✨│ai-help

[visual niche]

Thanks for answering. Is that enough?

[rare goblet]

Ayo? @visual niche level 1 !!!

[visual niche]

So what I'm asking is, the AI Cover programme doesn't matter, does it?

As far as I understand, I can only get clean vocals by installing this programme you mentioned.

[molten pecan]

You can either get them by using UVR5 or by searching them on YouTube

Though usually the ones from YouTube are on low quality, unless it's stated that they are the "official" vocals

So it's best if you get them yourself using the program

Si

[visual niche]

I understand, you helped me a lot, thank you very much.

[molten pecan]

Tenés la interfaz de color rosa?

[molten pecan]

You are welcome!

Feel free to ping me if you have any questions ^-^

[brittle wing]

sii

[visual niche]

If I have a problem, as long as it doesn't cause you any trouble.

[brittle wing]

pero oye, no puedo escuchar nada cosas de mi pc cuando tengo esa opcion activa :c

[molten pecan]

Hm que raro

Creo que eso es un problema de la aplicación en sí y no de Windows

Creo que vas a tener que cambiar una opción de w okada

Intenta preguntar en #🔍│help-w-okada

[brittle wing]

oh ya lo arregle uwu

[molten pecan]

Ah okay

Diviértete!

[timid pilot]

@molten pecan omg I hate this. when ever I talk with the voice changer, on obs I can hear myself (normal voice) then it right after. Its so annoying.

[molten pecan]

I think you can turn that off in OBS

Or on the Windows settints

[timid pilot]

know how I could?

[molten pecan]

No clue

Never used either OBS or w okada

Try asking on #🔍│help-w-okada

[brittle wing]

I had to reset my pc, and when i try to install the requirements-amd.txt it says:
"Failed to build fairseq
ERROR: Could not build wheels for fairseq, which is required to install pyproject.toml-based projects""

(Applio RVC Fork)

[rare goblet]

Ayo? @brittle wing level 2 !!!

[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

🇪🇸 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]

are you on windows

the requirements-amd file only applies to linux

requirements-dml is what you are looking for

Also, try reading the full error, not the last few lines

[brittle wing]

ohhh

thanks

oops

installing the said file didnt work...

how can i fix?

[sand iris]

reading is key

[brittle wing]

i did

[sand iris]

first line on your screenshot

[brittle wing]

i installed the microsoft visual c++

[sand iris]

Oh

[brittle wing]

didnt work

[sand iris]

Show me the components you have selected

[brittle wing]

wdyn

[sand iris]

You need like two optional ones

[brittle wing]

ah okay

[sand iris]

on the visual c++ tools there's optional tickboxes

[brittle wing]

one sec ill try and modify the installation

[brittle wing]

yeah ik

[sand iris]

👍

[brittle wing]

which ones to keep on?

ill send a ss one sec

[sand iris]

"Windows 11 SDK" and one other I can't recall from the top of my head

(yes, windows 11 sdk even if you are on windows 10)

[brittle wing]

which ones

the windows 11 sdk

and what else

@sand iris

[sand iris]

Huh the first two ones should be enough by themselves

[brittle wing]

okay thanks

[sand iris]

were they selected before ?

[brittle wing]

no i dont think so

:/

[sand iris]

Oh

[brittle wing]

the c++ wasnt downloaded automatically i think

[opal saffron]

How can I use it on the phone?

[violet heron]

I would use HuggingFace Spaces

More specifically

Ilaria RVC

https://huggingface.co/spaces/TheStinger/Ilaria_RVC

Create a HuggingFace account and duplicate the space to use

[opal saffron]

Ok boss

[brittle wing]

forgot

again

how to fix

😭

[sand iris]

lmfao

applio moment

[brittle wing]

fr

[sand iris]

pip install torch_directml

[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

[brittle wing]

ye thanks

just searched this up and found what u sent just now

🫶

[sand iris]

I swear Applio is silly

[brittle wing]

sure is

colab was way easier

[sand iris]

[brittle wing]

:3

@sand iris didnt work >.<

[sand iris]

It didn't find it ?

[brittle wing]

nope

[rare goblet]

Ayo? @brittle wing level 3 !!!

[brittle wing]

its odd

[sand iris]

It might be torch-directml rather than torch_directml

dash instead of an underscore

[brittle wing]

ah ok

is this done yet? or is it bugged

https://i.imgur.com/T6LIvCH.png

[rare goblet]

Ayo? @brittle wing level 1 !!!

[brittle wing]

its already installled

very odd

[sand iris]

How are you starting applio ?

I wrongly assumed you were using the system python

[brittle wing]

no, im using the go-appllio.bat

and selecting the runtime with AMD

[sand iris]

Oh lord

[obtuse stream]

wich pitch the best for speech voice? i want make same actor voice

[brittle wing]

did i do something wrong?

.<

[sand iris]

Not sure how they do things but

Nah it's on applio not on you

unsurprisingly

[brittle wing]

oh alr

xD

[sand iris]

you'll need to open a terminal in the same folder as applio

and run runtime\python.exe -m pip install torch-directml

[brittle wing]

ah alr

https://i.imgur.com/NMY8YcC.png

mine is repeating at this part

[obtuse stream]

wich pitch the best for speech voice? i want make same actor voice

[brittle wing]

anyone know what's with this?

still says already satisfied

[sand iris]

#🔍│help-w-okada

[brittle wing]

this is new tho

[sand iris]

Should be working now then

[brittle wing]

alr

lemme check

[vapid gust]

does this mean it will save all epochs?

[proper shale]

rvmpe is generally the best

[vapid gust]

can someone help me

how do i make it so all epochs save

with step

[obtuse stream]

Hop Length (lower hop lengths take more time to infer but are more pitch accurate): do you touch this?

[brittle wing]

-colab

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

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

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

[vapid gust]

can someone help me

[remote maple]

hey still having the same issue from yesterday. tensorboard is displaying a blank screen

[proper shale]

for rvmpe no

[obtuse stream]

ok i used the original setting 64

for rmvpe

[vapid gust]

i think u need to use firefox browser

[obtuse stream]

this

[proper shale]

It's fine

[obtuse stream]

thanks

[remote maple]

nope no change

[proud elbow]

hop length only works for mangio-crepe option

[obtuse stream]

how can I make sure that the sound is not too overtrain?

[proper shale]

Just test it

[obtuse stream]

how? should I always do it with a different epoch number? does the result depend on how long a sample I work with?

[brittle wing]

where do i import datasets on applio?

didnt use this for a while

[proud elbow]

Applio\datasets\YourModel

[brittle wing]

thanks

doenst work

huh

[proper shale]

...just use some samples and infer

and no, you can train the first 500 epochs then decide if the model sounds good enough or not

[brittle wing]

ummmm

how should i fix

list index out of range

[obtuse stream]

Is 8 minutes of sound samples enough?

[rare goblet]

Ayo? @obtuse stream level 4 !!!

[proper shale]

Would recommend 10+ mins but sure

I made a 8 min model and it sounded nice so

[brittle wing]

@proper shaledo u know how can i fix this?

[obtuse stream]

spectrogram? what do you mean

[proper shale]

I have no clue

Not a local user, sorry

[brittle wing]

[sand iris]

What the fuck

Now this doesn't make sense

[brittle wing]

exaclty

[brittle wing]

imma try and re open

works now

prob a launch error or smthng

[sand iris]

wait are you trying to train or infer

[rare goblet]

Ayo? @sand iris level 34 !!!

[brittle wing]

train

but it works now

dw

[sand iris]

Oh

You can't train on amd/intel sadly

[brittle wing]

oh damn

[sand iris]

It'll do it on CPU implicitly

[brittle wing]

yeah ik its on CPU

-colab

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

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

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

[rare goblet]

Ayo? @brittle wing level 4 !!!

[olive olive]

I trained my model on rvc but when i go to the file explorer in my rvc file i only see index training but not my pth

[quartz sky]

tysm 😭

[remote maple]

.pth will be in the weights folder

[olive olive]

I looked but he is not inside he is no where

[wooden isle]

-local

[azure marsh]

All Working Local Forks 🖥️

• Applio-RVC-Fork, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team Huggingface

Need some help? 🤔

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

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

[remote maple]

how are you training? local? collab?

[olive olive]

local

[remote maple]

what does the command prompt say

take a screenshot of one of the statements about an epoch

[olive olive]

I stopped everything but it always says success

[rare goblet]

Ayo? @olive olive level 1 !!!

[remote maple]

stupid question but you did hit start training right? not just train feature index

[olive olive]

i clicked on "one click training" then "train feature index" then i clicked on "train model"

i clicked on the three

[brisk hornet]

how do I duplicate a space

I'm on desktop-

oh

thank you!

[remote maple]

what type of local training are you trying?

[brisk hornet]

why is this being a pain..

[molten pecan]

I think it has to be a huggingface link

Not a weights.gg one

[brisk hornet]

damn.

[rare goblet]

Ayo? @vapid gust level 11 !!!

[vapid gust]

what u guys think

[cyan spear]

whenever i use rvc, the voice keeps cutting out and it takes way longer

ive never had this issue before

[proper shale]

what about d total

and decrease smoothing to 0.987

[vapid gust]

@proper shale

[proper shale]

hmmm, still going

[vapid gust]

ye

[proper shale]

but kinda stagnating

[vapid gust]

153 epochs

[proper shale]

I'd say it's gonna overtrain soon

[vapid gust]

its a 20 min dataset

💀

[proper shale]

Yeah still

[vapid gust]

and its studio ripped recordings

too

so its not like its not hq

[proper shale]

...doesn't change anything

[vapid gust]

really?

[proper shale]

it can and will overtrain at any time

[proper shale]

i mean... it kinda does

but also doesn't stop overtraining

sooner or later it'll happen

[vapid gust]

thats weird tho

because i thought

longer datasets

go for longer epochs

[proper shale]

^

and that 48k sample rate makes me worry a bit about quality

[vapid gust]

codename said

its fine

[proper shale]

might be a tad bit too noisy

[vapid gust]

to go with 48k

cuz i showed him a snippet of the

dataset

can i show the dataset

in dms?

@proper shale

[proper shale]

nah no need

if codename said it I'll believe him

I go for 32k always because it's safer

[vapid gust]

wait asec

[chilly vessel]

i tried to install rvc and it gav me this prompt

[rare goblet]

Ayo? @chilly vessel level 1 !!!

[chilly vessel]

C:\RVC-beta0717>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897
'runtime\python.exe' is not recognized as an internal or external command,
operable program or batch file.

Does anyone know what to do

[vapid gust]

@proper shale

[proper shale]

last part is probably not true

[vapid gust]

ye

[proper shale]

but yeah fair enough

[vapid gust]

can u check ur dms?

[marble forge]

when trying to process data for the model im trying to make i recieve around 50 of this exact error
During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\trainset_preprocess_pipeline_print.py", line 84, in pipeline
audio = load_audio(path, self.sr, DoFormant, Quefrency, Timbre)
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\my_utils.py", line 103, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: ffmpeg error (see stderr output for detail)

[proper shale]

how's the dataset folder look like

[vapid gust]

blud 💀

[rare goblet]

Ayo? @marble forge level 3 !!!

[proper shale]

what file thpe

type

[marble forge]

mp3

those

[vapid gust]

why in world

are u training

sound effects

[marble forge]

because why not

[vapid gust]

literally no gain from it

[trim ice]

Question, why does my RVC model sound so 'scratchy'? I used 400 epoches, RVCv2, 48k, rmvpe, pitch extraction, 30+ minutes of high quality training data without background noise or effects

[marble forge]

how do i fix it?

[trim ice]

Same training data through elevenlabs TTS, for reference:

[proper shale]

that's probably the audio you're using

[proper shale]

export to one wav with no spaces

pls

[marble forge]

ok

[proper shale]

• the pitch extraction you're using for inference

[trim ice]

As in, the audio I'm inferring, or the training data?

[rare goblet]

Ayo? @trim ice level 2 !!!

[proper shale]

the audio you're inferring yes

[trim ice]

... okay yeah maybe the eggman thing wasn't the best thing to test it with

[vapid gust]

that sounds pretty good loo

[trim ice]

it was just the first thing i thought of tho

[proper shale]

use elevenlabs to generate sample audio

[trim ice]

sadly elevenlabs can't do voice changing unless I were to hook it up with Whisper or something

[proper shale]

also 48k is kinda noisy

maybe try 32k?

[trim ice]

ah i thought RVCv2 (which only has 48k and 42k?) would be better

[vapid gust]

i think its the infering audio

thats ass

[trim ice]

(warning: quite vulgar but i'm pretty sure most people know what this is)

[vapid gust]

ye ur gonna need to mix it

[proper shale]

rvc v2 has 32k, 40k and 48

[vapid gust]

and hes screaming too

mic thumps

compression needed

eq needed

[trim ice]

oh i don't see that option, i'm using the original RVC ui

[proper shale]

use a newer rvc fork then

oorrrr just select v2 in the version setting

if it has rvmpe it definitely has v2

[trim ice]

applio?

[proper shale]

good luck trying to install that mess of a fork

mangio is better

[trim ice]

ah

[modern oxide]

rvc1006 is the newest currently

[tough kestrel]

yo so basically my client isn't working, as my mic isn't imputting anything but my mic does work

[trim ice]

do i need to mix the dataset audio too? it's ripped straight from the source game - no background noise or filter effects

[tough kestrel]

i also checked settings and it was allowed

for the app

so idk whats wrong

[proper shale]

check cmd

[trim ice]

I believe that's what I was using, I downloaded the pre-setup runtime

[vapid gust]

if ur dataset is studio recordings

then no need

[tough kestrel]

so what am i supposed to see

[vapid gust]

but if its uvr ripped

ur gonna need to mix it

[trim ice]

uvr?

[vapid gust]

UVR5

[trim ice]

ah

[vapid gust]

can i get a snippet

of the audio?

[trim ice]

i dumped it straight from the game

sure

[proper shale]

What is it saying

[tough kestrel]

also, nothing is recording,

[tough kestrel]

basically nothing

[rare goblet]

Ayo? @tough kestrel level 1 !!!

[proper shale]

send a screenshot just in case

[tough kestrel]

or did i cmd the wrong thing

[proper shale]

uh

the start_http thing

[tough kestrel]

ah

k

thx

[proper shale]

np

[vapid gust]

not bad

[proper shale]

looks fine

[marble forge]

error is still happening:

[trim ice]

Ah, did you mix it or just convert it to mp3?

[vapid gust]

mix

[tough kestrel]

hmm that sweird

[vapid gust]

u can hear it

cut out the low end

boosted the highs a lil bit

sounds a bit better now

[tough kestrel]

@proper shale

[trim ice]

Is there uh, any automated way to do that? Because I have 508 of these samples

[vapid gust]

honestly

if u can put all those 508 together into one single track

and mix it all at once

[proper shale]

gimme the entire thing

[vapid gust]

na

[trim ice]

(and just to clarify, there's not any other voices mixed into the 508, I used a pyannote.audio script to filter by voice and then manually listened thru them to ensure no bad ones made it thru)

[proper shale]

related to your network configs

[trim ice]

ah, okey! how uh, did you mix it? I assume Audacity?

[vapid gust]

fl studio

[trim ice]

ah

[proper shale]

wrong dataset path was inserted, most likely

[vapid gust]

the low end of ur dataset sounds real bad lmao

[tough kestrel]

so how do i fix that?

thx btw

[trim ice]

I don't have FL studio but my gf is quite skilled with it haha, I think I'll go ask her to handle that side of things

[vapid gust]

idk how she gonna mix 508 tracks

[proper shale]

That's the thing... I'm not sure

[vapid gust]

thats gnna take like a good day

[marble forge]

no it was not its something about ffmpeg

[tough kestrel]

hmmm

[trim ice]

mush them all into one big track?

[vapid gust]

ur gonna have to find some software to do it

cuz i dont think there is

[trim ice]

luckily I am a skilled programmer

[proper shale]

... that's exactly why its failing

[vapid gust]

nice

[sand iris]

spaces in filename

[proper shale]

wrong path

[sand iris]

it's spaces

[proper shale]

And that too

[trim ice]

I used a custom python script for the initial voice clip filtering, in fact

[marble forge]

yes however it fixed that

[tough kestrel]

also, do these files work on other software

[vapid gust]

if u know how to then truncate the dataset

as well

[proper shale]

what files?

[tough kestrel]

like the ones in #1159289700490694666

[tough kestrel]

the ones that you get there

[trim ice]

i should post the script on my github tbh, some1 else may find it useful

[trim ice]

clarify?

[proper shale]

yeah they work on rvc forks

[vapid gust]

cutting out silence ina audio clip

[smoky swallow]

Hi guys! I have a problem with RVC. "Running with the runtime Python, Please wait.".... I'm waiting 30 minutes and it didnt work. How can I fix that?

[vapid gust]

so no pauses in between talking etc

[tough kestrel]

alr thx

[rare goblet]

Ayo? @tough kestrel level 2 !!!

[trim ice]

gotcha

[obtuse stream]

Hop Length (lower hop lengths take more time to infer but are more pitch accurate): why dint use this in rnvpe?

[marble forge]

fixed it i think

[proper shale]

rvmpe doesn't use this

mangio does

[obtuse stream]

Why?

What will happening when i try use

[marble forge]

also what is the correct thing to look for inside of the tensorboard again

[proper shale]

It wont change anything

[marble forge]

like loss something

i forgot

[vapid gust]

loss/g/total

[proper shale]

G and D total

[marble forge]

ok ty

[vapid gust]

@proper shale

for d total

what am i examining for in the graph

[proper shale]

going down = good
stagnating whilst g/total is going up = warning

[proper shale]

yea its doin it

[vapid gust]

doing wat

[proper shale]

its still going

[vapid gust]

down

[proper shale]

if starts to plateau while g total is is going up

warning sign

possibly overtraining

[vapid gust]

i mean already if g total

starts going up

thats a sign already

[proper shale]

not really

i mean

it varies

[vapid gust]

going up without

any

down

movement

[proper shale]

it can go up slightly

and go down afterwards

so when it goes up, check d total

[obtuse stream]

Whats this?

[vapid gust]

its still got alot of time left

[proper shale]

if d total is going down ur alright

[vapid gust]

but im going to run out of gpu time soon 😭😭

[proper shale]

id say

stop training rn

[vapid gust]

naaaaa

[proper shale]

and export model

[vapid gust]

i want it to be accurate

af

[proper shale]

so u dont lose the model

u need the d/g files

[vapid gust]

wait

i pressed these settings

[proper shale]

it only saves stuff to weights and not ur drive

please

please

stop training rn and export

[vapid gust]

what

listen

listen

[proper shale]

so u dont lose anything

[vapid gust]

last time

i trained this dataset

it went to 326

epochs

then gpu time ran out

so i lost that model

[limber sinew]

happens every time I try to train.. whats going on here?

FileNotFoundError: [WinError 3] The system cannot find the path specified: 'C:\RVC\RVC-beta0717/logs/mute/0_gt_wavs/mute40k.wav'

[vapid gust]

but im at 240 epochs rn

[proper shale]

u need that file to train basically

[mellow moss]

@proper shale is it a good moment to DM you regarding the commissions?

[limber sinew]

oookaaaayy??? sooo.. what's going on here? I never deleted anything

[proper shale]

sure, keep in mind im kinda busy and might not be able to finish commissions for a little while

[proper shale]

perhaps preprocessing screwed up? idk

[mellow moss]

it's no worries, currently want to get more info about it

[proper shale]

okok

[vapid gust]

@proper shale

im gonna save

it

just in case

how do i do it so i can load it next time

[obtuse stream]

I used applio how can i check the train? I mean the graph

[limber sinew]

I mean, I tried to re-train 3 times. I get the same result

[proper shale]

export model, then when u come back u gotta insert the same Batchsize

and the same training variables

u can change save freq and total epochs to train

also only use the for resuming steps

[vapid gust]

wdym?

[proper shale]

there are steps with (for resuming) in rvc disconnected

i recommend using only those

[limber sinew]

I might have to download a new version or something

[vapid gust]

idk where is it

[limber sinew]

imma try that

[proper shale]

no need to preprocess or feature extract

[obtuse stream]

I used applio how can i check the train? I mean the graph

[proper shale]

again

[proper shale]

those

[vapid gust]

alright

what do i do here

[proper shale]

just run the step

itll export ur g and d stuff to ur drive

[vapid gust]

wow

l

lmao

[proper shale]

• latest pth in weights so u can test

[vapid gust]

what about this

[half cove]

You can do that whenever u want

[vapid gust]

should i do it rn?

[half cove]

If u want

[vapid gust]

wheres the index file?

or do i even need it

[half cove]

Train before u export

[obtuse stream]

I used applio how can i check the train? I mean the graph

[vapid gust]

why dont i have it

wtf?

[half cove]

Did u train index?

[vapid gust]

oh nvm

💀

[vapid gust]

now

now

which ones do i use

to infer

trained_IVF1920.index + s9750.pth?

[obtuse stream]

I used applio how can i check the train? I mean the graph

[limber sinew]

duuude.. I am hating this so much

2023-11-11 15:58:14 | INFO | main | "runtime\python.exe" infer/modules/train/preprocess.py ""C:\Users\raynb\OneDrive\Desktop\vali voice trainer"" 40000 8 "C:\RVC\RVC1006Nvidia/logs/Avali" False 3.0
['infer/modules/train/preprocess.py', 'C:\Users\raynb\OneDrive\Desktop\vali', 'voice', 'trainer', '40000', '8', 'C:\RVC\RVC1006Nvidia/logs/Avali', 'False', '3.0']
Traceback (most recent call last):
File "C:\RVC\RVC1006Nvidia\infer\modules\train\preprocess.py", line 11, in <module>
sr = int(sys.argv[2])
ValueError: invalid literal for int() with base 10: 'voice'
2023-11-11 15:58:15 | INFO | main |

[rare goblet]

Ayo? @limber sinew level 1 !!!

[limber sinew]

installed a new version, and now I'm getting a different error when I try to just process the data

[obtuse stream]

I used applio how can i check the train? I mean the graph

anyone??

[river egret]

I've been looking around for an explanation on how RVC works but I keep finding tutorials on how to use it.

What should I look for? A thesis? YouTube video?

[proper shale]

Try looking for their GitHub page

But I don't think they have a research paper of some sorts

[sand iris]

space in filename

"vali voice trainer" is causing issues

[limber sinew]

strange, didn't have that issue before

must be a new thing or something

[sand iris]

it has always been an issue

¯_(ツ)_/¯

Messes up with the way it's reading inputs

[limber sinew]

huh

alright then

thanks

[molten pecan]

-rvc

[azure marsh]

Hey, @river egret!

👇 Here are some useful resources to help you learn how to make ai covers

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

[half cove]

What did you try running?

[violet heron]

Did you wait for the other cells to finish

[silver anvil]

@molten pecan

[half cove]

They were joking dw

[vapid gust]

i just pressed run dependiesc

im trying to resume a model

[silent prism]

it says it cant unzip

need help

[violet heron]

What and where?

[copper pebble]

rghhh

resuming is a pain in the ass

that works like

2/10 times

[vapid gust]

nah because

my model trains till to like 326 epochs before gpu runs out

and its got way more potential

[olive olive]

why my model doesn't have .pth file ? i only have index

[river egret]

Thank you for giving me this but it seems that this resource is about making AI Covers, which I can already do.
I'm looking for an explanation on what is happening when "extracting features" or "training a model".
Already familiar with general AI training: Train models, evaluate which is best, carry those forward and retrain.

[olive olive]

use 7-zip

[silent prism]

when trying to download the model into rvc

[violet heron]

Which RVC?

[silent prism]

Aesthetic_RVC_Inference_H

[molten pecan]

Oh sorry I have no idea about making models and stuff

Try asking on #📑│making-models

[obtuse stream]

I used applio how can i check the train? I mean the graph

anyone??

[silent prism]

tryna download this

https://app.kits.ai/convert/shared/sidhu-moose-wala-jo1KkJr24R

i cant

it says download successful

but

cant unzip

@violet heron

[thin grotto]

Hello people I'm gonnas ask something dumb but when i wanna import a model it says that it use RMVPE as pitch extraction which voice changer type shoul I use when Importing ?

[violet heron]

The problem is that’s a kits model

[silent prism]

i found it in search models

[violet heron]

You can’t download kits models

[violet heron]

Stupid applio bot

[silent prism]

how can i find non kit model?

[violet heron]

The applio bot searches for kits models for some reason

Use the @shy zealot bot instead

Using /search

[silent prism]

yea nah, its not showing up

[vapid gust]

does anyone have the code

to fix this

?

[silent prism]

ohh my lawwrd

its showing up as error

when converting

can you show me a new rvc?

@violet heron

[violet heron]

You can’t use kits models on RVC??????

you can’t download kits models period

Your better off making your own model of that character or using a different one

[silent prism]

no

[rare goblet]

Ayo? @silent prism level 6 !!!

[silent prism]

its different

im using this

https://huggingface.co/akiyama1996/drake_clb/resolve/main/drake_clb.zip

and its not working

@violet heron

[violet heron]

On which RVC?

[silent prism]

applio

[wary jolt]

Guys who may recommend me applio for training vocal IA?

[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

[nimble zealot]

Does anyone use the aicovergen without web ui?

[still crystal]

how do i fix this i ran the feature extraction

[tender current]

Help

It gives me this message every time I try to run the downloader

ModuleNotFoundError Traceback (most recent call last)
<ipython-input-1-5883eeb53f11> in <cell line: 4>()
2 #@markdown Link the URL path to the model (Mega, Drive, etc.) and start the code
3
----> 4 from mega import Mega
5 import os
6 import shutil

ModuleNotFoundError: No module named 'mega'

I am new and don't really know what I'm doing

[brisk hornet]

why is this happening

[half cove]

you cant use weights download links

u need to upload to huggingface

[brisk hornet]

they used to work fine tho

what happened

[half cove]

they dont use huggingface links now

they use their own

and you need auth for it

so you cant use dowloaders

[brisk hornet]

pain

[half cove]

real

[brisk hornet]

Man...

this is a sad day for me

[tough kestrel]

this is kinda weird but pass through works but when its off it doesn't

like sounds go through with the lag that the voice changer has

with pass through

[worthy grove]

The dataset is a bunch of wavs

Anyone knows whats up?

[opaque needle]

Guys

Somewone Help Me Pls

[modern oxide]

can you send the screenshot of the ui

[olive olive]

I trained my model on rvc but the .pth don't show i only have the index what do i do

[simple viper]

Soooo for some reason RVC aint loading up just say wait web server

[modern oxide]

it's in weights folder which sometimes can be found in assets folder

[olive olive]

i check everytime the "weights" folder but it never show where is the assets folder ?

[modern oxide]

do you train locally?

[olive olive]

Yeah

[rare goblet]

Ayo? @olive olive level 2 !!!

[olive olive]

on grabio

[modern oxide]

what version/fork do you use?

[olive olive]

RVC-Beta

i downloaded the 7.zip file

on huggyface

[modern oxide]

https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006Nvidia.7z?download=true i recommend to install this version

assets should be inside the rvc folder

[olive olive]

Thanks i try now

[proper shale]

TensorBoard Installation Files

Use these to install TensorBoard and to launch it! Just drag and drop both in your RVC folder, then use TBI to install. After that, just use tbL to launch TensorBoard whenever you need it.

[proper shale]

I dont know how... :(

But hey at least this works

[wintry mulch]

Hey where do i download or use rvc?

Alright

also what is the different between train and infer?

[rare goblet]

Ayo? @wintry mulch level 1 !!!

[simple viper]

I still need help, my rvc crashed and now it wont load back at all. Just a white screen and its tryna to connect to web server

[rare goblet]

Ayo? @karmic oxide level 56 !!!

[wintry mulch]

ok

thank you

[simple viper]

ok

[rare goblet]

Ayo? @simple viper level 1 !!!

[simple viper]

It worked!!!!

Thanks sorry for using wrong channel

LAST QUESTION is there anything I need to fix after doing that or not?

[brittle wing]

Hey so.. I know how to seperate vocals and instrumentals but how do I seperate lead vocals from background vocals?

[olive olive]

can i change the type of extraction like "harvest" instead of "rmvpe_gpu" or it's only rmvpe_gpu because i tried on harvest and on asset then wieght .pth doesn't show

Oh ok

but why do the .pth file don't show ?

i checked but he is not inside i only have .gitignore and some preloaded .pth file

[brittle wing]

Hello May I ask if is there a new notebook for creating voice covers

[olive olive]

but i don't understand why because i don't make any mistake

[distant belfry]

how to reduce delay to near real time?

[olive olive]

then i click on what ? because i do "one click training" then "index"

ok i try

I still see no .pth in weights

no there is no error and i only have preloaded .pth file in assets and weight

save frequence is set to 50 and total is set to 300

[rare goblet]

Ayo? @olive olive level 3 !!!

[olive olive]

i don't see any epoch

produced

[empty dust]

yo wtf, i just wanted to change screenshot

Ok, one more time.
I'm trying to use model processing, but I only get this error. I fixed it before by processing a different model, but that seems to have stopped working

[upbeat fiber]

Can you download models from kits.ai?

[torn bay]

One message removed from a suspended account.

[rare goblet]

Ayo? @torn bay level 2 !!!

[torn bay]

One message removed from a suspended account.

One message removed from a suspended account.

One message removed from a suspended account.

One message removed from a suspended account.

One message removed from a suspended account.

One message removed from a suspended account.

One message removed from a suspended account.

[flint cloud]

old but thanks!

[rare goblet]

Ayo? @flint cloud level 1 !!!

[static gyro]

is the RVC disconnected RMVPE working?

[charred jewel]

What's wrong?

[rare goblet]

Ayo? @vapid gust level 12 !!!

[charred jewel]

I can't figure out how to get it to work, bc I trained it a little last night, tried to start it again, then it just gave me a bunch of errors.

[vapid gust]

honestly

resuming models on local is hard

to do

[charred jewel]

But it was only 100 epochs...

[vapid gust]

doesnt really matter

[charred jewel]

and I've resumed models that were 1,000 epochs

[charred jewel]

Ok

[vapid gust]

ig then its just a

glitch

[charred jewel]

I'm just gonna restart RVC

[vapid gust]

ye do that'

[charred jewel]

Alr I did, now I'm gonna try again.

[rare goblet]

Ayo? @charred jewel level 7 !!!

[charred jewel]

Is that good or bad?

[vapid gust]

idk

im training a model rn

[charred jewel]

ok

[vapid gust]

@charred jewel

do u have a good pc?

[charred jewel]

Yeah

[vapid gust]

ah

[charred jewel]

3060Ti, 32GBs RAM

Core i5

13600k

[vapid gust]

oh damnn

i got a i5

[charred jewel]

Nice

[vapid gust]

thats literally it

its a shitty laptop

[charred jewel]

Oh, then how are you training a model?

💀

ffmpeg error?

[vapid gust]

colab

takes quite long time

but

works

[static gyro]

RVC disconeccted seems to be working

[iron swan]

where is pth files in mangio

[charred jewel]

Ah

[charred jewel]

weights folder

[little owl]

how would i resume training on applio locally?

[rare goblet]

Ayo? @little owl level 1 !!!

[void yoke]

How to make songs out of the AI vocals?

[glad zealot]

covergen

[iron swan]

its not there :c

i can only find index but not pth

[solar dove]

can I use these models to tts?

[harsh ravine]

hey

im trying to make producer tag with a voice model but everytime i try to upload it or past model url it says error

[hardy kiln]

so how do i use the custon voices

like i have the "briar" one

and it doesnt work, all the default ones do tho

[empty dust]

How did you realise the model wasn't working? Do you have monitoring turned on? This is the third item under the microphone and output

[hardy kiln]

wat

yea

like i can hear myself with thhe other ones

but i cant with the custom ones

[empty dust]

Maybe the model itself isn't working? Send it here, I'll check

And try another custom model

[hardy kiln]

i have

ive done the markiplier one, and then 2 other

nonje work, but then wiith the same settings i try the default ones

and it works

the files are uploading

[empty dust]

Maybe... I don't know, are you sure you're adding them correctly?

[hardy kiln]

i think so?

[raw bluff]

bro how to I find Cristiano Ronaldo voice

[rare goblet]

Ayo? @raw bluff level 1 !!!

[raw bluff]

pls help me

[hardy kiln]

im doing the upload button

then chosing the pth and the index

[empty dust]

/search in the #🔍│find-models

Or weights.gg

[raw bluff]

I write but it doesn't come out

@empty dust

[hardy kiln]

maybe there is none?

[rare goblet]

Ayo? @hardy kiln level 2 !!!

[hardy kiln]

[empty dust]

Yep, I think so

[empty dust]

So... Hm...

[hardy kiln]

so does it work for u? if uve downloaded the files

[empty dust]

Yeah, it's working, just checked

[hardy kiln]

is that right settings or is something wrong? u recon

[empty dust]

Well, aside from the GPU, it's fine.

[hardy kiln]

LOOOL

[raw bluff]

Bro this does'nt download the ronaldo voice dude i want to download the sound rvc gui

[hardy kiln]

videocard wouldnt be problem cos like the other ones work

[empty dust]

Yeah, I know. That's why I said besides GPU. It's just the only thing that caught my eye

[hardy kiln]

i have 0 clue aye

[empty dust]

It's just that you even have the delay and response time displayed. Maybe try recording something in audacity or somewhere else?

I just noticed you don't have virtual cables installed

[hardy kiln]

i do

[hardy kiln]

ik ik

but the delay is same for all models

and the other ones work fine

its just not the custom ones

[empty dust]

Then don't forget to make the output with your virtual cable

[hardy kiln]

yea i have

yea no audio comes out at all its weird

just with custom voices tho

[empty dust]

All right, let me think about it

[hardy kiln]

do i have to like export to onnx?

or something

[empty dust]

Nope

[empty dust]

Try this for now

Colab Version

https://colab.research.google.com/github/w-okada/voice-changer/blob/master/Realtime_Voice_Changer_on_Colab.ipynb

[hardy kiln]

gpu version wouldnt matter no? cos all voices working but custom?

or am i wrong

[empty dust]

Computers and programmes should always work the same way, but they don't. So it's worth trying everything

I didn't think at all that anyone actually helps out in these chats on a regular basis

[hardy kiln]

im on rvc

im pre sure

[empty dust]

I think they understand more about it than I do, so yeah, best to go to that chat then

[hardy kiln]

ah

so any ideas? why the custom voices arent working or

[empty dust]

Because w-okada is the developer's nick on github, not the name of the repository?

[hardy kiln]

yea, ik my graphics card isnt best

[rare goblet]

Ayo? @hardy kiln level 3 !!!

[hardy kiln]

but

would that really effect it? as the normal voices are working just fine

[empty dust]

I call it RVC Voice Changer

Model processing

[hardy kiln]

wats that

[empty dust]

Oh, this one. I'd already forgotten about that damned piece of code

[hardy kiln]

is there any other possible reason other than graphics card?

@karmic oxide

[empty dust]

It's not so important to you

[hardy kiln]

im on 840 chunks and its 1000 delay

1000 isnt bad tho? is it

but wat im trying to figure out is why the custom voices dont work

grrr

[empty dust]

YO YEAH? Fucking hell, I should move it to an SSD then

yo wtf

I didn't think it was reading model data all the time, just loading it into RAM or VRAM

[hardy kiln]

yea i still get 0 audio

with the custom voices

im so confused

[empty dust]

Well, that's logical.

[hardy kiln]

kk

[empty dust]

Didn't know that either, deleted them as soon as I installed the app

[rare goblet]

Ayo? @empty dust level 6 !!!

[empty dust]

It's time to expand my knowledge

https://tenor.com/view/sam-lake-cheers-bottoms-up-coffee-tea-gif-7297956

omfg

EVEN HERE

I just want to use goofy ass gifs

[hardy kiln]

even normal voices arent working after i converted her to onxx

im so confused LOL

brb

gonna go install it on my brothers computer

4080 ti

[empty dust]

I know only one rule

https://tenor.com/view/sigma-rule-744-gif-25127400

yeah

Anyway, thanks

[little owl]

can someone help me with this? i’m trying to train locally using applio

[hardy kiln]

@karmic oxide the res is no 0 ms?

and isnt workin

wait nvm

normal voices work

yo

it works

tysm @karmic oxide

[jovial ridge]

hii guys

i need help with something

[azure marsh]

How to troubleshoot

• Give context: Throwing your error/issue alone will not help. Tell what step are you on, what were you doing before issue, your settings, your terminal, et cetera.
• Be relevant: You come to ask for help, not to complain. You can't ask for RVC help in voice changer help and vice versa.
• Be polite: DONT DEMAND FOR HELP OR I WILL FUCKING MURDER YOU. Ask politely, the server was not made for you.
• Don't be an attention whore: If people are ignoring your help, kindly ping them. If they are ignoring you, either they don't know how to solve it, they are busy, or you are an asshole.

WITH ALL DUE RESPECT, DO NOT BE AN ASSHOLE AND HAVE COMMON SENSE.

[empty dust]

Do something and it works

Don't thank me

[empty dust]

I like this guide

Can I use it?

-howtoask

[azure marsh]

How to troubleshoot

• Give context: Throwing your error/issue alone will not help. Tell what step are you on, what were you doing before issue, your settings, your terminal, et cetera.
• Be relevant: You come to ask for help, not to complain. You can't ask for RVC help in voice changer help and vice versa.
• Be polite: DONT DEMAND FOR HELP OR I WILL FUCKING MURDER YOU. Ask politely, the server was not made for you.
• Don't be an attention whore: If people are ignoring your help, kindly ping them. If they are ignoring you, either they don't know how to solve it, they are busy, or you are an asshole.

WITH ALL DUE RESPECT, DO NOT BE AN ASSHOLE AND HAVE COMMON SENSE.

[empty dust]

yeah

nice

[hardy kiln]

wait some work but others dont im now confused ahah

i converted them to onxx too

wait none are workin wat the flipp

huh

such a long download 😭😭😭😭

wat

ah

how do i get colab

i see

yea even the fucken default models

arent working

[empty dust]

Or steal your brother's PC

Also a good option

[hardy kiln]

facts

wait

how do i use colab version

?

[glad zealot]

which colab are you using

the link

[hardy kiln]

the one u sent

i got it i think

[glad zealot]

on the pins right?

[hardy kiln]

i got it

[rare goblet]

Ayo? @hardy kiln level 4 !!!

[hardy kiln]

the response time is so high tho

20000 ms

[glad zealot]

#🔍│help-w-okada

[loud berry]

guys I've made tutorial video

https://youtu.be/o2NbkP_YKxU?si=cOgSjZ87RXdeDN50

[astral jungle]

Hi

[empty dust]

Yo

[astral jungle]

Yo bro

[hearty jasper]

what makes ai sounds better? more filesize or more epoch?

[astral jungle]

I think it's epoch

[sand iris]

I mean

More filesize as in more length or higher quality?

[jolly crow]

Hello everyone, I wanted to make a model in the colab based on the guide from the docks on the gitbook from the creators of this server, the model was not up to training and time ran out. I started to continue training and after some time the errors started to come out (https://cdn.discordapp.com/attachments/965625051347034165/1173200504021594182/image.png?ex=6563171d&is=6550a21d&hm=21e5eadc46213c3e0791b048d4e888540bd42162897ca776ceee1f0d5ea5a99f& ) and the training has not started, can you help? Well, or tell me a normal way to train in some kind of cloud...

[rare goblet]

Ayo? @jolly crow level 1 !!!

[copper pebble]

i think its having some problems

[static garden]

its giving me this any idea?

Failed to execute cell. Could not send execute message to runtime: Error: await connected: disconnected

anybody?

[copper pebble]

what r u using

@static garden

[static garden]

google collab webpage thingy

[copper pebble]

yh which one? can u link me that?

[static garden]

https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=F6zbsuNs6xZF

[copper pebble]

ty

[copper pebble]

thats pretty old

i would link u the new one but its having problems

[static garden]

let me see that

[copper pebble]

https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS#scrollTo=Yj-_npuA_HRB try this one

dunno if it works well cuz as i said before

its having problems

[static garden]

so can i run all of this on default...settings are pretty overwhelming

[rare goblet]

Ayo? @static garden level 1 !!!

[copper pebble]

https://discord.com/channels/1159260121998827560/1160926856548200531

follow this

eventho their saying that RMVPE is actually broken

[fathom coral]

what is the best method for interference?

[tranquil cliff]

It seems that RMVPE pitch extraction on the "RVC Disconnected" colab (for training) works poorly; as an alternative use Mangio-Crepe or wait for the problem to be fixed.

Note: This message will be deleted once the issue is resolved

[tranquil cliff]

@copper pebble lo attacco?

[copper pebble]

o metti (yall use mangio-crepe) oppure (please use mangio-crepe)

[tranquil cliff]

Scusa ma il mio inglese non è da 10