask as many questions as you want

[timid valve]

ok got another one for ya

33 minutes of clear audio. do i need more than 300 epochs?

[molten pecan]

Very likely yes

[timid valve]

i know these are the most basic questions but i wanna hear it from an expert

[molten pecan]

But remember this

There isn't a way to know the right amount of epochs before training, as it depends on the length, quality & size of the dataset

It's best if you use TensorBoard. A tool for training AI models that can help you determine when to stop training

[timid valve]

I see

[molten pecan]

https://rvc-docs.github.io/Docs/ I included it in my website too

Feel free to check it out

[timid valve]

🙏

ohh i've seen people post those graphs

well now i know where they originated from

thanks again, super helpful stuff

[molten pecan]

Yup. It may look a little tricky but it's mostly sitting down and monitoring it

Have fun!

[timid valve]

https://cdn.discordapp.com/emojis/847563394314665984.webp?size=48&name=froggo&quality=lossless

[crude nimbus]

hey there! im running RVC off of a local install but when i try to use rmvpe for pitch extraction, i just get a connection error. It works perfectly fine with pm which is what im confused about. Im using a radeon rx 5700 xt if that helps.

[timid valve]

ok sorry to bother you one last time @molten pecan, but what's the difference between these two?

[molten pecan]

I don't train locally so I am not 100% sure. The second one seems to be useless, so use the Train model one

[molten pecan]

Oh no, I have bad news for you

AMD is not well compatible for AI training. Only NVIDIA

It's best if you train through the cloud

https://rvc-docs.github.io/Docs/ check the RVC Disconnected page

[crude nimbus]

oh i apologize, im not doing any training, just vocal pitch replacement. is rmvpe not supported with amd?

[violet heron]

RVC only uses CPU for inference on AMD (at least for windows and .pth)

[crude nimbus]

ohhh i see. so if i wanted to use my gpu for inference, would i just be better off using pm?

[rare goblet]

Ayo? @crude nimbus level 1 !!!

[violet heron]

I’m saying it will not use your GPU no matter what

I would still use RVMPE

[crude nimbus]

rmvpe gives me a connection error even just for inference though. how would i fix this?

[violet heron]

Send a screenshot

[crude nimbus]

heres my terminal and webgui

[violet heron]

Never saw that error before

Make a post in #1192011222023950368 and hopefully someone can help you with it

[crude nimbus]

got it. thank you for the help!

[dawn nebula]

yes

[fathom hearth]

can you import local voices?

[violet heron]

Yeah

.pth file goes in weights folder

And .index goes in the logs folder

[fathom hearth]

dont know what any of those are but thank you minecraft man

[timid valve]

would training a model on 300 epochs be the same as training two separate models using the same dataset with 150 epochs each and then merging the two?

[rare goblet]

Ayo? @timid valve level 2 !!!

[dawn nebula]

no

[winged skiff]

erm

what would be the best settings for me
I have a RTX 3050 Ti with 16 gigs of ram

[fathom hearth]

downgrade to the gtx 420 for the best performace

[timid valve]

i figured, thank you

[dawn nebula]

no worries

[winged skiff]

oh fr

[rare goblet]

Ayo? @winged skiff level 1 !!!

[fathom hearth]

fr fr

[rare goblet]

Ayo? @fathom hearth level 1 !!!

[fathom hearth]

100% real

no but fr, how do you upload your own local mp3s to model or create model with mp3

[violet heron]

Train section

[fathom hearth]

train section?

[violet heron]

If your using a local installation of RVC

there should be a ‘Train’ tab next to ‘Inference’

[swift merlin]

Does anyone know what setting in RVC Inference helps with voices fading in and out? Almost sounds like a heavy compressor on the vocal

[molten pecan]

Volume envelope

https://rvc-docs.github.io/Docs/

Check the inference settings page

[soft plover]

if my g/total starts going up but the d/total and mel is still going down do I stop training or do I keep going?

[swift merlin]

Thank you!

[dawn nebula]

keep going

wait

[soft plover]

o ok ty

[rare goblet]

Ayo? @soft plover level 1 !!!

[dawn nebula]

check g kl

[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

[scenic gull]

has this started overtraining around 18k steps?

smoothing is at 0.987

[dawn nebula]

check gmel

gkl

d total

[scenic gull]

uploaded 🙂

d total, g mel, gkl

[dawn nebula]

its not overtrained tbh

@proper shale what do u think

[scenic gull]

i read the guide, and it said i should take note when g total starts going up but d total continues to decrease

but im having trouble finding the sweet spot of exactly when its overtraining

[dawn nebula]

whats ur save freq

[scenic gull]

5

epochs

[cerulean hedge]

when i run mangio rvc this happens

[rare goblet]

Ayo? @cerulean hedge level 4 !!!

[cerulean hedge]

go-web btw

any solutions?

[north spruce]

does anyone know why waiting for web keeps showing up i deleted the storage settings and reinstalled

[cerulean hedge]

E:\RVC V2\Mangio-RVC-v23.7.0>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897

E:\RVC V2\Mangio-RVC-v23.7.0>pause
Press any key to continue

[hard swallow]

i wanna add peter griffin to my okada vc, but when i go to add the file it shows this

nvm i cant add a photo

any ideas how i can send an image of my issue

[scenic gull]

print screen then just crtl v doesnt work?

[hard swallow]

weirdly no

[scenic gull]

huh weird

[hard swallow]

wait

got it

[sturdy mountain]

idk if admins help so how this can get fixed

[glad zealot]

oh which channel are you using it

[cerulean hedge]

runtime\python.exe trainset_preprocess_pipeline_print.py C:\Program Files (x86)\Teardown\Mangio-RVC-v23.7.0\azaz\ 40000 3 C:\Program Files (x86)\Teardown\Mangio-RVC-v23.7.0/logs/azaz False
Traceback (most recent call last):
File "C:\Program Files (x86)\Teardown\Mangio-RVC-v23.7.0\trainset_preprocess_pipeline_print.py", line 8, in <module>
sr = int(sys.argv[2])
ValueError: invalid literal for int() with base 10: 'Files'

tf

what should i do?

[brave garnet]

RVC Colabs and Spaces

⠀

Google Colabs

⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.

AICoverGen-NoWebUI
Useful for making covers, doesn't include a UI, by Ardha, by Eddy, Hina and Gdr.

RVC Disconnected
To train new voice models, by Kit Lemonfoot.

EasyGUI
The OG interface, by Rejects.
⠀

Made with love by Ilaria 💖

[glad zealot]

uh not quite sure but try moving it to somewhere that doesnt have spaces in the file path

[cerulean hedge]

huh?

[glad zealot]

extract it somewhere that doesn have space

C:\Program Files (x86)\Teardown\Mangio-RVC-v23.7.0

your path rn have a space in Program Files

[chilly flax]

i have a problem

[glad zealot]

might help idk tho

[zinc anchor]

Yeah you can't have spaces in your file paths

It's weird

[glad zealot]

also instead in teardown damn

[cerulean hedge]

💀

[cerulean hedge]

its not actually teardown, i just named it for no reason

[glad zealot]

just try putting it in like C:\Mangio o

or somethiong

[cerulean hedge]

okay i will try

[cerulean hedge]

Thanks it fucking works.

[glad zealot]

noice

[cerulean hedge]

File "C:\Teardown\Mangio-RVC-v23.7.0\runtime\lib\site-packages\torch\optim\adamw.py", line 564, in _multi_tensor_adamw exp_avg_sq_sqrt = torch._foreach_sqrt(device_exp_avg_sqs)
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 2.00 GiB total capacity; 2.48 GiB already allocated; 0 bytes free; 2.55 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

[glad zealot]

try lowering the batch size

what is it at rn

[cerulean hedge]

1

[glad zealot]

aight ye i dont think you can

@sand iris meow

[cerulean hedge]

batch size per gpu?

[glad zealot]

ye

[zinc anchor]

Higher batch size

[glad zealot]

higher uses more tho

doesnt it?

[zinc anchor]

Idk

Never trained with 2GBs of VRAM

[glad zealot]

you might need to use online solutions isntead

[zinc anchor]

The default T4 on Colab has more

[glad zealot]

ye 15

[zinc anchor]

I think it's gonna be hard to train with 2GB

[glad zealot]

i dont even think it will work

maybe when nothing is running.. or on linux

[sand iris]

Training with 2GB will not happen unfortunately

[cerulean hedge]

oh.

any replacement?

[sand iris]

Anything related to the GPU is pretty much the exact same

[cerulean hedge]

i could use

[sand iris]

-colab

[azure marsh]

Suggestions for @cerulean hedge

☁️ 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-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[sand iris]

You can (ab)use google's colab platform to train models

[zinc anchor]

What's the minimum

4GB? 6?

[sand iris]

6

[glad zealot]

https://cdn.discordapp.com/emojis/789458743649632266.webp?size=48&name=KEKW&quality=lossless

[sand iris]

Never seen 4GB work

[zinc anchor]

Doable with a 1060 then

[glad zealot]

4gb might work on 1-2 batch

[zinc anchor]

I have 8GB but it's the RX 480

[sand iris]

does that even work though ?

[cerulean hedge]

damn my pc crashed when I put the batches to highest 😭

[sand iris]

I've heard batch size 2 is bugged

[glad zealot]

might

prob gonna be slow af tho so might as well just use colab

[cerulean hedge]

out of all of them which one is the best?

rvc v2 btw

[glad zealot]

id say easy gui for begginers

disconnected if you need more advanced options ig

[cerulean hedge]

I did it with easy gui last time but I could not find the index file

Where is the index file located btw?

[glad zealot]

logs

[cerulean hedge]

Thx

[glad zealot]

logs/<model name>/<index file>

pth file is in assets/weights/<pth file>

[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

[crystal lion]

hi yall

im trying to use mmvc for realtime voice changer but the bat just closes instantly

and the guide isnt really helping either atm

i already did this and it does show the correct path in the cmd

[umbral rose]

hey uh i ran the setup file

[rare goblet]

Ayo? @umbral rose level 1 !!!

[umbral rose]

installed the vac as well

but it keeps saying this

i typed the link into the browser but nothing loads

[proper shale]

yeah its... fine

[proper shale]

delete stored_setting.json and restart W-Okada

[umbral rose]

i deleted it and restarted it

it shows up

but the audio doesn't work

like it doesn't change my voice at all

[cerulean hedge]

for an audio of 11 minutes how much epochs would I need?

[umbral rose]

let alone pick my voice at all

[molten pecan]

You can't know the right amount, as it depends on the length, size and quality

Use TensorBoard instead

https://rvc-docs.github.io/Docs/ check Epochs page

[proper shale]

check cmd

[proper shale]

...uh

might be an OneDrive issue

move the MMVCServerSIO folder out of OneDrive

[umbral rose]

its on my desktop

[proper shale]

also delete model_dir and restart

think it should work by then

[proper shale]

but the path is inside OneDrive

[umbral rose]

fuck windows omg

[proper shale]

windows being windows ig

¯_(ツ)_/¯

[autumn laurel]

uh guys

how to use the voice changer on discord call?

[shut willow]

can someone help? i got an index and pth file or smth and idk what to do

[umbral rose]

the folder right

[proper shale]

you need a virtual audio cable - check https://rentry.co/VoiceChangerGuide

[autumn laurel]

tysm

[proper shale]

yeah

[proper shale]

yw ;)

[proper shale]

https://rvc-docs.github.io/Docs/ should explain everything you need to know!

[shut willow]

k ty

[proper shale]

yw :)

[shut willow]

wait so do i just put the pth file in google colab?

[rare goblet]

Ayo? @shut willow level 1 !!!

[shut willow]

ive done that for a phone guy voice

[proper shale]

u have to upload that zip in huggingface

[umbral rose]

god am i doing something wrong
there's just repeadet static noise

[proper shale]

the documentation explains it

[proper shale]

...give us cmd again

[shut willow]

ohh ok

[proper shale]

sorry I'm being repetitive lmao

[umbral rose]

no its fine

thank you lol

some people don't even help at all

its gone now

[vestal acorn]

what does this mean? this was made in applio colab

[proper shale]

oh

ur input must be ur mic

[umbral rose]

i don't speak cmd

[rare goblet]

Ayo? @umbral rose level 2 !!!

[autumn laurel]

why does my chrome says ran out of memory to open this page?

[umbral rose]

wait what

[proper shale]

also r u on AMD?

[proper shale]

yeah

[umbral rose]

no im on intel

i installed the intel one

[proper shale]

intel arc?

[umbral rose]

all i know is it's i5 11th gen

[proper shale]

😭

integrated graphics won't really work well

[autumn laurel]

hey

[umbral rose]

god damn it

thanks for helping

ima uninstall

[proper shale]

nah wait i got something for u

[autumn laurel]

.

[proper shale]

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

[proper shale]

i really don't know

should open normally, heck even my phone opens it

and it's like

the worst phone u can think of

[strange tinsel]

Oi

[umbral rose]

lol idk why but i hit the play button on the top left

should i have done that

[proper shale]

yeah lol

[umbral rose]

oh ok

i feel like a 5 year old

[proper shale]

eh dw i felt the same when trying to set up W-Okada the first time

like 4-5 months ago

[umbral rose]

thank you for helping again

[strange tinsel]

Tem como fazer pelo celular?

[proper shale]

sim!

se tu quiser suporte br e tudo mais entra no ai hub brasil

discord.gg/aihubbrasil

;)

[jade perch]

seems like I keep getting the access denied error while trying to download the models using Easy GUI. Tried different models, different alt urls aswell. Is this common?

[shut willow]

uhhh i dont understandd, what did uploading it to hugggingface do?

[proper shale]

yeah uh, that colab is outdated

and google drive links are broken

[jade perch]

oh damn, that sucks

[proper shale]

u need that so you can have a download link that u can paste

[shut willow]

oh

but how do i use my pth and index filles

[proper shale]

again, https://rvc-docs.github.io/Docs/

[shut willow]

ii checked and it wont give me propeer instructions

[vestal acorn]

what does this mean? this was made in applio colab

[proper shale]

😭 not sure what this could be

for the time being, i think u should use rvc disconnected

[shut willow]

oh wait i found instructions

[vestal acorn]

what is an array?

[brittle wing]

how am i supposed to know if its overtraining or not

[proper shale]

check the graphs

also

-overtrain

[azure marsh]

How do I know if my model is overtaining?

All-In-One Guide on how to make a good model

This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models

Credits: LUSBERT

Automated Overtraining Detection (AOD)

Will be available soon in #1159513888199540817

Credits: grvyscale

[proper shale]

^^^

the guide explains

everything you need to know

[autumn laurel]

hey is there an indonesian ai hub server?

[proper shale]

hmmm I don't believe there is

[autumn laurel]

ok thx

[proper shale]

check the indonesian language channel tho

[autumn laurel]

um where?

[proper shale]

#1160154650343903334

[autumn laurel]

ok

[brittle wing]

my graph doesn't seem to be showing anything useful tbh 😭

[shut willow]

now what do i do, i have my index and pth

[molten pecan]

Too tricky for beginners imo 😭

[brittle wing]

refresh

[shut willow]

?

[proper shale]

u paste the Huggingface download link in the download voice models section

[molten pecan]

Check the Ilaria RVC page of my website

https://rvc-docs.github.io/Docs/

[proper shale]

send the screenshots

[brittle wing]

😆

[brittle wing]

wait

[autumn laurel]

uh hey

[shut willow]

done

[autumn laurel]

what do i do after i get the virtual cable download link?

[molten pecan]

#🔍│help-w-okada

[proper shale]

u set up the input/output stuff

The input output settings should be like this in W-Okada:

Input - Your mic
Output - CABLE Input

And in Discord:

Input - CABLE Output
Output: Your headphones

or replace CABLE for Line 1 if ur using VAC by Muzychenko

[autumn laurel]

um i use the black hole

[rare goblet]

Ayo? @autumn laurel level 2 !!!

[brittle wing]

it has not overtrained yet right

[autumn laurel]

hey how to install the virtual cable i alrady got the link

[small umbra]

..how to install dez nuts in yo mouf?*

[shut willow]

now what

[proper shale]

yea doesn't feel like overtraining

[shut willow]

i downloaded the thingy

[proper shale]

but...

[autumn laurel]

pls tell me

[molten pecan]

It looks strange 😭

[shut willow]

oops nvm

[molten pecan]

Mode collapse?

[proper shale]

yeah, collapses

[brittle wing]

Exactly

[proper shale]

ov2?

[brittle wing]

it usually looks different

[brittle wing]

is it the pretrain type thingy in the first cell

[autumn laurel]

how to download the virtual cable from black hole (i already got the link) pls tell me IDK how ;-;

[proper shale]

yes

[molten pecan]

How long is your dataset?

[proper shale]

uhhhhh

[autumn laurel]

....

[proper shale]

get the VB audio one

[autumn laurel]

how

[proper shale]

or VAC by Muzychenko

[proper shale]

search it?

[brittle wing]

wait lemme check

[proper shale]

also, voice changer guide has links to download

[brittle wing]

about 10:40 mins

[molten pecan]

Are you using a batch size of 8?

[brittle wing]

its chosen cuz its the default settings

[brittle wing]

13

[shut willow]

idk what now im refreshing everything nothings happening theres nothing in the docs

[rare goblet]

Ayo? @shut willow level 2 !!!

[brittle wing]

😆

[crystal lion]

hello can soneone help me 💀

[molten pecan]

#🔍│help-w-okada

#1192011222023950368

[proper shale]

u should try the normal pretrains

[proper shale]

check the pins in help-w-okada

[crystal lion]

mwuah

[molten pecan]

Are you old enough to use Discord?

[shut willow]

yes

[brittle wing]

OMG i knew something was off with the settings in the first cell

[proper shale]

did u even insert the Huggingface link correctly in the place it tells u to

[molten pecan]

Use the original pretrain

[brittle wing]

now i have to retrain

so annoying 😩

[brittle wing]

thanks

[molten pecan]

yw! Have fun

[brittle wing]

🫄🏽

[shut willow]

\

[molten pecan]

Are you using Ilaria RVC from Google Colab or Hugging Face?

[proper shale]

Maybe delete the ?download true from the link and just name the model something else

Should work now

[shut willow]

ok

ohhhh when i copied the name it alsoo copied my username

hmm

[proper shale]

now just click on refresh right besides the model dropdown

[shut willow]

ok

done

[proper shale]

and check for ur model, should be there

[shut willow]

its there

[proper shale]

now it should be pretty straightforward

[shut willow]

ok

YES FINALLY

IVE BEEEN LOOKING FOR A GOLDEN FREDDY VOIICE FOR SO LONG

[quick kernel]

hello, someone can help me, why the song is not there? i refreshed 50 times but nothgin

[proper shale]

uhhhh you might be using an outdated colab

use this one; https://colab.research.google.com/drive/16LkwvFZeudTpUOsE_6bMjOq2qkxFo8Hr?usp=sharing

[quick kernel]

yesterday it worked

[proper shale]

it... doesn't work for free users

[quick kernel]

i try with this link, thanks 🙂

[proper shale]

yw :)

[molten pecan]

https://rvc-docs.github.io/Docs/

[molten pecan]

There's a guide for it here

;3

[quick kernel]

thank you!

[toxic sinew]

hi guys can someone explain me how to clone a singer's voice with macbook? what's the programme used? all the tutorials are used with windows

[quick kernel]

i cant find how i can add model to these version

[molten pecan]

What version?

Give full context pls

[quick kernel]

this

[proper shale]

ila rvc

[molten pecan]

https://rvc-docs.github.io/Docs/

Check the Ilaria RVC page

[quick kernel]

thanks i found 🙂

i dont know whats the problem..

refreshed a lot of times and nothing appear

[flint geyser]

yo

[proper shale]

actually u shouldnt refresh

[flint geyser]

anyone need help?

[proper shale]

u should wait until the file name appears there

its just how easygui is ig

also try refreshing the gradio page

[molten pecan]

The audio is not selected in "Choose the audio file"

Select it

[proper shale]

but they said it doesn't appear lol

[quick kernel]

now everything disappeared 😄

[quick kernel]

but i cannot select, nothing is there

[proper shale]

now u can try the refresh thingy only once

else, upload audio again, wait for it to appear on the dropdown

[molten pecan]

We should tell Ila about this error

[proper shale]

its just a gradio issue probably

and, she paused devving

[quick kernel]

i uploaded another song and its working , maybe something wrogn with my file... 😄

[proper shale]

maybe... hmmm

[worthy shoal]

Anyone got a VMT voice model

[rare goblet]

Ayo? @worthy shoal level 1 !!!

[oblique tusk]

hey guys i'm using realtime voice changer by w-Okada, is there a way to make it work across my pc ? like with any audio input ? like for example meet or whatsapp or any other app ?

I downloaded the VBcable program and it works like a charm here on Discord

but not on other apps

[calm solar]

When I try open the start_http.bat it doesnt work

😭

[proper shale]

if they do have input/output settings u can just change the input to the cable output

and it should work

[proper shale]

any more details? what happens exactly?

[calm solar]

So i downloaded the zip file and i double clicked on the start_http.bat and wwaited like 20 minutes and it didnt open so i tried again and it just wont open

[oblique tusk]

I changed it but it doesn't work, I have an android emu that has whatsapp and other apps on it and I changed the emu input settings and restart it and it still not working

[rare goblet]

Ayo? @oblique tusk level 1 !!!

[proper shale]

check #🔍│help-w-okada pins

[oblique tusk]

I tried whatsapp app on pc same thing, but whatsapp on pc doesnt have settings

[proper shale]

Probably just an app specific issue

You should try other things then

[oblique tusk]

so this is the right way no other ways right ?

i'll try and get back to you thanks ❤️

[proper shale]

yeah i guess there isn't anything else that can be done

[silk agate]

Hey guys sorry if this is the wrong channel, I'm just hoping to gain insight about the best practices for preparing a dataset. I've trained two models so far, one on 10 minutes of me reading from a book and the other was trained on vocals from my old songs. Neither of them really accel at quality cover songs. If the vocals on my target song are simple then the cover turns out great, but once vibrato or falsetto are introduced it kinda falls apart

[winged skiff]

Is there a fix to this or is it just a me problem 😭

[sand iris]

More and more common for some reason

Give this a shot, should work : #🔍│help-w-okada message

[winged skiff]

it happened for downloading torch too but worked after 5 retrys 😮‍💨

[winged skiff]

Its still not working unless im doing it wrong

[rare goblet]

Ayo? @winged skiff level 2 !!!

[sand iris]

what have you done rn ?

[winged skiff]

ive uninstalled RVC and reinstalled it to my desktop then I extracted the pretrain in the MMVC folder inside it

[sand iris]

does it look anything like this :

MMVCServerSIO/
    pretrain/
    [...]
    start_http.bat
    start_https.bat
    MMCVServerSIO.exe
    [...]

[winged skiff]

everythings in the MMVCServer Folder

[brittle wing]

help idk why i dowloaded evything but its just wont work

[winged skiff]

yes except MMVC.exe is above start_Http and start_https

[sand iris]

Can you re-screenshot the error ?

[sand iris]

Yeah the order doesn't matter, just wanted to illustrate

[winged skiff]

okok

[sand iris]

ok it needs more stuff than I thought

hang on a sec

drag those in as well

[winged skiff]

ok

[winged skiff]

when I let it through the firewall do I put it for both private and public?

[sand iris]

both I guess ?

[winged skiff]

well it opened but now I got a different issue

[sand iris]

yeah the default models won't work

[winged skiff]

ohhh

[sand iris]

Try importing a custom one, should do

[winged skiff]

ok

[worthy shoal]

I can't see start_http.bat what do I do?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[winged skiff]

ty dawg, I was getting brainrott trying to see what was wrong yesterday 😭

[crude rivet]

is there such a thing as "melody-guided TTS"? I want to pass in lyrics and a guide melody, and then have the AI sing those lyrics according to the existing melody. Is that even possible?

[storm kernel]

I think ur asking for too much there

[crude rivet]

I know 😭

I'm aware that Bark from Suno lets you put in lyrics, but there isn't any way to guide the melody of the generated vocals

[proper shale]

yeah that prolly doesn't exist

[crude rivet]

not yet at least 😅

[onyx shale]

Alright so I downloaded RVC, now what?

It's just a big file, so what do I launch?

[haughty pulsar]

mainline right?

[onyx shale]

Are you responding to me?

[rare goblet]

Ayo? @onyx shale level 2 !!!

[haughty pulsar]

yup

[onyx shale]

You should probably reply otherwise I won't get pinged, anyways I got the RTX one

[onyx shale]

And yeah it's not RVC-GUI it's the real RVC

[haughty pulsar]

yea sorry about that is it called something like RVC1006Nvidia

[onyx shale]

Yeah

[onyx shale]

Exactly what it's called

[haughty pulsar]

click go-web.bat

[onyx shale]

Okay

[haughty pulsar]

did you unzip it first?

[onyx shale]

Yup, it's in my documents

[onyx shale]

So it opened up CMD, do I need to type something or just wait?

[haughty pulsar]

wait

[onyx shale]

Okay it sent me to my web browser and it's on its webui. How do I create the voice models?

[haughty pulsar]

let me get a guide up really quick

[onyx shale]

Thanks 😄

[haughty pulsar]

https://rvc-docs.github.io/Docs/rvc/local/mainline-rvc/ this should be exactly what you need

[onyx shale]

Thanks again :D

[haughty pulsar]

[onyx shale]

So the guide says to enter the path of the training folder. Is the training folder the ZIP where I store all of the audio files of the character?

[haughty pulsar]

it is a normal folder that contains all your audio files.

[onyx shale]

I also have a version of it that's a regular folder

[onyx shale]

Perfect!

[onyx shale]

And second, when I copy the path there are quotes in it before and after. Should I keep those or will it mess things up?

[onyx shale]

And another thing what is the singer ID?

[haughty pulsar]

singer id you dont need to touch

[haughty pulsar]

i don`t think quotes are needed

[onyx shale]

Thx

[onyx shale]

kk

[onyx shale]

What does the GPU Index mean? The guide is using words I don't understand for that part

[haughty pulsar]

ok so go on task manager and check what your gpu says after

like gpu0

also under it i think it should show available gpus

[onyx shale]

Okay thx

[onyx shale]

What's the difference between CPU and GPU?

And turns out I have GPU 0 and GPU 1

[haughty pulsar]

simply
cpu = slow
gpu = fast

[onyx shale]

Ah okay

So which GPU do I use if I have 2?

[haughty pulsar]

one is probably intergrated gpu

[haughty pulsar]

doesn't it show available ones here?

[onyx shale]

It shows something there

[rare goblet]

Ayo? @onyx shale level 3 !!!

[haughty pulsar]

and what does it say

[onyx shale]

The same thing your screenshot says just in light mode

[haughty pulsar]

on gpu information

[onyx shale]

OHHH

0 NVIDIA GeForce RTX 2070 with Max-Q Design

[haughty pulsar]

use 0

[onyx shale]

Okay cool

[onyx shale]

How do I know whether to choose RMVPE_GPU or Crepe?

[spiral osprey]

If I make a voice model with 100 epouchs but the 2-hour dataset will it be good?

[haughty pulsar]

i recommend rmvpe_gpu

[dawn nebula]

u dont need that much tbh

[onyx shale]

kk

[haughty pulsar]

datasets that are too long can also be bad as rvc is bad with very long datasets

[dawn nebula]

yeah fax

[grizzled zealot]

relatively dumb question, would 4 gb of vram be good enough for training?

[dawn nebula]

not enough sorry

use colab

[grizzled zealot]

so i need 6 or more

[dawn nebula]

yeah

[grizzled zealot]

ok, thank you!

[real void]

Alguien me puede ayudar con un problema del modulador?

Lo que ocurre es que no se cambia la voz al darle start y ya lo tengo configurado

Help me pls :c

[onyx shale]

Quick, huge issue! I accidently clicked Train Model before downloading tensorboard! I got everything else done, but I have tensorboard and it won't work whatdoido

[haughty pulsar]

tensorboard isnt working?

[real void]

Can someone help me with a modulator problem?

[rare goblet]

Ayo? @real void level 1 !!!

[real void]

What happens is that the voice is not changed when I start and I already have it configured

[onyx shale]

Idk it's just showing that it downloaded. What now!? It says I need to keep an eye on it

But it's not helping!

I don't wanna overtrain it!

[haughty pulsar]

hm?

it should show a graph

[onyx shale]

It's because I started the process before downloading it okay? So how do I cancel the process?

How?

[haughty pulsar]

you can open tensorboard after you start training

[onyx shale]

But it's not doing anything!

[haughty pulsar]

send a screenshot

[haughty pulsar]

shift click the localhost link

[onyx shale]

???

[haughty pulsar]

http://localhost:6006/

[onyx shale]

???

[haughty pulsar]

it should open in browser

[onyx shale]

No, it's in CMD

So then how do I open in browser?

[haughty pulsar]

when you click the link

or type it

[onyx shale]

Link?

What do you mean?

[onyx shale]

It's a .bat

Ohhhhhhhhhhh

oka

sry