#✨│ai-help

1 messages · Page 26 of 1

proper shale
#

no waiot

#

where did you export it from

half cove
#

aint that for index training

cerulean dew
#

yes

half cove
#

seems like alot of people have this problem

#

hm

cerulean dew
#

im not the most tech savvy person do i dont rlly know how to answer this so ill just describe how i got everything to where it was
i started with my audio files which were mp3 and i converted them to wav files using cloudconverter and then downloaded them. after that i went to rvc disconnected and when i got to the point where the rvcdisconnected folder popped up i put a zipped version of that file into the folder

rare gobletBOT
#

Ayo? @cerulean dew level 3 !!! lfg

proper shale
#

shouldve reexported it in audacity

cerulean dew
#

so how far back do i need to go to do that, can i just do that with the wav file?

cerulean dew
#

Omg ur my savior tho it actually worked

cerulean dew
#

well tysm ive been trying to fix this for 2 days lol

#

its working now

#

but the epochs r taking forever since they r actually functioning now lmao

#

ill be waiting a while

gusty kestrelBOT
#

I have found 97 results for the search riggy...

Riggy

Owner: Kareem AQ
Uploaded: <t:1694444760:R>

Information

This model can be found on the Kits.AI platform, visit their page for more information.

#

💾 @lethal prism, sent you a DM with the model information!

vapid gust
#

why is there like that electronic parts

#

in some parts of my voice

#

and then

#

in some

#

its very clear

rugged loom
#

hello if someone could help me with this error

AttributeError: 'NoneType' object has no attribute 'tobytes'

#

thanks

#

this is full thing

<All keys matched successfully>
Traceback (most recent call last):
File "C:\Users\samus\Downloads\Mangio-RVC-v23.7.0\my_utils.py", line 96, in load_audio
ffmpeg.input(file, threads=0)
File "C:\Users\samus\Downloads\Mangio-RVC-v23.7.0\runtime\lib\site-packages\ffmpeg_run.py", line 325, in run
raise Error('ffmpeg', out, err)
ffmpeg._run.Error: ffmpeg error (see stderr output for detail)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "C:\Users\samus\Downloads\Mangio-RVC-v23.7.0\infer-web.py", line 292, in vc_single
audio = load_audio(input_audio_path0, 16000, DoFormant, Quefrency, Timbre)
File "C:\Users\samus\Downloads\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)

Traceback (most recent call last):
File "C:\Users\samus\Downloads\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\routes.py", line 437, in run_predict
output = await app.get_blocks().process_api(
File "C:\Users\samus\Downloads\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1349, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "C:\Users\samus\Downloads\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1283, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "C:\Users\samus\Downloads\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 2586, in postprocess
file_path = self.audio_to_temp_file(
File "C:\Users\samus\Downloads\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 360, in audio_to_temp_file
temp_dir = Path(dir) / self.hash_bytes(data.tobytes())
AttributeError: 'NoneType' object has no attribute 'tobytes'

rare gobletBOT
#

Ayo? @rugged loom level 1 !!! lfg

midnight mirage
#

RVC is now working, sorry for the inconvenience

jaunty halo
#

Guys

#

Can I ask something?

#

How many epochs I need to train a 30min dataset?

#

Studio acapella

daring raft
#

ayo, if my model doehave falseto singing in dataset, why the fck then when I run interference with me singing falseto, model sing it too??

placid storm
#

i didnt get the model after traing on rvc @violet heron

#

can some one help me

#

pls

azure marshBOT
#

RVC Guides (How to Make AI Cover)

Documentation
🇺🇸 English (main)

Translation by country

🇧🇷 Brasil (PT-BR)
violet heron
half cove
placid storm
#

how to save weights tho idk

#

how to do

placid storm
violet heron
placid storm
#

thsi one

#

or this

#

@violet heron

violet heron
#

But also the index

#

Do both, index train first

placid storm
#

what about this

violet heron
placid storm
#

ai

rare gobletBOT
#

Ayo? @placid storm level 8 !!! lfg

placid storm
#

thanks @violet heron

ancient cloak
#

@empty dust I'm tryna make smooth criminal but there are so many voice cracks

#

those are the settings

#

idk how to work them anyways

empty dust
#

Send the file you are trying to use

ancient cloak
#

alr

#

like the vocals?

empty dust
#

Yeah

hot lagoon
#

overtrain?

empty dust
#

Yeah, wait a minute

ancient cloak
molten pecan
#

hp5 is for harmonies. hp6 is for backup

rare gobletBOT
#

Ayo? @tough ice level 2 !!! lfg

molten pecan
#

Make sure you remove the reverb first!

empty dust
molten pecan
#

For the best results

#

First the reverb and then use karaoke

ancient cloak
rare gobletBOT
#

Ayo? @ancient cloak level 2 !!! lfg

molten pecan
#

It will give you both things

ancient cloak
#

I just use default settings inn uvr 5.6

molten pecan
#

Both the harmonies only and the lead vocals

#

Depends on what you select as your output

empty dust
ancient cloak
#

oh ok

#

so what do I do to fix it?

#

my gpu can run like anything also

ancient cloak
#

Ok thanks

#

wait but my gui doesnt match the one in the video

cursive magnet
#

is it available for iphone

molten pecan
cursive magnet
#

ai voice stuff

#

i’m a bit clueless sorry

molten pecan
cursive magnet
#

yeah

molten pecan
#

You can do it on mobile but it's a little slow

#

It's recommended to do it on a fast computer

#

For mobile you can use Ilaria RVC

ancient cloak
#

@molten pecan what settings do I put on UVR5.6 for the best quality?

cursive magnet
molten pecan
#

My bad I was searching the link

#

Good luck!

ancient cloak
#

bro wdym idk I thought someone would know about this lol

rare gobletBOT
#

Ayo? @cursive magnet level 1 !!! lfg

ancient cloak
#

Idk what to do since some covers im making have so much scratchiness and voice cracks.

molten pecan
#

You could try running the harmonies again with hp5 or hp6

hot lagoon
hot lagoon
#

can you use ROD in colab?

ancient cloak
#

What UVR5.6 settings do I use?

daring raft
#

does anyone know what feature did kits.ai used when they started? i made some cool models there but cant see any info ffs. Was it mangio crepe rvc 1 or rvc 2? what was it??

vapid gust
strong belfry
strong belfry
# vapid gust

also zoom in the graph a lot more please. set smoothing to 0.97

vapid gust
#

that’s a old pic

#

Of my old model

#

I’ll train another model soon I’ll give more pics

strong belfry
# vapid gust

that graph looks too smooth. lower it to like 0.95-0.97

vapid gust
#

again old pics

#

could u help me w why does the quality change in my inference

strong belfry
#

well for future note, use less graph smoothing. You are missing all the local minimum and maximums

vapid gust
#

alr

#

I’ll train sum soon

hot lagoon
#

how about mine?

vapid gust
#

dataset don’t look the best

strong belfry
#

youre gonna have to maximize the graph a bit to see

vapid gust
#

and ur at 80k steps????

hot lagoon
strong belfry
#

whats the graph smoothng at?

vapid gust
hot lagoon
strong belfry
#

try 0.95

hot lagoon
strong belfry
#

But looks like a 1hr+ dataset?

vapid gust
#

model won’t come out clean

hot lagoon
vapid gust
#

🤦🏿‍♂️

strong belfry
#

60k looks like a possible good point. if that is 81 epoch then you have like a 2hr dataset?

vapid gust
#

Regardless

strong belfry
#

Might need to continue training to see if the graph is actually reaching minimums.

vapid gust
#

Ur model isn’t good

#

So is ur dataset

strong belfry
#

There is no reason to say the model isnt good

vapid gust
#

Yk those greyed out orange lines behind the highlighted orange one

strong belfry
#

You have no information to determine that outside of the dataset size

vapid gust
#

I think someone told me that but

#

When they got super high and low peaks

#

It’s most likely that it’s not gonna come out as good

#

Idk it happened to me but idk how true it will be for him

#

Let’s see

strong belfry
#

Rather ask for samples instead of just saying it wont be good

vapid gust
#

Ye let’s just see if it’s true

#

but easily you can comparegood graphs with others no?

#

just as an example

hot lagoon
#

Where is the d?

strong belfry
#

Every model is different

hot lagoon
#

on god

strong belfry
hot lagoon
#

I will send like a demo when it hits 100 epochs is that?

strong belfry
#

I'd see a inference at 60k steps

hot lagoon
#

it would it be bad if I told you the Save frequency is at 50 💀

rare gobletBOT
#

Ayo? @hot lagoon level 6 !!! lfg

strong belfry
strong belfry
strong belfry
strong belfry
hot lagoon
strong belfry
#

no you cant go back to 50 and resume training. If you are at like 81 epoch, you should be able to cancel training, change save frequency to 5 and it should save at 85,90,95 and so on after you start the training again.

hot lagoon
#

true just check

ember lance
#

does rvc work with amd local?

#

can i use applio?

hot lagoon
brittle wing
#

whats the best colab or other cloud based Real Time voice changer

azure marshBOT
#

Suggestions for @brittle wing

All working colabs / 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.

molten pecan
#

-realtime

azure marshBOT
errant spindle
#

my rvc does not work even tho i have all the things since when i runthe .bat it says press any key then when i do it closes

#

idk

rare gobletBOT
#

Ayo? @errant spindle level 1 !!! lfg

errant spindle
#

i just downloaded the pkg from github

#

no i did not

#

i right click then i tried to run the .bat

azure marshBOT
#

RVC Guides (How to Make AI Cover)

Documentation
🇺🇸 English (main)

Translation by country

🇧🇷 Brasil (PT-BR)
errant spindle
#

what do i do then?

#

ohh

#

i got it downloaded

#

how do i run the darn thing

rare gobletBOT
#

Ayo? @brittle wing level 1 !!! lfg

crisp plover
#

Is there something that will assist my while singing live ? Does anyone know about a VSt or something RVC related that will make your voice sound way way better while singing live ?

rare gobletBOT
#

Ayo? @crisp plover level 4 !!! lfg

crisp plover
#

Please let me know i need in for tommorow :X

#

Can also pm me!

spare saffron
#

INFO:beta:Saving model and optimizer state at epoch 10 to ./logs\beta\G_2333333.pth
Process Process-1:
Traceback (most recent call last):
File "C:\Users\strin\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\torch\serialization.py", line 441, in save
_save(obj, opened_zipfile, pickle_module, pickle_protocol)
File "C:\Users\strin\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\torch\serialization.py", line 668, in _save
zip_file.write_record(name, storage.data_ptr(), num_bytes)
RuntimeError: [enforce fail at C:\actions-runner_work\pytorch\pytorch\builder\windows\pytorch\caffe2\serialize\inline_container.cc:476] . PytorchStreamWriter failed writing file data/2203: file write failed

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "multiprocessing\process.py", line 315, in _bootstrap
File "multiprocessing\process.py", line 108, in run
File "C:\Users\strin\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\lib\infer\modules\train\train.py", line 327, in run
train_and_evaluate(
File "C:\Users\strin\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\lib\infer\modules\train\train.py", line 640, in train_and_evaluate
utils.save_checkpoint(
File "C:\Users\strin\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\lib\infer\infer_libs\train\utils.py", line 153, in save_checkpoint
torch.save(
File "C:\Users\strin\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\torch\serialization.py", line 442, in save
return
File "C:\Users\strin\Desktop\Applio-RVC-Fork\Applio-RVC-Fork\runtime\lib\site-packages\torch\serialization.py", line 291, in exit
self.file_like.write_end_of_file()
RuntimeError: [enforce fail at C:\actions-runner_work\pytorch\pytorch\builder\windows\pytorch\caffe2\serialize\inline_container.cc:337] . unexpected pos 255835072 vs 255834960

rare gobletBOT
#

Ayo? @spare saffron level 1 !!! lfg

spare saffron
#

I have no clue what this means but Applio has worked for training models up until now.

opal saffron
#

How can I use ai on a Android?

modern oxide
opal saffron
#

Oh why nott

cunning gyro
#

[Voice Changer] generate new embedder. (no embedder)
[Voice Changer] Loading index...
Traceback (most recent call last):
File "voice_changer\VoiceChanger.py", line 301, in loadModel
File "voice_changer\RVC\RVC.py", line 84, in loadModel
File "voice_changer\RVC\RVC.py", line 153, in prepareModel
File "voice_changer\RVC\pipeline\PipelineGenerator.py", line 54, in createPipeline
UnboundLocalError: local variable 'inferencer' referenced before assignment

[Voice Changer] Model Load Error! Check your model is valid. local variable 'inferencer' referenced before assignment

#

help

errant spindle
#

how do i open the program i already have the files installed

errant spindle
#

rvc

modern oxide
#

Go-web.bat

errant spindle
rare gobletBOT
#

Ayo? @errant spindle level 2 !!! lfg

modern oxide
errant spindle
#

when i click run i go thru the firewall thing and it tells me to press key

#

and when i do it just close itself

modern oxide
#

Is go-web.bat inside the folder?

errant spindle
modern oxide
#

Hmm52634615_730845017316044_3880792

#

Did you install it from GitHub?

errant spindle
#

latest version

modern oxide
#

Have you installed requirements?

errant spindle
modern oxide
#

Like, runtime folder

#

You probably haven't

#

There should be a tutorial on git on how to install requirements

#

Or you can just download it from huggingface

modern oxide
#

If you download it from huggingface it already has all the requirements and you don't have to install them manually

errant spindle
#

when i try to run batch file it says press key and when i do it closes

modern oxide
#

That's weird

#

Can you send the screenshot of the folder?

errant spindle
modern oxide
#

Did you extract it into your disk C?

errant spindle
modern oxide
#

You were just launching it from archive?

errant spindle
modern oxide
modern oxide
#

No idea why that error happens

errant spindle
#

damn.

strange estuary
#

Hi guys can i make ai cover in my phone lol

modern oxide
strange estuary
#

Thanks btw

copper pebble
modern oxide
#

If yes then it's possible

copper pebble
#

ilaria rvc its having some troubles i heard

#

but maybe others work

copper pebble
modern oxide
#

Then I was a jun mod for 2 days

#

And then Ila promoted me xd

#

Because I found many racist/nazi profiles with blackcat

copper pebble
#

i see

modern oxide
# copper pebble i see

I hope they'll promote blackcat to a mod soon too, he worked way more in last 3 months than I did

outer isle
#

illaria not working

copper pebble
#

try to use other spaces

outer isle
tender basalt
#

my voice changer native client isn't opening

#

this number keep rising
last time it reach over 1000

sand iris
#

screenshot start of log

azure marshBOT
#

RVC Guides (How to Make AI Cover)

Documentation
🇺🇸 English (main)

Translation by country

🇧🇷 Brasil (PT-BR)
real tendon
#

I forgot to make index file while running on colab and the saves are all gone, except for .pth
is there a way to make index file from existing model?

proper shale
#

yea, just use the same settings (training variables on rvc disconnected), load up all the resuming steps (on rvc disconnected), then just train index

tender basalt
sand iris
#

Which version of voice-changer is that ? Also which CPU ?

#

We're trying to figure this one out, .16 introduced the issue, .16a supposedly fixed it

tender basalt
#

cpu i5 3470
version 1.5.3.16a

velvet beacon
#

Anyone pls share Google colab

velvet beacon
#

It's ok I got it for mobile

proper shale
#

that one is not gonna work

sand iris
tender basalt
sand iris
#

Do you have an AMD or Nvidia GPU ?

tender basalt
#

gtx 1060 3g

tender basalt
tender basalt
#

thx

shut sequoia
#

hi

floral vale
#

where do I download the app

rare gobletBOT
#

Ayo? @floral vale level 1 !!! lfg

floral vale
#

yea why not

floral vale
#

ok

floral vale
modern oxide
floral vale
#

where

floral vale
#

can't you just send it

modern oxide
#

literally all of these

#

are links of rvc

floral vale
#

oh it called RVC

floral vale
modern oxide
#

go-web.bat

floral vale
#

adminstrator?

#

it opened a website localhost ?

#

@modern oxide

modern oxide
floral vale
#

where do I setup my mic?

modern oxide
modern oxide
floral vale
modern oxide
#

no it's for training

floral vale
#

I want to use voice like minecraft villager and everyone hear me

modern oxide
#

why didn't you tell me that first 💀

floral vale
#

idk I downloaded a 5gb file bruh

#

but what should I do now?

modern oxide
#

download the other program called w-okada voice changer, i'll give you the link just tell me what gpu and cpu do you have

floral vale
#

wdym gpu cpu?

rare gobletBOT
#

Ayo? @floral vale level 2 !!! lfg

modern oxide
floral vale
#

Processor: 12th Gen Intel(R) Core(TM) i5-12400F 2.50 GHz

#

Grapichs Card: NVIDIA GeForce RTX 3060

#

Memory (RAM): 16.0 GB

#

System Type: 64-bit

#

is thats all
@modern oxide

modern oxide
#

download this and extract it the same way you did with rvc

floral vale
#

ok I take 5 minutes to download 3 GB file so Wait

floral vale
modern oxide
floral vale
#

I got so much files

#

there is start_http.bat and start_https.bat

#

what I click to start

modern oxide
#

no

#

not that

floral vale
dusk harbor
#

there is too many links on #1159949278270193734 which one do I need to download the voice changer? Also is the voice changer free?

sand iris
#

no

#

non

#

on

#

no

modern oxide
#

why

sand iris
#

the bat scripts are there for a reason

#

the mmvcserversio executable is just a python interpreter for the main file

floral vale
#

uhhhhhhhh

sand iris
#

you won't get the flags without running it through the bat scripts

modern oxide
floral vale
#

do I close this now it almost done?

sand iris
#

do not close it

floral vale
#

ok

sand iris
#

wouldn't be surprised either

rare gobletBOT
#

Ayo? @dusk harbor level 1 !!! lfg

sand iris
#

yeah the channels are confusing x)

floral vale
#

what is this

#

it opened that

dusk harbor
#

thank!

floral vale
# floral vale

@sand iris @modern oxide what I do it does not work I don't hear my self

dusk harbor
dusk harbor
#

gotcha, thanks!

sand iris
#

The code is even available

modern oxide
digital jay
#

smoke?

rare gobletBOT
#

Ayo? @digital jay level 1 !!! lfg

digital jay
#

twork

floral vale
dusk harbor
modern oxide
brittle wing
#

?

#

where does that mention anything about my issue

#

anyway i stopped the training cell and saved the model

#

seemed to go succesfully

#

gonna try again, hopefully it didnt break

#

uhhh does this look bad

#

smoothing at 0.987

proper shale
brittle wing
#

i stopped it yesterday

#

at epoch 175

proper shale
#

na what i meant was

#

i think its still not overtraining

brittle wing
#

uh yeah maybe

#

d/total and mel looks fine

#

and i think g is also fine

#

cuz

#

if i zoom in

#

it doesnt actually go up by a lot

#

ill wait until 200 epoch

#

oh and also any idea about this

#

does this mean itll disconnect (but not delete) runtime once the task is finished?

#

because when i stopped the task i just reconnected and the files from the runtime was still there

#

and it seems to have saved the model correctly

#

179 epoch

#

d/total is going down rapidly

#

mel and kl also

#

i guess its fine

violet heron
floral vale
rare gobletBOT
#

Ayo? @floral vale level 3 !!! lfg

floral vale
#

@modern oxide can you help me

violet heron
floral vale
brittle wing
#

@proper shale sorry for ping but this doesnt look right 😭

violet heron
#

You set okadas output as the cable

#

And the app you want to use as the cable

floral vale
violet heron
#

Actually not really

floral vale
#

wdym

violet heron
#

It’s needed when you want to use the voice changer on something like discord

floral vale
#

but how can I do it here

#

btw

#

I installed a villager sound but it does not sound like its what is wrong

violet heron
floral vale
#

download what

violet heron
#

Input on discord should be the virtual audio cable

violet heron
floral vale
#

I can't find that

#

there is only windows defualt and my mic

copper pebble
#

what settings did u put?

brittle wing
#

the rest i left as default

#

dataset is an hour long, consists of 6 .wav

copper pebble
brittle wing
#

this is what epoch 198 is looking like:

copper pebble
#

1h is wayy to much

brittle wing
#

do i have to start over-

copper pebble
#

yh u dont need 1h dataset for good models

brittle wing
copper pebble
#

even 5 minutes r good enough

brittle wing
#

would it be fine

copper pebble
#

if u wanna work with big dataset maybe u should try 15/max 20 min datasets

brittle wing
copper pebble
brittle wing
#

okay so

#

lets say i did

#

10 minutes

#

how long would training take

#

compared to an hour

copper pebble
#

once did 10 minutes 250 or 300 epochs and it took like 2 hrs

brittle wing
#

im training index right now

#

gonna try and see if it came out good or not

copper pebble
#

i think u need to retrain

#

with the new dataset

#

://

brittle wing
#

i have the 175 epoch model saved in drive

#

which was fine

#

in tensorboard at least

#

havent actually tried them yet

copper pebble
#

yh

#

lemme know how it turned out

brittle wing
#

rvc

#

rn

#

btw i have a ton of stuff now

#

in drive

copper pebble
brittle wing
#

what files do i need

brittle wing
copper pebble
brittle wing
#

do i need all

copper pebble
#

/the name of ur dataset

#

rvc disconnected has 4hrs runtime

brittle wing
copper pebble
#

with 1hrs and 450 epochs

brittle wing
#

uh

#

wat

#

im confused

copper pebble
brittle wing
#

what does "has 4 hours runtime" mean

#

is there like a limit on disconnected or something

copper pebble
copper pebble
#

u can run cells for just 4 hours

#

then it disconnect u

brittle wing
#

what happens after 4 hours?

#

does it like not work again on ur google account

#

and you have to make a new one

copper pebble
copper pebble
brittle wing
#

oh okay then

rare gobletBOT
#

Ayo? @brittle wing level 6 !!! lfg

copper pebble
#

but on that colab should be some resumes cells

brittle wing
#

isnt rvc disconnected like

#

recommended or something

brittle wing
copper pebble
#

kk

brittle wing
#

got them

#

finally

copper pebble
#

hope that it turned out well

brittle wing
#

im just slightly confused cuz i havent touched this stuff in a while

copper pebble
brittle wing
#

omg ilaria rvc

brittle wing
#

i have to upload it

#

to huggingface

#

first

#

aaaaaaaaaaaaaa

copper pebble
#

looks lil overtraing to me MJ

copper pebble
brittle wing
#

so

#

ok so

#

model is on huggingface yay

#

now

#

let me

#

test

copper pebble
brittle wing
#

wat

#

why isnt it working im confused

#

help im lost again 😭

#

what the fuck

#

sigh

#

im at the verge of giving up 😭

#

why is this so hard

copper pebble
#

did u refresh??

brittle wing
#

liike

#

bruh

#

why isnt it working

#

i refresh and my model doesnt appear

#

logs say it downloaded the thing

brittle wing
#

did

#

its not working

copper pebble
#

a

copper pebble
#

??

brittle wing
#

how am i supposed to do it ;-;

#

i paste in the link, i press the button

#

and then i go back to check, isnt there

copper pebble
brittle wing
#

oh wait

#

wat

#

i need to put it in a .zip...?

copper pebble
#

yh

brittle wing
#

this worked

copper pebble
#

i was abt to write it

brittle wing
copper pebble
brittle wing
#

a

#

sighhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh

#

fuck it whatever

#

this whole ai stuff is taking more time than expected 😭

#

i wish the guides were clearer on this...

#

a lot of things arent mentioned

copper pebble
half cove
brittle wing
#

someone fucking kill me

brittle wing
#

youre not the only one

half cove
brittle wing
#

if my model doesnt come out good imma be pissed LMAO

brittle wing
#

Thank you!

spare saffron
#

On Applio, when training, what should I put in the section "Provide the GPU indexes." should I just leave it at 0?

sand iris
#

Put it according to the listed GPUs below

#

Does it detect any ?

spare saffron
#

Oh, I see. I tend to look over things like that. It says the GPU I have is 0.

#

"0 NVIDIA GeForce RTX 3060"

sand iris
#

Put 0 then

spare saffron
#

Yeah, I did. Thanks!

#

Also, is there some sort of documentation on the settings I should use for my GPU? I'm just following the Training RVC V2 models guide made for Collab.

azure marshBOT
#

RVC Guides (How to Make AI Cover)

Documentation
🇺🇸 English (main)

Translation by country

🇧🇷 Brasil (PT-BR)
uncut field
#

anybody knows what the issue is here?

rare gobletBOT
#

Ayo? @uncut field level 2 !!! lfg

spare saffron
#

I've reinstalled Applio three times now and everytime it gets to Epoch 10 and tries to save at the checkpoint it gives me this error. Someone please help.

#

INFO:barnmetro:Saving model and optimizer state at epoch 10 to ./logs\barnmetro\G_2333333.pth
Process Process-1:
Traceback (most recent call last):
File "C:\Users\strin\Desktop\Applio-RVC-Fork\runtime\lib\site-packages\torch\serialization.py", line 441, in save
_save(obj, opened_zipfile, pickle_module, pickle_protocol)
File "C:\Users\strin\Desktop\Applio-RVC-Fork\runtime\lib\site-packages\torch\serialization.py", line 668, in _save
zip_file.write_record(name, storage.data_ptr(), num_bytes)
RuntimeError: [enforce fail at C:\actions-runner_work\pytorch\pytorch\builder\windows\pytorch\caffe2\serialize\inline_container.cc:476] . PytorchStreamWriter failed writing file data/2184: file write failed

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "multiprocessing\process.py", line 315, in _bootstrap
File "multiprocessing\process.py", line 108, in run
File "C:\Users\strin\Desktop\Applio-RVC-Fork\lib\infer\modules\train\train.py", line 359, in run
train_and_evaluate(
File "C:\Users\strin\Desktop\Applio-RVC-Fork\lib\infer\modules\train\train.py", line 656, in train_and_evaluate
utils.save_checkpoint(
File "C:\Users\strin\Desktop\Applio-RVC-Fork\lib\infer\infer_libs\train\utils.py", line 153, in save_checkpoint
torch.save(
File "C:\Users\strin\Desktop\Applio-RVC-Fork\runtime\lib\site-packages\torch\serialization.py", line 442, in save
return
File "C:\Users\strin\Desktop\Applio-RVC-Fork\runtime\lib\site-packages\torch\serialization.py", line 291, in exit
self.file_like.write_end_of_file()
RuntimeError: [enforce fail at C:\actions-runner_work\pytorch\pytorch\builder\windows\pytorch\caffe2\serialize\inline_container.cc:337] . unexpected pos 249166784 vs 249166676

solar plover
#

i aint reading all that

#

you might be out of space though

uncut field
#

so what to do now?

modern oxide
uncut field
#

yes i pressed the button to start training

#

and then that appeared

modern oxide
#

then just wait till it finishes

uncut field
#

oh thats it

#

?

modern oxide
#

yeah

uncut field
#

how do i know if its finished and where is gonna be my model saved?

modern oxide
#

if you're using disconnected you'll have to press the export cell, if you're using easy colab you'll find your model in your drive account

uncut field
#

yeah its the easy colab

modern oxide
uncut field
#

i found a folder called like the model i want to train

#

its gonna be there?

modern oxide
uncut field
#

hmm there is no weights folder 🤔

#

i am fuckin gblind

modern oxide
uncut field
#

how long can i expect for it to train a 50 epoch model?

modern oxide
#

50 epoch?

#

that's too low for a good model

uncut field
#

i set the save frequency to 50

modern oxide
#

like a couple minutes probably

uncut field
uncut field
modern oxide
winged fulcrum
#

whats the easiest way to train a voice model without installation on a pc?

primal barn
#

uhh colab?

#

ionno

winged fulcrum
#

alr

rare gobletBOT
#

Ayo? @winged fulcrum level 2 !!! lfg

brittle wing
#

-colab

azure marshBOT
#
All working colabs / 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.

uncut field
proper shale
#

I'd tell them to check KL

#

And Mel

obtuse glen
#

i need a good training notebook

rare gobletBOT
#

Ayo? @obtuse glen level 1 !!! lfg

proper shale
#

_ _

obtuse glen
#

same

proper shale
#

always the inanimate cartoon objects ppl fans with the inexplicable rage/j

proper shale
obtuse glen
obtuse glen
#

nvm

proper shale
exotic cliff
#

hi I'm new here, can someone help me to make my own models ?

obtuse glen
proper shale
#

run the export model step

#

they'll be on rvcDisconnected

obtuse glen
#

ok

exotic cliff
#

@proper shalethx

obtuse glen
#

help?

proper shale
#

G and D

obtuse glen
#

wait they have the ckpt in them?

proper shale
#

they are the ckpt

obtuse glen
rare gobletBOT
#

Ayo? @obtuse glen level 2 !!! lfg

obtuse glen
#

but it is one

#

so it doesnt load the file in the gui

uncut field
#

ahhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh

proper shale
#

use the amen break e90 s810 one

proper shale
#

D/G are training only

gleaming forum
#

Hello, I am not very informed about the RVC, in fact I do not know what version I have, what I do know is that it is going very badly in terms of ms everything sounds very robotic, I do not know if the version of the program you install is old or idk, I have rtx 4070 and hyperX quadcast.
help:( srry if I didn't explain, my English is not good.

dense dirge
#

ma

copper pebble
#

@brittle wing how did the model turn out??

rugged sage
#

I just installed RVC_GUI, I run the .bat file and the cmd prompt opens up with "Running with the runtime Python, please wait..." and nothing happens. How do I fix it?

gleaming forum
#

😦

copper pebble
rain coyote
#

hello guys, i need a model for my rihanna-diamonds ai cover.But whenever i try it with different models,it stays bad.Like do u guys have any recommendations for high note songs?

rain coyote
#

whenever i try to reach high notes,it's getting interrupted

rare gobletBOT
#

Ayo? @rain coyote level 1 !!! lfg

sand heart
#

Illaria RVC has quite literallynever worked for me

#

i dont know why it just never id

#

it takes forever to process

#

and if i refresh anything

#

i will get

#

nonstop errors

copper pebble
sand heart
#

is there any other ai cover service i can use

#

cuz i have tried illaria rvc on multiple occasions

copper pebble
#

if u didnt try to duplicate the space

sand heart
#

its never worked

#

how do i do that?

#

OH

#

IT FINALLY WORKEDD

#

AFTER SO LONG

tidal robin
#

I want to delete all of this how

molten pecan
#

I thought you were using RVC

tidal robin
#

whats better?

molten pecan
#

I'm not familiar with w-okada

tidal robin
#

what diffrend about rvc and wokada

#

?

molten pecan
#

Try clicking that grey Edit button

tidal robin
#

This only pops up ik

molten pecan
molten pecan
tidal robin
#

i found out\

#

its in the fdolder

molten pecan
#

Congrats!

tidal robin
#

xD ty

molten pecan
tidal robin
#

Works! ty for ur time tho!@

prime hemlock
#

HOSTING INDUCTIVEGRUB HELP DESK

regal umbra
#

@half cove

#

does anyone know what this file is?

half cove
#

use added

regal umbra
#

what do you mean

half cove
#

wait is that the only files in the model?

#

should have a pth

regal umbra
#

yeah

#

@modern oxide could you tell me what it is please, i'm worried

half cove
#

just use the pth and the index

regal umbra
#

there's no pth tho, it's just those 2 files

half cove
regal umbra
#

oh

half cove
regal umbra
#

i mean my gf is the one using the program but she's not in here, and we're worried about that file before anything else

#

@mighty parcel do you maybe know what it is?

molten pecan
half cove
molten pecan
#

You are going to get banned

regal umbra
#

what

molten pecan
#

No sex life allowed here

regal umbra
#

was i not supposed to ping one of the mods for help, i didn't mean to break any rules

mighty parcel
mighty parcel
#

its harmless, and whoever left it in is just dumb

#

no need to fear over it, it's just a log file leftover, and is what is read when you want to see the graphs for the training stats

regal umbra
#

thank you!!

mighty parcel
#

people are just goofy and upload files they dont need to

meager dome
#

I just can't find any index file

mighty parcel
#

you have to click the train index button, it is not ran when clicking train model

meager dome
#

wait what

rare gobletBOT
#

Ayo? @meager dome level 1 !!! lfg

meager dome
#

😭

#

where is the train index button

proper shale
#

in the GUI?

brittle wing
molten pecan
brittle wing
meager dome
#

ok I found where, but how do I upload the trained model files?

#

I trained it a month ago and forgot to get the index file

proper shale
proper sorrel
#

hey uh this is probably a stupid question but how do you train models-

proper shale
#

mainly just name and rvc version and sample rate

azure marshBOT
#

Hey, @proper sorrel!

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

Documentation
🇺🇸 English (main)

Translation by country

🇧🇷 Brasil (PT-BR)
molten pecan
proper shale
molten pecan
proper sorrel
#

i was tryna follow this and its weirdly confusing to me-

#

my brain is smol

proper shale
#

It's ok - first time is always a lil bit complicated haha

proper sorrel
#

Well i used to do it on the webui!

molten pecan
proper sorrel
#

but yknow-

proper shale
meager dome
#

I don't remember a thing from the last time I trained it

proper shale
meager dome
#

like my question is where do I put the trained model's files before making the settings for getting the index file

proper shale
#

oh ur using...

proper sorrel
proper shale
#

if this works to train index then ill be surprised

meager dome
#

it's asking for the path

proper shale
meager dome
proper sorrel
#

Are there any suggestions on what i should use to train a model?

rare gobletBOT
#

Ayo? @proper sorrel level 1 !!! lfg

proper shale
#

like extracted the zip

proper shale
#

that one actually works

meager dome
#

so the zip has to be extracted

#

ok

proper shale
#

if you've already done this step the path should be /content/dataset/

meager dome
#

I didn't do that step so I don't see it

proper shale
#

yeah you should've

meager dome
#

can I do it now?

rare gobletBOT
#

Ayo? @meager dome level 2 !!! lfg

meager dome
#

like the model's trained, I just don't have the index file

#

I trained the model one month ago

proper shale
#

yeah you can do it

#

make sure everythings correct

#

just in case

#

and it should work

meager dome
#

ok then how do I do it

proper shale
#

now just... train index

#

if you have everything correctly set up it shouid work

meager dome
#

I don't have everything correctly cuz it trains the index in 1.5 seconds

proper shale
#

oookay that definitely isnt normal

meager dome
#

yeah

proper shale
#

...or maybe ur dataset is really short

#

highly unlikely tho

meager dome
#

it s one hour

#

the problem is I don't know where to upload the dataset

proper shale
#

to the dataset folder in ur google drive

#

if ur still using that colab

meager dome
#

im still using colab, yeah

#

yeah but it should be a zip file or?

proper shale
#

zip yeah

obtuse glen
meager dome
#

look, this is the zip, it has only wam files

errant spindle
meager dome
#

now I put the path here

#

but the train index doesn't work

#

the train model does

#

it's training it again

proper shale
#

delete spaces

meager dome
#

it doesn't

tame mica
#

"baza de date"?

meager dome
meager dome
#

that's how I named it

proper shale
#

yeah basically

#

dont use spaces in dataset zips

meager dome
#

OH

#

I renamed it

#

now it gives me some error

proper shale
#

oooh no wait

meager dome
meager dome
proper shale
#

you need to extract it first

#

using the extract dataset stwp

#

stop the webui for a while and do that

#
  • make the 'dataset' folder in Google drive and move your zip there
serene tree
#

yo anyone help me download the voice changer lol

meager dome
#

ok, bear with me for one second

#

so I made the folder and I'll move the zip here

#

ok, I uploaded it

#

I asume I have to run this cell

rare gobletBOT
#

Ayo? @meager dome level 3 !!! lfg

proper shale
meager dome
#

ok, now?

#

I still get some error

#

maybe I should rerun the cell with gradio

#

that's what I did

proper shale
#

asks for the folder and not the file, btw

meager dome
#

you re so right

#

now another error

proper shale
#

that

meager dome
#

or maybe it isn t an error

proper shale
#

is not an error

meager dome
#

oh yeah

#

I thought abt that

#

😭

proper shale
#

it's ok it's ok

meager dome
#

is this an error?

proper shale
#

i mean

#

no

#

wait i

#

no-f0-todo might be a problem

meager dome
#

i think it is cause it still won't train the index

proper shale
#

did you check for the model to have pitch guidance?

meager dome
#

is it that important?

proper shale
#

tick that if you haven't

proper shale
meager dome
#

where should I tick

proper shale
#

its that broken 😭

proper shale
meager dome
#

it was ticked

proper shale
#

oh then probably the wavs u have are screwing up

meager dome
#

the model was trained just fine

#

why would the index not be the same

proper shale
#

should've done it right there

#

when you could

meager dome
#

you're right

#

maybe that's the problem

#

wait I just saw something

#

this is the right training folder path

#

cause look

#

or nevermind

#

don't mind it

proper shale
#

i found it out now

#

it should be /content/dataset only

meager dome
#

yeah I put it like that

#

I unticked the

#

and now I am training the model all over again

proper shale
meager dome
#

if I stop the training and click on train feature index it still doesn't work

#

this is so hard

azure marshBOT
#

RVC Guides (How to Make AI Cover)

Documentation
🇺🇸 English (main)

Translation by country

🇧🇷 Brasil (PT-BR)
errant shadow
#

Yo @cobalt oyster

gloomy smelt
#

quick question are these settings good for covers with 300 epochs rmvpe models

#

With clean audios n stuff

cerulean dew
#

RVC GUI keeps giving me this error when im trying to convert things how do I fix it?

#

looks like this on the GUI itself

hexed dew
#

whats the ideal number of epochs to use for a singular wav file?

rare gobletBOT
#

Ayo? @hexed dew level 1 !!! lfg

cerulean dew
hexed dew
#

around 2 mins but the sample has almost all phoentic sounds

cerulean dew
#

Probably go for somewhere between 250-300

hexed dew
#

I dont want to overfit the model

#

Ah okay it was set to 500 at default am I just able to use the checkpoints at 250?

daring dust
#

theres is any way to make voice with AI with mobile?

cerulean dew
hexed dew
cerulean dew
#

idk ive never tried interrupting the training to do that

violet heron
#

Use google colab, I recommend the RVC Disconnected colab

lyric agate
violet heron
azure marshBOT
#

Suggestions for @lyric agate

All working colabs / 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.

lyric agate
hexed dew
#

Hi I trained a model but when I download it from the colab link and load it into rvc gui I get this:

cerulean dew
hexed dew
#

I used another model but it worked fine

#

Just the model I made just now has the issue

cerulean dew
#

i dont think thats what causes it tho bc I had this issue on a model that was working fine just the other day

hexed dew
#

When I try load the model in it says "Please select a model and input audio file"

#

I think there might be an issue with the actual zip file itself

cerulean dew
#

hmm i just tried my different models, all of which have worked at least once, some i have made at least 5 covers with and none of them r working when im having this error

#

so maybe we r experiencing a different error

hexed dew
#

For a model youve trained can I see the pathing?

cerulean dew
#

mine seems to be related to available space, but i dont know how to clear the space necessary to continue making covers

hexed dew
#

Yeah Im unsure how to delete models on RVC gui either

cerulean dew
#

show u the pathing

hexed dew
#

just send me a screenshot of the folder path and contents

#

Maybe because the zip file has a folder in itself

rare gobletBOT
#

Ayo? @hexed dew level 2 !!! lfg

hexed dew
#

Im not sure

cerulean dew
rare gobletBOT
#

Ayo? @cerulean dew level 4 !!! lfg

cerulean dew
#

idk i literally got this model to work yesterday

#

and there were no issues

#

but now rvc gui is just being stubborn

#

and wont let me do literally anything

hexed dew
#

how did you create the model? using rvc disconeected or

cerulean dew
#

yes

hexed dew
#

how many epochs did you use

cerulean dew
#

300

hexed dew
#

but I cant even load my model in

cerulean dew
#

i think i might have fixed it?

#

i still dont know exactly what the problem is but i accidentally have a ton of different version of the rvc folder downloaded and i think only one of them works

#

idk why thats the case

#

torch.cuda out of memory

#

thats what the code says

#

idk how to fix that

hexed dew
#

thats probably a gpu problem

blissful pawn
#

how do i run the models?

#

sorry if its a dumb question

cerulean dew
#

i keep seeing that online but i have no idea how im meant to fix that

hexed dew
#

try to run on cpu instead of gpu

cerulean dew
#

how?

hexed dew
#

on rvc gui