✨│ai-help

[proper shale]

This really helps

[lucid crow]

But I think it's not on the cmd

Cause the app runs

Smoothly

[flint vapor]

Maybe there is something on cmd, if you send we will see if all is ok 😄

[proper shale]

Cmd tells when there's stuff wrong, so it'd be nice if you sent it

[eager pelican]

huh for what

noo its misinformation

almost everything is wrong stuff like batch size and the g and d and shit

[glad zealot]

fix it :D

cause idk about those

[eager pelican]

i lsot the code to it :D

i need to contact rentry or smth on removing it smh

[glad zealot]

NOOO :D

[eager pelican]

YES :D

[glad zealot]

make a new one ig :D

[eager pelican]

no misinformation on AI hub not like the server is already the center of it

[eager pelican]

too lazy

[glad zealot]

nyooo

[prisma fern]

Exception: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from http://www.nvidia.com/Download/index.aspx

smh

i have updated the latest driver version

[eager pelican]

also no the images are not dead :D
smh this is why i keep saying google docs is better
https://docs.google.com/document/u/0/d/1VQRwU8NiFCKh3aOOEPM7wwhG1zAhsuNy/edit?usp=docs_home&ths=true&rtpof=true

[prisma fern]

Exception: Found no NVIDIA driver on your system.

can anybody help

[flint vapor]

Are you using the right runtime on colab?

[brittle wing]

Applio rvc colab "no cached install found" error. Whats that?

[forest sluice]

Local or Colab? If you're using local, what's your GPU?

[brittle wing]

what are the minimum requirements for rvc?

[flint vapor]

Depends, the most important thing is the gpu, but you should consider also cpu, ram

[flint vapor]

gpu with a good vram also

But if you don't have requirements you can use also colab

[brittle wing]

i mean, i have an amd radeon hd 8690m, 2 gb vram, 12 gb ram and an intel i7-4610m

[flint vapor]

Unfortunately is not good for local version

[brittle wing]

if that helps

[flint vapor]

You can use colab

[brittle wing]

not even for inferencing?

[proper shale]

overfitting

[eager pelican]

that guide is wrong asf 🗿 it doesnt explain overfitting properly and its gotten many mistakes

[flint vapor]

Amd have some problems with local rvc, idk if inferencing will work finely

[proud elbow]

local inference with cpu is possible but slow, even your old cpu would be much slower

[flint vapor]

I recommend to use colab for inference

like ilaria rvc

if you don't have requirements

[brittle wing]

is huggingface a better alternative to google colab?

[flint vapor]

Huggingface spaces can have some problems

colab is more active

[brittle wing]

i don't want to get disconnected halfway

[flint vapor]

With inference is difficult you get disconnected

you should stay a lot for get disconnected

[brittle wing]

plus, idk, making the dataset, letting colab run it all, making sure it doesn't close midway thru the training

i just want that out of the way

[proud elbow]

the HF spaces can only use cpu, but still slower than average modern cpus

[brittle wing]

so, idk, kinda want models for free and such

[flint vapor]

If you use only for inference i recommend ilaria rvc

If you use for training i can recommend RVC Disconnected or paperspace but on paperspace you need to pay

[flint vapor]

You can make the dataset without colab

[brittle wing]

yea, but idk, might be noisy and stuff

[rare goblet]

Ayo? @brittle wing level 2 !!!

[flint vapor]

Only when you need uvr

[brittle wing]

i use rip n mix

cuz idk how to get uvr to run on my gpu

and again, same issues like sd and rvc: not enough vram

[proud elbow]

UVR can run in cpu mode but slower

[wispy lava]

is it gonna make much diffrence in speed if i increase batch size in rvc training?

[flint vapor]

If you use all these services everyday i recommend you to use a trial with paperspace or colab

But you can run uvr also on cpu

[flint vapor]

Unfortunately not

[brittle wing]

idk, this whole ai convenience is kind of an elitist thing to me

[proud elbow]

minimum recommended gpu is 6 gb vram

[brittle wing]

from the outside and such

[wispy lava]

so what diffrence will it make?

[half cove]

usually use 8

[wispy lava]

it was at 4 by default, i increased it to 6 cuz idk if my gpu can handle it easily

thought it might even slow down the training

[half cove]

then just stick with 6

[proud elbow]

try 8 if u have 8 gb vram

[flint vapor]

Is not for speed but you should use a good batch size because of voice artifacts or problems with final result

[wispy lava]

btw if i got a 10 min audio and a 8gb vram card do yall recommend caching the training ?

[flint vapor]

But is not about speed

[wispy lava]

like that other option

[flint vapor]

You can use 8

of batch size

[wispy lava]

is it gonna speed it up at all?

[flint vapor]

Is not changing so much i think is always same

but 8 is like default option

for not risk some problems with the final result

[half cove]

it does affect training time

[rare goblet]

Ayo? @half cove level 147 !!!

[wispy lava]

4 was default

[flint vapor]

also 8

Is for not exaggerate so much

because can affect the quality of model

[wispy lava]

ahhh alr

[molten pecan]

https://aihubdocs.github.io/en/rvc/local/mainline/ here's a usage guide

[wispy lava]

btw how long would you say it could take for each epoch with a 8gb card?

[flint vapor]

if your dataset is 10 minutes i think 20 seconds per epoch

[unique nebula]

Hi people, I would like to know how to get a model of our streamer?

[flint vapor]

according to my experience

[wispy lava]

yeah alr thats exactly the speed it is going 💀

[flint vapor]

How many epochs you set?

[wispy lava]

100

i think 50 is enough

[proud elbow]

500 for default pretrain or 300 for Ov2 regardless of dataset length & batch size

[wispy lava]

idk

tho

[flint vapor]

Oh ok ok, also 100 is ok if you not using ov2

[wispy lava]

tf is ov2

[proud elbow]

new pretrain

[flint vapor]

Oh is a pretrain

[wispy lava]

if its not the default option im nto using it then lol

[flint vapor]

No is not default, ov2 should be downloaded externally

[wispy lava]

oh lol

what does it do exactly?

pretrain?? what does that mean

[proud elbow]

ov2 is also available in rvc disconnected colab

[wispy lava]

?

[flint vapor]

It can improve some audio quality

[odd shale]

pretrain is basically a pack of various voices trained so you can train a model.

[wispy lava]

Ooooh! i see

so basicly with pretrain its gonna need less epoch?

for the same quality?

[flint vapor]

Generally with ov2 pretrain if you have a short dataset you can use more epochs and if you have long dataset few epochs

and can improve a bit the voice quality

[odd shale]

Nah, it's far more recommendable to use short datasets with OV2

[flint vapor]

If i'm not wrong

[flint vapor]

Yes true

[odd shale]

If you have a normal sized dataset, then go for RIN or regular pretrains.

[wispy lava]

btw, how many epoch do you recommend for 10 minute and 20 minutes of audio?

[flint vapor]

With 10 minutes you can use default pretrain or rin pretrain

[odd shale]

30 mins of dataset?

[wispy lava]

no like for each

[flint vapor]

Idk if there is a specific epochs you can use

around minutes

[odd shale]

No there's not

[wispy lava]

mhm

[flint vapor]

With 20 minutes of dataset i used 250 but is your choice

[proud elbow]

500 for default pretrain

[flint vapor]

there isn't a specific one

[odd shale]

Epoch count before overtraining will always vary desuwa.

[rare goblet]

Ayo? @brittle wing level 2 !!!

[wispy lava]

bro idek how to use pretrain 💀

[odd shale]

You both let me talk to him.

Sonya, to use an pretrain just select or 32k, 40k or 48k of samplerate.

[flint vapor]

I want to help you but i'm not into colab sorry

[proud elbow]

it actually varies between 300-500 in my cases

[odd shale]

At least for your cases.

But that's not the case in Sonya's case

Sonya, i recommend you to check Tensorboard.

[wispy lava]

oh yeah im also confused about the samplerate aswell and what it does

i chose 48k

[odd shale]

Are you sure your audio samplerate is 48k?

[wispy lava]

ihnfc

[odd shale]

?

[wispy lava]

"i have no fucking clue"

[flint vapor]

You should see before the frequency of dataset, because if not 48k it can create some fake frequencies and make the audio bad

you can analyze with spek

[odd shale]

Then DON'T start training and download Spek

[wispy lava]

Spek?

[odd shale]

Yes.

[flint vapor]

Yes is a spectrogram tool

[odd shale]

Kakao let me handle this

[wispy lava]

bruh i almost finished training ngl

its at 70

30 more to go

[flint vapor]

https://www.spek.cc/p/download

[flint vapor]

Sure 😄

[odd shale]

You haven't even checked tensorboard?

[wispy lava]

Whats tensoreboard 😭?

[odd shale]

It must be a joke.

[flint vapor]

https://aihubdocs.github.io/en/rvc-resources/epochs-overtraining--tensorboard/

[wispy lava]

oh yeah to see the training status?

smth?

[odd shale]

YES

[flint vapor]

Yes, for check if you are having overtraining or other similiar problems

[odd shale]

Also for checking if your model has overtrained or not.

[molten pecan]

thanks for sharing the docs Kakao =3

[wispy lava]

well the RVC i have doesnt have a tensorboard smth builtin idk, i used smth before to train another model it was for tortoise i believe, that one had tensoreboard this one didnt

so yeah idk,

if theres something i should download manually i didnt

[molten pecan]

you can download it. read the docs thoroughly and then ask questions

[odd shale]

I was trying to help him but it seems Sonya didn't know about various RVC slangs.

[flint vapor]

You are welcome! 😄

[odd shale]

Una pregunta Julia.

[wispy lava]

never said i was an expert in ai 😭

[odd shale]

Tu siquiera has entrenado tu primer modelo?

[molten pecan]

claro. ya hice varios

[flint vapor]

@wispy lava I recommend to read all docs after you learn all you can proceed with all the training 😄 i say this for not risk overtraining with your models and low quality problems

[odd shale]

No has publicado alguno?

[wispy lava]

my last braincells are struggling to focus on reading the docs

so yeah i cant do that 💀

[odd shale]

Then how we are supposed to help you then?

[molten pecan]

aún no. lo haré en el futuro cuando se arregle el Colab

el TensorBoard anda roto ultimamente

[odd shale]

Y te han salido bien los modelos?

Sabes como limpiar audio?

[odd shale]

Yo uso RVC en paperspace y el tensorboard me funciona bien.

[wispy lava]

alr imma try to understand it and do what it says

[molten pecan]

obvio Leo 😭

[flint vapor]

I want to help you about this but unfortunately i don't have some competences in colab

[odd shale]

Demuestralo.

[molten pecan]

eh?

[flint vapor]

I can help more about local and general questions about train and inference

[odd shale]

Sabes como eliminar ruido vocal como clicks de boca?

Con RX10

[odd shale]

Caiste en el bait

[molten pecan]

con un plugin de FL Studio supongo

o iZotope

aún no me he topado con ese problema

[odd shale]

Izotope es RX10

?

[wispy lava]

nothing

[tacit badger]

Hi

[odd shale]

Spanish is our native language.

[brittle wing]

who is willing to do some datasets for me for free?

[tacit badger]

Applio says: FileNotFoundError: [Errno 2] No such file or directory: 'D:\\Applio\\assets\\audios\\dobryi_den_output.wav' when I processing audio, I got error

[wispy lava]

Pensé que este era como un canal solo en inglés.

[odd shale]

Sadly no one.

[flint vapor]

You can request on model request but for free is difficult that someone can make them

[brittle wing]

idk

i don't want to make it too noisy

[tacit badger]

it's shows double slashes on directory, but... I set it wrong

[odd shale]

You'll must learn how to clean audio then.

[brittle wing]

idk

[odd shale]

Idk what?

[brittle wing]

if how i clean my audio is clear enough

[odd shale]

I didn't understand

[flint vapor]

You need to listen them very well many times

[tacit badger]

heeyy

you hear me!!

[odd shale]

What?

[tacit badger]

@odd shale
I mean this

idk

[odd shale]

1 - Don't ping staffers unnecessarily when you can simply reply to their messages

[tacit badger]

sorry for pinging, yup

[odd shale]

2 - Why there are double slashes on your file path?

Yo no pregunté eso en tono de que yo no sabía.

Le pregunté a Julia que si sabía hacer eso.

[tacit badger]

okay, but... what should I do?

[flint vapor]

Mangio on my D driver works well

[tacit badger]

hmm, strange

maybe my Python version is uncompatible

[flint vapor]

This is the only error you have?

[tacit badger]

yup

[flint vapor]

You have applio v3?

[tacit badger]

I not use for model training, because I have didn't GPU, Intel Core i5 only CPU

[tacit badger]

v3.0.7

[flint vapor]

is strange because you have also double slashes idk why

[tacit badger]

I too think so...

maybe I using incompatible OS, but idk...

but I using Windows... and Applio running correctly on my machine

because Windows using DOS path, Linux/Mac/BSD Unix path

[proud elbow]

meanwhile you can try original RVC or mangio and see if it should work on yours

[tacit badger]

okay

and one q

why in Applio no RIN_E3 and Ov2 pretrains?

[proud elbow]

find and download it from #📰│dev-updates

[tacit badger]

okay thanks

I downloaded 2 files and put on \train folder?

ahh, pretrained base model on pretrained model on Mangio, thanks

[wispy lava]

yo guys

whats the diffrence between D_1900.pth and G_1900.pth

[proper shale]

both r training files

[gentle sundial]

What do I do with the audio output file path on EasyGUI? It keeps giving error messages.

[brittle wing]

how fast would rvc disconnected be able to train a model?

[rare goblet]

Ayo? @brittle wing level 3 !!!

[proper shale]

it varies on dataset length and batch size

normally, more than 2 hours. I'd stop before an hour or two and export model to train on a later date

[brittle wing]

at around 200-300 epochs

and idk, a dataset of 20 minutes

with a recommended batch size of 2 on colab, according to the notebook

[proper shale]

bruh what

never use batch 2 unless you have a really short dataset

[flint vapor]

2 of batch size is wild

[brittle wing]

"cpu_threads: Specifies the number of CPU threads to be used during the training. Colab only has 2 so don't even bother changing it."

[proper shale]

that is not batch size

[brittle wing]

taken right from the notebook page

[proper shale]

that is cpu threads

and yeah that should not be changed

[brittle wing]

oh, ok

does a batch size of 16 work for colab?

[proper shale]

batch size is for the gpu specifically, its in the training step

[proper shale]

yeah but I'd just recommend using 8

less prone to screw up anyway

it's in the training step

[brittle wing]

i thought that the maximum batch size colab can handle is 20

[proper shale]

just scroll down on it

[proper shale]

yeah

[gentle sundial]

Can anyone help me with this? Why is this issue never covered?

[proper shale]

but you don't need batch 20

never ever

[proper shale]

honestly bro i got no clue...

[brittle wing]

and by how much would the speed be affected if i reduce the batch size from 16 to 8

[proper shale]

scalars tab, search for total and you'll get g/total and d/total

then you can see g/mel and g/kl, search for those individually

[proper shale]

like half speed (?), but way less prone to messing up

[brittle wing]

i hate this

[gentle sundial]

I'm talking about the inference part by the way. I tried to look everywhere and there's no clear answer to fix this.

[proper shale]

yeah this has been talked about a lot

[proper shale]

since ive been out of the community for a bit, idk how to solve

mainly g/total, kl and mel

if g/total goes up, check kl and mel

if those r rising for a while, yeah probably overcooking

glad i could help ;)

[latent kettle]

ohh .. how much batch size for 30 or 30+ do you recommend

[proper shale]

8 or 12 should do, preferably 8

[latent kettle]

and for 1 hour ?

[proper shale]

you don't need an hour of data

😎

[latent kettle]

and tell me how can I convert my 32khz dataset into 48khz ? and do thing will affect on voice model?

[rare goblet]

Ayo? @brittle wing level 2 !!!

[proper shale]

Oh yeah your smoothing's probably at 0,6

go to 0,987

or 0,9 if ya want

[proper shale]

you can't

just use 32khz always

[latent kettle]

will it effect model quality?

[proper shale]

literally, 32khz is best

it wont have as much noise as 48khz, plus some datasets arent that clean

[latent kettle]

bacically it depend on your dataset and length and .. g loss and d loss as well as mel kl graphs..

[proper shale]

like 2-5 hours, but if you're on colab i highly recommend to stop training before that, exporting the model and then training again at a later date

and yes training length does vary

yeah

probably like an hour or two after training

like... just look at the tab every 5 mins, maybe refresh TensorBoard and see what's going down

[latent kettle]

will it sound poor... basically I want to ask that what if I train a model o 48khz .. or in 32 khz .. the output quality of models will same or different

[proper shale]

it wont sound poor lol

[latent kettle]

what does it mean . can you explain?

[proper shale]

it will not sound bad compared to 48khz

in fact it might just be better because it'll cut off all the noise above 16khz

and most mics are capped around 15khz

so yea

[latent kettle]

what does the term 'target _sample_rate' mean

[proper shale]

sample rate of the rvc model

[latent kettle]

yeh .. but in technical term ...

[proper shale]

...it just means sample rate of the rvc model

it'll output at that range, 32khz

[molten pecan]

https://aihubdocs.github.io/en/ check the sample rate page to learn =3

[latent kettle]

thanks.. That's what I need

[molten pecan]

yw

[wispy lava]

yo guys

what do i do with D_1900.pth and G_1900.pth files

how do i use them? how do i turn them into a "small file"

[glad zealot]

look in the weights folder

that D and G are used for training the model

[wispy lava]

oh wtf

i didnt know it was there

💀 thats crazy

alr thanks

i thought i had to create my own file with those files

[glad zealot]

nah it should auto make them

[wispy lava]

oh btw incase i somehow lose that file is it possible to actually make that small file with the 2 training files?

[latent kettle]

it really helped.. thanks a lot again

[glad zealot]

youd have to resume training to get the new epoch save

resume from where it last saved basically

[wispy lava]

mhm yeah sounds simple enough

k thanks

[molten pecan]

np

[brittle wing]

Guys, answer the question, please.
does it make a difference which site I train the RVC model on?
if yes, which site is better to use for training?

[rare goblet]

Ayo? @brittle wing level 1 !!!

[latent kettle]

I don't think it effects on model quality.. your settings can . such as sample rate batch size or the type of pich extraction algorithm

[hushed herald]

can someone explain to me how to install RVC please ?

[latent kettle]

yes ..

[latent kettle]

https://aihubdocs.github.io/en/rvc/local/mainline/

[split latch]

anyone knows why this happens?

[whole ruin]

can someone help and explain why i cant see the gpu settings? im only left with the cpu settings and for the F0 detection i can only use these 2 prompts ( rmvpe_onnx and crepe_tiny )

[molten pecan]

https://aihubdocs.github.io/en/ try following the guide here

RVC Disconnected one

On the left

Click the folder symbol

Can you send a ss?

@brittle wing open the content folder

[rare goblet]

Ayo? @brittle wing level 3 !!!

[molten pecan]

Try restaring the colab

[latent kettle]

is runtime is disconnected

I don't think.. you have to train it again

[molten pecan]

@brittle wing check the docs and go to the retraining chapter

Read the switching method

Follow the instructions

You are good ;3

Shit. @half cove what should they do?

[half cove]

screenshot

f

[wispy lava]

Something went wrong
CUDA out of memory. Tried to allocate 39.36 GiB. GPU 0 has a total capacty of 8.00 GiB of which 2.13 GiB is free. Of the allocated memory 4.69 GiB is allocated by PyTorch, and 155.55 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

man wtf????? 39gb 💀????

[whole ruin]

can someone help and teach me how to use the voice changer?

[molten pecan]

#🔍│help-w-okada

Reminder: for Okada or real-time Voice Changer help, ask in #🔍│help-w-okada

[low tundra]

i just want to know what app do i use
so i can use the voice changes
and is there a specific app i need to use for it to work or no ?

[molten pecan]

In real-time as you speak? Or through audio files?

If it's the latter, check our docs

https://aihubdocs.github.io/en/

[low tundra]

ok

[whole ruin]

sry im new new

[molten pecan]

np

[low tundra]

uhhh i want the app that i will use the files on

not how to make a voice file

is there an app or no ?

[molten pecan]

Yes. Check the AI cover page

You'll find the starting point

[low tundra]

ok ty

[rare goblet]

Ayo? @low tundra level 1 !!!

[low tundra]

....

[mighty fog]

why the og easy gui doesn't work

[wild vapor]

JSONDecodeError Traceback (most recent call last)
<ipython-input-9-6c7f6ebd9716> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:

5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

what does that mean 😭

[odd shale]

Which colab are you trying to use?

Send link.

[mighty fog]

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

[rare goblet]

Ayo? @mighty fog level 1 !!!

[odd shale]

Yeah.. I think that colab doesn't work anymore since months.

Use Ilaria RVC better

Unless you wanna train a model.

[wild vapor]

link?

[mighty fog]

it's impossible because i even made an ai cover like 1 month ago

but it's laggy i admit

[odd shale]

https://colab.research.google.com/drive/16LkwvFZeudTpUOsE_6bMjOq2qkxFo8Hr?usp=sharing

You both use this.

But it's not for training

[wild vapor]

thanks

[rapid niche]

I haven't trained a model in a while, are people still using google collab or is there something else?

[rare goblet]

Ayo? @rapid niche level 1 !!!

[odd shale]

RVC Disconnected, Applio or Paperspace

[woeful sedge]

does webui give you enough time to inference voice?

[half cove]

yes

[mighty fog]

how to you import the models on the collab ?

[odd shale]

No, you must import them from the GUI

[half cove]

i dont know if the main is fixed rn

i dont think so

[odd shale]

Ilaria RVC is broken?

[mighty fog]

how do you do 😭

[half cove]

rmvpe download fucked up

use the one in pins until its fixed

[glad zealot]

that part is fixed, a new problem arises lmao

[split latch]

.

can some one help with this

[glad zealot]

can you try it as just datasets/<dataset folder> ?

[dawn nebula]

space in model making

[split latch]

ohhhhh

I guess my pc OCD did not help me

thanks

[glad zealot]

a fck im blind i was looking for a space there

[brittle wing]

Can someone help me in installing Mangio Forked RVC on mac m1??

[glad zealot]

ye

just another case of me being blind

[brittle wing]

idk...

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[dawn tendon]

Can someone pls help me?

[tranquil cliff]

Hi

What do you need help with?

[dawn tendon]

No url link

[split latch]

can't get rid of the reverb, is it gonna be an issue?

I mean i could but the model will sound bad

[rapid niche]

Can someone help me please? It loads my dataset fine but then loads it again and says no such file or directory. Idk what’s wrong

[desert niche]

@molten pecan what batch size should I use for an almost 5 hour dataset?

[fluid matrix]

holy shit dont use 5hours

[molten pecan]

oh jeez, I don't think you need that long of a dataset

10 mins is fine, as long as it's good quality

[desert niche]

okay

I got it from a twitch vod

I'll cut it down

[fluid matrix]

if you have that much, dont use more than 30min in my opinion

[desert niche]

noted

[fluid matrix]

using streams means you have a lot to clean up first, dont just train the audio lol

[desert niche]

I did clean it up

with uvr

lol

took like

20 something minutes

[fluid matrix]

i doubt those 3h are all useable

[desert niche]

yeah, I'll cut it down to 10 minutes

[fluid matrix]

well if you said you did clean it up, just make sure to cut it down

yeah

[desert niche]

what batch size should I use for 10 minutes?

[fluid matrix]

nah i would do more than 10

10 is not a lot

[desert niche]

oh

okay I'll do uh

ig 20

minutes

[fluid matrix]

10 is fine like she said but thats like the bare minimum

[desert niche]

I see

[fluid matrix]

are you training locally or do you use a colab?

[desert niche]

locally my friend

[fluid matrix]

whats your gpus vram?

[desert niche]

lemme check rq

[desert niche]

NVIDIA RTX A4500

[fluid matrix]

you have that locally? 😭

[desert niche]

yeah

is that bad

[fluid matrix]

that thing is like 2k

how 💀

good for you ig

you should be fine with 8 then

[desert niche]

okay thanks

[fluid matrix]

np

[dawn tendon]

Can anyone tell me why it doesn't download?

[flint vapor]

Check if the link is still valid

[dawn tendon]

That is weird

Because i looked and it did download

So what was that?

[flint vapor]

Open the zip check if the files are ok

[dawn tendon]

Alright

[eager stratus]

yo sorry if this is the wrong chat but which version should i download if i have a nvidia graphics card and a ryzen cpu?

[fluid matrix]

i assume you mean for the voice changer? read this: https://rentry.co/VoiceChangerGuide

[eager stratus]

yee

alrr ty :D

[dim tree]

Is there any way to fix the delay? It's only a few seconds but it still messes with me.

[vague glade]

how do i use ai model PACKS?
i can only select one model from the pack.

using applio

[rare goblet]

Ayo? @proven cypress level 1 !!!

[pale nimbus]

@brittle wing use ilaria RVC

[glad veldt]

I have over 450 audio files for my dataset

is there a way that I can merge them into one without losing quality?

[brittle wing]

maybe a silly question but how do I download the voice changer?

[vague glade]

yep, use an audio editing software and import EACH ONE audio file into the software

[flint vapor]

https://rentry.co/VoiceChangerGuide

[vague glade]

30-50 epochs, maybe?

[half cove]

[proper shale]

after like an hour

yeah dw

what would be worrying is if it started rising, and didnt go down after 10-20 epochs

*and didnt go back down

as you said, it did go back down, so you're fine

[desert niche]

is 1 second training bad, the dataset I used is 10 seconds and the batch size I used was 8

[agile robin]

i need google collab for cover

[desert niche]

and pretrain was Ov2

[agile robin]

can you help me

[azure marsh]

Suggestions for @agile robin

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[flint vapor]

You can reopen

You are using colab or local?

Unfortunately on colab idk how to help because i use only local version, you have any links in the colab terminal for tensorboard?

[smoky root]

best realistic female voice model?

[tacit grail]

ok idk why but when i put the folder address in the index file part it doesnt say theres an index file in it even after refreshing

using applio btw

[odd shale]

There's no best realistic model.

Ask for one in #1159289738314919936 channel

Paid requests are allowed too.

[tacit grail]

oh nvm i figured it out now

[agile robin]

didnt work

😭

i try all of them

i nneed just for cover

[molten pecan]

https://aihubdocs.github.io/en/ check the AI cover page

=3

[verbal basin]

whats wrong with easygui rn?

getting a file reading error

when the file is correct

[agile robin]

thanks but i gennerelly used this and now its giving error

[molten pecan]

This what?

Please give a lot of context or we can't help you

[agile robin]

when i downloand model giving erorr

[flint vapor]

What model are you downloading, sure that is still available?

[agile robin]

yes i used alwas

always

same model

illaria

[sand iris]

retry

it errors the first time

...even though it works most of the time

[agile robin]

?

[agile robin]

didnt work i try 1 hour

always say ERROR - The download failed. Check if the link is valid.

but i know

still valid

didnt work 😭

i think ilaria is broken

do you have

another

cite

okay thanks btw

its so confuse i ddont understand this covergen

yes i did after?

song input? so i need vocals but say yt link

yes

yes after

so directly yt link not vocal or something

[rare goblet]

Ayo? @agile robin level 5 !!!

[agile robin]

really im thankful for care about me

im trying it now if it doesnt worj illl write back

last question

di o need to do anytihng pitch?

just i know this for example if male to female -12

but this have two picth swtting

but if male to male no need change right?

ready now how to downloand i dont see

yeah i understand

[sand iris]

[agile robin]

yess i understand

[verbal meteor]

hi, can i ask a stupid question here?

[sand iris]

JoeShhh

[agile robin]

i cant

find

the song

[odd shale]

What is it?

[verbal meteor]

at step two in the google collab it says this:

---

JSONDecodeError Traceback (most recent call last)
<ipython-input-15-bbb88814eba7> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:

5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

and it won't run

can anyone tlel me what i'm doing wrong?

[odd shale]

On which colab?

[verbal meteor]

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

[odd shale]

Yeah, i think that colab is broken

[verbal meteor]

ooh that explains it

[rare goblet]

Ayo? @verbal meteor level 1 !!!

[verbal meteor]

thank you very much, could you recommend me a better one?

[agile robin]

i cant findd

[fluid matrix]

that colab is really really old

[agile robin]

noyhing have

in

[fluid matrix]

@verbal meteor https://aihubdocs.github.io/en/

use the ones listed here and follow the guide

[verbal meteor]

thank u!

[fair tundra]

can someone help me use voice audio cable thing so i can use on disocrd please

[quasi gate]

somebody can help me? i getting this error, i did the Feature Extraction

<ipython-input-31-75389f92609d> in <cell line: 27>()
     26 listdir_res = list(os.listdir(feature_dir))
     27 if len(listdir_res) == 0:
---> 28   raise Exception("No features exist for this model yet. Did you run Feature Extraction?")
     29 
     30 try:

Exception: No features exist for this model yet. Did you run Feature Extraction?```

[odd shale]

-realtime

[azure marsh]

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

[north torrent]

yo, its my first time using collab so i might be a little dumb, but i keep getting this message at the end of step 2(extract features) i know i already did it but it just keeps popping up

[agile robin]

yes

oh i findd

im so thankfull

[north torrent]

shoot, forgot to send the error message im getting

error preprocessing data... Make sure your data was preprocessed.

[quasi gate]

someone?

i can't make .index file

[gray garden]

Do you need python to make a model/dataset in colab?

[sand iris]

No

[gray garden]

i feel so dumb at this

[rare goblet]

Ayo? @gray garden level 2 !!!

[molten pecan]

https://aihubdocs.github.io/en/ I made beginner-friendly, dumb-proof guides =3

enjoy

[gray garden]

I am not proud to admit at being a visual learner .

thank you for guide

[molten pecan]

yw =3 any questions you lmk

[fair tundra]

/guides realtime

[azure marsh]

🎙️ Realtime Voice Changer (Local)

Suggestion for @strong shard

[molten pecan]

@fair tundra please use #🤖│bots to not clutter the chat

[bold magnet]

can someone help me with training

[rare goblet]

Ayo? @bold magnet level 2 !!!

[bold magnet]

after the feature extraction thing the training doesn't work

[urban mulch]

one question what do you all use to use vocies i downloaded vocie ai but all the voices are not combatible

[thin stump]

I can't read this. Do I use 32k for my model?

for training

[analog obsidian]

anyone can tell me what really causes mode collapse besides silence and noise?

[thin stump]

@molten pecan Sorry for the ping, I think this should be updated and include the download link to Spek

https://aihubdocs.github.io/en/rvc-resources/audio-formats--sample-rate/#sample-rate

[molten pecan]

use 40k

[thin stump]

Thanks, will do

[molten pecan]

I think Ilaria Audio Analyzer is enough for determining the sample rate

though we'll still add an advanced dataset cleaning guide in the future nonetheless =3

including other software

[thin stump]

It was confusing at first but I see that it's in 20khz for audacity too so just checking

[proud rivet]

hi! I got this to work last night and it's wonderful, but I can only do it live. How can I add a pre-recorded .wav or .mp3 file to the program? can someone point me in the right direction to learn about this? Thanks!

[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

[thin stump]

Are you talking about AI covers? You can't use a wav or mp3 of just someone talking. It needs to train off of a reference to create a model

[molten pecan]

https://aihubdocs.github.io/en/ here's everything you need to know =3

check the AI cover page

[proud rivet]

thanks, I actually was reading that and that's why I came here. Because there is a step in the instructions I either don't understand or it skips. It tells you how to extract the vocals (check) and get the voice model (check) but in the inference section, it doesn't tell you how to add it to the program.

[molten pecan]

which page exactly?

AI Cover one?

and which program?

please give full context 🙏

[glad tide]

i have 4060ti16g and i5 12400f and i didnt get any voice why

this what i install

[molten pecan]

for okada/voice changer stuff ask in #🔍│help-w-okada

[proud rivet]

yeah the AI cover one and this is the part where I get stuck. I"m using Realtime Voice Changer Client (maybe that's the wrong program?)

[steep valley]

Hello I downloaded a Chester index and pth file where do they get imported

[molten pecan]

the voice changer is for changing your voice in realtime, as you speak through a microphone

the ones explained in the docs are for inferencing audio files

[proud rivet]

oh my bad 😅 Let me try to find the right program for this then

[rare goblet]

Ayo? @proud rivet level 1 !!!

[molten pecan]

the AI cover page redirects you to it

Mainline RVC

=3

[molten pecan]

check the AI cover page
https://aihubdocs.github.io/en/

[proud rivet]

okay thanks! I think the link to RVC is broken but I found it on the side

[brittle wing]

Hi! I'm trying to use the IIaria RVC space and when I try to put the link to download the model it says error. I am using the link that is posted. Do I have to extract everything from the zip and then use a link from there? How would I find a link in there? Please help and thank u

[molten pecan]

really? can you send said link?

i'll take a look at it

[molten pecan]

are you using Colab or Hugging Face?

[proud rivet]

https://aihubdocs.github.io/en/essentials/how-to-make-ai-cover/ the underlined RVC doesn't go anywhere for me. I don't know if that's meant to be a link

[brittle wing]

hugging face

[proud rivet]

where it says"To convert the the voice into your desired person, we'll use the best software for this, RVC." RVC is underlined like if it was supposed to be a link

[molten pecan]

ah that's not supposed to be a hyperlink

check step 3

[molten pecan]

you are using an outdated version

https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/ here's the updated one =3

[proud rivet]

Oh i see my bad again 😅 I downloaded it and now i'm installing it. Thank you sooo much

[molten pecan]

yw! have fun ^^

[plush compass]

In Colab, is ov2 only available at 40k?

[crystal thunder]

Hello

[molten pecan]

buenas noches

[crystal thunder]

Buenas

Donde está el repositorio de las voces que se pueden hacer?

[molten pecan]

querés hacer modelos de voz?

[crystal thunder]

No no

Las voces disponibles

[proud rivet]

donde dice "Voice-models" no?

[crystal thunder]

No si, pero según yo había una tabla de excel con links y eso

[molten pecan]

hay varias formas de buscarlos. la más fácil es en https://weights.gg o #1175430844685484042

https://aihubdocs.github.io/es/guías-populares/modelos-de-voz/

si necesitás más ayuda lee nuestros docs =3

[crystal thunder]

Thank you

[proud rivet]

como no tienen a Alejandra Guzman?! 😂 😅

[brittle wing]

ohhhh okay thank u 🙂

[rare goblet]

Ayo? @brittle wing level 1 !!!

[molten pecan]

yw =3 have fun!

[brittle wing]

Bro! I was training a model, everything was fine, I managed to download the .pth file, but when I went to download the file that starts with added, colab disconnected, and now I can't download anything else D:

[rare goblet]

Ayo? @brittle wing level 4 !!!

[brittle wing]

Does anyone have a tip for me to train models without interruptions? I spent two hours to train just 250 epochs.

[modern surge]

-local

[azure marsh]

🖥️ Local stuff

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

Need some help? 🤔

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

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

[thin stump]

What do I do with logging interval? Do I leave it alone

k leaving it as default

[young halo]

How much time does it takes for the GPU to restore?

On RVC Disconnected

[low shard]

In google colab around 12-24 hours

You can also use alt accounts

[young halo]

What if I want to resume a training? ಥ⁠‿⁠ಥ

[thin stump]

Do you know where that training progress gets saved just in case I get kicked out?

[forest sluice]

Look inside the rvcDisconnected folder on your Drive

[gritty cargo]

is for train or change the voice

[glad veldt]

how do you decide the best hop length to set for the model

[thin stump]

I only see added_IVF

[molten pecan]

it's for inference. but the docs contain tutorials for everything

[molten pecan]

https://aihubdocs.github.io/en/ go to the RVC Disconnected page and read the Retraining chapter

specifically the Switching method

[molten pecan]

It's recommended to use Crepe only when the dataset is high quality

if that's your case, a value between 20 - 128 is good

leaving it too low might lead to voice cracks

in my experience, I tried up to 60 without problems

after it I did, but likely due to my dataset

[haughty flame]

Hola!
Usar la computadora (pispeando páginas de internet) mientras se deja entrenando el modelo puede causar algunos crashes?
Y usar tres monitores consume más menoria VRAM y por ende también perjudica al rendimiento del entrenamiento?
Soy muy nuevo en esto, perdón si son preguntas obvias.

[thin stump]

Local use or google colab?

[haughty flame]

Local use.
Im sorry, which captcha?

[molten pecan]

pues con versiones modernas de RVC, no se suele crashear. Evita usar "RVC GUI"

nunca he escuchado a alguien que se le crasheara mientras usaba RVC moderno. excepto cuando uno lo hizo en una laptop de gama media

[spare shale]

T-T

[fierce cedar]

hola, tengo un problema (nose como funciona esto)

[molten pecan]

Los captchas aparecen en la versión en la nube de RVC. Si usas RVC local, no tendrás captchas

[spare shale]

What exactly?

[frigid night]

can you still get links for huggingface to use with llaria? im just trying to do something quick and cant find hank hill in the voice models section anymore

[quasi solstice]

Alguien tiene el colab para hacer epoch? Tengo el colab para hacer inferencia (canciones) de ia hub

[spare shale]

https://colab.research.google.com/github/poiqazwsx/RVC-Notebooks/blob/main/Ilaria_RVC_Tempfix.ipynb

[molten pecan]

yeah ofc. https://aihubdocs.github.io/en/essentials/voice-models/ here are more methods to search voice models

[frigid night]

spanish is absolutely diabolical

[molten pecan]

sent the wrong link lol there you go

[fierce cedar]

que

[molten pecan]

https://aihubdocs.github.io/en/rvc/cloud/rvc-disconnected/ acá tenés una guía de uso completa si te sirve

recuerda no usar una cantidad semi-arbitraria de epochs, si buscas los mejores resultados

[fierce cedar]

yo la verdad es que e intentado usar el rvc pero cada que quiero usar una voz me sale error y al hacer el tercer paso no me sale el link, que hago?

[molten pecan]

qué paso exactamente? y qué RVC?

da más contexto porfa

[thin stump]

Can I just close colab and resume tomorrow?

[fierce cedar]

voy

[quasi solstice]

Tengo la voz ya hecha de 350 epoch pero quiero llegar a los 500

[rare goblet]

Ayo? @fierce cedar level 1 !!!

[thin stump]

it won't let me export

[molten pecan]

as long as you donwloaded the necessary files, yes you can

[thin stump]

the added_IVF?

[quasi solstice]

Pero como viven cambiando los colab nose ahora como seguir por eso pedía. Gracias!!

[molten pecan]

in the Retraining chapter I explained the necessary files

[thin stump]

ok

[molten pecan]

de nada. recuerda que la cantidad perfecta de epochs depende según el tamaño, calidad & duración del dataset

por lo que no hay una forma de saber la cantidad correcta de antemano

https://aihubdocs.github.io/es/recursos-de-rvc/epochs-sobreentrenar--tensorboard/ lee esto para aprender más

[quasi solstice]

Sii sii muchas gracias x pasar la página!!! Esto buscaba 🥰🥰

[rare goblet]

Ayo? @quasi solstice level 1 !!!

[molten pecan]

de nada. diviertete =3

[fierce cedar]

me sale esto: Traceback (most recent call last)
<ipython-input-5-0b6d81064fb3> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:

ya probe varios modelos y es mas de lo mismo

[molten pecan]

que raro. jamás vi un error así

[haughty flame]

Mil gracias por las respuestas!

Tengo un modelo de voz entrenándose con dos archivos de 7 minutos muy limpios.
NVIDIA GeForce GTX 1660 Ti
Core i7, 32 de RAM Y 16 de VRAM con tres monitores.
Le puse 4 en Batch size per GPU.

Algo que recomienden cambiar? Va super lento, lo cual no me molesta, puedo dejarlo días porque puedo dejarlo por la madrugada, pero si es innecesario y puedo acelerar el proceso sin perjudicar el resultado mejor.

[molten pecan]

diría que preguntes en #🌍│español

[fierce cedar]

no me digas eso

[molten pecan]

mañana cuando los chicos se despierten te podrán ayudar seguro

son activos

[fierce cedar]

ok gracias

[molten pecan]

intenta usar un batch size de 8

diría que uses 4 cuando el dataset es corto (cerca de 2 minutos o menos)

[haughty flame]

Ah, buen dato.
Se puede cambiar el batch size en medio del entrenamiento?

[rare goblet]

Ayo? @haughty flame level 1 !!!

[molten pecan]

habrá que pausarlo, cambiar el valor, y reanudar. la cosa es que no si esto va a arruinar el modelo

el batch size define la cantidad de información que se procesará al mismo tiempo

según el batch size que pongas, RVC dividirá el dataset en varias piezas

nunca he intentado esto, así que no sé cual será el resultado

[haughty flame]

Mejor lo dejo así y ya cambio para el próximo modelo de voz. 😬

[molten pecan]

sip. ten paciencia y suerte

[haughty flame]

Qué genia. Gracias 🙏🏻

[molten pecan]

de nada =3

[thin stump]

@runic zinc More doesn't mean it's better. There's numerous factors like the dataset, the pretraining method, was the audio cleaned before training? https://aihubdocs.github.io/en/rvc-resources/epochs-overtraining--tensorboard/#tensorboard

[hot lagoon]

Can someone help?

[spiral osprey]

easy gui and rvc v2 disconnect doesn't work for me.

[clear lily]

is this normal?

[forest sluice]

Yes

Look at the logs on the Colab tab

[clear lily]

the file is 5 GB 😭

18?!?!

ITS 18 GB

how do i stop it lol

[forest sluice]

18GB what?

[clear lily]

the file that its exporting on is 18 GB.

20 now

[forest sluice]

if you're training on Colab, don't mind about it

[clear lily]

im not

[forest sluice]

Ok, then terminate the cmd

[clear lily]

alr

jesus

is it because i dont have it on my c drive?

[hexed hound]

me tira el mismo error

[brisk dew]

hello, i think i have a problem with mangio rvc, can somebody help me?

[brisk dew]

i'm trying to train an model, but it doesn't seem to train and when i look at "weights" folder nothing is there, i looked at logs too and i only could train the feature index, but the model itself i can't train i think

and for some reason the program says that he couldn't recognize my gpu, i think it is because from amd

[proud elbow]

try check the option "save only latest G and D files to save disk space"

[clear lily]

ok

[clear lily]

Ok so it says it's 270/13. Why does it say that

But it's staying at 30 Mb and not growing

[proud elbow]

you should get more files in weights folder, which are the actual model file instead of the G and D file

[clear lily]

idk what that means lol

[brittle wing]

I'm training using this link and 2 times it's spontaneously disconnected after a few hours. Like it just stopped with no popup box. What am I doing wrong? https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw

[brisk dew]

?

[glad zealot]

it takes a while before it saves a model

i think by default its 25 epochs before it gets saved

[brisk dew]

i don't think the problem is waiting, like, if i look at the terminal, it sends only txt and stops

[rare goblet]

Ayo? @brisk dew level 1 !!!

[glad zealot]

can you screenshot the console?

[brisk dew]

yeah, but wait a minute because i closed it a while time ago

[gritty radish]

i made a dataset and it has a bunch of audio files in it but how do i get a index and pth file

[glad zealot]

you need to train it on rvc, you can do it locally(your own hardware) or online

[gritty radish]

mkay

[brisk dew]

@glad zealot sorry for taking to much time, but this is the only message that appear on my terminal and i can't find any pth file only index, i did 10 epochs just to show you what happen and to not take too long time

[glad zealot]

do you have a gpu?

[brisk dew]

yes and it doesn't recognizes

[glad zealot]

whats your gpu?

[brisk dew]

rx 6600

oh wait, i don't know if this is my gpu

[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

[glad zealot]

you can cehck in task manager

if you're on windows

i forgot how on linux

[brisk dew]

ok, let me check then

i'm not too much on computer so i was thinking like gpu is a video card

[glad zealot]

yes gpu is the graphics card

you can also try training online if it doesnt want to train locally

[brisk dew]

ok i have a gpu

[rare goblet]

Ayo? @brisk dew level 2 !!!

[brisk dew]

oh ok

i think it is because i use rx and it is from amd, not nvidia

[brisk dew]

i'll try then, do you have any suggestion where can i do it?

[glad zealot]

-colab

[azure marsh]

Suggestions for @brisk dew

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[glad zealot]

you can try applio, rvc disconnected and easy gui

[brisk dew]

ok, thx so much

[forest sluice]

Save only latest epochs

Enable that option in the GUI

[forest sluice]

You should the weights folder alongside the logs folder

You'll see the models in *.pth

The G_* and D_* files will grow in size instead of creating more backups if you select that option

[rare goblet]

Ayo? @strange reef level 1 !!!

[proper pivot]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[strange reef]

does anyone know how to copy a link zip file to a rvc space? ive been trying to copy a few links from huggingface.co but it's always a error

these r the zips but idk how to get the official links 😭

[modern surge]

-local

[azure marsh]

🖥️ Local stuff

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

Need some help? 🤔

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

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

[proud elbow]

copy link on the file's LFS or download icon, that should contain "resolve", also make sure the model repo itself is set public

[inland coral]

I went to this site and I need to download this audio file, how do I do it? These are the materials.

There is a file called ckpt

I can't take photos

help please

?

[forest sluice]

What site? Huggingface?

[inland coral]

yes

I cant take photos

Should I send you a link?

[forest sluice]

Give me model link

[rare goblet]

Ayo? @inland coral level 1 !!!

[inland coral]

https://huggingface.co/selcukkubur/rafadan-hayri/tree/main

[forest sluice]

That's Stable Diffusion, not RVC

[inland coral]

Understood thanks

I'm using translation, sorry :)

[modern surge]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[lament wagon]

Can anyone find a way to send me directly to the install RVC page for this
https://aihubdocs.github.io/en/rvc/cloud/inference/ilaria-rvc/
Because the link page is gone

[forest sluice]

https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/ Moved to a new section

[lucid crow]

hello moderators

can someone help me with rvc i cant hear myself

[forest sluice]

Your config looks good as I skimmed

[forest sluice]

Maybe try right on Discord?

[lucid crow]

doesnt work

[forest sluice]

ServerIO Analyzer next to the AUDIO section?

[lucid crow]

0

[lucid crow]

what will i do?

can u dm me ?

[forest sluice]

Click start there on the voice changer app to record your voice, click stop to see the results:
in is for you input audio (a.k.a your original microphone voice)
out is for the changed voice

[lucid crow]

still none/

[forest sluice]

Now, just use your normal mic with a voice recorder, and hear if your mic produces any sound in the first place

[lucid crow]

@forest sluice

[forest sluice]

Maybe just try to search by voice on Google

[lucid crow]

can u dm me

[lucid crow]

what is this

[forest sluice]

You go to google, use "search by voice" (click on the microphone button), then say something randomly

[sand iris]

Can you hear anything rn ? or is it just vcable ?

[lucid crow]

none

i cant hear

[lucid crow]

what should i

search

[forest sluice]

speak anything randomly

make sure to allow microphones

[lucid crow]

it worked.

on a test

on google

but for rvc i cant hear/

@forest sluice

[brittle wing]

@sand iris

[sand iris]

alright

[lucid crow]

@sand iris

can you help me with mine it seems @forest sluice went off

[sand iris]

give it a brew install libxslt and that should be the end of it

[brittle wing]

ok 🙂

[sand iris]

the function it can't find is from that library

[brittle wing]

Warning: libxslt 1.1.39 is already installed and up-to-date.

[lucid crow]

gahdur for my case pls

[rare goblet]

Ayo? @lucid crow level 7 !!!

[lucid crow]

i need help

[sand iris]

yes

head to #🔍│help-w-okada

[brittle wing]

.

[lucid crow]

huh?

[sand iris]

hmmmm~2

[lucid crow]

rvc im habing a problems

[sand iris]

realtime help goes there

just to make things cleaner

dw I'll help soon

[lucid crow]

im not having problems with okada.

the rvc,

[loud pond]

why is RVC not changing my voice

[manic sapphire]

Hey uh can someone explain how to download

[loud pond]

i literally sound the same its not voice changing

[sand iris]

time for plan B then !

STATIC_DEPS=true pip install lxml

oh wait a sec

[manic sapphire]

Hey where to download

[sand iris]

not yet

[lucid crow]

Why no one helps me.

[manic sapphire]

Idk anything

[lucid crow]

So unfair

[loud pond]

PC type?

[brittle wing]

pls wait 🙂 be a little patient

[sand iris]

mate I'll be there in a sec

[manic sapphire]

Ye

[lucid crow]

ok

[manic sapphire]

Im new in this server

[rare goblet]

Ayo? @manic sapphire level 1 !!!

[manic sapphire]

And idk what to do

[loud pond]

me next or whenever you get time just ping me thank you 🫶

[sand iris]

STATIC_DEPS=true pip install lxml==4.9.2 --force-reinstall --no-cache-dir
@brittle wing

[loud pond]

-realtime

[azure marsh]

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

[sand iris]

...should fix it once and for all

[brittle wing]

ok 🙂

[manic sapphire]

Oohh

[lucid crow]

im waotong

[manic sapphire]

Ty

[sand iris]

Not sure why it can't find the shared functions despite xslt being installed.. prolly a version mismatch or something

[lucid crow]

@sand iris can i dm u

[manic sapphire]

So if i need a voice i just go to voice model right?

[lucid crow]

finallly ouch s here

[sand iris]

what do you need ?

[forest sluice]

I'm currently trying to think of the cause of the error

[loud pond]

no..

[lucid crow]

hi can u hear yours?

[loud pond]

you download the voice changer

[lucid crow]

to whom?

me or?

[loud pond]

Could you help me after

[manic sapphire]

Oh

[sand iris]

head over to #🔍│help-w-okada too 🙃

[manic sapphire]

Im trying to download the voice changer first i guess

[forest sluice]

And one thing: Is your GPU AMD?

[brittle wing]

i do sh run.sh right?

[sand iris]

ya

...It should no longer try to get shared functions, and use it's own now

[brittle wing]

should i restart my mac too? (pic from https://docs.aihub.wtf/vocal-cloning-guide/rvc-v2)

[rare goblet]

Ayo? @brittle wing level 3 !!!

[sand iris]

if it doesn't work then yeah

[brittle wing]

ok

[sand iris]

...though just relaunching the terminal would be enough

[brittle wing]

ok

ahh. moment of truth. i am running sh run.sh

[lucid crow]

yes

[brittle wing]

same error 💀