✨│ai-help

[faint bloom]

how 2 download

[rugged stirrup]

Download the .index file

[rugged stirrup]

Right click on it and it will give you options hit download

[rare goblet]

Ayo? @faint bloom level 3 !!!

[rugged stirrup]

Yes

[faint bloom]

now what do i do with it

[rugged stirrup]

Now go to the assets folder click on the down arrow and you will see a "Weights" folder

[rugged stirrup]

Download the top one

That one should likely be fine

[faint bloom]

ok what do i do with the 2 files

[rugged stirrup]

Well you use RVC Infer locally

[faint bloom]

ok

[rugged stirrup]

If you want to use it on colab right now

Go to the inference tab and you can load your model and upload your audio and inference

[faint bloom]

i think it work

i used this to train

i cuildnt find the original voice thingy they used in the movies

[rugged stirrup]

Is this your dataset?

[faint bloom]

i think

ye

[rugged stirrup]

Did you try to inference with your model?

Wait

[faint bloom]

in the movie these are realy the only times these guys speak

[faint bloom]

how to do that

[rugged stirrup]

This is your Entire dataset?

[faint bloom]

ye

[rugged stirrup]

I told you you go to the inference tab

[faint bloom]

im there

[rugged stirrup]

It seems like you didnt clean it very much also 26 seconds is very short

[rugged stirrup]

Now upload a audio file to the big box that said drop file here or something like that

[faint bloom]

the problem is theres not much dialouge of them from movies

[rugged stirrup]

So you probably shouldnt expect the model to sound very good

[rugged stirrup]

Well you can try to inference and see what it sounds like if it sounds good to you then it sounds good to you

[faint bloom]

sounds good enough, works until i somehow find out what they used back in the 90s

[rugged stirrup]

Ok

[faint bloom]

wait a second

this makes for an interresting robotic voice

[faint bloom]

ty for your help

[rugged stirrup]

Np

[faint bloom]

also one more question

does it save if i open it again at a later date

[rugged stirrup]

It doesn't

If you close the google colab it will delete it so thats why you have to download the model if you want to have it online where you can download it upload it to Google Drive or Huggingface

[faint bloom]

how do i do that

[rugged stirrup]

You create a Huggingface account https://huggingface.co

[faint bloom]

ok got 1

now?

[rugged stirrup]

You go to the top right where your profile picture is

And click it and click new model and it will allow you to name it and upload your model make sure to release it under the "Openrail" license

[faint bloom]

ok

now?

[rugged stirrup]

What does it look like for you

Idk what your on

[rugged stirrup]

K thats good now hit create model on the bottom left

[faint bloom]

when can i make models

[strong fulcrum]

For RVC, what do I put in the .env file? Where does .pth go?

[rugged stirrup]

What?

[rugged stirrup]

Idk why its saying that

[strong fulcrum]

https://pypi.org/project/rvc/

it says if i have already downloaded models, specify them in the .env file

[proud elbow]

if it's your first time, try using vpn and/or another browser

[rare goblet]

Ayo? @proud elbow level 28 !!!

[strong fulcrum]

and also, i've installed rvc, but when I run, i get error that says rvc_infer is not found

[rugged stirrup]

You put the index file in the "Logs" folder and you put the .pth in the "weights"

Idk what that guide is telling you but I think you should put the pth and index file in logs and weights

[rugged stirrup]

Which version of RVC do you have

[strong fulcrum]

uhh

[rugged stirrup]

Do you have mainline or mangio? or some other version

[strong fulcrum]

should be 0.3.5

[rugged stirrup]

That is very outdated

[strong fulcrum]

i'm on mac, and doesnt seem mainline works for mac since there is only a bat file and says nvidia

[proud elbow]

not even cpu mode? might try applio

[rugged stirrup]

Ohh

[strong fulcrum]

i just did pip install rvc

[rugged stirrup]

So idk if there is any version for mac

[strong fulcrum]

applio and rvc_cli were giving me errors

[rugged stirrup]

Instead try easyGUI which is on colab

[strong fulcrum]

said that there were certification verification error or something

[strong fulcrum]

i'd like it locally however

[rugged stirrup]

Actually wait

You can run RVC locally with mac download it here: https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006Nvidia.7z

And run the .sh file

[strong fulcrum]

there was a sh file?? i didnt see it

[rugged stirrup]

Yeah

Scroll down a little

[strong fulcrum]

can't download it for some reason

[rugged stirrup]

Did you click download anyway

[strong fulcrum]

could u give me the file link instead?

[faint bloom]

another browser dont work neither does vpn

[rugged stirrup]

Or continue to download

[rugged stirrup]

Are you in another country?

[strong fulcrum]

when i open the link, it immediately downloads a .download file instead of the actual file

[rugged stirrup]

Strange

[strong fulcrum]

i'll just look for it manually

[rugged stirrup]

https://huggingface.co/lj1995/VoiceConversionWebUI/tree/main

There is the normal link with all versions

You should get the newest version

[strong fulcrum]

ye lookin thru here rn

[faint bloom]

yes

[strong fulcrum]

ohhhhh i think i know why it's just giving a .download file. im probably out of space lmao

[faint bloom]

from germany to north america

[rugged stirrup]

Hugging face should work in Germany

So idk what happened with you but maybe try Google drive

[strong fulcrum]

also what's the difference between rvg and rvc?

rvg just a tts?

[rugged stirrup]

Idk what RVG is

[faint bloom]

what do i do in google drive

[rare goblet]

Ayo? @faint bloom level 4 !!!

[rugged stirrup]

You mean GPT SoVITS?

[rugged stirrup]

Select the index and pth file and right click on one of them when both are selected and a popup should pop up and click send file to zip

And once its zipped upload it to Drive

Or you can just drag both files in without zipping if you dont know how to zip a file

[faint bloom]

where do i drag em

[rugged stirrup]

Upload them to drive

[faint bloom]

ok i have the zip saved

opps

welp its in my drive thingy

[rugged stirrup]

Well now you can use easyGUI https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw

When installing RVC is done open the easyGUI and it will give you a Gradio link click on it and the UI should open when your in the UI click on the download models tab

[strong fulcrum]

finally cleared enough space on my drive lol

[faint bloom]

ok now?

[rugged stirrup]

Are you on the UI?

[faint bloom]

ye

[rugged stirrup]

Go to the download models tab

Is your file zipped?

[faint bloom]

yes

[rugged stirrup]

Right click on your zipped file on Drive

And make sure its public or anyone with link can download

Once its on that setting paste the link into there and click download model it should download and go to the inference tab

And once your on the inference tab you click refresh and you should see your model pop up and also make sure to select your index file too

[faint bloom]

how to get link

[rugged stirrup]

You right click the zip file

In drive

[faint bloom]

got it

[rugged stirrup]

Make sure its anyone with link can view

[faint bloom]

alr works

tysm

[rugged stirrup]

Great

[umbral acorn]

does anyone know how to remove the voice background of this audio?. because its not music so idk how to remove...

[rare goblet]

Ayo? @umbral acorn level 4 !!!

[rugged stirrup]

Use mvsep karaoke models

[umbral acorn]

alr thanks i'll try

[shut remnant]

where can i find the .pth file in training in disconnected ?

[rare goblet]

Ayo? @shut remnant level 1 !!!

[rugged stirrup]

Open colab files and then open the click the dropdown button on the RVC file and you should scroll down and see a file called "Assets" click the dropdown on that and you should see a file called weights click the dropdown button on that and you will see your pth files in there

[strong fulcrum]

@rugged stirrup welp the sh file apparently got permission denied

[rugged stirrup]

Idk what to tell you then

[strong fulcrum]

i tried chmod and still no work

[strong fulcrum]

how about the other one? rvc_cli or applio?

[rare goblet]

Ayo? @strong fulcrum level 5 !!!

[rugged stirrup]

Applio does work on apple

So you can try that

[strong fulcrum]

Do you know anything about this error? huggingface hubert ssl.SSLCertVerificationError: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get local issuer certificate (_ssl.c:1007)

[rugged stirrup]

Sadly no I don't

[strong fulcrum]

this happened when I ran both applio and rvc_cli

[rugged stirrup]

But applio does work for apple

[shut remnant]

i've put the save frequency to 50, and it's already on 400, but i still can't find a single pth

[rugged stirrup]

You have to check your RVC folder

[strong fulcrum]

is applio supposed to automatically install the dependencies and requirements in the setup/install?

because i got gradio not found error. so i tried manually installing the requirements in the .venv, then got a package error

[rugged stirrup]

Yeah all you have to do is run a bat file It will likely give you some options just fill out the options or whatever I don't really know how to use applio

[rugged stirrup]

On applio?

[strong fulcrum]

yes

[rugged stirrup]

Strange

[strong fulcrum]

so i ran install.sh then run.sh then it gave error gradio not found, so i went in and activated .venv to manually pip install the requirements.txt, then got a package error

Requirement already satisfied: markupsafe~=2.0 in ./.venv/lib/python3.10/site-packages (from gradio==4.17.0->-r requirements.txt (line 30)) (2.1.5)
Collecting orjson~=3.0 (from gradio==4.17.0->-r requirements.txt (line 30))
  Using cached orjson-3.10.0.tar.gz (4.9 MB)
  Installing build dependencies ... done
  Getting requirements to build wheel ... done
  Preparing metadata (pyproject.toml) ... error
  error: subprocess-exited-with-error
  
  × Preparing metadata (pyproject.toml) did not run successfully.
  │ exit code: 1
  ╰─> [6 lines of output]
      
      Cargo, the Rust package manager, is not installed or is not on PATH.
      This package requires Rust and Cargo to compile extensions. Install it through
      the system's package manager or via https://rustup.rs/
      
      Checking for Rust toolchain....
      [end of output]
  
  note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed

× Encountered error while generating package metadata.
╰─> See above for output.

note: This is an issue with the package mentioned above, not pip.
hint: See above for details.

[rugged stirrup]

Ok go to the applio download page and and click download What does it download for you

[strong fulcrum]

it doesnt download. it just brings me to the github releases page

[strong fulcrum]

so what i did was just clicked on "code" then downloaded the repo

[strong fulcrum]

the applio download page brought me here: https://github.com/IAHispano/Applio/releases

[rugged stirrup]

https://docs.applio.org/faq/rvc/

click oh wait that's not it

https://docs.applio.org/

There

click that and click download

[strong fulcrum]

i don't see a download button

[rugged stirrup]

sorry that was the wrong link again

https://docs.applio.org/

[strong fulcrum]

that's the same too lol

[rugged stirrup]

Wait that's the exact same thing I thought Wait hold on https://applio.org/

there there oh my gosh

stupid copy and paste on a phone wouldn't copy the correct link

[strong fulcrum]

so this one? ^

[strong fulcrum]

this is where i was earlier. the donwload button just redirects me to the github releases page

[rugged stirrup]

Yes

[rugged stirrup]

are you sure

[strong fulcrum]

Yeah

[strong fulcrum]

Just brings me here

[brittle wing]

where can i download VAC

[rugged stirrup]

I don't know what to do sorry

[rugged stirrup]

look up retrieval based voice conversion web UI on Google and you will find a GitHub for it download the package and wait for it to be done then extract it and open it when it's extracted and you will scroll down and find a file called goweb.bat

[strong fulcrum]

Ok, thanks for ur help anyways

[rugged stirrup]

Np

[rugged stirrup]

Bot

[proper shale]

https://vac.muzychenko.net/en/download.htm

[proper shale]

https://huggingface.co/IAHispano/Applio/tree/main/Compiled

just download the latest version there

[rare goblet]

Ayo? @brittle wing 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

[storm viper]

i know english can somebody come to unlimited vc help with voice changer

pls

hello

[low shard]

For help abt wokada realtime voice changer ask in #1192011222023950368 or #🔍│help-w-okada

[wary bluff]

where can i find MDX23C on UVR?

[timber spruce]

why can't i upload any voice-models to easygui

[low shard]

Send colab link

[timber spruce]

https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw#scrollTo=DZDKirCM0F9g

[low shard]

Okay thats the updated one, send the model download link lemme see

[timber spruce]

https://huggingface.co/vinnystop/caseohv2/resolve/main/case2.zip?download=true

the one i first wanted to do

then i tried some other voice models to see if the issue was up to the model

[rare goblet]

Ayo? @timber spruce level 1 !!!

[timber spruce]

but it ain't working

just does this whenever i try to do any model

however when i use THIS it just works

[low shard]

lemme check rq inside the model files

[wary bluff]

Appreciate it. Would you say this is the best vocal isolating model?

[low shard]

yea, the only problem of it is you need a good gpu for it, if you dont it will take tooooo muchh time, if you dont got a good gpu i could suggest you another good model

[wary bluff]

ah okay then i might need a different one. Which other one would you suggest

[low shard]

those are the models i suggest you, the other ones than the mdx23c one are also good (ofc not as good) and can run even on less powerful pc

btw theres a google colab for mdx23c model if you wanna try it

[low shard]

yea the model looks fine, you are using google colab gpu right?

[timber spruce]

seems like so

yeah i am

[wary bluff]

yeah i've been primarily using Kim vocal 1 and 2 but i'll definitely check those others out. Thanks

[timber spruce]

let me try this

nope

that doesn't work either

( if you're confused i just enabled this on )

i'm new to this type of stuff so uhh i don't know what it does so i'm just experimenting

[low shard]

yea i just seen too

the weird thing is in the colab output it says downloaded

but in the ui it doesnt show it so yea

there might be a problem with easygui

ima tell the creator rq

@quick trench hey you may wanna check out this issue

[timber spruce]

exactly

that's what i've noticed

it says it's good

but it displays it isn't

so the fix is to just not use easygui for now?

[low shard]

lemme check the no ui inference cell and see if that works first

[timber spruce]

sure

[rare goblet]

Ayo? @timber spruce level 2 !!!

[low shard]

i guess for now you could use other colabs, as easygui is having some problems, you could check out the other colabs on our docs https://docs.aihub.wtf, you can find them inside the cloud folder inside the rvc folder

[timber spruce]

alright sure

thanks for helping

[low shard]

yw, if u get other issues ask here

[timber spruce]

aight

[low shard]

yw

[marble arrow]

help, always getting error with this colab
https://colab.research.google.com/github/Eddycrack864/Ultimate-Vocal-Remover-5.6-for-Google-Colab/blob/main/Ultimate_Vocal_Remover_5_6_for_Google_Colab.ipynb#scrollTo=TXNqGZcWJmW-

[dense ermine]

does anyone know why i keep getting JSONDecodeError in google collab?

ive been trying multiple different ones and keep getting the same error

mainly this one https://huggingface.co/C0ttontheBunny/PlanktonOv2/resolve/main/PlanktonOv2.zip?download=true

please help

[rare goblet]

Ayo? @dense ermine level 1 !!!

[pastel oak]

you could be using an outdated colab

check pinned messages for the docs guide and select one of the colabs from there

[dense ermine]

alright thanks ill give it a shot

[dense ermine]

which one is it i cannot find it

[pastel oak]

i recommend ilaria or disconnected

[dense ermine]

it works thanks again!

[visual pumice]

on long vocal notes like "oooooo" my ai model makes a "ooouuurrrrr" or whatever, it like doesnt make a stable note.

[low shard]

ye thats outdated lol

[dreamy falcon]

did anyone have a problem where their .logs, despite being selected - don't actually work?

[visual pumice]

or it will be "youuuuuuuu" and the ai makes it "youuuooouooooo"

[finite galleon]

Try lower the index rate

Or feature ratio

[visual pumice]

index at a solid 0

i dont use index files, ever. they never work or sound good imo

[finite galleon]

Try setting protect to 0.5

[visual pumice]

ill try that

[finite galleon]

And you using rmvpe right?

[visual pumice]

rvmpe gui

oh wow that sounds way better

man why do they make the default values the ones that sound the worse

every youtube ai cover would sound phenomenal

best preset ever

Oh my god.

why did i not do this earlier??? omg

there was this one part of this song that the ai would always pronounce atrociously incorrect and its actually doing it correctly now

[rare goblet]

Ayo? @visual pumice level 6 !!!

[storm rose]

the old google colab link i used to train models with seems to not work anymore, can anyone send the most recent working one please?

[low shard]

use one of those in our docs https://docs.aihub.wtf/ (check inside the cloud folder, inside the rvc folder)

[desert niche]

-local

[azure marsh]

🖥️ Local stuff

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

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

[storm rose]

thank u soo much

[low shard]

yww

i personally use

rvcdisconnected tho

which is also in the docs

[storm rose]

oh, i'll try that then 🫶🏻

[brittle wing]

Can someone there train someone's dataset cause I don't have time

[low shard]

for rvc ai voice model requests, please ask in #1159289738314919936 , not here.

[smoky linden]

Could some one help me with this error is missing 'inf' file or driver package corrupted...

[stone shell]

lol name buddies

[rare goblet]

Ayo? @stone shell level 1 !!!

[brittle wing]

I didn't know thanks

Is advanced rvc inference incompatible with huggingface?

I got this error when i try to download a model

[low shard]

Yw

[low shard]

Dont use that hf space

May be outdated

U need an hf space

Or google colab?

[brittle wing]

google colab

applio got banned

[low shard]

Cus hf space runs on cpu which is way slower

And colab got a daily free gpu

[low shard]

What?

[brittle wing]

applio colab got banned

[low shard]

It didnt wdym?

[brittle wing]

I got error when training

[low shard]

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb

[low shard]

what error? disallowed code ?

[brittle wing]

gpu disconnecting

unexpectedly

[low shard]

Yea but what exact error

Like GPU backend

Or disallowed code?

Or inactivity?

[brittle wing]

Inactivity maybe

I will try to retrain in Applio

so i can see if it actually got banned

[low shard]

Inactivity its because you left it afk then, not bc it got banned, check https://discord.com/channels/1159260121998827560/1213586188091523075 for that theres a bypass

[low shard]

Yea send a screenshot if you still get an error while training and following the guide i js sent you

[split frigate]

just moved rvc to new storage and it took about 95gb!
That is extremely high, how can I clean it up

[dawn nebula]

clear up the model checkpoints

they are large

[split frigate]

weights?

[dawn nebula]

these

if you have no intentions to train more models

[split frigate]

what if i deleted models itself

models thats i already putted in w okada

[wild sundial]

can someone help me?

I want to do an ai cover but I can't find the Google colab link

[little ocean]

guys it says i have to upload an image but all images are nit workiing in mvioce .ai

[brittle wing]

Does applio train models fast???

Cause I still haven't tried training a model with it

[wispy burrow]

How long it takes depends on the GPU you have, how long your Dataset is and Batch size.

[brittle wing]

Uh how do I get a fast GPU for free

[low shard]

are you doing it locally or on colab?

[wild sundial]

guys

[brittle wing]

Colab

[brittle wing]

The person's dataset is too big

[low shard]

then you cant do nothing about it, you are already using the T4 Free daily gpu, just wait

[wild sundial]

what I need to do when appears that

[low shard]

outdated colab

use the newer ones

[wild sundial]

oh thanks

[low shard]

check them inside the cloud folder inside the rvc folder in our docs https://docs.aihub.wtf

i personally use ilaria rvc (which is in that doc dw)

[brittle wing]

For next model

[low shard]

?

google colab offers only the T4 gpu daily for free

[brittle wing]

Even my first model didn't have that big of a dataset

This person is too greedy

[low shard]

gpus arent free, they cost alot

or are you talking about a commission you did to a model maker ?

[wispy burrow]

But for how long can you use it?

[brittle wing]

I'm training someone else's model cause they're too lazy to do it themselves

[low shard]

well, technically 12 hours a day max, but it also depends on the usage, so for example tasks like rvc/wokada, you can for 4 hours, it varies

[low shard]

someone is paying u?

[brittle wing]

Nope for free

[low shard]

what exact request are u doing

like a friend

or someone in this server

[brittle wing]

Someone on the internet

And I'm trying to

Find someone who can train it half, me the other half

[low shard]

well, training a model takes time

[smoky linden]

How can i fix my gpu not showing it just shows cpu gpo0 gpo1 gpo2 gpo3

[low shard]

are you talking about wokada realtime voice changer?

[smoky linden]

yea

[low shard]

for that, please ask help in #1192011222023950368 or #🔍│help-w-okada

[smoky linden]

ok

[brittle wing]

I KNOW

I just want a way to do it faster

[wispy burrow]

What Model does the person want? Maybe it already exists?

[rare goblet]

Ayo? @wispy burrow level 21 !!!

[brittle wing]

Of a specific youtuber

[low shard]

there isnt , google colab offers only that gpu for free

[brittle wing]

Isn't applio faster at training

[low shard]

you could search:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models

[brittle wing]

Nahh

[low shard]

just to see if its already made lol

[brittle wing]

They want their dataset to be trained

[brittle wing]

He wants his dataset to be trained

Not someone's model

[low shard]

oh rip

[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

[versed linden]

my training is continuing from the wrong checkpoint, any fixes?

[brittle wing]

Use the fixed version

[rare goblet]

Ayo? @brittle wing level 7 !!!

[brittle wing]

https://colab.research.google.com/drive/16Q44VBJiIrXOgTINztVDVeb0XKhLKHwl

https://colab.research.google.com/drive/16Q44VBJiIrXOgTINztVDVeb0XKhLKHwl

[wispy burrow]

Did you save the D & G files when you stopped training?

[versed linden]

yes, they are in the logs folder

[wispy burrow]

How often did it save? after every how many Epochs?

[strong fulcrum]

yea i downloaded the latest, 3.0.9, but the problem i was having was with running it

[versed linden]

5

[strong fulcrum]

i was getting errors in the setup when i ran the sh install file

[wispy burrow]

At what Epocg did you stop? And at what Epoch did it continue training?

[versed linden]

I stopped at 55 epochs, and it continued from 15

[sly ferry]

i need help installing rvc

if someone is able to help that would be great

[brittle wing]

Id it overtrained?

Is*

Also where was the guide for overtraining

[smoky torrent]

Which model in UVR to split an octave in a vocal?

[rugged stirrup]

You should probably do the lowest point first which is here

Like if you zoom in near there you could find the lowest point and you should save a model near there if you have any

[brittle wing]

WHAT

Is it finished

[rugged stirrup]

Hold left click and highlight the area that is near where I screenshotted and zoom in all you do is hold left click highlight it a little and release left click

[brittle wing]

Where is the guide

[brittle wing]

Did I finish training?

[rugged stirrup]

It looks like you finished but idk because idk how many epochs you did

[brittle wing]

350 for a 82 samples dataset

[rugged stirrup]

Use a model in this area

[brittle wing]

Well 350

[rugged stirrup]

Is that near that area?

[brittle wing]

Yes

[rugged stirrup]

Take a picture of how many steps it has

Which means just take a picture of the name of the pth file

[brittle wing]

The steps on the graph???

[rugged stirrup]

No the name of the model it has the amount of steps on it

Are you using RVC disconnected?

[brittle wing]

Where do I find those steps

[brittle wing]

Yes

[wispy burrow]

Don't you have one that is slightly earlier than the last one?

[brittle wing]

Nope but I have it at 350

[rugged stirrup]

Well you go to the colab files click the dropdown in the RVC folder go to the assets folder and click the dropdown and then you will see the weights folder

[wispy burrow]

How often did you save?

[brittle wing]

10

[rugged stirrup]

So then you can take a picture of the name and it will have the steps next to the name

[wispy burrow]

Then Epoch 330 or 340 could be better.

[brittle wing]

I can't find it

[rugged stirrup]

Its at the left below the keys button

[brittle wing]

Ohhh

No

[rugged stirrup]

It should look like that

[brittle wing]

I can't find the RVC folder you're talking about

[rugged stirrup]

And you should click the dropdown on the left next to the RVC folder

[brittle wing]

I know it already

[brittle wing]

What rvc folder there's no such

[rugged stirrup]

It should make an RVC folder

Refresh it

[brittle wing]

Which one amongst these

[rugged stirrup]

Mangio RVC Fork

[brittle wing]

Next?

[rugged stirrup]

Click the small triangle button to the left next to it

[brittle wing]

I know already just tell me which folder next

[rugged stirrup]

Now you will see a weights folder

[brittle wing]

BINGO

Which one I choose

[rugged stirrup]

Take a picture of the pth files

[brittle wing]

I'm seeing these for the first time

[rugged stirrup]

Extend it fully so I can see the names of it like the full names

[brittle wing]

I did it just didn't load early

[fallow sparrow]

i have a question, so i have a chromebook instead of a pc, how do i make ai song covers on my chromebook?

[brittle wing]

How do I extend it on a tablet

[rugged stirrup]

There should be a little thing to the right that you can pull out

[brittle wing]

Are they all .pth files?

[rugged stirrup]

Yes the files in the weights folder are all pth files

[brittle wing]

So basically the one e40???

Is the model that isn't overtraned

[rugged stirrup]

Maybe try epoch 320?

If you have that

[brittle wing]

Was my model truly overtrained according to the graph

[rugged stirrup]

It doesnt look overtrained it looks like you can choose a model from the circled area in the steps it might give you decent quality

[craggy quarry]

"File is not a zip file" - NO ONE ON THIS SERVER SOLVED THIS YET!!! Why is that??? I've been looking for solution for hours, and also, IT IS a .zip file!!!!

[rugged stirrup]

Do you mean the audio file isnt in the zip file?

[brittle wing]

But was it overtrained there

Should I continue or end it

[rugged stirrup]

Usually the lowest point or near the lowest point is the best

[rugged stirrup]

You should stop the training

[brittle wing]

Well I DID?

So the 350 one is okay???

[craggy quarry]

When I'm about to import the model that I made, I paste the shared Google Drive or Mega link for the zip file containing the model, it says: Error, file is not a zip file

[rugged stirrup]

Do 320?

[brittle wing]

Uh

Are you sure

[rugged stirrup]

Yeah

[rugged stirrup]

Do you have the index and pth file in the zip?

You also have to make sure inside the zip the pth and index arent inside another file cause that will mess it up

[brittle wing]

Which one is the correct.pth file

[fallow sparrow]

i have a question, so i have a chromebook instead of a pc, how do i make ai song covers on my chromebook? i want to make an ai cover where lowtiergod sings stronger than you

[rugged stirrup]

There is no "correct" pth file just experiment with each one and whichever one sounds best to you keep the rest you can delete

[rugged stirrup]

You have to use colab

I don't think there are any ways to use RVC inferencing on chromebook

[brittle wing]

Uh well okay

[craggy quarry]

The files it got inside:

G_104800.pth
D_104800.pth
config.json

It used to work like this back in May.

[rare goblet]

Ayo? @craggy quarry level 1 !!!

[brittle wing]

But the model is finished right there's no need to waste any more time

Right?

[rugged stirrup]

G and D are for continuing training make sure its the pth file and index

[brittle wing]

Generative & discriminative

[rugged stirrup]

Ik that

[rugged stirrup]

You can try to inference cause you stopped training

[sly ferry]

i need help installing rvc

[rugged stirrup]

What OS do you have

[sly ferry]

windows

[brittle wing]

Okay thanks but I don't need to spend any more time training it right?

[rugged stirrup]

What GPU do you have

[rugged stirrup]

No you dont need to train it any more

[sly ferry]

how do i check?

[rare goblet]

Ayo? @sly ferry level 1 !!!

[brittle wing]

EVEN BETTER

[rugged stirrup]

Just test the models and whichever one sounds the best to you save

[rugged stirrup]

Open task manager

And go to the performance tab

[sly ferry]

Amd radeon

[brittle wing]

I trained it at 700 before but it didn't sond bad but I'm glad I learned about overtraining now

[rugged stirrup]

Oh you need the AMD version here: https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006AMD_Intel.7z

[rugged stirrup]

Great

[sly ferry]

thank you

[rugged stirrup]

When its fully downloaded unzip it and you will see the goweb.bat file just run it and you have it open

[sly ferry]

ok thanks!!!!!

[craggy quarry]

i fixed the problem, just needed to upload it to HuggingFace and get the direct download link! But now, the model is outdated! It's RVC V1!!! :DDD And says use a model trained by RVC V2, but well I don't wanna train a model again for 1 week 😭

[rugged stirrup]

Alr then

[sly ferry]

im not really as good as tech as alot of people are, how do i unzip a folder?😢

[rugged stirrup]

Right-click it and there will be a menu that pops up you will see extract file click that and then another menu will pop up and click extract all

[sly ferry]

thanks

[strong fulcrum]

after pip installing rvc, i thought I got it working, but i got this error. anyone know why?

    raise ValueError(f'mutable default {type(f.default)} for field '
ValueError: mutable default <class 'fairseq.dataclass.configs.CommonConfig'> for field common is not allowed: use default_factory```

[craggy quarry]

Interesting. I got only these 2 files, and no more and the config. Do I have to close the model some way?

[hardy terrace]

Eveyrtime i go to paste thelink into the second step it never workss? I have tried many times with different voice models but still nothin??

[rugged stirrup]

Paste the link in what?

[hardy terrace]

The second step where it says "Download the mdel"

model

[rugged stirrup]

You arent supposed to use G and D for inference

[rugged stirrup]

What are you using easyGUI?

[hardy terrace]

Yes

[rugged stirrup]

Which link are you using

[hardy terrace]

https://huggingface.co/C0ttontheBunny/Portalmodels/resolve/main/GlaDOSOv2V2.zip?download=true

That one

[rugged stirrup]

Then it should work

[craggy quarry]

💀 ⚰️

[hardy terrace]

I could show a ss?

I tired again and is says error

[rare goblet]

Ayo? @hardy terrace level 1 !!!

[rugged stirrup]

You have to use the .pth file and index which are products from training with the G and D

Pth is core parameters and index helps with accent

[rugged stirrup]

Yeah you can

[visual flint]

I'm using this RVC Colab page (https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing), but I get to Step 2 (downloading the voice model) and i get a red error message saying "JSONDecodeError # File exists, proceed with creation of creds and client". I've bought some Compute Units from Google Colab but the problem still occurs. Any ideas?

Screenshot: https://imgur.com/Y0VHjkl

[rugged stirrup]

Outdated colab

Heres the new version: https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw

[hardy terrace]

Are there any yt videos on how to use this new version?

[rugged stirrup]

Click the install RVC play button and once its done with a checkmark click open GUI

[hardy terrace]

Alright

[visual flint]

having a look now thanks! if i have any questions about how to get it up and running ill ask

[rugged stirrup]

You just click the Install RVC play button and once thats done click open GUI

[marsh trout]

guys

guyz

[visual flint]

is this usable on mobile btw?

[rugged stirrup]

Yes it should be useable on mobile

[stiff tusk]

now i have the word band link do i just paste it here?

[rich sphinx]

btw, any models you guys wanna suggest that just do a genuine male to female voice (no anime this and that, just normal)

[rare goblet]

Ayo? @rich sphinx level 1 !!!

[visual flint]

is there a text to speech on this btw or just speech to speech?

[rugged stirrup]

Only speech to speech

[visual flint]

Ah fair enough. I don't suppose u know if there's a free text one available? Thanks for the help regardless, i'll probably still use it!

[rugged stirrup]

You could try this: https://on4t.com/text-to-speech

[brazen gorge]

Album could not be found, please check your search values and try again.

Search value: 'model:markiplier youtube:https://youtu.be/Kq914-2jmMs?si=F1st5fl3TpcJKGCh'

[rugged stirrup]

300 character limit though

[brazen gorge]

Album could not be found, please check your search values and try again.

Search value: 'model:markiplier youtube:https://youtu.be/Kq914-2jmMs?si=F1st5fl3TpcJKGCh'

[rugged stirrup]

@stiff tusk #🤖│bots

[stiff tusk]

my bad

ill stop

jus tryin to figure it out ill see if theres a setting to make it not appear mm

[visual flint]

That's fine, just got a great example out of it lol. Now that I have the GUI up and running, can i just continue to use the gradio.live link forever or will i need to launch a new session via the colab each time I use it?

[rugged stirrup]

You will have to relaunch because colab has a limit

If you want to have true unlimited run locally

[stiff tusk]

which channel should i use for the wordbank link / is there a way to have the bot message only show for me

[rare goblet]

Ayo? @stiff tusk level 1 !!!

[rugged stirrup]

If you want to make Ai covers use astra labs or the weights bot

[stiff tusk]

idk what the weights bot is but ill look at the astra labs channel ty

[blazing turret]

does anyone know how to download models off of weights.gg

i cant figure it out

[visual flint]

Last question, thanks for all the help btw. When i want to launch a new session, i open the Colab page, do i have to install RVC again or can i skip to opening the easyGUI?

[rare goblet]

Ayo? @visual flint level 2 !!!

[blazing turret]

i see where it says zip file but theres no option to download it

[rugged stirrup]

Yes you have to install RVC again

[visual flint]

great thanks

[blazing turret]

i cant figure it out

have you used weights?

[rugged stirrup]

What are you trying to do

[blazing turret]

im just trying to download a model

from weights

[rugged stirrup]

Download it and extract it

[blazing turret]

where do i download it?

[rugged stirrup]

You click 3 dots and it should have an option to download a model

[blazing turret]

oh my god your a life save

[rugged stirrup]

Its to the right

[blazing turret]

thank you sm

[rugged stirrup]

Np

[brittle wing]

-colabs

[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

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

[heady hare]

can someone help me i cant hear anything and idk how to do the settings

[low shard]

are you using the realtime voice changer?

[heady hare]

i dont know

this is how it looks like

oh

yeah

[heady hare]

realtime voice changer client

[low shard]

yea, for that ask in #🔍│help-w-okada or #1192011222023950368 then, sorry but i dont use voice changers so idk how to help you, but an helper will help you in sometime in that channels

[heady hare]

i asked in the forum

[bold nexus]

can someone help me? i cant downland the model and i get this error:
ModuleNotFoundError: No module named 'mega'

[low shard]

oke wait for an helper to come

[heady hare]

@low shard it also shows me a really big CMD whenever i open it

is that normal

like

a command prompt

and if i close the command prompt it also closes the application

[brittle wing]

how to use the voicemodel???

i downloaded spongebob

[low shard]

Yes, it uses cmd its normal, just idk how to help your problem as i dont use it

[heady hare]

alr

[low shard]

Please read in our docs https://docs.aihub.wtf

[brittle wing]

i dont even have the program

[sly ferry]

when i try to open the batch file it says windows protected my pc

[rugged stirrup]

Just click run anyway

[tidal cairn]

can someone refer me to rvc links for google collab

trying to find active ones

[unique elk]

do i have to click allow to "Allow this notebook to access your Google credentials?"

[rugged stirrup]

Yes

[unique elk]

does it actually like

[rugged stirrup]

No

[unique elk]

get access to my google credentials

ok

[rugged stirrup]

It just needs access to the drive

[unique elk]

oooh i see

ok thank you

[rugged stirrup]

Np

[sly ferry]

is this one for songs?? im trying to use it for that

[rugged stirrup]

Yes you have to put the pth file in the weights file and the index in the logs file

[brittle wing]

when opening start.http it shows me cmds and then pops away

[rare goblet]

Ayo? @brittle wing level 1 !!!

[sly ferry]

how do i add my song

[rugged stirrup]

You copy the path of your mp3 or wav and paste it in the

[brittle wing]

kimid

[brittle wing]

.

[sly ferry]

thats not what mine looks like

[rare goblet]

Ayo? @sly ferry level 2 !!!

[rugged stirrup]

#🔍│help-w-okada

[unique elk]

what do i do once ive started "open the easygui"?

[rugged stirrup]

Click the Gradio link

[unique elk]

it opened this tensorboard thing but it says no dashboards are actvie

ooooooh

[rugged stirrup]

It only activates when you train a model

[unique elk]

ok thank you again :)

[rugged stirrup]

Np

[unique elk]

uuuuh last question hopefully

[rare goblet]

Ayo? @unique elk level 1 !!!

[unique elk]

how do i put in my model voice

[rugged stirrup]

You go to the model download tab

[rare goblet]

Ayo? @rugged stirrup level 18 !!!

[rugged stirrup]

And paste your drive or huggingface link into it and click download

[unique elk]

ooooooooooh

ok thank you again again

[rugged stirrup]

Np

[sly ferry]

what do i do here?

[rugged stirrup]

That is used for real-time inferencing if you want to use it just select the model by using the button on the right and make sure your microphone is selected and you also need VB cable for that

https://download.vb-audio.com/Download_CABLE/VBCABLE_Driver_Pack43.zip

[pastel oak]

I would recommend VAC, VB-Cable causes too many problems

https://software.muzychenko.net/freeware/vac470lite.zip

Run "setup64"

[rugged stirrup]

I use VB cable usually

[jovial otter]

each time i put my model voice, it's always said 'failed to download file'

[rare goblet]

Ayo? @jovial otter level 1 !!!

[bold nexus]

pls help i cant downland the model

Easy GUI (for RVC v2, with crepe) (with improved downloader)

ModuleNotFoundError Traceback (most recent call last)
<ipython-input-1-bb7ca959319e> in <cell line: 4>()
2 #@markdown Link the URL path to the model (Mega, Drive, etc.) and start the code
3
----> 4 from mega import Mega
5 import os
6 import shutil

ModuleNotFoundError: No module named 'mega'

[pastel oak]

old colab, use one of the newer ones

check pinned message, open the docs guide

i recommend ilaria or rvc disconnected

[bold nexus]

ty

[blissful bough]

Hello I need help but i cant send pictures

I get an error that says this:

RuntimeError: The expanded size of the tensor (12800) must match the existing size (12400) at non-singleton dimension 1. Target sizes: [1, 12800]. Tensor sizes: [12400]

Im trying to train a model further that ive already trained

Could someone help please

[rare goblet]

Ayo? @blissful bough level 1 !!!

[pastel oak]

whats ur gpu

whats the colab link

[blissful bough]

i have an rtx 2070 super

and im not using colab im using the rvc website thing

Mangio-RVC-Web it says

[blissful bough]

https://www.youtube.com/watch?v=ixB9oalT3cQ&t=1236s i used this tutorial

so you know what kind of rvc im using

[pastel oak]

i looked into the history of this channel for anyone thats had this issue and theres not rlly been a solution

you can try:

when selecting the samples and versions, you must always reclick v1, then v2, then reclick the correct sample rate

see if that works

[blissful bough]

thanks ill try

nah it didnt work

[pastel oak]

ya no clue for the fix

id recommend using Mainline instead of mangio crepe

its in pinned messages, docs guide

[blissful bough]

alr thanks

[marsh perch]

What is "dataset" ? Maybe that's a stupid question, im sorry

[blissful bough]

i fixed it

[low shard]

check https://docs.aihub.wtf/rvc/resources/datasets/#introduction

[marsh perch]

So, is this just the voice that you train, but everyone call it by a very smart word as "dataset" ?

[pastel oak]

how?

[low shard]

well its basically the 'samples' of the voice you wanna train in rvc context

[blissful bough]

basically just renamed the model (to create a new one)

[rare goblet]

Ayo? @blissful bough level 2 !!!

[blissful bough]

because apparently you cant train the same model further if it has the same name

[fair vortex]

hi, does anyone know why vb cable no longer works, any solutions ?

[rare goblet]

Ayo? @fair vortex level 1 !!!

[pastel oak]

gotcha tyy

[pastel oak]

vb cable shit, download vac instead

https://software.muzychenko.net/freeware/vac470lite.zip

Run "setup64"

[rare goblet]

Ayo? @pastel oak level 21 !!!

[pastel oak]

uninstall vb-cable first and restart pc

[marsh perch]

Understood, atleast now i can understand what people mean by saying "you have a bad dataset" thanks

[fair vortex]

thank you

[low shard]

yw

be careful about the cleaning

of your dataset

the dataset in rvc context needs to be pure cleaned vocals

[marsh perch]

Ye, i tried to clean it well, but the model appeared very bad

Maybe its because the dataset is only 18 secs

[low shard]

its better if you get a longer dataset for better results, and also be sure its very well cleaned, also could depend on the epochs, in case it overtrains

[calm wyvern]

does anyone have any way to get rid of the robotic sound i get.

I try everything to remove background noise, but the leftover noise either whitenoise or just robotic residue interfers with my training

[wispy burrow]

What are you using to isolate the voice?

[marsh perch]

What is the best way to clean the dataset ?

[calm wyvern]

uvr5

[rare goblet]

Ayo? @calm wyvern level 1 !!!

[low shard]

check this https://docs.aihub.wtf/rvc/resources/vocal-isolation/ , if you have a good pc use uvr locally, if you dont use the colab

[marsh perch]

Does it work good with AMD ?

[wispy burrow]

I've had some good results with UVR.
It can isolate thevoice, remove echoes , reverb etc

[marsh perch]

Is rx570 4gb and R7 2700x enough to work with UVR ?

[low shard]

it should https://github.com/Anjok07/ultimatevocalremovergui/releases/download/v5.6/UVR_v5.6.0_setup_opencl.exe

[marsh perch]

Yep, alr dowloaded that version

But i don't really understand this\

That's why im asking

[wispy burrow]

I'm not sure if AMD is suppoerted but it can work on CPU if it isn't.

[low shard]

it can work for cpu, but will be wayy sloww, if you want you could try the google colab of it using the free daily gpu

[boreal sapphire]

can any guy guide me

im having some problems with it

[low shard]

with what

[boreal sapphire]

i installed this voice ai

first i can listen to myself and second my voice is not constant with the changer

its lagging basically

[low shard]

do you mean https://voice.ai/ or the realtime voice changer from our server ?

[wispy burrow]

What UVR models did you use?

[boreal sapphire]

yes

[rare goblet]

Ayo? @boreal sapphire level 1 !!!

[pastel oak]

i love it when you ask a x or y question and people say yes

[low shard]

i mean, the link i sent or the one u downloaded from the server ?

[boreal sapphire]

no from website

[low shard]

btw, you like helping people?

[calm wyvern]

wdym by that?

[low shard]

we arent affiliated with voice.ai, its paid, we use wokada realtime voice changer which is all free, and can use rvc voice models from our server

-realtime

[azure marsh]

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

[low shard]

if you want you can use the wokada one

but we dont offer support over voice.ai

[boreal sapphire]

oh how can i ddowlnoad it

[low shard]

https://rentry.co/VoiceChangerGuide

[boreal sapphire]

alr lemem uninstall this one

[low shard]

follow this guide, if you get any error, please ask in #🔍│help-w-okada or #1192011222023950368 but not here as its an help channel for rvc which is for making ai covers

[boreal sapphire]

alr tysm

[low shard]

sadly i cant help you about wokada as i dont use voice changers (i never call ) but you will find helpers in the help channel ready to help you

yw

[boreal sapphire]

btw im on a site

how to dolwnoad it from here

[boreal sapphire]

.

[pastel oak]

follow the steps, its very straight forward

[boreal sapphire]

its saying doolwnoad nvidia

[pastel oak]

if you have nvidia you click on the link for nvidia. if you have amd you click on amd

this is about GPU

[boreal sapphire]

oh ok

its 3gb

[pastel oak]

yes

[wispy burrow]

There are different models that can be used. Some are focused on different things.
For example, UVR-MDX UVR-NET-Voc-FT-to isolate vocals or DeEcho-DeReverb for removing echoes/reverb

[shrewd eagle]

Are there python libraries I can use to integrate RVC into my pipeline?

[fair vortex]

uhm, how do i uninstall it :)) i tried it already but it seems to be still in my pc and i think thats why vac doesn t work

[pastel oak]

search for "add or remove programs" on your windows search bar of your pc, open the thing, uninstall VBCable

[calm wyvern]

yeah i do use them

[fair vortex]

done, vac still doens't work on discord, do you know why ?

[pastel oak]

can you send screenshot of wokada and discord input/output

also maybe because you had both installed at the same time it could have messed up tbh

if you havent restarted pc yet do that, if you did then maybe reinstall vac

but lets see your wokada and discord first

[fair vortex]

@pastel oak

[pastel oak]

wokada Select output: Line 1

and on discord

Input: Line 1

[fair vortex]

thank you man

it works now

🤍

[pastel oak]

i recommend you use

Audio: Client

I think chunk 96 is going to be tough with 1060, you probably have to raise it but you do you
You could use 8k as extra probably for bit better quality

[fair vortex]

thank you for advices

[half arrow]

Hey guys, any help on this? JSONDecodeError: Expecting value: line 1 column 1 (char 0)

I keep getting this and Im a bit stuck on how to fix

[oblique solar]

Same issue as @half arrow

[rose bear]

what does epochs mean?

[pastel oak]

youre probably using an outdated colab, check pinned messages, open docs guide and choose one of the newer cloud/colab ones. i recommend ilaria or rvc disconnected

[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

[vernal lintel]

which

[oblique solar]

thank you, it was the outdated one, Ilaria works like a charm

[dense ermine]

does anyone know how to make AI covers with multiple people singing?

[oblique solar]

I would recommend separating individual singers into separate tracks and convert them one by one with voices you want

[rare goblet]

Ayo? @oblique solar level 1 !!!

[dense ermine]

how would u recommend i do that?

[oblique solar]

if you dont have the separate tracks available, and if they do not sing over each other, cut the track into pieces in Audacity?

[dense ermine]

okay thanks

[proud elbow]

try separate individual voice into its own track, duet singing together (esp different gender) may be hit or miss, and group one may even be less possible

[dense ermine]

will give it a shot and tell u how it comes out sometime

[dense ermine]

@low shard how would u suggest i do this

[brittle wing]

So uh about this updated new RVC Collab, can it work on mobile?

[low shard]

if the people are talking at different times, just uvr mdx23c model so it separates the vocals and instrumentals of the song, and then you manually check the vocals and cut out in 2 different files the 2 different parts where the voice talks, then use different rvc models on the different 2 files and then put this all back together

if its 2 people singing at the same time, well you could firstly separate the vocals and instrumentals with uvr mdx23c, then use the vocals as an input in uvr and use the hp karaoke 6 model so it separates it into 2 different files (1 for a voice, 1 for another) but may not be perfect

[low shard]

yup

[brittle wing]

Oh goodie

Speaking of which, can it also let you copy and paste the model link without downloading it? Like does it depend on which collab you're using?

[low shard]

wdym without downloading it? just get the download link of the model you want, then paste it in the part where it asks you for the download link based on which colab you are using, then just click download, you need to do that to use a model

[brittle wing]

Well I'm saying on the old Collab, you can just copy and paste the download link and it will work, you're saying I have to download it now?

[low shard]

which colab are you using btw

[brittle wing]

Ok

I got it working but how do I use my downloaded model? Do I copy and paste from my drive?

Nvm I'm looking at the instructions from the website

[low shard]

you need to go to download tab, paste the download model link, and then click download

[brittle wing]

Ok

It worked so now what?

Do I download it again? Or copy and paste it or press voice models then it will allow me to use it?

[low shard]

okay now you can use it, please check from here https://docs.aihub.wtf/rvc/cloud/applio-colab/#2-select-voice-model , its the full guide of the colab

[brittle wing]

It worked

[olive tiger]

why is it super choppy in actual calls

[pastel oak]

do you use VB-Cable or VAC as virtual cable

First says "CABLE Input" the second "Line 1"

[olive tiger]

VB

wym first says

[pastel oak]

people dont know the difference so im telling you what the output is called for both

VB is bad, get VAC and try again

https://rentry.co/VoiceChangerGuide#virtual-audio-cables-mandatory

uninstall Vb first, restart pc then get vac

[olive tiger]

i have both

[pastel oak]

if it still does issues, go to step 18 and do the task manager thing

[proud elbow]

click refresh in inference page so it will show up in the model list (u can also check the logs folder to make sure)

[dense ermine]

I keep getting this error message "raceback (most recent call last):
File "/content/Ilaria-RVC/app.py", line 1, in <module>
import subprocess, torch, os, traceback, sys, warnings, shutil, numpy as np
ModuleNotFoundError: No module named 'torch'" when i start to download the tunnel gradio on ilaria

?

[proud elbow]

have u followed the guide for settings (f0, threshold, chunk, extra, etc)

[dense ermine]

please help

[rare goblet]

Ayo? @dense ermine level 3 !!!

[tawdry bridge]

Hi, I have a question.
I currently have a 21 minutes dataset, already separated the vocals with UVR and cleaned the set with RX10 and several plugins.
Still there's a part of the dataset, around 1 minute and a half, where you can hear that a part of the music wasn't able to be removed, as you can hear some parts of the song really intertwined with the voice.
Would it be better to just remove this section of the dataset, or should I train with it?

[proud elbow]

try BS roformer in MVSEP (and probably combined with MDX23C)

[olive tiger]

i did all of it now theres js demonic whispering

[rare goblet]

Ayo? @olive tiger level 1 !!!

[pastel oak]

💀 what

[olive tiger]

theres like weird noises

it sounds like demonso

😭

[pastel oak]

send screenshot of wokada and discord input output

[olive tiger]

wokada?

[pastel oak]

pls use rmvpe

crepe is olddd

[olive tiger]

ohhh okok

[pastel oak]

use chunk 96 and extra 16k

you could go even lower in chunk probably

try again now

[tawdry bridge]

ill check that, thanks a lot

[olive tiger]

i did all that its js picking up the sound of my ppc fans

pc

and making it sound odd

[pastel oak]

did you select the correct input as microphone

ohh

ServerIO Analyzer

set that to your headphones too

[olive tiger]

huh

im confused mbmb

[pastel oak]

how

[olive tiger]

like where

[pastel oak]

on the right is something called serverio analyzer

big

[olive tiger]

OH

i see it mb

[pastel oak]

set that to headphones

[olive tiger]

it is

[pastel oak]

and on audio for input make sure its your main mic and not some webcam thats next to your fans or something lol

[olive tiger]

my fans are js hella loud

😭

lemme put a towel on my pc

💯

[pastel oak]

you can enable sup2

[olive tiger]

shi u right

[pastel oak]

and echo maybe not sure if itll help in this case

[olive tiger]

IT WORKS

tsymm

[rare goblet]

Ayo? @olive tiger level 2 !!!

[brittle wing]

I already got that and yep, the updated new Collab is better

Hey that reminds me, does the weights website voices work on those Collabs?

[rare goblet]

Ayo? @brittle wing level 6 !!!

[proud elbow]

as long they provide the huggingface direct download link like you had

[tawdry bridge]

thanks a lot again, a single pass of the BSRoformer did the trick
now I want that to be my to-go algorithm, I'll wait for it to be implemented on UVR or similar.

[rare goblet]

Ayo? @tawdry bridge level 1 !!!

[short tide]

The updated website is so confusing now, Help🥲

[swift kettle]

guys pls help . i keep being stuck in step 2, whatever the link put in they said error

even google drive or huggingface

[foggy pebble]

How can I make a parody of a song where I change the lyrics but have the same singing?

[thin stump]

Channels like There I ruined it used AI voices but they mixed it in a DAW. I found this pic on someone's Fiver page

[foggy pebble]

Thanks!

[rare goblet]

Ayo? @foggy pebble level 1 !!!

[thin stump]

I posted the link on #1159290752195633273 since I don't want to advertise by accident

[raw wedge]

I got disconnected from Google colab for using Applio saying it is not intended for such purposes

Is there anyway to get around this?create new account?

[rare goblet]

Ayo? @raw wedge level 2 !!!

[old oracle]

-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

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

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[ionic yew]

is there a google colab tutorial to make GPT-SoVITS models?? because I'm confused with the file directory for the dataset. cus all i found is how to do it locally but otherwise everything else is in Chinese.

[ionic yew]

don't click it

[frosty lichen]

how do i use the new ilaria?

like how do i add a voice

[long hinge]

Entry not found

[proven hill]

in the new release?

[proven hill]

same method, either huggingface or you can drag and drop a zip

[frosty lichen]

mk

[long hinge]

Yep.

In github Hugging Face.

I fixed it it was corrputed Data.

[proven hill]

the link was fixed immediately

[long hinge]

Oh alright.

Thanks, llaria!

You're the best!

[proven hill]

awww, no worries!

[latent kettle]

https://github.com/TheStingerX/Ilaria-RVC-Mainline/releases/tag/beta1.0

training is now possible on mac ? with ilaria RVC mainline beta 1?

[proven hill]

trainign on mac no, but on colab yes

[latent kettle]

training on mac when? or never 😭

[sand iris]

Needs :
a) apple to make MPS (that allows GPU accel for this kind of AI stuff) behave properly
b) rvc training code to be slightly more convenient to deal with, instead of being one giant black box

[proven hill]

doesnt have CUDA, so until i dont go crazy and find a way for AMD and Intel cards..

also what my fav frechie is saying

[latent kettle]

is there are any chances in future for training on mac?

[sand iris]

¯_(ツ)_/¯

[latent kettle]

i think Mac is useless for Ai stuff

[proven hill]

yes

[sand iris]

Apple being apple, it's down to someone knowledgable willing to implement it & have the damn hardware on hands

[proven hill]

if i enable training for CPU, yes, but good luck with that

[opal mica]

Guys someone help how do i change my voice to brimstone in valorant

@proven hill

@sand iris

any mod?

[pastel oak]

stop randomly tagging people

do you have the voice changer installrd

[opal mica]

Saw-y-

[rare goblet]

Ayo? @opal mica level 1 !!!

[opal mica]

im--- ne---w

[pastel oak]

what gpu do you have

[opal mica]

amd

radeon

r7

why tho

[pastel oak]

i dont think that gpu is gonna be strong enough for the voice changer

youll have big delay

you can still try

[opal mica]

W

[pastel oak]

https://rentry.co/VoiceChangerGuide

follow AMD steps

[opal mica]

Aew

[pastel oak]

and you need brimstone voice model, if someone created it you can use it

[opal mica]

How

tutorial

?

i used search

[pastel oak]

tutorial is the guide

just read

[opal mica]

and found and download

Where id tutolrial

[pastel oak]

https://rentry.co/VoiceChangerGuide

[opal mica]

#tutorial

[fluid zephyr]

So what exactly is the advantage of ilaria RVC vs this ---https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI

[indigo aspen]

can somebody explain what these pretrains are?

[fluid zephyr]

I know it says itsfaster and everything but I'm guessing it's a whole like.. revamp right?

[fluid zephyr]

Just different pre-train models, I personally use OV2

Is it like a whole new RVC AI trainer and stuff? Like with brand new code

[indigo aspen]

does it like affect the model output? like will it sound different?

[finite galleon]

Yeah a lil bit

[fluid zephyr]

Yes, ov2 makes it sound more realistic and better

It depends what voice you're training though some sound better than others on OV2

Also

So ilaria RVC is like a whole new thing, damn I just checked I'm so dumb 💀, (I'm sorry ilaria)

[hard gust]

Does Ilaria's RVC fork has any training-helper features? Like Applio's overtraining detector

[blissful skiff]

anyone know what the passthrough button does?

[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

[proven hill]

yes, new thingie from before :3

also no youre not dumb!
just ask if you have any curiosity <3

[fluid zephyr]

Thank you!

I'm gonna try it out soon