✨│ai-help

[brittle wing]

and is this good?

the audio quality?

[high night]

Make sure it doesn't have any music or artificacts

[loud berry]

@proper shale please help

[solemn bough]

I have some problems with Ilaria RVC. It converts vocals, but doesn't let me hear it or download it. What do I do?

[rare goblet]

Ayo? @solemn bough level 1 !!!

[frigid night]

mine just doesnt convert 💀 its going on 200 seconds now

[spiral osprey]

How do I make a cover more faithful? Is it changing pitch or something? How can I make it more like the original?

[brittle wing]

how do i get ai model maker role?

[rare goblet]

Ayo? @brittle wing level 2 !!!

[wispy burrow]

Go to #1159514067187277865 and submit your Model for validation using the /submit command there.
If your model is approved, you'll get the role and can post your model in #1159289700490694666 .

[lapis echo]

训练结束, 您可查看控制台训练日志或实验文件夹下的train.log , what this?

[daring marsh]

Curious since I get like 50 res on most default voices but when I download models from here it spikes up to 1000 and the output gets really choppy, any ideas on that?

[wispy burrow]

#🔍│help-w-okada

[daring marsh]

thanks

posted there too

[rare goblet]

Ayo? @daring marsh level 1 !!!

[wispy burrow]

It's just the training logs.
Everthing the cmd displays during training should be written in the training log.

[rare goblet]

Ayo? @wispy burrow level 18 !!!

[spiral osprey]

How do I make a cover more faithful? Is it changing pitch or something? How can I make it more like the original?

[rare goblet]

Ayo? @spiral osprey level 3 !!!

[agile hollow]

What's wrong?

[glossy coyote]

does anyone know what this is?

[lapis echo]

I started training and it appeared after 6 seconds

[violet heron]

Are you logged in?

[agile hollow]

Yes

[molten pecan]

Try duplicating the space

[agile hollow]

Already done

[molten pecan]

Are you using Brave?

Or an AdBlock?

[agile hollow]

I don't know

I don't have both of them

chat really went 😴

[rare goblet]

Ayo? @agile hollow level 2 !!!

[molten pecan]

Hm then that's odd

I would try using another web browser

[agile hollow]

https://tenor.com/view/ok-and-ok-and-whole-sqaud-gif-22248092

[molten pecan]

Or another huggingface space

There are a few of them out there

[proper shale]

I don't knoe

[glad zealot]

anybody tried the new applio?

the updated

[proper shale]

wait wha

[glad zealot]

theres this now

and idk if its a bug on colab or not but it seems to not save weights?

heres the setting that i have

[solemn bough]

Damn why no inference spaces stopped working

[high night]

Colab? What colab are you using? Modified by you?

[glad zealot]

weights is empty

[glad zealot]

yup

[solemn bough]

They literally just error out

[high night]

You are crazy

[glad zealot]

its training but no weights being saved somehow

[brittle wing]

guys help

[proper shale]

Python is not installed to path

[brittle wing]

where do i install it

to which path

[proper shale]

Just have to tick "Add to path" in the installation process iirc

And you select the drive in which Applio is in

[brittle wing]

alr thx thx

[proper shale]

Np! Hope I helped ;)

[glad zealot]

@proper shaleknow how to fix mine?

[proper shale]

Honestly no

[glad zealot]

same

[proper shale]

So weird

[glad zealot]

they updated it

12 hours ago

i should have forked it

[brittle wing]

here?

[glad zealot]

yup

[brittle wing]

and do i click customize installation or just install now

[glad zealot]

i can probably train this still using disconnected or kaggle training but naaah

[pulsar wyvern]

where do people find voices

[glad zealot]

i need to wait till tomorrow to use kaggle

[wispy burrow]

That's normal, it created the txt file so it can log the training session.

[pulsar wyvern]

i wanna do brad pitt but i cant find his voice

[glad zealot]

you can check weights.gg or check #1159289700490694666

if not there then you can try making a request

[brittle wing]

??

[frigid night]

how fuckin often does Ilaria error? ive been trying to do this for like an hour and a half and it keeps popping an error after a couple of minutes

[rare goblet]

Ayo? @frigid night level 1 !!!

[pulsar wyvern]

thx

[rare goblet]

Ayo? @pulsar wyvern level 1 !!!

[pulsar wyvern]

i have a problem. i get error, when ever i choos "D" or "G"

[glad zealot]

ignore D and G

just get those with the name of the character taht you make

[pulsar wyvern]

so only like this?

[pulsar wyvern]

do i even need to put them in the folder then?

[harsh laurel]

Aside from that what is a recommended amount of epochs?

[brittle wing]

guys someone help...

its still the same

[pulsar wyvern]

what is the difference?

[high night]

@copper pebble help this guy

[copper pebble]

shit i got it wrong

[brittle wing]

help me...

idk what the hell is going on

[copper pebble]

give me a mom

[brittle wing]

sure

[copper pebble]

hope i can help

[brittle wing]

don't worry

[high night]

Mom?

What does that mean

[brittle wing]

moment

[high night]

Oh

[copper pebble]

lmfao

dunno a shit abt these things tho

[brittle wing]

damn

what do i do then

[copper pebble]

type C:\Users\me\path\to\python.exe

[brittle wing]

in run?

or in cmd

[copper pebble]

cmd prompt

try

[brittle wing]

the system cannot find the path specified

[rare goblet]

Ayo? @brittle wing level 3 !!!

[glossy coyote]

-rvc

[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

🇫🇷 France

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

🇩🇪 Germany

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

🇮🇹 Italy

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

🇯🇵 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

[vale oyster]

Aicovergen clean vocal better than my manual UVR, any recommend?

[copper pebble]

https://initialcommit.com/blog/python-is-not-recognized-as-an-internal-or-external-command#:~:text=Specify the full location to python.exe,-You learned previously&text=One way to fix the,\to% try this im in school rn

[brittle wing]

alr thx alot

C:\Users\dorot\AppData\Roaming\Microsoft\Windows\Start Menu\Programs\Python 3.9

this is the location

[glad zealot]

@proper shalefound fix for the training.. it does train now, but it doesnt seem to save the model at checkpoints

it only saves when it stops traning

[proper shale]

oh

that's

fine?

but does this affect ur colab

(I still haven't tested it)

[glad zealot]

for training no

but for testing the models while training yes

since i cant get the models

it doesnt get saved

[proper shale]

huh...

[glad zealot]

REEEEEEE

[proper shale]

is this because applio updated

[glad zealot]

yup

[proper shale]

in GitHub

[glad zealot]

yup

[proper shale]

😭

smh my head

[glad zealot]

lemme try if i can get an older version

aha got it

welp since i cant train using this rn imma just steal someone's kaggle acc and use it

[rare goblet]

Ayo? @glad zealot level 28 !!!

[glad zealot]

#1159289700490694666

[proper shale]

or weights.gg

[glad zealot]

oh hey look

i didnt need to get the specific version lol

they just have it

[proper shale]

Yee

[brittle wing]

YESD

YES

OMG

I DID IT!!

[glad zealot]

nice

btw can you try using training?

it should work on yours

but try saving every 10 epochs or something if it saves anything on the weights folder

since applio is bugging on colab imma use this kaggle version instead

[brittle wing]

how tho..

im fresh as fuck..

it will be my first ai model

[brittle wing]

Trying to install easiergui from mangiokalotweaks that i found on the aihub guide for my local rvc folder but i keep getting this error:

File "D:\RVCv2\RVC1006Nvidia\RVC1006Nvidia\EasierGUI.py", line 22, in <module>
from i18n import I18nAuto
ImportError: cannot import name 'I18nAuto' from 'i18n' (unknown location)

D:\RVCv2\RVC1006Nvidia\RVC1006Nvidia>pause
Press any key to continue . . .

As the guide said i clicked on the install_easyGUI.bat and when i tried to run the run_easiergui.bat file it gave me that error on the command prompt whenever i tried to run it

[proper shale]

Have you tried the harder GUI though

[brittle wing]

The original one? Seems a bit too complicated for me and i prefer having a simpler gui

[glad zealot]

the mangio one?

[proper shale]

EasyGUI just has a few issues and honestly, you get used to the 'harder' GUI

[brittle wing]

yeah mangio kalo tweaks one

[brittle wing]

Is it built for the colab version?

[proper shale]

which one

[brittle wing]

The easierGUI by mangio kalo tweaks

[rare goblet]

Ayo? @brittle wing level 1 !!!

[brittle wing]

I'm following this guide

[proper shale]

I think that's a local version of the easyGUI

[plain drift]

hello guys i have an error when i hit the start train button

[brittle wing]

Yes i'm installing the local version, correct

[glad zealot]

why not use the one of colab by rejeks? it doesnt have gui tho

[proper shale]

So yeah kinda built from the Colab version

[proper shale]

what does terminal/cmd say?

[brittle wing]

I wanna find a gui which is simplified and makes the job simpler

Have any recommendations?

[plain drift]

File "X:\AI TOOLS\Mangio-RVC-v23.7.0\infer-web.py", line 1014, in click_train
set([name.split(".")[0] for name in os.listdir(gt_wavs_dir)])
FileNotFoundError: [WinError 3] The system cannot find the path specified: 'X:\AI TOOLS\Mangio-RVC-v23.7.0/logs/mi-test/0_gt_wavs'

[rare goblet]

Ayo? @plain drift level 1 !!!

[glad zealot]

easy gui on colab using gui seems to not be working atm sadly

but using it without gui seems to work with no issue

[plain drift]

how can i send a screenshot ?

[brittle wing]

Aha. right now im using webui running it on my localhost, is there an alternative "easygui" for that usage?

[plain drift]

i have also given a specific path too previously but its shows the same

[brittle wing]

The options often confuse me while tweaking the settings

[proper shale]

you didn't preprocess

[plain drift]

i did but nothings shows up except the count down but nothing like that pre process success of that sort

[proper shale]

oh

your folder has a space on it

and that

completely breaks everything

[plain drift]

i dont see any spaces cause i copied as path and removed the inverted commas

[brittle wing]

you probably want to rename your "AI TOOLS" folder because that's where your space is

[proper shale]

^

[plain drift]

oh alright Lets try that

[proper shale]

then try renaming the path inserted in the first step

and then process data + feature extract again

[brittle wing]

guys how long can it take

[faint violet]

Hello, I have an error with applio when I use the "Download and separate audio tracks" category, but I don't know how to fix it.

[brittle wing]

did it just start?

no..

already like 5 mins

[plain drift]

this one worked.. Holy thank you man

[brittle wing]

so what do i do??

[brittle wing]

I'm glad the problem was solved for you!

[brittle wing]

try refreshing?

[proper shale]

You screwed up

[brittle wing]

how tho

how i screwed up

[proper shale]

Should've kept CPU processes at 2

[brittle wing]

oops

[proper shale]

Anything else screws RVC up

[brittle wing]

alr imma do it again

[proper shale]

Glitches haha

[proper shale]

Happy to have helped you a bit too! :)

[brittle wing]

and here we staaart agaain

[plain drift]

I have this error now and i dont find the trained model anywhere in the folders

[rare goblet]

Ayo? @plain drift level 2 !!!

[plain drift]

I dont know which path its showing error for

[brittle wing]

yo mj help... @proper shale

it's like this again

[proper shale]

Screwed up again I think

Because uh

It didn't find the preprocessed files

And it didn't even find mute40k.wav

[craggy heath]

i have a question, wasnt there a website with all the voice models

[proper shale]

There is - weights.gg

[craggy heath]

tysmm

[brittle wing]

mj help

[proper shale]

Its doing it

I think? idk

😭

[brittle wing]

how much does it usually take??

i started it like 3-4 mins ago

[proper shale]

depends on how long ur dataset is

[brittle wing]

the audio has 30 mins overall

[proper shale]

oh

should take a while it

ig*

[brittle wing]

like 1h?

[proper shale]

nahhh

like 5-10 mins

[brittle wing]

more?

uff

alr

damn ok ok

bro...

its still going

[plain drift]

any solutions you can suggest me

?

[brittle wing]

ffs how do i fix it

When i try to add models using the correct guide it keeps getting stuck into this loop

[proper shale]

Honestly I don't even know, since it isn't saving the preprocessed stuff

[somber parrot]

complete guess but try removing the spaces from the path?

[brittle wing]

why i get "Error" when i press convert on Mangio-VC-Web?

[woeful turret]

how do i make my own ai model? im just starting

[rare goblet]

Ayo? @woeful turret level 1 !!!

[brittle wing]

https://discord.com/channels/1159260121998827560/1159636677178953802

[woeful turret]

thank you

[brittle wing]

No problem

[brittle wing]

pls

[proper shale]

you probably did something wrong

[brittle wing]

i spammed convert and now work

lol

[rare goblet]

Ayo? @brittle wing level 1 !!!

[plain drift]

can anybody send me the apiloo version github link

[proper shale]

its in #1159949278270193734

[brittle wing]

I don't know how to correctly set custom models locally in my rvc folder for it to show up

[rare goblet]

Ayo? @brittle wing level 2 !!!

[wispy burrow]

ModelName.pth in weights. Or logs/weights if it's Applio...
added_IVF*.index in logs/ModelName.

[brittle wing]

pth files show up but the indexes don't

[minor nexus]

Hi, what would be the link to using RVC to create covers?

[brittle wing]

there is no weights folder in my rvc

[brittle wing]

#1159949278270193734 message

[plain drift]

same issue i havent found any solution yet

[wispy burrow]

What RVC are you using?

[brittle wing]

RVC1006Nvidia

[plain drift]

is that applio?

[brittle wing]

No just the original one

haven't tried applio yet.

[wispy burrow]

Ah The Regular RVC.

[brittle wing]

Correct

[wispy burrow]

Still there should be a weights folder.

[plain drift]

my issue is i get all the processees done and i dont find the trained model in any of the folders including weights with the original

[brittle wing]

i dont train my models i just want to import pretrained models

[brittle wing]

Weird, i reinstalled from the zip file

[wispy burrow]

For Applio, the weights folder is in the logs folder.

[brittle wing]

Should i just create the weights folder myself?

Still doesn't appear after i restarted the webUI.

[brittle wing]

I just found the weights folder in assets, weird.

I thought it was at the root of the RVC1006Nvidia folder Lol

YES IT WORKS!

still can't wrap my head around why this happened

[ionic scaffold]

the wordband bot isn’t working, I keep getting the application didn’t respond

[wispy burrow]

It normally is.
I don't even an assets folder in any version of regular RVC I have...

[woeful turret]

its been stuck like this for like 30 minutes and nothings downloaded

[wispy burrow]

Ah...Applio...
You're probably missing some requirements.
https://docs.applio.org/installation/requirements/

[fickle tundra]

is this overtrained?

i rly regret setting the checkpoint save to every 50 epochs

[hot lagoon]

Couldn't click due to name 'PID' is not defined

[proper shale]

possibly

[woeful turret]

i check and i have them all

[woeful turret]

deleting python and trying again cuz it might be messin it up

[fickle tundra]

heres a better view of it btw

(its going back down so it might be good)

but then again idk

[plain drift]

"Starting Applio with runtime for Intel CPU support (you must have Nvidia support installed)...
The system cannot find the path specified.
Press any key to continue . . .
I have this issue

[wispy burrow]

Run Go-Applio-manager.

[rare thorn]

how do i find a model id

[wispy burrow]

And then
[3] Download NVDIA Runtime

[pulsar wyvern]

wkat os best? normal rvc or applio?

[jolly edge]

What do i have to fill here?

[pine ivy]

hi, i've tried to use google colab to change my voice, but it isn't working. is there any solution?

[woeful turret]

i give up on this because whoever made this made it incredibly annoying to find out what you did wrong to make it not work

[plain drift]

exectly what i did led me to this

[wispy burrow]

Agreed.
What version of python do you have?

[woeful turret]

3.9.8

[wispy burrow]

Go-Applio-manager.
Not Go-applio.

[woeful turret]

im going to detonate myself with a C4

[plain drift]

alright thanks

[wispy burrow]

I wish this got updated...

[woeful turret]

why does it do this every time

[wispy burrow]

What's the full path for the input audio?

[woeful turret]

C:\Users\***\Documents\1_videoplayback_(Vocals).wav

[wispy burrow]

Seems to be good...

[hot lagoon]

what shoud I do with the rmvpe_gpu just leve it as it is?

[woeful turret]

apparently the gaw dam wav file was corrupted or something and i just had to change it out

[rare goblet]

Ayo? @woeful turret level 2 !!!

[wispy burrow]

I have no anwser for you but just out of curiosity, what are you using?

[brittle wing]

i keep getting the tobytes error when trying to use applio @wispy burrow do you think you can help me with this?

temp_dir = Path(dir) / self.hash_bytes(data.tobytes())
AttributeError: 'NoneType' object has no attribute 'tobytes'

it was working fine before but then suddenly it conked out

[rare goblet]

Ayo? @brittle wing level 1 !!!

[wispy burrow]

Ugh...I hate those.
Not really sure what causes them...
What are you trying to do?

[woeful turret]

nevermind.

[brittle wing]

just trying to do model inference

it was working fine up until i switched the model then suddenly everything just broke

my audio pathway is fine

@wispy burrow wait uhhh i figured it out

it only works with indexes????

it doesnt work when i dont use an index

now it works even without an index??? what the helllllllllllll

[echo creek]

Some people argue that overtraining doesn't exist. In that case, can i assume that increasing the number of epochs will lead to better results? regardless of overtraining?

[brittle wing]

i aint an expert but im p sure

[low shard]

nope, overtraining does exist

[wispy burrow]

Well, at least it works now.

[low shard]

these are the epochs you should use based on your dataset lenght (btw the tabel are the epochs suggested generally, its not like a rule, sometimes it could depend, i wanted to share it to everyone because i know it helped me and some other model makers, dont take it too seriously but only as a reference)

its in italian but the left part is indicating the lenght of the dataset, and the right one the epochs to use

[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

🇫🇷 France

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

🇩🇪 Germany

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

🇮🇹 Italy

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

🇯🇵 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

[wispy burrow]

Hmm? Shouldn't it be less Epochs for bigger Datasets?

[echo creek]

but, some argue that if the dataset is long and of high quality, overtraining may not exist.

[wispy burrow]

Aren't Steps more important?

[low shard]

honestly i never heard about that before, idk where you heard this but like in rvc v2 overtraining does exist and could happen

[low shard]

not really, more epochs based on more long is your dataset

i would advice you to do like that tabel, it helped me in making my 15th models

[wispy burrow]

Someone should pin it.

[low shard]

ofcourse it could depends by how clean is the dataset , but generally it can be useful, me and other model makes i know use it

let me pin it

[fickle tundra]

i think this may be overtrained

[stiff cliff]

Why does generating on Hugging Face take so f*cking long?

[fickle tundra]

well i have a 40 something minute dataset and mine looks like it's overtrained at like 210 epochs

[proper elbow]

help?

Exception in thread Thread-1:
Traceback (most recent call last):
File "threading.py", line 980, in _bootstrap_inner
File "threading.py", line 917, in run
File "C:\Users\hp\iCloudDrive\RVC1006Nvidia\RVC1006Nvidia\gui_v1.py", line 653, in soundinput
with sd.Stream(
File "C:\Users\hp\iCloudDrive\RVC1006Nvidia\RVC1006Nvidia\runtime\lib\site-packages\sounddevice.py", line 1800, in init
_StreamBase.init(self, kind='duplex', wrap_callback='array',
File "C:\Users\hp\iCloudDrive\RVC1006Nvidia\RVC1006Nvidia\runtime\lib\site-packages\sounddevice.py", line 898, in init
_check(_lib.Pa_OpenStream(self._ptr, iparameters, oparameters,
File "C:\Users\hp\iCloudDrive\RVC1006Nvidia\RVC1006Nvidia\runtime\lib\site-packages\sounddevice.py", line 2747, in _check
raise PortAudioError(errormsg, err)
sounddevice.PortAudioError: Error opening Stream: Illegal combination of I/O devices [PaErrorCode -9993]

it say this while trying to run it

aka audio convertion

[fickle tundra]

mind you it's prob going up then back down like it did last time i thought it was overtraining

[low shard]

so it went overtraining at 210 epochs?

[proper elbow]

Help?...

[fickle tundra]

yep (this is an update at 217 epochs)

[sand iris]

icloud drive ?

[proper elbow]

i name things funny

it not actual icloud drive

[sand iris]

Oh

[low shard]

btw the tabel are the epochs suggested generally, its not like a rule, sometimes it could depend, i wanted to share it to everyone because i know it helped me and some other model makers, dont take it too seriously but only as a reference

[fickle tundra]

and my dumb ass set checkpoint save to every 50 epochs so i can't even get it at the best moment before overtraining

[sand iris]

The realtime gui is uhh... not really stable

I've only seen it have issues lmao

[fickle tundra]

oh ok then

[fickle tundra]

the lil dot is the 200 epoch save (which was actually a pretty good spot for it to save but i would have preferred just a tiny bit later)

[low shard]

could i dm u for smt?

[fickle tundra]

sure

[proper elbow]

awww

[tranquil cliff]

Hi

[plush pond]

mfw local training

this is me trying to import 19 minutes of sound data

oh yeah, oddly enough, none of my resources are being used as much

probably because the rig is good? idk

[fickle tundra]

i think your computer is just shit /j

[rare goblet]

Ayo? @fickle tundra level 5 !!!

[fickle tundra]

like anytime i try to make a model i smell electronic burning smell but it still trains

[plush pond]

I'm also trying to pass that time by watching anime (azumanga daioh)

damn.

[fickle tundra]

and also, thats about the amount of time that it takes me to train a whole model with a 15 minute dataset

and thats just you trying to upload a dataset

[fickle tundra]

mind you, i dunno what it's like straight up uploading, cause i usually put the dataset zip on google drive then put the link in the resources section and it does it in like 5 - 10 seconds, so maybe try doing that

[plush pond]

nah, it said not to upload zips

or maybe it doesn't apply to the second option

[fickle tundra]

what one are you using? cuz im using applio

[plush pond]

mangio

[fickle tundra]

ahhhh okay, maybe just try uploading a zip to drive and paste the link in there anyway, might work, you have been waiting like 50 minutes for it to load so it's work a shot

[plush pond]

64 now

oof

[fickle tundra]

wait i was wrong it was 64

[fickle tundra]

but yea like bro i could train like 3.8 models with 1 minute datasets with the amount of time you have been waiting for ur 19 mins of audio to upload

[brittle wing]

where can i train a model in the cloud with colab or something similar?

[plush pond]

ah, the GPU limit on colab was lifted

thank fuck

[tranquil cliff]

Rvc disconnected v2

-this is a guide in Italian and English for rvc disconnected- https://discord.com/channels/1159260121998827560/1160926856548200531

[hard gust]

Guys, someone knows why Applio with DML infers everything with autotune for me? the autotune checkbox is unchecked ofc

[fickle tundra]

I thought we were talking about local-

[plush pond]

and then it turns out I gotta wait for 5 hours if I'm going for 400 epochs
since every single one of them is being processed for ~45 seconds

[rare goblet]

Ayo? @plush pond level 2 !!!

[plush pond]

fuc

okay, nah, this doesn't even need "help" anymore

[brittle wing]

wdym spaces from the path?

[sand iris]

There's at least one file name that contains spaces

That usually causes a lot of issues with the current RVC stuff

[visual jungle]

so uuh where do i find the trained feature index?

[sweet zealot]

In the same folder as the model

[visual jungle]

oh found it

for me it was in the logs for some reason

[snow jacinth]

how tf do u inference a modely wodely i dont wanna make one i wanna make it talking singing

im like stupid but like

show me where

[torpid valley]

hey guys, anyone using life time voice? how do i give it more time to produce higher quality sound? cuz all i get is mumbling and other shit

[brittle wing]

file?

you mean the audio?

the mp3 to train?

@sand iris

[sand iris]

yeah

[brittle wing]

OHHHH

they have a heck of spaces

[wraith jungle]

I have my voice model created from a couple of months back, which RVC Colab do people recommend to use nowadays/anyone have a guide for just mixing a song + voice I was using Applio before but there seems to be lots of new forks

[brittle wing]

What’s the different between G and D?

help now

[rare goblet]

Ayo? @brittle wing level 4 !!!

[brittle wing]

it says preprocessing completed..

but its still not done right?

@sand iris

[sand iris]

go on with the next steps

It should be good

[brittle wing]

u sure?

what now??

what pitch extraction algorith?

and how much epochs

[spiral osprey]

How do I make a cover more faithful? Is it changing pitch or something? How can I make it more like the original?

[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

🇫🇷 France

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

🇩🇪 Germany

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

🇮🇹 Italy

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

🇯🇵 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

[wet ingot]

hey what is the best headset to use?

[brittle wing]

-rvc

[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

🇫🇷 France

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

🇩🇪 Germany

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

🇮🇹 Italy

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

🇯🇵 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

[timid imp]

Hey ! I don’t know if anyone can help me on this but I’ve been trying to make an AI cover but whenever I try to drop my audio acapella file in Mangio-RVC-Web it’s stuck on " Uploading…" If anyone ever had this issue or know how to fix it I could use some help ^^

[civic goblet]

does anyone know about the colab

[misty cosmos]

heyyy, im having the same problem.... i cant use a voice that doenst has an index can u help me

[rare goblet]

Ayo? @misty cosmos level 1 !!!

[glad zealot]

No idea.. it's supposed to work without index tho

[civic goblet]

hey do you use the colab to make a model

[viscid moss]

yes, rvcDisconnected

[civic goblet]

can you help?

it says "os" is not defined

[viscid moss]

to make a model? Sure

[civic goblet]

yeah

i got my .wav files

i'm stuck at the error

[viscid moss]

yeah, in a folder in ur GDrive, rigth?

let me show u, wait

[civic goblet]

nah i downloaded it but got it to the colab

somehow

[autumn hill]

ok aicovergen.

[viscid moss]

Have you already done that?

[civic goblet]

i was on preprocessing data

not there yet

[viscid moss]

can u show me the error 👀

[viscid moss]

💀

hmmmmm try collab?

[civic goblet]

hol up

[viscid moss]

ok

[autumn hill]

i hate using the colab but yeah guess thats the best option

[civic goblet]

is opera a chromium based browser

[viscid moss]

U can use Hugging Face but it's too slow or u can use the WordBand bot

[viscid moss]

I think that's not the problem

[civic goblet]

just asking

running dependencies

i forgor

[viscid moss]

don't worry

[viscid moss]

[civic goblet]

it taking too long

wifi bad

[viscid moss]

nah, it's just collab issue. Just wait

[placid light]

Colab thinking that this code is disallowed when in reality it isn't because it has no webui. 💀

[civic goblet]

true

[viscid moss]

oh, the bot's AICoverGen have that warning

Use my version (spanish)

wait

[placid light]

But I don't speak spanish. bruh

[viscid moss]

https://colab.research.google.com/drive/1qezaBvCn4eSG1U2L4a64tEHTynbIu8eN?usp=sharing#scrollTo=DtHla7cvRRqh

It's just minor changes

It's literally the same with spanish titles

lol

and no warning

[placid light]

I'll just follow it the same way as AICoverGen?

[viscid moss]

yep,the same way

[placid light]

Does it support Google Drive links for downloading models?

[viscid moss]

only Hugging Face and Pixeldrain

'cause it's direct download

[glad zealot]

The one with webui doesn't have that warning lol

[placid light]

One problem I have though is that I am going to use a private model, and second is that PixelDrain doesn't exist anymore.

[civic goblet]

DUDE COLAB RN

[viscid moss]

ya, the problem is this....

A banned word

[glad zealot]

The webui one doesn't have a banned word

[placid light]

So, in order to circumvent it, just rename it to something unrelated. lol

[viscid moss]

My noob fix lol

[glad zealot]

So it does work atm lol

[civic goblet]

it's stuck on mount drive

[glad zealot]

The one of the pins

[glad zealot]

You can check what I did it's on the pins on this channel

[viscid moss]

ya, but i decided zip all and unzip later in collab

[placid light]

"Collab"? 💀

Anyways, okay.

[glad zealot]

Mine has 2 versions, you can use the original or my fork but ye

[civic goblet]

bit weird tbh but "collab" it is.

[glad zealot]

Not much difference except for the pitch control

[viscid moss]

I see

[glad zealot]

Applio colab works too :D

[civic goblet]

why is it taking so long for it to mount drive

[placid light]

Yours has clone and install in one cell.

Which makes it easier for me.

[glad zealot]

Mine does :D cause there's no reason to separate them

The download mdx models can also be in 1 cell ngl

[placid light]

Wait, I thought Colab only bans webui's, but also RVC itself too?

[glad zealot]

Nope

Not at all

They just ban specific names

[placid light]

So what does AICoverGen mean to them?

[glad zealot]

No idea lol

But it does give warning

[placid light]

How to upload my created model to hugging face?

[viscid moss]

Exactly!

U can find the banned word in this HF space

https://huggingface.co/spaces/mofu-team/ggl-chk

[fathom remnant]

Uhh
TypeError: stat: path should be string, bytes, os.PathLike or integer, not list
Help?

[viscid moss]

@glad zealot

[glad zealot]

Ohhh nice lemme check that

[placid light]

lol

[viscid moss]

Literally

Google sucks

[placid light]

Can you help upload my own model to HF?

[viscid moss]

Locally?

[fathom remnant]

Yes and just trained localy

[viscid moss]

-upload

[azure marsh]

Tag suggestion for @placid light

huggingface.co upload Guides

🇺🇸 English guide

Link: https://rentry.org/fdg_guide_new
Credits: FDG

🇮🇳 Hindi guide

Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes

🇹🇷 Turkish guide

Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes

[glad zealot]

Not my method but damn they are so close at patching mine

[viscid moss]

Google is smart and stupid at the same time

[fathom remnant]

Look near the bottom the whole string AiCoverGen is banned

[glad zealot]

I'm not using that name tho lol

[fathom remnant]

Yea but funny

[viscid moss]

hmmmm First time I see that, I don't know what to do

[fathom remnant]

Its only the last one also

[glad zealot]

My method just clones a repo but also renames it while doing it

So it doesn't ever use the name

[placid light]

I just wanted to upload the model on HF, not sharing it on AI Hub because it's a private model.

But anyways, thanks.

[fathom remnant]

I did accidentally use it with the wrong index file though

[viscid moss]

W Okada 💀

[glad zealot]

I don't use that also lol

[viscid moss]

Interesting

[glad zealot]

I mean technically I do but not at the same time

[viscid moss]

ya, I understand

[glad zealot]

It's in the files but the cell is not executing it

I mean it's executing it but it's not using the same name

[viscid moss]

Have you tried encrypting the names? 👀

[glad zealot]

It is

Lol

[viscid moss]

but in base 64

[fathom remnant]

  File "C:\Users\OneDrive\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\gradio\routes.py", line 437, in run_predict
    output = await app.get_blocks().process_api(
  File "C:\Users\OneDrive\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\gradio\blocks.py", line 1346, in process_api
    result = await self.call_function(
  File "C:\Users\OneDrive\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\gradio\blocks.py", line 1074, in call_function
    prediction = await anyio.to_thread.run_sync(
  File "C:\Users\OneDrive\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\anyio\to_thread.py", line 31, in run_sync
    return await get_asynclib().run_sync_in_worker_thread(
  File "C:\Users\OneDrive\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\anyio\_backends\_asyncio.py", line 937, in run_sync_in_worker_thread
    return await future
  File "C:\Users\OneDrive\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\anyio\_backends\_asyncio.py", line 867, in run
    result = context.run(func, *args)
  File "C:\Users\OneDrive\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\lib\infer\modules\vc\modules.py", line 318, in vc_single
    if os.path.exists(file_index)
  File "genericpath.py", line 19, in exists
TypeError: stat: path should be string, bytes, os.PathLike or integer, not list``` this is traceback

[placid light]

Obfuscating the names is a smart move.

[glad zealot]

Ahh nah thats the first thing they patched it think

At least most of the stuff I saw were using it

[viscid moss]

I was part of the Applio team, and thanks to base 64 encryption we managed to avoid being banned for 1 month

[fathom remnant]

Ehh its corrupted "boo hoo" e4900 then

UH OH

IMA RESTART IT

[viscid moss]

[fathom remnant]

Might be a gpu overflow

[viscid moss]

Might

[fathom remnant]

It was

Restarting fixed

[viscid moss]

U need this baby

[fathom remnant]

?

[viscid moss]

RTX 4090

[fathom remnant]

i have 4060.

[viscid moss]

ohhh

[fathom remnant]

it was open too long.

[viscid moss]

how long is your dataset?

[civic goblet]

i was afk

[signal otter]

@viscid moss what Is a .json and a pth. What i need to do with this two files, to do a ai cover. What program i need to use?

[viscid moss]

.json???

[signal otter]

Yeah

[viscid moss]

The files that u need are .index and .pth

👀

[signal otter]

That's strange the model i download only haves .pth and file .json, not the .index

[civic goblet]

alright, it's becoming too slow

[viscid moss]

what's the model?

[signal otter]

Is the Hajime Hinata from danganronpa (english dub)

[viscid moss]

lemme check wait

[civic goblet]

alright eddy i finished dependencies what next

[rare goblet]

Ayo? @civic goblet level 7 !!!

[viscid moss]

rvcDisconnected, right?

[civic goblet]

yeah

[signal otter]

@viscid moss if can't find. Is on the Request model chat

[viscid moss]

send me the link

[fathom remnant]

me?

[viscid moss]

set the name of the model and sample rate

[viscid moss]

ya

[civic goblet]

what's sample rate?

and where is it

[viscid moss]

32k, 40k, 48k

-colab

[azure marsh]

Google Colabs

RVC

• AI-Cover-Gen-No-UI (inference only)
• RVC v2 Disconnected (training only)

🌐 Translated colabs

• PT-BR: -colabs_br

Audio Separation/Isolation

• MDX Net Colab
• UVR arch colab 5.0.2/c1.04
• KaraFan colab

[civic goblet]

do i set training variables

[viscid moss]

replace the name and set 40k, it' better

[viscid moss]

ya

[fathom remnant]

8 secs (all voicelines that are not just noises)

[signal otter]

@viscid moss
https://mega.nz/file/BKRgUJRa#JgfUjk38o3meHnxr1tn1i9sMf8_mVUZmax4o1_eTDAo

Here it is

[civic goblet]

now preprocessing

[viscid moss]

nice

[viscid moss]

checking

[civic goblet]

load dataset, do i put my zip in there

[fathom remnant]

Only sounds ok though might have been the audio i genarated

[civic goblet]

or nah

[viscid moss]

Ok, I see the problem. It is an SVC model

Not RVC

💀

literally

-svc

[azure marsh]

https://media.discordapp.net/attachments/1089308888085577808/1118602866156896356/IMG_6652.gif

[civic goblet]

wtf it made a folder itself

[viscid moss]

It's normal, the files will be saved there

[viscid moss]

hmmmm very small datasets have problems with the index

[civic goblet]

so i put the zip file in, now put name in?

[viscid moss]

ya

replace zipfile.zip

for example

eddysvoice.zip

[fathom remnant]

No it had been open for too long

applio

[viscid moss]

oh

[civic goblet]

it worked let's gooooo

[civic goblet]

feature extracting

ill save them files

[fathom remnant]

How do you get verified?

[civic goblet]

chat

a fuck lot

[fathom remnant]

ahh

[viscid moss]

LVL 5

[signal otter]

@viscid moss what's the diference between svc and rvc?

[rare goblet]

Ayo? @signal otter level 2 !!!

[civic goblet]

so what should i set the stepcount

[viscid moss]

SVC is the ancestor of RVC. SVC is no longer used, it has become obsolete

[viscid moss]

skip that

run the index

[civic goblet]

just play it or no

[viscid moss]

andthen train

[fathom remnant]

Man i love step count 1 (dont)

[viscid moss]

for resuming

[civic goblet]

for training what should i set my settings

[viscid moss]

Try 25,500, 20

[rare goblet]

Ayo? @viscid moss level 25 !!!

[civic goblet]

nothing else

[viscid moss]

train the index first

[civic goblet]

just run it?

[viscid moss]

ya

and then train

[civic goblet]

alright

[viscid moss]

https://tenor.com/view/anya-forger-anya-spy-x-family-anime-salute-salute-gif-okey-dokey-gif-25700022

[civic goblet]

tensorboard is

somewhat working

[viscid moss]

can u show me?

[civic goblet]

yeah hol up

[viscid moss]

Do you know how to use tensorboard?

[civic goblet]

no clue

[viscid moss]

ok, send me a pic and I teach u how to use

[viscid moss]

oh, so fast

but I need the tensorboard pic

[civic goblet]

mac is a BEAST

epoch 40

[viscid moss]

ok, change inactive to scalars

[civic goblet]

ok

[subtle cypress]

Hi, what is the link?

[civic goblet]

done

[viscid moss]

in the config icon activate the check to reload the tensorboard automatically

[civic goblet]

-colab

[azure marsh]

Google Colabs

RVC

• AI-Cover-Gen-No-UI (inference only)
• RVC v2 Disconnected (training only)

🌐 Translated colabs

• PT-BR: -colabs_br

Audio Separation/Isolation

• MDX Net Colab
• UVR arch colab 5.0.2/c1.04
• KaraFan colab

[subtle cypress]

Thanks

[viscid moss]

sendme another pic

U're seeing graphs now

[civic goblet]

yep

[viscid moss]

unmark "Ignore outliers......"
set the smoothing to 0,95
and in the search bar of tensorboard type "g/total"

[civic goblet]

loss/g/total?

what about it

[viscid moss]

ya

[viscid moss]

You must train until the graph starts to go up and does not go down again.

[civic goblet]

how to train it, wdym

[viscid moss]

Check this

https://rentry.org/rvc_making-models

[viscid moss]

just watching the graph

U're training ur model right now

U just hace to stop the train when the graph rises

-overtrain

[azure marsh]

How do I know if my model is overtaining?

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

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

Credits: LUSBERT

Automated Overtraining Detection (AOD)

Will be available soon in #1159513888199540817

Credits: grvyscale

[civic goblet]

so just stop it when it rises

[viscid moss]

ya

and then

run the next cell

[civic goblet]

which is what?

[civic goblet]

alright

[viscid moss]

And that's it

Ur model will be saved in ur GDrive

The .pth and .indedx file

[civic goblet]

and how to convert to hugging face?

[viscid moss]

just zipping the pth and index and upload to ur HF account

check this

-upload

[azure marsh]

huggingface.co upload Guides

🇺🇸 English guide

Link: https://rentry.org/fdg_guide_new
Credits: FDG

🇮🇳 Hindi guide

Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes

🇹🇷 Turkish guide

Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes

[civic goblet]

noice

bro automaze looks like it will kill me

[viscid moss]

hahahaha

[civic goblet]

it's like 35 ish

[brittle wing]

how do i use the index

[viscid moss]

I see

[viscid moss]

RVC? Realtime?

[brittle wing]

rvc

[civic goblet]

dude it ain't going up

[viscid moss]

What are you using?

[viscid moss]

how much epochs?

[civic goblet]

500 u said

[brittle wing]

ngl idk can i dm u

[civic goblet]

right?

[brittle wing]

ill send a ss

[viscid moss]

ya, but right now epoch

[civic goblet]

181

[viscid moss]

i think is a false positive

[civic goblet]

???

[viscid moss]

continue watching the graph, it may go down again

[brittle wing]

can i dm u

[viscid moss]

Sure

[brittle wing]

cuz idk wat im doin idk how to add the index

[civic goblet]

200 epoches

[fathom remnant]

Yea on my new one it went up for a LONG time but then went down again

[rare goblet]

Ayo? @fathom remnant level 5 !!!

[fathom remnant]

Nice

[civic goblet]

verified

[fathom remnant]

I wonder what the limit to tts is

Becasue Ima be testing that with this one

[viscid moss]

https://whyp.it/users/22854/eddycrack-864
Guys, wdy think?

These are the models I have made (they are still missing), most of them are commissions lol

[civic goblet]

it's not budging

[viscid moss]

[civic goblet]

epoch 250

[fathom remnant]

GPU?

[viscid moss]

He's using colab

Tesla T4

[fathom remnant]

Hmm

Thats not good

[civic goblet]

detected it

so

im doing this is the morning

[rare goblet]

Ayo? @civic goblet level 8 !!!

[civic goblet]

it's 12 am

im not giving a frick anymore

I'll remember you all in therapy.

[viscid moss]

U can ue AICoverGen or the dicord bot

[whole scaffold]

how can i use the AI models? gardio.live doesnt work anymore

[fathom remnant]

#1159949278270193734 has all the software for doing such things

[whole scaffold]

yeah thats just links not all of us are smart enough to figure most of that stuff out

[fathom remnant]

The #1159513888199540817 will be of use to you

also https://docs.aihub.wtf

[plain drift]

I have this error I have reinstalled nvidia runtime, updated and reinstalled but this the error im getting

[frigid wasp]

I know, but I was using it for AI song covers

For some reason, certain links cause the program to spit out an "There's been an error" message.

This happens when downloading models for the RVC program

[final schooner]

RVC Gui is a fast and easy way to make voice covers, great for beginners.

[plain drift]

does anybody have a solution

[rare goblet]

Ayo? @plain drift level 3 !!!

[lament eagle]

Why is it always at 0%

is this broken or still training

[somber parrot]

I believe that's just how many steps into the next epoch it is, so it's printing out every epoch

logger.info(
    "Train Epoch: {} [{:.0f}%]".format(
        epoch, 100.0 * batch_idx / len(train_loader)
    )
)```

[lament eagle]

does it mean it prints out nothing when it's [0%]

or what

[somber parrot]

it means your log interval is the same as the epoch interval so it will just show 0%

by default the log interval is 200 so your model would have 200 steps per epoch

[lament eagle]

ohhh

I mean I'm retraining with the same dataset rn

[somber parrot]

it should be the same thing happening unless you changed the dataset or log interval

which you don't need to do, the 0% isn't anything good or bad, it just shows at what point it's at currently

[glad zealot]

Nah that's fine you don't need to worry about it

[brittle wing]

https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=OVQoLQJXS7WX

Why does this Easy RVC Gui won't work?

[rare goblet]

Ayo? @brittle wing level 1 !!!

[agile hollow]

I tried to make Alastor's Game AI cover with Caine's voice, and everyone was perfect but refrain broke AI

[young hearth]

using direct ML (using AMD) and rvmpe algotirthm gives me error

pm works completely fine, but rvmpe just crashes

[graceful obsidian]

ur gpu? also what does the log say, error wise

quite possibly you're out of ram for inference, if other f0/feature methods work just fine

( was the case for me )
try to confirm if that's the case by trying to infer something short, 5-10 seconds

[proper shale]

Certain models had their links taken down because of the owners, maybe that's the issue?

[stoic cloud]

Nah it's not, that's just the progression of each epoch, but it's mostly 0% because it doesn't update

[graceful obsidian]

because the logging displays 2 things:

1. Epoch that's just now has started the training
2. Epoch that's already done training + it's stats and so on
   So, that's normal. dw about it

[tranquil cliff]

Hi

[last plank]

Is it possible to use any of these voices as a windows narrator?

[astral jungle]

it has been banned

[astral jungle]

Hi

[high night]

Is it possible to use any of these voices as a windows narrator?

windows

narrator

voices

these

possible

[glad zealot]

make a model of it?

[astral jungle]

Hello @high night

[errant sapphire]

how do you use a rvc v2 model

[astral jungle]

I use https://colab.research.google.com/github/hinabl/AICoverGen-Colab/blob/main/Hina_Mod_AICoverGen_colab.ipynb#scrollTo=NEglTq6Ya9d0

[glad zealot]

hey its mine OwO

[high night]

this is made by me guys

had a hard time modifying it

ngl

[brittle wing]

yo checkmate

[civic goblet]

do i use "commit changes to main" for hugging face

[brittle wing]

can you help me??

[high night]

wsg brtoher

[brittle wing]

i need help with training

i completed the step 1

what now??

[civic goblet]

colab?

or locally

[brittle wing]

local

[civic goblet]

dang

[brittle wing]

what do i do now???

[high night]

honestly

i dont know shit about that

but wait

i got something

[high night]

https://discord.com/channels/1159260121998827560/1159618293594861579

[civic goblet]

wait what files do i upload in hugging face

[brittle wing]

yeah ik i installed it

[high night]

https://discord.com/channels/1159260121998827560/1159618293594861579https://discord.com/channels/1159260121998827560/1159618293594861579https://discord.com/channels/1159260121998827560/1159618293594861579https://discord.com/channels/1159260121998827560/1159618293594861579

follow this guide now

[astral jungle]

Oh thank you for creating this wonderful thing

[high night]

no problem

[civic goblet]

checkmate do u know what files u need to upload

[high night]

where

[civic goblet]

in hugging face

[civic goblet]

so what files?

[high night]

audio files

[civic goblet]

types?

pth and index?

[brittle wing]

anyone knows how long will it take???

[vital raft]

Someone who could help with how to use it on dsicord

[astral jungle]

What to use?

[civic goblet]

the files aren't uploading

[brittle wing]

guys

how much epoches do i set???

and what save frequency

[astral jungle]

Upload to Huggingface?

[civic goblet]

yeaa

the pth has an error

[civic goblet]

im dead confused

the D_23333333

[astral jungle]

Maybe file D_23333333 broken

[civic goblet]

so what do i do

[astral jungle]

Do you have an index file yet??

[civic goblet]

uploaded? cuz yes

[astral jungle]

If you already have the index, then you only need to train the pth file

[civic goblet]

the G_233333333 is uploaded

uploading*

the index and other pth is uploaded

[astral jungle]

You upload each file, right?

[civic goblet]

the 4 of em?

[rare goblet]

Ayo? @astral jungle level 10 !!!

[astral jungle]

You should create a zip file and include pth and index files @civic goblet

[civic goblet]

yeah

just put the zip in raw

[astral jungle]

Can you take a screenshot for me to see? @civic goblet

[civic goblet]

it aborted

[astral jungle]

Ồ

[astral jungle]

I know why

[civic goblet]

why?

[astral jungle]

Please do so and re-upload

[civic goblet]

so apache

[astral jungle]

Yup

[civic goblet]

done

[astral jungle]

Okay

Then upload the zip file

[civic goblet]

ok

so upload it

hashing

[astral jungle]

Please wait 1-2 minutes

[civic goblet]

uploading

[rare goblet]

Ayo? @civic goblet level 9 !!!

[civic goblet]

wifi garbage

never using an IoT router ever again

it didn't work

aborted

[astral jungle]

How many MB does a pth file weigh?

[civic goblet]

idk but the zip was 1.3 gb

[astral jungle]

Oh then the pth files cannot be used

[civic goblet]

damn can i fix it

[astral jungle]

Where do you train models?

[civic goblet]

colab

[astral jungle]

Can you give me the link?

[civic goblet]

for it?

i used disconnected

[cobalt oyster]

Hi @astral jungle

[civic goblet]

-colab

[azure marsh]

Google Colabs

RVC

• AI-Cover-Gen-No-UI (inference only)
• RVC v2 Disconnected (training only)

🌐 Translated colabs

• PT-BR: -colabs_br

Audio Separation/Isolation

• MDX Net Colab
• UVR arch colab 5.0.2/c1.04
• KaraFan colab

[astral jungle]

You should retrain the pth file

[astral jungle]

Hello bro

[civic goblet]

ima just make a new one

hhgregg mascot

[cobalt oyster]

What is the best way to remove vocals

[civic goblet]

nvm

[cobalt oyster]

Which one is better Google colab or vocal remover website

[civic goblet]

im doing it at 300 epoch this time

[astral jungle]

Wait

[cobalt oyster]

After helping @civic goblet help me I will wait

[astral jungle]

https://vocalremover.org

@cobalt oyster

[astral jungle]

I think you should use EASY instead of RVC V2 Disconnected

[civic goblet]

where is it?

[astral jungle]

Because when EASY finishes training, it will automatically save to the drive

[last dagger]

any softwares to test models? I don't want to test them on a website but if it's possible locally

[astral jungle]

But I don't know if it has been fixed yet

[civic goblet]

where is it

[astral jungle]

Because I haven't been able to train these past few days

[slim hatch]

colab or mvsep

[cobalt oyster]

? Could u repeat

[civic goblet]

nah not using easy

[astral jungle]

https://colab.research.google.com/drive/1khsjzEigBogtkA6ZNL3gNXZqmjUAKkM1?authuser=1#scrollTo=LnMrksQI1lnW

[cedar jolt]

guys is there anyway to change huggingface file into drive?

[plain drift]

Im having this error can anybody help a brother out

[civic goblet]

im using disconnected cuz i actually know somehow

[brittle wing]

Guys i need a help

https://ibb.co.com/4W4y2YB

(p.s i can't post images so i post link There's image)

[cedar jolt]

.

[civic goblet]

should i delete logs

[astral jungle]

How many epochs have you trained?

[civic goblet]

300

[astral jungle]

It has now reached over 100 epochs?

[civic goblet]

i wasn't doing it now

yesterdat

[astral perch]

Does a mixed dataset with singing and speaking result into a better or worse voice model

[plush pond]

Is it normal for trained epochs to show "0%" constantly? When I was making my first two models, the percentage numbers would be random

Each of the epochs do take time to train, from 35 to 50 seconds, so I'm assuming it's still gonna work

[astral jungle]

Are you using Disconnected?

[plush pond]

yeah

Dataset is 19 minutes long

[astral jungle]

I don't know because I don't use Disconnected much

You can try switching to EASY