✨│ai-help

[rugged stirrup]

Is this raw output through RVC?

[brittle wing]

No

[rugged stirrup]

What is it then

The dataset?

[brittle wing]

It's a conversion from the webui with removed lead vocals and unbreverbed

And I filtered out the instrumental

[rugged stirrup]

Try to disable that option

[brittle wing]

What option

[rugged stirrup]

So you dont have to filter out the instrumental

[rugged stirrup]

The option to combine it all into one

Do you use Aicover gen?

[brittle wing]

There's no such option

[rugged stirrup]

Oh

[brittle wing]

Yes most frequently

[rugged stirrup]

Well try to use another colab or something

[rugged stirrup]

Use easygui or something for a moment

[brittle wing]

What...

[rugged stirrup]

So you can inference cleanly

[rugged stirrup]

Do you run locally?

[brittle wing]

Wait I have some samples of the 320 epochs one.

[rugged stirrup]

Alright

[brittle wing]

Raw samples

[brittle wing]

Now I'll inference the other one

[rugged stirrup]

There are quite a lot of artifacts in it

And you maybe trained the wrong sample rate?

[brittle wing]

The song is full of sfx

[rugged stirrup]

What sample rate did you use?

[brittle wing]

40k

[rugged stirrup]

what was your audio?

[brittle wing]

Normal Acapella dry

[rugged stirrup]

use a program like spek to check your actual audio sample rate

[brittle wing]

On Android?

[rugged stirrup]

No

Windows

[brittle wing]

Everything's fine with the dataset

[rugged stirrup]

There are online ones although they dont work as well

[brittle wing]

I don't have a computer to check

[brittle wing]

Uh

Judging from the previous samples just tell me which is the better one

[rugged stirrup]

Give me your model

[brittle wing]

I'll give you both.
https://drive.google.com/file/d/1M1rEj1y_2TCl7YnF42GU77gcuPIpL0kH/view?usp=drivesdk
https://drive.google.com/file/d/1KqU6CDaZWkl9LRzs79fOk0WsTpcgg7k1/view?usp=drivesdk
Nope they're still on my drive.
I'll upload the better one on Huggingface

[rugged stirrup]

DM it to me or something

Ill try to check it

Make sure its a huggingface link

Or actually

send it here

yeah

[rugged stirrup]

what are 2 links for?

[brittle wing]

They are still in my Drive

[rugged stirrup]

Oh

Nvm

[brittle wing]

One is the overtrained and the other one is the lesser epochs

[rugged stirrup]

Ok let me try it

[brittle wing]

Okay

[brittle wing]

It's meant for rapping

[rugged stirrup]

Ill just test it with a ballin cover sample

colab internet sucks

[brittle wing]

I'm curious who sings this song

Okay test em and tell me.

But probably you don't have a general idea of what jhope sounds like since you don't listen to BTS.

[rugged stirrup]

BTS raps?

[brittle wing]

Yes jhope is all singer & rapper

All-rounder

And dancer

[rugged stirrup]

Well ill just try to tell if theres background noise or artifacts

[brittle wing]

If you don't have a general idea of what he sounds like that's probably why

Cause I do and the "overtrained" one wasn't that overtrained

[rugged stirrup]

Idk why but its failing to download

Well since you have it on your computer probably try to use it and put an acapella through it

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

Its easy to use

[ivory solar]

@sand iris

[sand iris]

Yup, that's the outdated one

-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

[ivory solar]

ohhhh

thanks

[brittle wing]

Because you have to download and put the model in pixel drain & copy the link?

What colab is this?

[brittle wing]

How do I inference with easy gui

[rugged stirrup]

easy GUI

[brittle wing]

Cause I only see training cells

[rugged stirrup]

What? there is a install RVC button and a run GUI button

[brittle wing]

What do inference with - song or dereverbed lead vocals?

[brittle wing]

Ah okay

[rugged stirrup]

yeah

[proud elbow]

the latest one has the non ui inference cell

[brittle wing]

Does it have options to regulate accent and index

[rare goblet]

Ayo? @brittle wing level 9 !!!

[stark gulch]

how can I send pictures i need help

[brittle wing]

I downloaded models but it doesn't give me the option to choose them from the drop-down menu

I have refreshed like 2-3 times

[brittle wing]

Uh well but it doesn't let me pick a model I refreshed several times

[rugged stirrup]

That's what happened to me I don't know

[bitter notch]

what is path to dataset?

[rugged stirrup]

try to upload your file to hugging face

[rugged stirrup]

You go inside the data set file you create a file inside the data set file You put your mp3 or wav file into the file that is inside the data set file and then you copy the path of that file that is inside the data set file and you put it into the preproces data path thing

[brittle wing]

Oh well

[rugged stirrup]

just use the model you think sounds best in your opinion

[trail vector]

I wonder, is there any api for generating tts using rvc ?

[rugged stirrup]

RVC itself can't be used for TTS generation but you could use Microsoft TTS or whatever you want and inference it through RVC

[bitter notch]

how do i copy the path?

[rugged stirrup]

You right click on it and it should say copy path

[trail vector]

mmmmmmmmmmmmmmh good idea

[bitter notch]

it says there is no file or directory

[brisk quarry]

https://tenor.com/view/nuh-uh-beocord-no-lol-gif-24435520

[forest token]

is it possible at all to run inference on a Mac M2? I've tried installing multiple GUIs and all of them, even the ones that supposedly support mac, fail on installing/running

[rugged stirrup]

did you put the mp3 or wav file in there

[brisk quarry]

how do i resume training a model that already exist

[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

[sand iris]

What have you tried until now ?

Installing on mac is kind of a pain for various reasons

mostly apple inconsistencies, devs lacking hardware to test on, and poorly written installers

[rugged stirrup]

type in the name of the model and click train but make the epoch count higher than the epoch count it is For example if it's 300 epocs and you want to do 100 more you have to do 400 epocs

[hasty crow]

@sand iris applio is broken
ImportError: cannot import name 'Option' from 'typer' (unknown location)
make: *** [Makefile:20: run-applio] Error 1
im on paperspace and i can't run it
i've re-installed everything many times but it seems to be broken

[sand iris]

¯_(ツ)_/¯

Haven't dealt with applio in a while
Don't quote me on that, but they mostly focus on windows

[hasty crow]

this is a nightmare honestly

[wind grove]

How?

[hasty crow]

colab works fine but my paperspace notebook wont...

[finite galleon]

Shouldnt colab and paperspace be the same

[sand iris]

yesn't

[hasty crow]

well not exactly

[finite galleon]

Since you just use the same .ipynb

[sand iris]

if you rely on gdrive integration it wont' work

[hasty crow]

im not on a ipynb

[sand iris]

if you rely on colab quirks it won't

[hasty crow]

i unzipped the folder

[sand iris]

You're doing raw terminal ?

[hasty crow]

yeah

im following LolelSpaces guide

[sand iris]

Might want to go the manual way honestly

[hasty crow]

https://docs.google.com/document/d/1lIAK4Y0ylash_1M2UTTL_tfA3_mEzP0D2kjX2A3rfSY/edit#heading=h.jber1ci0vzjk

[sand iris]

ditch the makefile

[hasty crow]

but everything worked fine like yesterday

[sand iris]

python3.10 -m venv .venv
source .venv/bin/activate
# Get applio code via git
cd Applio
pip install -r requirements.txt
python app.py

kinda like that

[hasty crow]

what even happened

[sand iris]

¯_(ツ)_/¯

[hasty crow]

should i send you my run-install code will it help you?

resolve

[sand iris]

good idea

note that I can't test it atm

so any applio specific quirks I won't be able to guess

[hasty crow]

may i send you a we transfer link with the files cause the text is massive ?

[sand iris]

sure

the files should fit on discord though

[brittle wing]

How do I get the model url

Cause I copied& pasted and it failed to download.

[rugged stirrup]

I don't really know on Google drive been forever since I used it I always use hugging face now

[brittle wing]

I uploaded it to Huggingface

But where do I take the correct url from I forgot how

[rugged stirrup]

If you uploaded it to hugging face there should be a little download button next to your model right click it and it will open a little menu and click copy download link

[brittle wing]

https://huggingface.co/elliyonline/JackinTheBoxv2_320epochs/blob/main/jhopeJackInTheBoxv2.zip
???

Is it this one

[rugged stirrup]

Yes

Yes

[brittle wing]

Nvm done

[rugged stirrup]

Alright

[hasty crow]

ok so here are the files https://we.tl/t-oz7JkqXO39

they are from applio 3.0.8

and the error for me is something with temper not finding a function

[forest token]

Seems like https://www.tryreplay.io/ works fine, would recommend

[hasty crow]

how do i do that if makefile is not fixed?

[brittle wing]

Final answer

Both have artifacts I think but ...

Maybe it's cause of the lead vocals

[rugged stirrup]

how about you ask a model master or a quality control

[sand iris]

I just got back

how the hell is that error happening

[hasty crow]

what else do you want to know about my procedure

[sand iris]

https://github.com/search?q=repo%3AIAHispano%2FApplio+typer&type=commits

no mention of typer anywhere

...nevermind

[hasty crow]

this is the exact error code

[brittle wing]

Where

[hasty crow]

Traceback (most recent call last):
File "/notebooks/Applio/app.py", line 1, in <module>
import gradio as gr
File "/usr/local/lib/python3.9/dist-packages/gradio/init.py", line 12, in <module>
from gradio.cli import deploy
File "/usr/local/lib/python3.9/dist-packages/gradio/cli/init.py", line 1, in <module>
from .cli import cli, deploy
File "/usr/local/lib/python3.9/dist-packages/gradio/cli/cli.py", line 4, in <module>
from gradio_client.cli import deploy_discord # type: ignore
File "/usr/local/lib/python3.9/dist-packages/gradio_client/cli/init.py", line 1, in <module>
from gradio_client.cli import deploy_discord
File "/usr/local/lib/python3.9/dist-packages/gradio_client/cli/deploy_discord.py", line 3, in <module>
from typer import Option
ImportError: cannot import name 'Option' from 'typer' (unknown location)

[brittle wing]

How to do qualify control?

Where to find a model master?

[rugged stirrup]

ping somebody that is quality control like waste of RVC zero or Diablo commissions DM?

just ping once

[sand iris]

I don't really have the time to troubleshoot this, but that's some wonky shit

[hasty crow]

where can i find help

[sand iris]

Probably version mismatches

[hasty crow]

docs are useless

chat gpt too

iahispano mods are sleeping for good

[brittle wing]

Okay but how

[rugged stirrup]

You just ping them on discord like any other person

[brittle wing]

Dm?

[rugged stirrup]

You just @ them

[rugged stirrup]

no

[brittle wing]

Okay what's their name

How do I find them

Username

[rugged stirrup]

just say at waste of RVC zero or Diablo commissions DM or a model master like lista the dancer or something one of those

And you can ask them to review your samples and they can tell you their opinion

[brittle wing]

Lisa the dancer?

[rugged stirrup]

I think that person's not on right now

[brittle wing]

What's their username

[rugged stirrup]

so you probably need to do quality control

[brittle wing]

How

To do that

[rugged stirrup]

You at their name

[proud elbow]

how about try submit your model in #1159514067187277865
might get some useful feedback

[brittle wing]

But I don't wanna post my model there...on Discord I'll delete my account

With what program do I do quality control?

I already had this role on the account I deleted

[brittle wing]

It doesn't allow me to add image

[proud elbow]

discord mobile issue?

[brittle wing]

Could be

[half cove]

what model pretrain did you use here btw?

[keen mica]

is this graph a good start or is it bad ( I heard if its too curved or some shit its bad)

[hardy terrace]

When I try to past the link for the model I want to used for some reason it doesnt work

[rare goblet]

Ayo? @hardy terrace level 2 !!!

[hardy terrace]

Yo?

[terse shard]

What's the current recommended RVC software?

[proper shale]

seems to be a good start

[proper shale]

there's applio and mainline ig

applio has cool features, mainline is more focused in stability

it really depends on what u want

[proper shale]

give us more details

what are you using?

what error does it give?

any logs?

[trail marlin]

how do i drag the zip into voice?

[ebon rover]

my voice changer doesnt work

[proper shale]

can you give cmd/terminal logs? any extra details?

[ebon rover]

i mean i downloaded everything but when i speak it doesnt change my voice

i clicked start

[proper shale]

yeah check cmd and see if there's any Tracebacks, or anything that indicates some sort of error

[ebon rover]

Ok

[brittle wing]

How do I find all of the deleted ai models from the old server that used to be from deleted accounts
Aside from huggingface, applio and weights

[rare goblet]

Ayo? @brittle wing level 1 !!!

Ayo? @split comet level 1 !!!

[split comet]

im using mac

[latent kettle]

-realtime

[azure marsh]

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

[rocky lantern]

Hey so I'm in need of some help.

Problem that I'm having is that my voice changer program keeps on saying "Pipeline is not initialized." Immediately after that it says "Waiting generate pipeling..."

and it keeps on doing that until I press stop

I don't get to hear the things I'm saying because of that.

[rare goblet]

Ayo? @rocky lantern level 1 !!!

[rocky lantern]

nvm I don't hav RVC

I just realized, my fault.

[pastel oak]

delete folder "pretrain" and "model_dir" and run it again

[rocky lantern]

Alright, I'll try that

[pastel oak]

voice changer help is in #🔍│help-w-okada

for tha future

[rocky lantern]

Yeah I realized I was in the wrong channel. Thanks tho! I'm still booting up the vc so we'll see if it works.

[unique root]

do i need a good quality mic to have a decent effect??

[pastel oak]

anything it picks up it converts, so if you cant understand anything from your mic, the voice changer will have issues

[rare goblet]

Ayo? @pastel oak level 22 !!!

[pastel oak]

a high end mic is not needed but a $5 mic might cause issues

[foggy pebble]

What's AI voices?

[limpid quiver]

Artifical Intelligence generated voices ig

You tell AI generate it, it generates, and then you use

[foggy pebble]

i can't match the pitches singing as the input

I just want to replace song lyrics to make a parody for friends

[thin stump]

What song if you don't mind me asking? I don't think it'll change a thing even if I knew what you are talking about

[foggy pebble]

Secret Agent Man

[thin stump]

since all the singing is your skills

[foggy pebble]

yeah i can't sing which is why using my voice is not a good input

too off pitch and key

[pastel oak]

You can try practising pitch correction with another tool (e.g. FL Studio) if youre decent enough at singing

cant promise itll get you the results you want

[foggy pebble]

is there free autotune to another voice track?

[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

[desert niche]

Ov2 32k

[magic oar]

Hi guys,someone can tell me why my audios getting interferences , its look like it go too high in the treble

[rare goblet]

Ayo? @magic oar level 1 !!!

[thin stump]

I forgot the part of the fiverr page where the guy said he had 10+ years of experience as a rapper 💀. Anyways, a reason people make era models of a particular album or song is because the singer is going to sound different

[thick cloak]

im following the guide to train a model but the tensor borad webpage doesnt show me any graph

[rare goblet]

Ayo? @thick cloak level 1 !!!

[thin stump]

Click on the Scalars and it'll show

[thick cloak]

i already did that

no graph

also the stop training button isnt there

its still the train model

[half cove]

main ones work right?

i think this is a ila rvc code issue

not a colab one

[thick cloak]

anyways tensor board doesnt work so how can i stop training if the button isnt there for me

[half cove]

stop training isnt on tensorboard?

[thick cloak]

i mean on the rvc webpage

[pastel oak]

its bugged, you just wait for the next weight file to be saved and then manually close the cmd and web

[thick cloak]

alright

anyways tensor bord shows me the graph now

so i guess its fixed

it did only start doing the graph a bit into the modles training is that still fine

or is it cooked

i dont see the models checkpoints in the weights folder

nevermind i found it

[thick cloak]

i get errors when i try to use the model

[rare goblet]

Ayo? @thick cloak level 2 !!!

[desert niche]

dunno

[proven hill]

no local works

[half cove]

but i tried on local...

😭

should i do a fresh install or something?

[proven hill]

try

[thick cloak]

@half cove this is my models folder after training, which files do i need to take from it

[half cove]

this ur logs folder correct?

[thick cloak]

yes

[half cove]

pth should be saved in assets/weights

or in models

one of them

[thick cloak]

ok i found it in weights

i took the pth file

is there another one i need

[half cove]

you need added index and modelname-pth

[thick cloak]

this right?

[half cove]

yes

[thick cloak]

ok i got both files now

where should i put them to use the model

[half cove]

there should be a model folder

[thick cloak]

i created a folder to put the two files in

[half cove]

hm?

do you want to use them or make them public?

[thick cloak]

just use them

[half cove]

what fork are you using

i assumed it was ila rvc but maybe not

[thick cloak]

im using the one thats pinned in #✨│ai-help

this one

[half cove]

what is it called

the folder

[thick cloak]

i put the files back becasue under selecting the voice model the older ones popped up

[half cove]

what was the zip called

[thick cloak]

RVC1006Nvidia

it errored

@half cove

[half cove]

what cmd say

[thick cloak]

nevermind the path was wrong to the file

i fixed it

worked

[half cove]

[thick cloak]

how much should i have in my dataset for training the model

[rare goblet]

Ayo? @thick cloak level 3 !!!

[brittle wing]

Is this overtraining?

[thick cloak]

for the training thing, where you select which one you want to use. In the tutorial i read it says that you cant use harvest

is harvest broken

[static matrix]

ok

[thick cloak]

or what is the best to use besides it

[half cove]

please use rmvpe

[thick cloak]

ok

[static matrix]

not sure

[brittle wing]

Is it overtrained?

[clear snow]

i need help making training a model. i never did one before and i have 3 hours of audio

[thin stump]

Can you post your k and d/total?

[brittle wing]

I closed the colab

[thin stump]

Your dataset doesn't need to be that long. It could be 20-40 . minutes since you need to check if there's no noise

Ok

[half cove]

k total??

[thin stump]

I woke up from nap lol. It's kl 🤣

[blissful fulcrum]

guys how do i adjust algorithm to mangio-crepe

[brittle wing]

Should I reopen it

[thin stump]

Now you know next time. I think we have a thread pinned at #📑│making-models that explains the charts better

[brittle wing]

Really

[thin stump]

#1213509354343637065 message

[brittle wing]

Yeah but I can't find it

[thick cloak]

i cant get tensor board to work reliably so what would be a good epoch for training with like 10 minutes of training audio @half cove

i see a lot of models are around 200

[thin stump]

I linked it

[half cove]

im guessing the venv install is broken

ive heard people saying

[half cove]

you gotta use tensorboard for accurate results

[thick cloak]

so for it to run i just need to have its tab open before i start training right

and when i do start it should detect?

[brittle wing]

I didn understand anything

[rare goblet]

Ayo? @brittle wing level 2 !!!

[brittle wing]

Didn't

Look at my graph and tell me

It's the same graph two different photos

[thin stump]

I have. Seeing KL and d/total graphs can explain your downward slant of g/total a bit more.

Because if you have a sudden drop on g/total and it seems too good to be true, you check if the other charts are going upwards

[brittle wing]

So it's probably not overtrained?

This is what you're trying to say?

In simple words

[thin stump]

Yeah you can continue training if you aren't satisfied with the pth file you downloaded

Test your model if you haven't already

[brittle wing]

It isn't overtrained?Cause someone else yesterday saw the graph and said yes

[misty stone]

help

[thin stump]

I won't know without seeing your KL and d/total graph

[misty stone]

okada shows white skreen

screen

it said server wait

[thin stump]

Because your g dip looks normal

[brittle wing]

Okay what do I type to get KL graph

[misty stone]

can anyone help me

[thin stump]

It's just kl

[brittle wing]

Loss/d/total.

And KL?

[thin stump]

Yup

[misty stone]

can anyone help with server wait problem

[thin stump]

https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/#resuming-

[rare goblet]

Ayo? @thin stump level 20 !!!

[brittle wing]

I already know how to resume

I know everything about training aside from this what you told me

Also I have 82-85 samples, 15 seconds each how many epochs are enough for this dataset how do I calculate it

[lethal shale]

Heyy helperss I need some help lol I get nan error with rvc on google collabs

Is there a way to make models?

Someone said something like Applio but I have no idea

[thin stump]

Depends on the batch size and dataset quality

[lethal shale]

You need TensorBoard to calculate it

[brittle wing]

What is batch size?

[thin stump]

Play it safe and do 8, 4 is for short dataset like 2 minutes

and also datasets less than 6 minutes should use ov2 pre training

[brittle wing]

I can do big datasets too my first model WAS with 250 samples

[brittle wing]

Nope all the samples are from a specific album

There must be all I can't pick & choose.

[thin stump]

https://docs.aihub.wtf I can't link a specific page

[lethal shale]

I get nan error

On google collabs

[misty stone]

hey

can anyone help me

what do i choose for input device like the virtual one right

[brittle wing]

How much do I set the smoothing for the other graphs?0,987?

[thin stump]

Yeah

[brittle wing]

Okay I'm re-resuming

[thin solar]

how to make my model sing and it not sound trash

[thin stump]

Could you post your graphs if it's working? I can tell you where to stop

[brittle wing]

can somebody help me run rvc on macos??

[gleaming summit]

I dont really know if this is the right channel to ask but is there an application that can use RVC Models and use them as TTS which inputs into your microphone?, been searchin for smth like that since im a mute and I cant stand the damn microsoft voices anymore

[brittle wing]

man im fucked nobody gonna help me with this one

[brittle wing]

Look.

[rare goblet]

Ayo? @brittle wing level 3 !!!

[brittle wing]

It's the same graph from the photos at 350 epochs

[thin stump]

Nice keep going

[brittle wing]

It isn't overtrained???

Howww

[thin stump]

That is not overtraining

[brittle wing]

When will it beee

[rare goblet]

Ayo? @snow scarab level 1 !!!

[thin stump]

I think g could go down a bit more. But you need to zoom in

[lone ore]

hi, i’m new, please tell me how to make an AI-COVER🙏

Which site do you need to go to, what to do, etc.?

[thin stump]

If it keeps going up and down like it's stacked on top of each other then I would stop

[brittle wing]

400 epochs?

Will my model be finished there

[thin stump]

What's the step count? I see that 19k is kinda your lowest dip

You're the only one that can see it since you can zoom in your chart

[brittle wing]

How do I zoom

Without zooming in too much

[thin stump]

Press either of the tiny chart buttons

[brittle wing]

Which buttons I'm sorry

The blue ones?

[thin stump]

The blue ones 🟦

#1159289738314919936

[thin stump]

#🔍│help-w-okada it's for the real time voice changer

[brittle wing]

Which one out of the Three

[thin stump]

Either one. I double click them

I'm on mobile so I can't point out which

[thin stump]

https://docs.aihub.wtf/

[lone ore]

thx🤝

[safe sphinx]

can someone send ilaria rvc link

[brittle wing]

How about now 370 epochs

If it's overtrained I'll pick 360 from wieghts

[thin stump]

So lowest dip is at 20k?

[brittle wing]

What does it mean

[thin stump]

For g/total

[brittle wing]

Yes

[thin stump]

Past that dip is the overtraining point

[brittle wing]

With my model yes

[brittle wing]

It's overtrained?

When it was at 350?

Was it overtrained

[thin stump]

No, I mean just use the step count nearest before the overtraining point

[brittle wing]

I don't have lesser than 350 epochs

Is it overtrained now?

[thin stump]

On the dip, what's the step count. Hover over it

When you stop training, use the one closest to it on your weights folder

[brittle wing]

Explain in simple words

Well

I couldn't download it at 340 yesterday I have 320

the audio i want for inference wont show up in the box no matter how many times i refresh or how long i wait

[thin stump]

You need to put your mouse on the lowest dip on the tensor board for g/total so it can show the step count

[brittle wing]

How I do that with a phone

[thin stump]

Oh that makes sense

Maybe around 21k steps I guess if you can't tap at it

[brittle wing]

Bow it's even lower

370 epochs

[thin stump]

Ok 370 then

[brittle wing]

Is it overtrained here

At 370

Cause it was low at 350 already

[thin stump]

Are you talking about 22-23k steps? Then don't use that

[brittle wing]

20k!

The 350 one was at 20k

[glass cypress]

How do I tell what batch size to use while training?

[thin stump]

I think you can use the 370 if that's the newer dip

Download both and compare

[brittle wing]

THIS IS THE ONE.350 EPOCH

[brittle wing]

Which one do I pick

320
350
370

[thin stump]

😅 I'll get on my pc

[brittle wing]

Is the 21k steps as shown in the photo(350 epoch) the correct one???

[brittle wing]

Oh ok

[rare goblet]

Ayo? @brittle wing level 4 !!!

[brittle wing]

Is 21k steps ok?

Cause this is where the 350 one is at

[thin stump]

Should be

[brittle wing]

Cause it's at 21k

So that's the one?

[thin stump]

Yes that's the lowest dip I see

[brittle wing]

That's the 370 epochs model.

I just resumed it now ...

[brittle wing]

This is from yesterday 350 epochs.

[quartz perch]

Hey guys, I want to make a model for a specific song, if i make the dataset me singing the song or pronouncing the lyrics correctly and slowly, will that help fix the problem with pronouncing words ?

[rare goblet]

Ayo? @quartz perch level 1 !!!

[thin stump]

Ok i'm just saying 370 is better since there's nothing out of the ordinary for kl and d/total

kl being constant is okay

[brittle wing]

So the last trained one is good not overtrained?
Should I quit training THIS model?

Is it ready

[glass cypress]

I'm getting this error when I press the train model button in ilaria

[thin stump]

Yeah, you can quit training

[brittle wing]

Another user claimed it was overtrained at 350 epochs.

[thin stump]

350 is slightly underbaked but whatever. I think it's going to sound similar

they didn't ask for kl and d/total for context

[brittle wing]

Underbaked???

[thin stump]

that's a difference

not THAT underbaked. just a little

where it won't even matter

[brittle wing]

It still needed training after?

[thin stump]

370 is basically your lowest dip so that's where you would normally stop. I'm not stopping you from using 350 epochs either

[brittle wing]

So the final one is at 370

[thin stump]

yes

[brittle wing]

I already used the 350 epochs

But the 350 one needed training as you said

[thin stump]

350 is plenty already, but if you want it on point then 370

[brittle wing]

370 is overtrained or not

[thin stump]

but it's not like it'll affect the model quality that much. No 370 is not overtraining

380 and above is overtraining

[brittle wing]

So this is the end.

I can post it

[vale radish]

everytime i try to make a cover in the gui it just crashes after 10 seconds

[brittle wing]

Cause I'm planning on making another model

And I don't wanna waste any more time on this one.

[thin stump]

Sorry for confusing you. You can post it in #1159514067187277865 if you're confident

with a sample audio

[vale radish]

Can someone please help me

[thin stump]

https://docs.aihub.wtf/extra/model-maker-role/#model-maker-role

[brittle wing]

It's not my first time actually this model is jhopes jack in the box era model that I previously trained at 750 epochs from my deleted account.But I realized it was overtrained at 750 epochs

So I decided to train it again.

How do I ask the admins to remove one of my model posts

From deleted user.

Also the dataset on this one was quite good

[thin stump]

That's great then. I would just title it V2 - Jhopes Jack in the Box Era with a different huggingface link (when you upload it on #voice-models)

[brittle wing]

The next one's gonna be better

[brittle wing]

I'll upload it on Huggingface too.

Inference samples and everything.

[brittle wing]

yo why is my voice lagging

idk why

its bugging a lil

did u install

vb cable

make sure u got these

the up one is ur mic

then

down as cable input

and ur discord setting input should be as output

and ur headphones

like the cable output or wtv on input

and ur headphones on ur other

[rare goblet]

Ayo? @brittle wing level 1 !!!

[thin stump]

you can now

[brazen crystal]

I'm testing the CLI version and I have a problem. Using standard RVC, training does not pose any major problems, although in older versions of RVC I had the opportunity to choose the number of threads during feature extraction. Now this is not there, so I cannot reduce chop length to 64 due to 16gb of ram. The CLI version also does not have the option to reduce the number of threads, and in addition, the length of 128 also causes memory overflow. Is there any way to do this?

[muted meteor]

Whenever I put in the link provided for an rvc model it doesn’t work. Am I supposed to download it first?

[glad zealot]

You need to copy the download link not the link to the model

[muted meteor]

I’m confused because the link to the model makes it download

[glad zealot]

If you got the link from #1175430844685484042 it should be the download link and if don't work idk why

[pale thunder]

speaking of rvc models, when i use the beatrice model im fine, but when it comes to RVC models, the sound wont come out. any help?:(

[glad zealot]

You might want to convert to onnx, might make it work

[pale thunder]

okay, ill check that out

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[astral jungle]

Can you give me the Vietnamese TTS link?

[regal radish]

hi

[rare goblet]

Ayo? @regal radish level 1 !!!

[haughty zodiac]

alright guys it s not detecting my gpu !!! @odd shale

[regal radish]

i put a model at the voice changer but it doesnt works

[regal radish]

and it only happens with a few models

[glad zealot]

What does it say now when you have 0?

[glad zealot]

What specifically happens? In the console

[regal radish]

like

i upload the model

it starts loading

then it finishes

and the model doesnt shows up

stays as blank

[haughty zodiac]

thats hella weird because i already put a 0 in it before and didnt work

maybe it was a bug

[glad zealot]

Ye what does it says in the console

[regal radish]

nothing

[glad zealot]

Works now?

[regal radish]

it just closes

[regal radish]

wait u mean on the cmd?

[glad zealot]

Yes

[regal radish]

im using the colab version

[glad zealot]

It should still be there

[regal radish]

yeah

[glad zealot]

To where you got the link

[regal radish]

oh

ok lemme see

[glad zealot]

What does it say when you upload model

[haughty zodiac]

it doesnt say its using it

or maybe it is

[glad zealot]

It is using, to verify go to your task manager and look at the gpu section when you are training

[regal radish]

this

[glad zealot]

Performance

[glad zealot]

You're uploading the pth or Puting the download link on the model downloader on colab?

[regal radish]

uploading the pth

[glad zealot]

When you are training, the part that says gpu should put the graph higher

[glad zealot]

Tf

[regal radish]

yeah

[haughty zodiac]

alright ill wait for epoch one, this just takes a little longer than rvc

[glad zealot]

A wait you are using a sovits model not an rvc model

Different programs and usage

[regal radish]

ohh

how do i use it then?

[glad zealot]

If it's a recent model, atm it only works with tts and kinda hard to use, you seem to only need voice changer tho so just look for the same character but for rvc

You need a program called gpt-sovits

To use the sovits models

[regal radish]

this one is a sovits too?

[finite galleon]

No thats rvc

Rvc consists of a .pth and a .index while gpt-sovits consists of a .pth and a .ckpt

[regal radish]

well

[regal radish]

then why it doesnt works when i try it?

i thought only sovits that wouldnt work on realtime voicechanger

[finite galleon]

Whats the problem and which program you using

[timid raft]

My voice chat aint workin on discord

[brittle wing]

How do I get the correct Huggingface link

How it was done I forgot

[regal radish]

w-okada, colab version

[rare goblet]

Ayo? @regal radish level 2 !!!

[regal radish]

the problem is that i upload the model and it doesnt works

[regal radish]

i upload it, starts loading and doesnt shows up

[finite galleon]

Was iit the hina mod

[brittle wing]

How do I get the right model URL I forgot how it was done it must contain "download=true" in the url

[half cove]

download = true is only a marker

[thin stump]

not on PC again so someone could help you. You copy the link to the download button. It's on the guides too

[half cove]

the download bit is if it has resolve

[brittle wing]

I have this problem

[regal radish]

yeah

[brittle wing]

"File is not a zip file"

[finite galleon]

I am not sure then

[finite galleon]

Change the blob in the middle of the link to resolve

[regal radish]

awww :C

[brittle wing]

does anybody know why my voice is laggy

i got a high end and everything

[brittle wing]

Thanks

Still error

[finite galleon]

Change the %20 to an _ maybe

[regal radish]

doesnt another helper knows how to fix it?

[brittle wing]

That's how it should be formatted

[brittle wing]

yo can u help me out 😔🙏🙏

my voice is lagging or bugging

[finite galleon]

Yeah they know more about the voice changer colab than me

[finite galleon]

If you are talking about the voice changer then idk

[brittle wing]

alr ty

it sounds normal when i hear it

but weird when i test it

with toher

others?

https://huggingface.co/elliyonline/j-hopeJackInTheBoxErav2/resolve/main/jhopeJackInTheBoxErav2 by armyshope.zip?download=true
Done

It was in the download button

[haughty zodiac]

its gpu

[proud elbow]

it seems working now, then check the console window if the training is ongoing

[rugged stirrup]

Would 32k do for this? the audio if 44k but not sure if I should do 30 or 40k

[proud elbow]

yes 32k

[rugged stirrup]

Oh ok

[rugged stirrup]

Can someone tell me if this is normal? it went down too quick and then is going up like im overtraining it only at 100 epochs

[visual fiber]

hiyo, anyone having issues with step 2b in RVC. I keep getting"load model(s) from hubert_base.pt
move model to cuda
no-feature-todo

I have tried fresh uninstall and reinstall and only have one .wav in my data set folder.

[rugged stirrup]

How is the D going down but the G is going up?

[visual fiber]

Never mind, I figured out that my step 2a was wrong. I was entering the path exactly to the dataset rather then the folder itself for RVC to search inside.

[haughty zodiac]

how do i resume training after i tested out the model? do i just let it run? this is much slower than rvc but does sound good for 15 epochs

[rugged stirrup]

You just type the name of the model in the train tab and click the train button

Atleast thats what you do in mainline

[haughty zodiac]

i was using the inference on here

[haughty zodiac]

on rvc you just let it run

[rugged stirrup]

Ohh so you use applio to inference?

[haughty zodiac]

whats mainline? is it different than applio

[rugged stirrup]

its normal RVC

[haughty zodiac]

im using it to train too

[haughty zodiac]

just updated?

[rugged stirrup]

No

[haughty zodiac]

why they change the name?

[rugged stirrup]

Your wanting to continue training right?

[rugged stirrup]

Man applio's UI is confusing idk how to use that

[haughty zodiac]

ITS easy

[rugged stirrup]

Not for me

[haughty zodiac]

it took 2 hours for 15 epochs , i wouldve been at 150 for rvc its a litle slow but im a stick it out n see the quality

[rugged stirrup]

Well when its done tell me how good the quality is cause 2 hours for 15 epochs is massive trade-off for a little quality boost if any

Thats 15 epochs?

[haughty zodiac]

yeah

[haughty zodiac]

my voice too

[rugged stirrup]

Oh wow

[haughty zodiac]

yeah quality is good no filters on the track too

[rugged stirrup]

Well thats nice but I probably wont switch to it

[haughty zodiac]

it got a option for pre auto tune !

[rugged stirrup]

Nice

[haughty zodiac]

8 iTS a 30 minute file

im gpu using my 2060 super, rvc 23.7 is alooot faster, but ill still train with this to see the quALITY

[hoary loom]

Anyone know How to use rvc? I’m kinda new I wanna make an ai song

[soft stratus]

Can you help me

[rare goblet]

Ayo? @soft stratus level 9 !!!

[hoary loom]

I don’t know anything tbh I’m new

[soft stratus]

Ok

The error came on the model information and i want anyone to fix it @low shard

[half meadow]

So, if I wanted to change the language of the interface, how would I go about doing so?

[forest crescent]

what voice changers do you guys recomend

[fluid topaz]

which python version should i use when im going to use RVC mainline, i reinstalled pc and forgot about that, been a long time

[uncut bolt]

i keep on getting this warning: ttributeError: module 'huggingface_hub.constants' has no attribute 'HF_HUB_CACHE'

[haughty zodiac]

OK now we going faster lowered the batch size to 4

[soft tiger]

i need the latest RVC local software for voice changer for AMD

[rare goblet]

Ayo? @soft tiger level 4 !!!

[soft tiger]

@tired kraken

[gusty kestrel]

@feral river, I have found 2 results that match your search!

pasteldiffusion

Jett Valorant

• Uploaded: <t:1693267200:d>
• Likes: 0

Epochs

500

Technology

RVC

Algorithm

Crepe

Requested by yorutastic · Mar 31, 2024, 4:33 AM

[tired kraken]

Not the guy to ask, wait and be patient until other people are online. Or just use help commands in #🤖│bots

[soft tiger]

i mean you do have a helper role

it was worth a shot my guy

[brittle wing]

Anyone know why the avatar photo doesnt get accepted, i chose 400x400, 350x350 and 512x512, none of them worked

[fallow wing]

I'm 9 years old

[haughty zodiac]

so applio 3.10 is more accurate than rvc 27.3 right?

[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

[wispy burrow]

That should be Applio 3.10 I think.

[brittle wing]

Is there any UVR model that removes background SFX sounds?

[rare goblet]

Ayo? @modest ibex level 1 !!!

[graceful obsidian]

you mean the official, upstream repo rvc amd release?

[soft tiger]

like i had a good software but it seems out of date

[graceful obsidian]

well, there's few things about amd release

it's kinda broken

[soft tiger]

ik amd is butt compared to nvidia

[graceful obsidian]

vram usage is increased compared to cuda / Nvidia. Pitch issues too

you're better off using it in cpu mode tbh

unless you want, like mentioned, pitch issues and out of ram errors

[soft tiger]

nah fuck all that

cause look

i just need help finding this software

lemme show you

[graceful obsidian]

I use it in cpu mode too, am an amd user

[soft tiger]

the latest version

[graceful obsidian]

so you want newest release in general

it's the 1006 version dude @soft tiger

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

[soft tiger]

hold on

lemme show you

[frozen vault]

This happens when I run the application, it closes when I press any key, how can I fix it?

[opal mica]

which voice changer to download for use rvc

@dusty rampart

[graceful obsidian]

https://github.com/w-okada/voice-changer
or rvc's native voice changer

[dense moon]

whats the latest rvc versions rn? its been like a few months since i last tried and failed miserably on first attempt cuz the character i wanted to clone had very little dialogue xd

[snow scarab]

Hi all, Im training a model with TrainingV2_NoUI.ipynb, however, its just goofy, 1 epoch is around 0.6 sec to finish, and the final AI is just goofy and not working. While I made 4 AI already before, with the same settings and kinda same lenght datasets and same quality datasets... So idk what Im doing wrong now.....

[opal mica]

hey

how to downlaod

i downloaded

but how to open

[graceful obsidian]

read the github page.
iirc, either "voice-changer-native-client" or "MMVCServerSIO" executable

[sand iris]

It's start_http.bat

[opal mica]

brah

[graceful obsidian]

doesn't matter really. if it opens it opens
my memory can fail hence I recommended the page reading.
I don't use w-okada but native rvc's

[opal mica]

its 3 gb

the w okada

i watched a video

if it works

its gonna be fun

[sand iris]

Which video ?

I've yet to see one that doesn't mislead one way or another

[opal mica]

okk

[sand iris]

Just make sure you downloaded the right version for your machine

[opal mica]

https://www.youtube.com/watch?v=aVfoC1EHlVs&t=28s&ab_channel=Fofamit

[sand iris]

Else you're gonna have a bad time

[opal mica]

okk

ty

[brittle wing]

how can i use the rvc models in real time likein discord what do i need to download

[opal mica]

please tell me ?

thanks!

[sand iris]

Do you know which GPU your machine has ?

[brittle wing]

amd rx 7800 xt i have

[opal mica]

ya

i know

[sand iris]

Link to realtime voice changer for AMD GPUs :
https://huggingface.co/wok000/vcclient000/blob/main/MMVCServerSIO_win_onnxdirectML-cuda_v.1.5.3.18a.zip

nvidia : https://huggingface.co/wok000/vcclient000/blob/main/MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a.zip

[opal mica]

fk

i think i just downlaoded the nvidia one

lemme download the amd one

[sand iris]

¯_(ツ)_/¯

[rare goblet]

Ayo? @sand iris level 83 !!!

[brittle wing]

can you help should i just download it ?

[sand iris]

Yeah you're gonna need the program to run it

[opal mica]

how to download btw

[sand iris]

there's a guide pinned in #🔍│help-w-okada

[brittle wing]

why is it 3Gigabytes ?!

[opal mica]

this is mine

[sand iris]

Lots of dependencies (+ it bundles cuda on the amd version somehow and that takes up most space)

[opal mica]

oops

[sand iris]

Oh you're gonna have a bad time with those

[opal mica]

lmao

[opal mica]

lmao

ikr

[sand iris]

they're really slow for this

[opal mica]

ah sh

will i be able to run valorant with this

[brittle wing]

does my graphics card need to be fast for the voice changer ??

[rare goblet]

Ayo? @brittle wing level 1 !!!

[sand iris]

Yes, but yours is definitely fast enough

[opal mica]

yeah

[brittle wing]

i dont think when you have int graphics

[broken cloud]

Just curious why it's not picking up my GPU to use it to process the vvoice?

[opal mica]

idh int

i hvae amd

[sand iris]

@graceful obsidian how's the OG RVC's voice changer on amd btw ? I've yet to get it working, but I'm getting mixed feedbacks about it

[brittle wing]

is the file a virus or why is it 3 gigabytes ?

[graceful obsidian]

It's better for me

I an an amd user yea

[opal mica]

hey will i be able to run valorant with it?

[sand iris]

There is a lot of stuff bundled with it. Python & stuff is notoriouly big in filesizes

[brittle wing]

what amd gpu do you have ?

[sand iris]

plus the amd version bundles some nvidia stuff for some reason... Which takes up most space

[graceful obsidian]

oldie, 570 4 gig

but yea, I prefer the native

[opal mica]

look at this shine

[graceful obsidian]

esp considering there's fcpe

[sand iris]

You'll get horrendous delays, but I guess since you have two GPUs, one could do voice change and the other run the game

the delay will be horrendous though

[graceful obsidian]

it's not the quality levels of rmvpe raw conversion wise

but it's def the way to go for if it's realtime

[opal mica]

the voice delay right?

[sand iris]

yup

[graceful obsidian]

( make no mistakes, normal users won't notice it )

[opal mica]

btw do i need to do something to have them to work for 1 each

[brittle wing]

wjen my programm is downloadet should i just reun it and then ?

[sand iris]

voice changer just got fcpe as well, but it ain't available on amd there...

[sand iris]

there's GPU selection on the program

[graceful obsidian]

with something like 570 4 gig I can go to as low as I'd say, 0.8 to 0.5 sec of delay

[opal mica]

okk

[graceful obsidian]

but it's fine tbf, it's not a deal breaker

[sand iris]

unzip the archive, & run the start_http file

[opal mica]

my file is downloaded in 2 or 3 minutes how to open it?

[brittle wing]

:( so long

[opal mica]

lmoa slow internet

sad

[brittle wing]

i have 100 mbits per sec

[opal mica]

same

my internet is fast

but

my pc is bad

my pc is good

but my gpu is bad

[broken cloud]

For "GPU" it has my CPU but isn't picking uo my GPU to use as the processor?

[opal mica]

you have to select it he said

[rare goblet]

Ayo? @opal mica level 4 !!!

[broken cloud]

Yeah but it doesn't pick it up to select

[proud elbow]

Nvidia gpu or not?

[broken cloud]

AMD 7900xt

I've tried it all out and it works but it's using my CPU instead of my GPU

[brittle wing]

is that normal ?!

[sand iris]

windows freaks out when it runs software from unknown devs

"more info" allows you to run it anyways

[brittle wing]

and when i dont want the voicechanger anymore is it easy to uninstall ?

[brittle wing]

is it easy to uninstall ?

[broken cloud]

any ideas?

[rare goblet]

Ayo? @broken cloud level 1 !!!

[sand iris]

remove the folder you unzipped & it's gone

doesn't touch anywhere else on the system

[brittle wing]

and what is it doing in the screenshot ?

[brittle wing]

.

[sand iris]

downloading the AI files it needs

[proud elbow]

for RVC or voice changer?

[rare goblet]

Ayo? @proud elbow level 29 !!!

[sand iris]

the zip only contains code, not the needed models

[brittle wing]

is it done ?

[rare goblet]

Ayo? @brittle wing level 2 !!!

[broken cloud]

I assume for voice changer. It isn't picking up my GPU

[sand iris]

any errors above that ?

Never mind it should have launched now

[brittle wing]

i closed it

[proud elbow]

oof, u should download the directml instead

[sand iris]

Oh, the command prompt window has to stay opened

[brittle wing]

should i open the start folder again ?

[broken cloud]

Ah, alright, uninstall quite easy via the windows apps/utilities?

[proud elbow]

just delete entire folder

[broken cloud]

Awesome just did it, really appreciate it ! :))

[brittle wing]

should i open the http start folder again ?

[sand iris]

yes!

[brittle wing]

which one

[sand iris]

http

https is for offloading to another pc

[brittle wing]

what do i need to do now ?

[brittle wing]

wzhat settings do i need to change ?

[sand iris]

https://rentry.co/VoiceChangerGuide

full guide there

[opal mica]

yo

mine almost downloaded

lmao

am gonna need some help

[brittle wing]

should i download the audio cabel now ?

[sand iris]

if you plan on using it on other apps ya

[opal mica]

this is hell

no cap

lamo

lmao*******

btw

fired

stay online i need your helo

hello

its downloaded

now what to do

[opal mica]

yo

[gusty gust]

can someone help me to find location for my new model please!

[opal mica]

yeah

[brittle wing]

how do i upload a model ?

[opal mica]

yeah i need to know that too

theres some data error there

[brittle wing]

can you help me with the settings ?

[opal mica]

which one?

[brittle wing]

http

[opal mica]

done

it showed some command prompt

now the command prompt is also gone

now what to do

you already did this right?

[brittle wing]

@sand iris do you know why ia cant choose gpu 0 just nothing happens

[violet pecan]

yo how do i fix the sound not coming in like after i switched the mic input in discord to virtual cable

[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]

someone needs help?

[brittle wing]

yes

idk what i should choose here

[brittle wing]

.

[proven hill]

#🔍│help-w-okada

[violet pecan]

yeah like ion know why the voice changer didnt work in discord after i chose virtual audio cable as uh the mic

[rare goblet]

Ayo? @violet pecan level 1 !!!

[proven hill]

did you restart your pc after installing it?

[violet pecan]

yeah i did

it didnt work

[proven hill]

[glad zealot]

whats your voice changer and discord audio settings?

[violet pecan]

discords i put it as the defaults

idk

[glad zealot]

you need to set your output in the voice changer as cable

output = cable input
monitor = your headphones

on discord
input = cable output

[violet pecan]

wait, its lagging

it works but its lagging

[foggy sentinel]

helo

[brittle wing]

wtf

[foggy sentinel]

what ??

[violet pecan]

it stopped working now after the program crashed

T^T

[opal mica]

hey

i need help

after 20 mmins

i need help guyssss

[violet pecan]

same

[ripe jacinth]

is there a fork of the colab version where everything is just stored on google drive?
including the python dependencies (those take sooo long to install!!!)

[foggy sentinel]

i'm here what do you need ?.

[rare goblet]

Ayo? @foggy sentinel level 1 !!!

[sand iris]

No, else it could break at any moment which isn't good

plus that's also a loooot of space for a gdrive

[opal mica]

hey fired

@sand iris

this happens

when i try the video u said

[pastel oak]

Do f0: rmvpe or rmvpe_onnx

[opal mica]

that when command promt close by itself

it sais failed

[sand iris]

ok what the fuck

[violet pecan]

do where

[sand iris]

what's the folder where you extracted it ?

like full name

[opal mica]

kk

[pastel oak]

on f0 det..?

[sand iris]

C:\whatever\something\MMVCServerSIO