✨│ai-help

[brittle wing]

cheers man

this was really weird

[quasi dagger]

Indeed

Good luck training!

[brittle wing]

thank you!

[pallid cedar]

Damn

Applio is down

After downloading the model, a message "Error Connecting Time blah blah" appears

Anybody know?

[ancient needle]

rror opening Stream: Illegal combination of I/O devices [PaErrorCode -9993] what do this mean

[ancient needle]

This been happening since I got it and its annoying as shit

[rare goblet]

Ayo? @ancient needle level 1 !!!

[surreal tendon]

I'm trying to train something with Applio and some sort of error comes up. Everytime i go to extract features, something is spat out in the console (about like, threads and hop-lengths?) and multiple elements crash, forcing me to terminate, messing with my graphics driver and other apps like discord (ping me when you reply so i see it lol)

[rare goblet]

Ayo? @surreal tendon level 1 !!!

[surreal tendon]

i'm 99% sure Applio crashes after this because nothing happens in the console

It's odd because I managed to train another ai model earlier... but now it just doesnt work

[unique mantle]

I have a problem that in TernosBar there is only 1 dot at grath and its doesn't update. It may work with low chance. All I found was about code and nothing from client side. Any solution to make it work always. Maybe some file is not created?

[charred pendant]

Getting Error below when trying to do step 2

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

[proper shale]

use the colabs in https://docs.aihub.wtf/ - that one is outdated

[charred pendant]

bet

[lapis wraith]

Can sombody help me I got the voice changer on and stuff but I keep hearing myself

How do I turn that off

[odd shale]

-realtime

[azure marsh]

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

[odd shale]

Read the rentry guide.

[quartz ridge]

Hello can someone help me setup this voice model?

Lilas Ikuta (SnowieV3.1 JP-HuBERT) (1,290 Epochs 96,750 Steps)

I think I need to download this for it to work

https://huggingface.co/MUSTAR/Snowie_V3.1_JP_hubert/tree/main

https://discord.com/channels/1159260121998827560/1231585086668017838

Im trying to upload it to Voice.ai

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[barren canopy]

Everything still looking good?

[crystal gull]

🤔 Is all this info valid for a gpu like a 4070

Is there better tactics per gpu?

[barren canopy]

Choose Local if you have a 4070

But no tactics don't really change, you'll just be able to converge the model faster

[crystal gull]

Alright so I should go along with following christopher's video then ey?

[barren canopy]

No

Use Applio instead

And then follow Christopher's video

[crystal gull]

Gotcha

[barren canopy]

Mangio seems to be outdated & gives a ton of issues randomly

[crystal gull]

Just applio on github?

Believe I've found the right thing

[barren canopy]

Yes

[crystal gull]

What is "RMVPE" and why do voice creators flex that they used that in the process of making it?

Haven't heard of it but from a few people

[barren canopy]

https://arxiv.org/abs/2306.15412

[crystal gull]

Ah

Gotcha thank you

[barren canopy]

But tl;dr

RMVPE outperformed other methods in terms of accurately figuring out a speakers pitch, even when there's background noise messing things up. So, it's like a super reliable way to get the right notes from a song, even when it's a bit noisy

[crystal gull]

I'm not too involved in the ai voice aspect, how long would it take for something like 500 epochs on a 4070?

[rare goblet]

Ayo? @crystal gull level 4 !!!

[crystal gull]

Ofc obviously most likely depends on a lot of different factors

[barren canopy]

Depends on dataset size mainly

I have 20 mins, 3080 and 200 epochs takes approximately 100 minutes (30 secs per epoch)

[crystal gull]

Oh that's not bad, what would you suggest to do for finding datasets? Are there specific places you'd look, or would you just try to find long talking videos on youtube and cut it up from there

[crystal gull]

I think I'll give it a shot with a 15m data set, it's relatively clean. And has 1 single voice, should I split it up into parts or could I just keep it as a single

[barren canopy]

In my experience it doesn’t affect performance of single audio versus batch of audios, just make sure there’s no abrupt cutoff mid sentences

[crystal gull]

pffff okay so I should go through it and trim it to just be words

and no cuts

owwie rvc is fucking destroying my ram

Oh shit.. This is actually fire, this is only 20 epochs and I didn't trim any of the background out or cuts in speaking

Probably explains the background fizzling

[proud elbow]

depend on dataset length and batch size

[crystal gull]

Right

[quartz ridge]

Can u guys help me?

[rare goblet]

Ayo? @quartz ridge level 1 !!!

[crystal gull]

https://dontasktoask.com

[proud elbow]

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[crystal gull]

Ah you've already got a command ey, nice.

[quartz ridge]

Welp im new here so I have no idea

I dont mean to be annoying or rude

[quartz ridge]

But yeah basically I just cant figure out how to get this voice model to work on voice.ai

[crystal gull]

No problem just don't ask to ask, spit your question out outright

[proud elbow]

just ask something specifically, even chatGPT wont give you helpful answer unless u do it

[crystal gull]

Oh ffs ew, my voice im training sounds ok through a female tts but through my voice or a male tts it sounds horrible

[proud elbow]

voice.ai is paywalled and not recommended over RVC forks

[crystal gull]

^

[quartz ridge]

Oh I see

Then what should I use then?

Im just trying to use that model for singing

[surreal tendon]

Learnt I can send images now (probably because I leveled up lol) so I'm gonna append to this issue a screenshot of my applio console to be more helpful and the first process that crashes (there's like six)

[proud elbow]

links and guides right there https://docs.aihub.wtf/

[quartz ridge]

Alright thanks

[proud elbow]

if u know how to code, find a code in an "extractf0" file if not wrong, and change code that sets "num_of_processes" to 2 or 4 instead of all of the cpu cores

[surreal tendon]

ill try that right away
i'll find the file eventually lol

that last line there, yeah?

[proud elbow]

yeah right

[surreal tendon]

done, i'll try running it again

there we go! that seemed to work!
it doesnt throw up errors anymore!

ayyyy!

thank you so much!

[proud elbow]

yeah it also happened on a pc with Core 14700 and surprisingly someone's with ryzen 7950x when trying to f0 with crepe+hop length

[crystal gull]

C:\Users\BumiMandias\Downloads\Applio-3.1.1\logs\Mary | epoch=237 | step=4977 | time=20:37:36 | training_speed=0:00:07 | lowest_value=23.31536865234375 (epoch 195 and step 4085)

pretty bad dataset but not bad timing ey

[proud elbow]

and changing num_processes made it work

[surreal tendon]

absolute wizardry but so long as it worked lol

i got an 11th gen i5-11400H
although it worked for my first model so i was very confused

[quartz ridge]

So im using Applio but im not sure how to do this:
How to use: 1 Download G and D pth files from Huggins 2 Place it in the pretrained_v2\custom pretrained folder if you have Applio, upload it directly (but you can also send it to the folder) 3 Paste the path to the pretrain in the fork (standard fork or mangio fork) 4 You're ready to burn!
@proud elbow

[gentle drum]

can someone direct me to the most simple model for me to convert voice to voice and make a cover of my song in a celebrities voice

[proud elbow]

RTX laptop? it wont be capable of training unless with 6+ gb vram

[surreal tendon]

yeah i have an RTX 3050

no clue if it's the 6bg or 8gb version but y'know

[rare goblet]

Ayo? @surreal tendon level 2 !!!

[thin stump]

@open fossil We have guides here https://docs.aihub.wtf/essentials/how-to-make-ai-cover/

I recommend IlariaRVC or weights.gg since that's on the cloud. I personally use tryreplay.io

[gentle drum]

thank you! i'm literlly the most beginner you could possibly be, but i'm trying to figure this out so your help is greatly appreciated lol

[rare goblet]

Ayo? @gentle drum level 1 !!!

[gentle drum]

lol

[thin stump]

np, all the voice models are free to download on weights

but yeah we can't really help you out with mixing since most of it is self-taught https://www.youtube.com/watch?v=rSxLF3kkBw0

[gentle drum]

is it possible for me to make my own rvc model? I want to do beyoncé, but i can’t find one I like and I kind of want it to be modeled on a recent song of hers

[thin stump]

It's possible to make it for free https://docs.aihub.wtf/essentials/how-to-make-voice-models/. Most of us use RVC disconnected to train https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/#rvc-disconnected

And if you don't want to go through all that, then you can choose a model master to work with in#1191429836321849435. Paid of course

[barren canopy]

I've been listening to it for too long, can someone tell me which version sounds the best?

[grim bay]

Should I remove reverb or noise first on the separating vocals?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[grim bay]

Which one is better? Denoise then deverb or vice versa?

[thin stump]

yeah it's too complicated to explain to a person without scaring them off. The right answer is to lurk more. Not everyone is like that though

[rare goblet]

Ayo? @brittle wing level 19 !!!

[thin stump]

so i hit them with the default. I don't like it either

[hot seal]

guys if trained a model using crepe, and then used it on W okada should i put the F0 Det to crepe tho or waht?

[obtuse stream]

anyone help me how to ubdate the new version?

[grim bay]

Should I remove reverb or noise first on the separating vocals?

[tough fiber]

I have 5 audio files, each 30-40 minutes long. Should I merge them and train on the combined file, or should I place them in a data folder separately and start training?

@quick jungle can u help me

[quick jungle]

Watchu want? :3

[tough fiber]

that was bot lol

@zinc anchorryanz can u help me

no one onlie

[fluid matrix]

[fluid matrix]

if you are sure that you finished editing your dataset, then it doesnt matter

either have all the files in the same folder or combine them into one file in that same folder

but make sure the folder itself only includes your dataset, nothing else

[tough fiber]

oh okey. i saw somewhere It said it would be better if you keep them separately, so I wasn't sure and wanted to ask.

also which batch size should i choose

16 is good

[fluid matrix]

what are you using to train

[tough fiber]

mangio rvc fork

[fluid matrix]

colab? locally?

[tough fiber]

oh

paperspace

a4000, p5000 , rtx 4000 or rtx 5000

gpus

lets said locally

[fluid matrix]

14 i think

but that works

[tough fiber]

bet thanks <3

[fluid matrix]

https://docs.google.com/document/d/1lIAK4Y0ylash_1M2UTTL_tfA3_mEzP0D2kjX2A3rfSY/edit#heading=h.noo2awd6f570

make sure to read that to be sure

good starting point

[tough fiber]

yea iknow already they didnt mention batch size or smt

[fluid matrix]

pretty sure it did

yeah its in the settings

14

should be the same, but why do you want to convert it to be 48k

[tough fiber]

oh yea i just saw the picture

[fluid matrix]

also that depends on what sample rate you choose when training

yeah it should be the same sample rates

thats why im asking

idk if it makes sense to upscale like that if you are going to train with 32k for example

then you should stick to that sample rate, but tbh 48k is not that great in my experience

not a lot of benefit

well you want to train in 48k

but your audio should actually be 48k

check with spek or rx idk

whatever you use

for example if most of the frequencies are at about 20k then use 40k and not 48k

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[grim bay]

-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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[proven jasper]

-ai humanizer

[low shard]

?

[sacred gull]

Heya guys, how to install (which folder) npy file from model pack?

[stiff goblet]

Applio Fork has full of bugs and glitches. Mangio RVC Fork still the best

[opal rune]

anyone admin can help me for make myself sound plugin ? Pls pm to me, cause ı cant do plugin with guide video.

[tame mica]

?

[proper shale]

https://docs.aihub.wtf/ has guides on model creation, strongly recommend reading it

[lament briar]

I am using the Mangio RVC Fork and try to seperate the vocal but as i click on convert it just continues good but then as i am watching in the folder where the finished file should be, i cant find it anywhere

can somebody help me

[opal rune]

re u have any guide video for this link

[tame mica]

don't use the pre installed vocal isolator on rvc. use mvsep or uvr instead

[lament briar]

do you have a installation link?

[rare goblet]

Ayo? @lament briar level 1 !!!

[lament briar]

for the app

[proper shale]

No. The link includes written guides

[opal rune]

thanks a lot

[rare goblet]

Ayo? @opal rune level 2 !!!

[tame mica]

mvsep.com or github.com/Anjok07/ultimatevocalremovergui

[lament briar]

thanks bro

[gilded peak]

Is there anyone here who can make a jupyter notebook to train rvc models because trainning on colab has some problema. if there was a jupyternotebook where we could do the things we do in the rvc2 disconnected notebook. for the jupyter session I am trying to use lighting.ai.

[rare goblet]

Ayo? @gilded peak level 1 !!!

[dusky zealot]

Anyone know why this error might be happening when I click on convert audio? (I'm using Ilaria RVC)

[fluid matrix]

u sure the audio path is correct and the cmd is still open

it should tell you whats wrong

[cold hamlet]

best colab (with ui) for covers?

not training

[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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[cold hamlet]

aight thanks

[rare goblet]

Ayo? @cold hamlet level 1 !!!

[cold hamlet]

aight

i cant download a model

[marble forum]

Best settings for normal rvc?

[cold hamlet]

i even tried from public index

[proud elbow]

google drive doesnt give direct link. try the huggingface link instead

[cold hamlet]

where do i ge the huggingface for certain models from

i get the link from the google sheet

[proud elbow]

download it locally then reupload to your huggingface account

[cold hamlet]

alright

how long is this gonna take

😭

[rare goblet]

Ayo? @cold hamlet level 2 !!!

[cold hamlet]

like

upload as a file not a link?

tf is this now

appolio colab works with no errors

why is it giving me gpu errors

i havent even used it today

what the hell

no

whats t4

[proud elbow]

colab gpu (see the status in top right)

[cold hamlet]

i cant find it

[proud elbow]

well, click menu Runtime -> change runtime type -> select T4 or make sure of it, then Disconnect and delete runtime and start again

[cold hamlet]

yes yes its connected

it worked

i tried my alt

[cold hamlet]

how do i upload it on my hugging face account.

[dusty forum]

why do i keep getting this error:

JSONDecodeError Traceback (most recent call last)
<ipython-input-16-c61cad373f60> 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)

[odd shale]

Which colab are you using?

[dusty forum]

v2

[rare goblet]

Ayo? @dusty forum level 1 !!!

[odd shale]

I mean, give a link.

Are you trying to do inference with a model?

[dusty forum]

https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=Y13Eh9r_g8f-

[dusty forum]

im just trying to download it

[odd shale]

Use Applio colab instead.

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb

First execute the Start Applio cell first and then Install Applio one and click on the public link

[rough tinsel]

umm i wanted to make a singer's 1000 epoch zip file model
can i get to know how do i do it?

[odd shale]

Do you wanted to train a model to 1000 epochs?

Umm.. Bud, more epochs doesn't mean more quality.

[rough tinsel]

yes

then?

[odd shale]

Depending on your dataset, you must check the Tensorboard to know when your model overtrains.

[rough tinsel]

huh

[pastel oak]

high epoch number doesnt mean model is good

[rough tinsel]

so how to make it good?

[pastel oak]

check pinned message, open docs guide and on extra resources it explains tensorboard

[rough tinsel]

including more datasets into it?

[pastel oak]

nah

open the guide and do that thing i wrote

[rough tinsel]

umm can you mention it?

[odd shale]

Give him the docs.

[odd shale]

Having a clean dataset and avoid overtraining it.

[rough tinsel]

oh

[pastel oak]

https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/

[rough tinsel]

so taking voice from songs will make it good?

[odd shale]

If you know how to isolate the vocals properly and how to delete reverb, backing vocals and harmonies.

Then yes.

[rough tinsel]

i think i should ... be expert in that first

will do that

thanks dea for the docs and info

[odd shale]

You're welcome.

[hidden dew]

How do I determine the number of epochs my model should have when training?

[odd shale]

You can't.

[hidden dew]

???

So how do I know how to set the number... do I just test it until its good

[cold hamlet]

tf

how t

[odd shale]

I mean, just put for example, 500 epochs with 15-20 epochs for each save point and check tensorboard, when you see it's overtraining, stop training and go for the save point that matches with the lowest point or gets near it

[hidden dew]

ahh alright tysm

i never learned to use tensorboard

[cold hamlet]

what should i do

[hidden dew]

is there an info sheet

[odd shale]

https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/

Use this doc

Read it.

[hidden dew]

thanks

[odd shale]

Maybe the model u were trying to download wasn't packed properly or it's something to do with the name.

Including you must change "blob" for "resolve" on the model link

[cold hamlet]

ok i did

works thanks

[odd shale]

You're welcome buddy.

[hidden dew]

Is there somewhere in tis server i can post screenies

[cold hamlet]

here

u can

[hidden dew]

nvm

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[ancient needle]

why do i keep getting error9993 when nothing i downloaded is illegal???

[low shard]

What rvc are u using wdym

[ancient needle]

rvc1006nvidia

[ancient needle]

sounddevice.PortAudioError: Error opening Stream: Illegal combination of I/O devices [PaErrorCode -9993] this error

[low shard]

Sorry but i honestly dont know much about local and never seen that error before

[strange quest]

is applio down?

[errant plume]

@odd shale hey how many epoches are recommended for1hour dataset?

[odd shale]

There's no specific epoch amount for X length of dataset.

You must check tensorboard to see when to stop due to overtraining

[errant plume]

okay

thx for the info

[strange quest]

nvm she was just having a moment

[ancient needle]

damn, thanks tho

[near zealot]

Could someone help me, I get this error in Google colab
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
chex 0.1.86 requires numpy>=1.24.1, but you have numpy 1.23.5 which is incompatible.
pandas-stubs 2.0.3.230814 requires numpy>=1.25.0; python_version >= "3.9", but you have numpy 1.23.5 which is incompatible.
torchtext 0.17.1 requires torch==2.2.1, but you have torch 2.0.1 which is incompatible.
torchvision 0.17.1+cu121 requires torch==2.2.1, but you have torch 2.0.1 which is incompatible.

Successfully installed all packages!

[brittle wing]

Make one of singer Paulo Londra in Spanish xf

[left scaffold]

What fix? Connection errored out.?

[sturdy sleet]

where can i find applio for amd

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[devout marlin]

so, I am installing rvc v2 but when I do the second step I get this error, what can I do?

JSONDecodeError                           Traceback (most recent call last)
<ipython-input-5-29d1edf57ac0> 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)```

[upbeat void]

me2

[devout marlin]

wdym

[upbeat void]

me too

[devout marlin]

oh, k

[rare goblet]

Ayo? @devout marlin level 1 !!!

[thin stump]

Outdated colab https://docs.aihub.wtf/

@alpine breach https://colab.research.google.com/signup

[devout marlin]

ty

[gloomy zephyr]

so, the inference list is empty when I click on it

any fix?

[gloomy zephyr]

anyone have any idea what a dataset should look like when training? is it a single file, or multiple audio files each of 5-10s clips?

[frosty shard]

hey all, im getting a json error when I plug in the link to my voice model and run it inside the google collab

any help?

[frosty shard]

can u resend it? im being taken to a sign up page

[thin stump]

you can search up colab credits on google and it's the first result

I use pay as you go because I don't make that much models

[frosty shard]

so what link am i supposed to go to for the non outdated collab?

[thin stump]

https://docs.aihub.wtf/essentials/how-to-make-ai-cover/ either use the cloud or local

[frosty shard]

thank you

[thin stump]

I use tryreplay.io because it all looks confusing to me. Just preferences

[gloomy zephyr]

does anyone know how to extract seperate vocals from a duet song? YT doesn’t have any tutorials on that, only vocals and instrumentals

[rare goblet]

Ayo? @gloomy zephyr level 1 !!!

[frosty shard]

has anyone made a cover of bury the light?

[thin stump]

#🔍│help-w-okada

I meant post your question there or #1192011222023950368 so it'll be less likely to be ignored

and rvc is retrieval vocal conversion, which is just the same thing as Okada except it's realtime

[empty wigeon]

Why dont i get the "running on Public URL" once it's done downloading in the "download the model" section ? I tried the one with Drake and pressed the little download button but it turns red and says that it has been executed. Then i press download on Step 3 but i dont get the link. What is happening?

Im using this one https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing

And this is the error i got in Step 2: Download the model

Any suggestions?

[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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[empty wigeon]

Yes. Found that link in a youtube video

[hot seal]

guys dose mangio-crepe 1 hop length worth for it?, and how it mades a change in the model

[empty wigeon]

What does this mean? It automatic created a folder called "rvcDisconnected" inside my Google drive with the folder "Experiment_name" in it. But i still get an error?

Now it gives me an error at this one. I've downloaded everthing all the way down so far

[proud elbow]

are you trying to resume from preprocessed files before or start over?

[empty wigeon]

i dont have that one

[rare goblet]

Ayo? @empty wigeon level 2 !!!

[empty wigeon]

I used a model from this discord. A trained model i think with the huggin link

Not in my google drive no

im basically just trying to get another voice on a track that i've seperated vocals and instruments. You know, switch the voice with another one

Oh... Im really trying to learn this stuff thats why im confused about what to use. So i should be using the last link u sent?

[proud elbow]

so I didn't realize that you wanted to use an existing model to infer, not your own dataset to train

[empty wigeon]

Im sorry. Thats my fault. Thought you all know what i meant xD

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[empty wigeon]

I got the public url and uploaded my audio file with vocals only, and got the model as well but when i click on "generate" it only downloads like 50% and then gives me this error ?

Uploading the zip file and press "upload model" but i keep getting this error

Colab link?

What part of it?

Nvm. I dont know how but it seems like it worked

[lapis tinsel]

Hi everyone

I created a model, but it says that an error occurred when I could not create an index. Is an index necessary or not necessary for my model?

Please help me

Who is active?

🥺😭

[crude bolt]

index is not necassery

load the pth and try if it works

[gloomy zephyr]

a big problem I have right now with the models I have is that the vocal ranges aren't very flexible, it can't go into the low octaves or into falsetto

is there anything I can do to not make the voice crack? by transposing the pitch of the voice, or am I out of luck if I don't have some high/low pitched clips of the character available?

[timber narwhal]

hi, what is the best rvc model for male to female voice?

or should i create a model myself or what?

[gloomy zephyr]

you use the model of the voice you want to change to

and then adjust the pitch if it doesn't sound right

[timber narwhal]

I'm not looking for a specific character's voice, im just looking for a female voice. Which model I should use in that case?

[gloomy zephyr]

there should be some default ones

[timber narwhal]

I've heard that default ones are very bad

So do people use the default ones in general?

[rare goblet]

Ayo? @timber narwhal level 1 !!!

[gloomy zephyr]

uh

I'm not too sure in terms of voice changer

but I do do voice cloner

[timber narwhal]

oh what's that

[gloomy zephyr]

you can use it to change the voice of dialogue/make AI covers

[timber narwhal]

sorry am i in the wrong channel

I want to change my voice to a female voice to troll people but i don't want it to be so obvious like anime girl sound

[thin stump]

#🧬│ai-chat message and we have free e-girl samples in #1191429836321849435 . They aren't necessarily bad, probably generic

[timber narwhal]

thank you

[thin stump]

from Ilaria or MRmodz

[gloomy zephyr]

can someone help now

I need help or advice

either is appreciated

[thin stump]

you can't fix it, it's like that because it wasn't trained to do high ranges in the dataset

[gloomy zephyr]

well I'm training my own

[rare goblet]

Ayo? @gloomy zephyr level 2 !!!

[gloomy zephyr]

my question is

do I have to add clips with a higher range

or can I manipulate the clips I have

to a higher pitch

and bypass the restriction of having to search for ridiculously high pitched clips

[timber narwhal]

I couldn't find any free models from mrmodz

[thin stump]

Don't manipulate clips since that can degrade the quality (when training the model) You want to keep it natural where it's not considered screaming. But if you're talking about post production mixing, then do whatever you want

[thin stump]

check ilarias shop

[gloomy zephyr]

the problem is that the character I'm trying to train is from a game

and there are limited audio clips

and the voice only goes so high

which is a little troublesome in terms of creating the dataset

[thin stump]

Yup, it's like that. You can try to find more samples from the same VA who uses the same voice for every character they voice

[gloomy zephyr]

I think that'll just result in multiple voices

because VAs do have different personality for different characters

[thin stump]

in some cases. But I think rvc is just trained to identify soft and high pitches including the accent https://discord.com/channels/1159260121998827560/1227927569890803713

there's no perfect model imo with rvc v2

even model masters can make 💩 once in a while

[gloomy zephyr]

okay, I've found a second character that's by the same VA

and this is perfect because this character's range is much lower

except

it sounds so different

let me just

[thin stump]

you'll have to review some of it then

[gloomy zephyr]

hmmmm

I think you can hear the difference

I'm not sure if I can use this

is this fine?

[thin stump]

Nope, two different voices

[gloomy zephyr]

yeah that's what I thought

sigh

so what else can I do here

try not to get out of the range of the model?

[thin stump]

use a different song to do covers with? Unless that has to be one you want to use

[gloomy zephyr]

this song is one of the hardest to do because it goes from low to high between like 3 octaves

which is why I wanted to do it so bad

cause it would be epic

to actually pull off

[full fossil]

hey i got this error:

TypeError: Cannot read properties of null (reading 'path')
TypeError: Cannot read properties of null (reading 'path')
at http://127.0.0.1:18888/index.js:2:3246075
at m (http://127.0.0.1:18888/index.js:2:1536451)
at Generator.<anonymous> (http://127.0.0.1:18888/index.js:2:1537797)
at Generator.next (http://127.0.0.1:18888/index.js:2:1536880)
at e (http://127.0.0.1:18888/index.js:2:1543652)
at s (http://127.0.0.1:18888/index.js:2:1543855)

Pls help?? I tried to "Export to onnx" and ERROR CAME UP?

[proud elbow]

u mean like from A3 to A6?

[gloomy zephyr]

uh, it's Saikai by Lisa and Uru

I'm not sure the exact notes

[ember cliff]

JIKEV REQUIRES IMMEDIATE ASSISTANCE

[rare goblet]

Ayo? @ember cliff level 1 !!!

[gloomy zephyr]

but it goes from pretty low to really high

[gloomy zephyr]

#🔍│help-w-okada

[full fossil]

oh ok thx

[rare goblet]

Ayo? @full fossil level 1 !!!

[proud elbow]

just see the vocal spectrogram

[raw palm]

The error is conditional

[gloomy zephyr]

...how

[proud elbow]

like 440 hz translated to A4

[gloomy zephyr]

it just looks like that

[nova fern]

Is there anyone can suggest me how to fix this problem?

[thin stump]

outdated colab? You must have pulled it off from youtube https://docs.aihub.wtf/

[nova fern]

I'll try it out. Thanks!

[proud elbow]

this song? another duet hell to deal with

[gloomy zephyr]

yep

[rare goblet]

Ayo? @gloomy zephyr level 3 !!!

[gloomy zephyr]

I’m tryna get that duetted

separating the two vocals are another issue altogether

[proud elbow]

even BVE can't separate many duets

[gloomy zephyr]

I think the voices are distinct enough

can you toss me the separation model

[proud elbow]

try do it yourself

[gloomy zephyr]

uh I kind of can’t

the chorus is sang together

[ember cliff]

Help jikev

[crude bolt]

have you figured it out ? @proud elbow

[rare goblet]

Ayo? @crude bolt level 8 !!!

[proud elbow]

what?

[crude bolt]

i was just curious - i found a song in youtube called the first take - is it that song ?

OH

nvm i tagged wrong person - mb mb 😦

[grand iron]

-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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[crude bolt]

RVC-HFv2, by r3gm Huggingface Spaces

[empty wigeon]

Can someone please explain to me why i can't merge the vocal into the audio file? Im using an existing model and im using the Hina_Mod_AICover colab. I've uploaded the instrumental file as audio and i've upload the voice model. But when i click "generate" it's only the instrumental part that plays

[steel forge]

You need to upload the vocals, not the instrumental. This is "singing voice conversion", it doesn't make up lyrics for an instrumental

Also no need to type in bold all the time

[empty wigeon]

Lmao correct. Literally did it like 3 hours ago and i already forgot

[empty wigeon]

sorry. bad habit

[tacit tinsel]

do you use any other method than plain hearing to decide which audio sounds the best after proccesing?

[rare goblet]

Ayo? @tacit tinsel level 2 !!!

[tacit tinsel]

is there any tool for data?

[proper shale]

Mmmm. Not sure there is

After inferencing, you mean?

[tacit tinsel]

i am basically trying out different models and methods to remove audio from song

thought theres a better way to find out which is best other than listening

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[low shard]

So, the best models for separating instrumentals and vocals? Those ones are in our docs https://docs.aihub.wtf/

[rare goblet]

Ayo? @empty wigeon level 3 !!!

[old oracle]

-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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[stiff goblet]

Which step should I consider as best one ? @proper shale

[vestal shell]

my voice is very very delayed

like a lot, and is laggy

[fathom geyser]

How i can fix this?

I'm getting this everytime i'm training a model

[rare goblet]

Ayo? @fathom geyser level 2 !!!

[fathom geyser]

I need to fix this error

Can someone help me?

[proud elbow]

which colab?

[fathom geyser]

easyGUI

[proud elbow]

no much idea, try rvc disconnected instead if u want

[fathom geyser]

Where?

[proud elbow]

-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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[reef tusk]

can you still download applio and train locally?

[vagrant mortar]

I'm going to ask something about RVC technology, specifically the platform that uses that technology, which is Applio. Generally, when I clone a voice from a different language, just as when I make the inference and conversion of a voice from my language (Spanish, Russian, Portuguese) to that different language (let's say English), it is not very noticeable that it is difficult for it to speak my base language. (the accent comes out more from the base audio that I provided the final inference). Is there a way for the original character's accent to be properly preserved, even after making inference? Should I disable pretrained or custom pretrained or should I touch another specific setting? Why does this happen?

How to Maintain an Accent When Cloning a Voice? Help me!

[proud elbow]

note that there is tradeoff between "indexing accuracy" and chance of artifacting for many average models

[noble anvil]

can someone please guide or have the tutorial to install RVC v2 on Macbook Pro? Please help

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[arctic socket]

whats the uh

download link

for rvc

and can it do live audio change?

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[proper shale]

yeah just use this one
https://rentry.org/VoiceChangerGuide

[hidden dew]

how do i know what my batch size should be

or should i leave it at 8

default

[proper shale]

keep it at 8 yeah, if u have a smaller dataset (~5 mins) use 4

[hidden dew]

okayy

im using tensorboard right , but how do i know which epoch count is which step

like

the step is 1k+ while the epoch count in the console is like 180

[proper shale]

pick the virtual audio cable if possible, else set the output to the communications variant of the virtual cable

[proper shale]

just check the weights folder, it should be saving your models with the epoch and step counbt

[hidden dew]

okay

ohhhh

tysm

[proper shale]

yw :)

[hidden dew]

i thought the step and epoch number is the same

[lapis wraith]

Can anybody help me I got some voices for exmaple an egirl voice but it doenst sound that good what do I need to chnage or does anybody have some good voices? pls

[proper shale]

https://rentry.org/VoiceChangerGuide

This could help you with your settings

In short, change your extra, mess with tune and index

Experimentation is key

[lapis wraith]

Alright thank you 🙂 , do you also know where to find some good voices except here in the dc servers? I don't find a link or something

[proper shale]

https://weights.gg

[lapis wraith]

thank you, really helped me out

[sinful lotus]

@proper shale

[proper shale]

yw

[proper shale]

delete pretrain and model_dir folders and restart w-okada, keep the command prompt window open so it doesnt stop the downloads n stuff

[fossil solstice]

help me i cant find the vcable i installed everything needed

please i beg

@proper shale

[proper shale]

have u restarted yet? vb cable requires that

[fossil solstice]

yes

[rare goblet]

Ayo? @fossil solstice level 1 !!!

[proper shale]

mm

it should be in the input/output parts then...

[sinful lotus]

What could be the reason for the intermittent sound only on WhatsApp?

[fossil solstice]

hm ok

[proper shale]

frankly, not sure

[fossil solstice]

@proper shale my chrome keeps allowing michrophone but discord says no

help

damned thing

ok i found cables

do i use output im confused

[sinful lotus]

What could be the reason why the application translates the voice of those we chat with on Discord?

[fossil solstice]

it keeps bloody saying ive blocked discords microphone usage wth

will it work in games like roblox still?

[proper shale]

yea it might

[fossil solstice]

hmm gotta find headphones w a mic

[proper shale]

if ur using on chrome just go on the permissions part of it

[vivid vigil]

I have trained a custom voice model and i have multiple. Pth files , some are G and some are D. Which one do i need to upload to okada voice changer?
Morover, i also donot have index file.

[proper shale]

and then just set it up like this:

The input output settings should be like this in W-Okada:

Input - Your mic
Output - CABLE Input

And in Discord:

Input - CABLE Output
Output: Your headphones

[fossil solstice]

i have and nothing works

whats w okada

[proper shale]

...the voice changer

[fossil solstice]

o

[sinful lotus]

.

[fossil solstice]

this is way too complex

[proper shale]

wrong output/input settings

or mic bleed

[fossil solstice]

i wastold its ez

[proper shale]

The input output settings should be like this in W-Okada:

Input - Your mic
Output - CABLE Input

And in Discord:

Input - CABLE Output
Output: Your headphones

Also enable Sup2 + Increase S Thres @sinful lotus

[sinful lotus]

When that part is wrong, isn't the sound supposed to go away at all?

[proper shale]

Well, you do kinda have a point

Just make sure your stuff is right tho

[fossil solstice]

i have no headphones w a mic.///

does this work on mobile'

or ipad

[robust halo]

@leaden hollow

Why is it lagging?

It's stuttering.

Ion get it.

[lost pewter]

is not an accepted origin. (further occurrences of this error will be logged with level INFO)

[leaden hollow]

The wrong mj 😭😭😭😭 @proper shale

[gusty kestrel]

@onyx grove, I have found 1 results that match your search!

hazza11

A Random Old Man I Found On YT

• Uploaded: <t:1692230400:d>
• Server: AI Hub
• Likes: 0
• Lang: EN

Epochs

500

Technology

RVC

Algorithm

Rmvpe

Requested by disaccharides · May 15, 2024, 9:11 PM

[manic plaza]

Alright so how do I make the RVC2 notebook work on kaggle

[upper furnace]

help petrein make problem

[proper shale]

You probably didn't format the dataset correctly - make sure that:

1. dataset folder name is the same as the experiment_name in the Set Training Variables step.

2. you're not compressing the files as .rar. this causes issues.

3. in case you did all of those, it could probably be an audio issue, try a re-export in Audacity as a .wav with all audios you used.

[upper furnace]

waw

@proper shale

[tacit tinsel]

Hi when I use Rvc c2 on colabs under Preprocessing I get this error "Error: Extracting is shut down because hubert_base.pt does not exist, you may download it"

where do i have to put this file?

I downloaded it but I dont know in which folder to put

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[vivid vigil]

Can someone here help me train a model on my voice on there gpu?
Issue is that my current laptop does not have gpu and im unable to train the model.
Thanks

[proper shale]

Make sure you've ran the Dependencies steps...

[proper shale]

Just use Colab

https://docs.aihub.wtf/ has Colab links & guides

[manic plaza]

Alright so how do I make the RVC2 notebook work on kaggle? I get these errors... and then aria2 doesn't run
(ping me when you reply please.)

[tacit tinsel]

i started from the top again and everything was smooth until "Preprocessing and Feature Extraction"

my dataset is imortet from the zip but at the preprocessing i get these errors

i really dont know what to change

[proper shale]

...huh

did it just... not clone repo? that's extremely weird

[tacit tinsel]

it cant find some directories but idk why :S

[proper shale]

at that point i suggest using applio for training

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb

guide for it: https://docs.aihub.wtf/rvc/cloud/applio-colab/

[tacit tinsel]

ok thanks, i will try it

[spring orchid]

-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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[empty wigeon]

What Voice Conversion settings do you all use? Can't figure out what they all do

[tacit tinsel]

is there any training software that I can use with and amd?

i only found options for nvidia

T_T

[nimble vapor]

need a bit of help

[manic plaza]

Any alternatives to google colab for RVC voice training? Can't get kaggle to work with any of the notebooks

[bold ember]

Whenever I run sovits I get this message can someone please help me?

C:\Users\User\Downloads\sovits\GPT-SoVITS-main>runtime\python.exe webui.py
The system cannot find the path specified.

C:\Users\User\Downloads\sovits\GPT-SoVITS-main>pause
Press any key to continue . . .

[proud elbow]

paperspace, and vast.ai, though needs some effort to set up and port into it

[nimble vapor]

how can one fix this

[junior dune]

i need help

What do I do after i download a hugging face file?

[tired swift]

i have an issue with my cmd/python after doing this "git clone https://github.com/SociallyIneptWeeb/AICoverGen
cd AICoverGen
pip install -r requirements.txt"

at the end it gives me two errors " ERROR: Failed building wheel for fairseq" and "ERROR: Could not build wheels for fairseq, which is required to install pyproject.toml-based projects"

i can dm a screenshot if that would help

i am attempting to install and use AICoverGen

[rare goblet]

Ayo? @tired swift level 1 !!!

[tired swift]

if anyone could help me i would appreciate it tho🙏

[tacit tinsel]

after installing the applio local version i get this error " ModuleNotFoundError: No module named 'gradio' "

the installation was compelte tho. Any ideas?

[manic plaza]

Thanks

Sadly I don't know how to code other than javascript, and it doesn't really use javascript.
Can someone please give me a working kaggle / paperspace notebook? Thanks

[proud elbow]

all the RVC code uses Python as well as the PyTorch framework, and actually Python should be quite easy to learn.

[manic plaza]

Thanks for the tip

[wraith quest]

ModuleNotFoundError Traceback (most recent call last)
<ipython-input-7-7f286ebd62ee> in <cell line: 4>()
2 #@markdown Link the URL path to the model (mega, Drive, etc.) and start the code
3
----> 4 from mega import mega
5 import os
6 import shutil

ModuleNotFoundError: No module named 'mega'

---

NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.

To view examples of installing some common dependencies, click the
"Open Examples" button below.
pls guys help

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[magic hare]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[low shard]

is this the ilaria rvc colab

click runtime, disconnect and delete runtime, and re run everything

[magic hare]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[lament briar]

Hello i just created a voice but does somebody know how to do text to speech with that voice?

[low shard]

you can use applio tts or ilaria tts that are already built in, else, you can use any tts generated audio (better if its same gender and language of the modle u wanna use) and then use it as an input in rvc

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[brittle wing]

C:\AI Vocal Training\Mangio-RVC-Fork-23.7.0\Mangio-RVC-Fork-23.7.0>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897
The system cannot find the path specified.

please help me

[cyan fractal]

What's rvc?

[proper shale]

retrieval based voice conversion, the thing we use for ai voice conversion

[proper shale]

rename folder to "AIVocalTraining", with no spaces

and try again

[cyan fractal]

oh alr thx

[brittle wing]

I still got the same message 😦

[brittle wing]

C:\AIVocalTraining\Mangio-RVC-Fork-23.7.0\Mangio-RVC-Fork-23.7.0>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897
The system cannot find the path specified.

C:\AIVocalTraining\Mangio-RVC-Fork-23.7.0\Mangio-RVC-Fork-23.7.0>pause

[proper shale]

[brittle wing]

lol

[proper shale]

i think you should just

use either mainline

or applio

mangio is sadly outdated at this point

[brittle wing]

Applio given me huge errors

Traceback (most recent call last):
File "C:\AIVocalTraining\Applio-3.1.1\env\lib\site-packages\requests\compat.py", line 11, in <module>
import chardet
ModuleNotFoundError: No module named 'chardet'

[proper shale]

are ya using the prepackaged one?

[brittle wing]

not sure, i believe the zip files

[proper shale]

I'd just recommend you to use Mainline at this point

[brittle wing]

Thank you, it works. How do I get Tensorboard for Mainline?

[proper shale]

https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/

has the .bat installer n stuff alongside graph analysis tutorials

[brittle wing]

I appreciate your help, thank you

[proper shale]

You're welcome

[wild forge]

What does index file mean on applio

[cold tinsel]

my sample rate is 22050, so which sample rate do I choose for training?

Well it's from the output info:
Sample Rate: 22050
Duration: 22.99 minutes
Number of Samples: 30421605
Bits per Sample: 44100

[proud elbow]

find another better source

that's like crappy old game audios

[lapis tinsel]

Hi everyone

I have a dataset of 20 minutes but it's not in zip format, but in MP3 format, is it ok if I create a model with it?

[proper shale]

just use 32k

[proper shale]

Yeah, just format it according to what you're gonna use to train

[lapis tinsel]

Don't know the names of the platforms that cut it, mine was a full MP3

Or does it work fine even if I upload it like this?

[proper shale]

You don't need to trim it every 10 seconds like some people have claimed, RVC does that automatically

Just zip it if needed, like if you'll use RVC Disconnected

[lapis tinsel]

Can I just zip this one file?

[proper shale]

It's not necessary, depending on what you'll use

That's why I said, zip it if it's needed... it's mostly not needed

[lapis tinsel]

Most thanks

[proper shale]

You're welcome! If you have any more questions feel free to ask em

[cold tinsel]

Thanks. I will try 32k. I don't really understand how the sample rate is so low since I isolated many of the vocals from flacs

[rare goblet]

Ayo? @cold tinsel level 1 !!!

[proper shale]

Wait what models did you use to isolate?

[cold tinsel]

I used the MDX23 colab and used BSRoformer ep_368_1296 and VitLarge and the output format was FLAC

[proper shale]

Mm

[cold tinsel]

and the hq reverb by foxy

[proper shale]

FoxJoy doesn't have any SR limits n shit... I think it's still 44kHz, but either way use 32k

[errant plume]

@proper shale is it safe to connect to local runtime in rvc disconnect in the middle of the session cyz its taking time for the dataset to be imported and iwanna save my gpu runtime

[brittle wing]

How do I use/apply custom Pre-Trained Models to train on my Ai Vocals in Mainline?

To quick the training process*

[proper shale]

Frankly... it won't speed up the process in terms of epochs per second

[errant plume]

resume the training by adding the g and d files and enter the stepcount

[proper shale]

huh?

wdym local runtime

[proper shale]

Most of the time, the original pretrains will do things just fine

[errant plume]

this one

[velvet peak]

can someone etach me how to use rvc on local please

[velvet peak]

i forgot how it was and i reseted the pc so i also need the download link

[brittle wing]

Oh ok, I heard great things about Pre-Trained models like Titan.

[proper shale]

frankly, never used that option

[proper shale]

https://docs.aihub.wtf/

[errant plume]

someone told me that there is a way to save gpu runtime but i forgot it.. do you know how to do that?

[proper shale]

nope

[velvet peak]

idk how to download on github 😭

[proper shale]

i mean... you can save GPU runtime when training... just stop training earlier, save and come back tomorrow for example

[proper shale]

it shows you exactly where to click

[velvet peak]

im blind okay?? 😭

[errant plume]

@proper shale so connecting to the local runtime will disconnect the session?

[velvet peak]

oh found it

[proper shale]

again, I'm not entirely sure what this option does or why it would be useful

[errant plume]

okay

[velvet peak]

i have to download all of the 22 assets??

[cedar adder]

hi, does anyone have any alternatives to DeReverb vocal on cloud besides the UVR by Not Eddy because it's not been working recently for me.

[proper shale]

there's always MVSep

https://mvsep.com

create an account btw, has lower queues

[cedar adder]

thanks also what model do i use for DeReverb?

[proud elbow]

foxjoy reverb HQ

[proper shale]

Either Foxjoy's dereverb or the UVR one

I've had better results with FoxJoy's though... it just doesn't fight echoes

[cedar adder]

thanks guys appreciate it

[rare goblet]

Ayo? @cedar adder level 1 !!!

[errant plume]

@proper shale can u kindly check ya dm

someone pls help me im getting this error at the feature extraction part

[unique blade]

where can i find my index file in Applio?

[proper shale]

logs folder

[unique blade]

all i see is th epth files

[proper shale]

you didnt train index then

[unique blade]

i did and it said index genrated

[proper shale]

well, it should be in the logs folder

or maybe on assets

istg if Applio changed this imma rage hard

[unique blade]

i cant faind it

ill train the index from colab

[green zealot]

What does this mean and what's the difference?

[proper shale]

ContentVEC seems to be better at detecting what's being spoken than HuBERT, but try out and see.

These are the ones that detect speech (and what phonemes should be used) when you infer your audio

[green zealot]

Thank you so much for your answer

[proper shale]

You're welcome :)

[unique blade]

yeah i did and didnt faind shit

anyway i made the indix with disconcted

ok

[lament briar]

What is a tts generated audio?? and can i find the Text to speech function in the Mangio-RVC-v23.7.0 Folder?

[low shard]

Nope mangio doesn't have it built in

With tts generated audio i mean that you can use any text to speech tool and just use it as an input

[lament briar]

oh okay i understand thank you very much!

[low shard]

You can see some tools in https://discord.com/channels/1159260121998827560/1212368971127590922 if u need them

Also yw

[royal patrol]

What to put in the CHUNK and EXTRA menus?

[lament briar]

do you maybe know the link for bark on the huggingface site?

[low shard]

U mean https://huggingface.co/spaces/suno/bark ?

[spring orchid]

rvc disconnected still broken ?

[lament briar]

yeah thank you very much

[rare goblet]

Ayo? @lament briar level 2 !!!

[low shard]

Yw

[low shard]

Wdym

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[spring orchid]

yeah i guess ill wait a fix

maybe its the titan pretrain on rvc disconnected

[lament briar]

does somebody know where i can get the Adam/freya ai voice from elevenlabs?

i think i got ip banned from elevenlabs because i used it too often

[spring orchid]

-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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[covert nest]

i need help can i need someone to cover my song

[upper furnace]

help datset petrein

[covert nest]

2

[rare goblet]

Ayo? @covert nest level 1 !!!

[covert nest]

@upper furnace

[upper furnace]

yes?

a ok

[brittle wing]

Should I train Ai vocals with my CPU or GPU?

My specs: Ryzen 9 5900x + RTX 2080 Ti

Also during training, what should my temps be in average?

I know Training Ai can be intense with hardware, I Don’t want my PC catching on fire.

[wraith coral]

Can anyone help me with voice train

[odd shale]

GPU.

[tacit tinsel]

i had the same problem yesterday

I used applio desktop version with fake gpu enabled and trained a model.
After training in logs folder, I got a bunch of .pth files. Most are like this "train_20e_20s.pth" I guess its the epochs but I have 2 huge pth files called D_2333333.pth and G_2333333.pth. What are they ?

[meager pollen]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[odd shale]

These are the pretrained files that get created every time you hit a save point on the model training

You can use those D and G files (they must coincide with the .pth on step count) to continue training from a save point.

[tacit tinsel]

oh ok so if my model is fully trained i can just delete them right?

one more question. After i finished the training successfully, I tried to generate an index, but i got this error.

"Failed to train index: need at least one array to concatenate
If you are running this code in a virtual environment, make sure you have enough GPU available to generate the Index file."

[rare goblet]

Ayo? @tacit tinsel level 3 !!!

[tacit tinsel]

is an index important to have?

I guess I cant generate one because I used the fake gpu option ? :S

[bright wren]

is there any way to give the vm any form of emotion or vocal queue like clearing their throat or laughing or anything?

i don't want to use my models for just songs but they sound so damn monotone even in elevenlabs rip

[bleak urchin]

where do you donwload rvc? im new

[tidal vault]

what is the doug walker model called

please help what is the nostalgia critic model called

[bright wren]

#1159513888199540817

[proper shale]

Mmm not sure anything can do that rn at least

At least any TTS...

[bright wren]

womp womp

[proper shale]

Maybe Bark but honestly it sounded super low quality, and there isn't really any way you can use it from my knowledge

But I'm most likely wrong

[bright wren]

i don't think so

I tried bark and i couldn't figure out how to open any form of GUI to even try and use it

Suno doesn't support it anymore either

[proper shale]

Just wonderful

https://elevenlabs.io/docs/speech-synthesis/prompting#emotion Hmmmm

[blissful scaffold]

colab not working?

says no gpu availeble

[cold cave]

-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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[primal crow]

Hey guys. I'm about to cry/hj, how do I simply make and export a voice model .pth file. I have the wav for training and I have voice to voice RVC GUI for Windows, but I need to make a custom voice model (Free), help?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[low cliff]

RVC GUI is not working

[spring orchid]

rvc disconnected and applio no ui not working for me

[low cliff]

same

[bright wren]

Seems like putting examples like this helps a bit with it

"Are you sure about that?" he said, confused.
"Don’t test me!" he shouted angrily.

many thanks for the link

[proper shale]

Yw :)

[nocturne mural]

colab down

[tardy bone]

Anyone know how to fix or at least mitigate volume clipping in this?

There are parts where Ozzy's voice gets muted somewhat

[barren canopy]

Lower aggression on your voice remover when you're seperating the stems, maybe?

[tardy bone]

mhm

[brittle wing]

How complex is this software to use?

[brittle wing]

rvc disconnected aint working

[verbal peak]

What can i useto use my mic on discord

like the voice changer

[rare goblet]

Ayo? @verbal peak level 1 !!!

[glossy kelp]

assuming this has happened before, how long does it stay down, typically?

[brittle wing]

if someone knows how to fix it ping me

[glossy kelp]

Same

[brittle wing]

cause i use to use rvc disconnected before and now its not working

[nocturne mural]

I can't give you an exact time, since like gradio it was supposed to be fixed in 1 day but it ended up taking more than a week.

[nocturne mural]

mounted the drive?

[brittle wing]

yes

it doesn't work anymore

i use to be able to train models

[nocturne mural]

weird Model/${null}/Model

try restarting the environment and make sure everything is ok

[brittle wing]

i did it aint working

i tried with different accounts too

[nocturne mural]

exactly how to load your backup model

🐈

[spring orchid]

colab gpu still down?

[rare goblet]

Ayo? @spring orchid level 10 !!!

[brittle wing]

like how do i save it?

[nocturne mural]

no

[spring orchid]

so its working now?

[brittle wing]

https://status.cloud.google.com/incidents/xVSEV3kVaJBmS7SZbnre#Dh3y8L8YSyVY8gnSdWiZ found this

[nocturne mural]

yes

[brittle wing]

old version of rvc disconnected was much better

[nocturne mural]

a

[spring orchid]

not working yet ;-;

[nocturne mural]

Are you using an old version?

[brittle wing]

nah

[nocturne mural]

what

✅

i dont speak english but well

[brittle wing]

yeah rvc disconnected is messed up cause it saves a weird folder

[spring orchid]

maybe its of some servers of america

[nocturne mural]

I remember the colab loads your backup by a .zip and does not read your drive, you have not edited it?

or have I never come across the cell that does that?

🐈

[brittle wing]

no i load the dataset from a zip

[brittle wing]

yeah im not affected

[spring orchid]

lucky, give me your powers

[brittle wing]

but rvc disconnected aint workin tho

[nocturne mural]

well, unfortunately, I don't know how to solve it anymore, you will have to wait for someone who is experienced with this colab.

[lofty heart]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[lofty heart]

-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
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[spring orchid]

@nocturne mural hasta aqui llega el applio no ui colab y ahi se queda permanente no se si a alguien mas le pasa, pero ayer funcionaba bien ahora no se que paso

[silk brook]

chart went backwards ?

[nocturne mural]

te refieres a entrenar?

umm

ya que ya hizo el backup de tus archivos

intenta stopear la celda y ejecutarla de nuevo

a ver si sale log del rvc en si

[thin stump]

That might be an mode collapse. That's why it needs to be 0.987 smoothing too

[brittle wing]

On Mainline for Ai vocal training.

Are custom pre trainers speed up the process for longer data sets?

Worth it?

Or just use there default?

[silk brook]

its set to 0.987

what is a mode collaprse?

might not have been at that point though

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[brittle wing]

When training Ai, what’s your guys GPU temps usually at?

[finite crystal]

I haven't make an Ai Model for very long time now. how do I make one?

[brittle wing]

I guess, Titan for example. I heard great things about it for vocal singing?

[rare goblet]

Ayo? @brittle wing level 7 !!!

[thin stump]

There's a new one that came out #1240679287036579901 message

I haven't listened to the samples though

[brittle wing]

oh, thanks for clarifying

[brittle wing]

Ill probably give it a try, it says 32k? however I have 48k datasets, wonder if it still works? they have it at 60 epoch, i also wonder if I can go higher?

[thin stump]

Nope, just wait for Titan large or use the original pretrain

[brittle wing]

Bet, I appreciate your help.

[thin stump]

another few hours/week spent training ig. That's why simplecup had united do the 32k training for titan

[sturdy stratus]

hello how to fix rvc gui high ms

[rare goblet]

Ayo? @sturdy stratus level 1 !!!

[proper shale]

it is the best sample rate

[proper shale]

hm? what's your GPU

[sturdy stratus]

My com i7 3770 and gtx 1050ti It's bad, right🥹

[proper shale]

Mhhhh, it's just outdated y'know

I don't like saying that but

Yeah

You could try using Colab, though:

https://colab.research.google.com/github/hinabl/voice-changer-colab/blob/master/Hina_Modified_Realtime_Voice_Changer_on_Colab.ipynb

[sturdy stratus]

Oh So if I want to upgrade my computer How should I upgrade?

[proper shale]

Imma be honest with you, that i7 3rd gen ain't gonna help you a lot

So, in my view, you should try to get at least a Ryzen 5 3rd gen or 5th gen if possible, and an RTX 2060 or 3060 if possible

[sturdy stratus]

Thx you And if it's an Intel brand?

[proper shale]

I don't really recommend Intel if you're going budget

But maybe the i3-12100F or the i5-13600 if possible

[sturdy stratus]

Oh thx ☺️

[proper shale]

Yw! Hope you can get that PC 🔥

And btw, I think it's best for you to do your own research and see what are good budget CPUs and good budget GPUs

Just avoid AMD GPUs if you want to use AI though...

[proud elbow]

any CPU using your mobo socket unless too old.

(intel & AMD don't really matter on low end)

[steel forge]

mainline rvc will work just fine on a 1050ti, but realtime will struggle a bit. especially with a game open

[wanton merlin]

hello, i just finished training my model with easyGUI📱- 4/20/24. i ran out of gpu usage in google collab but the training is finished. i can't train index and download from the gradio.live web.
is there a way i can download the pth file from google collab?

INFO:root:Saving model and optimizer state at epoch 250 to ./logs/hattavoice/G_2333333.pth
INFO:root:Saving model and optimizer state at epoch 250 to ./logs/hattavoice/D_2333333.pth
INFO:hattavoice:saving ckpt hattavoice_e250:Success.
INFO:hattavoice:====> Epoch: 250 [2024-05-17 09:14:50] | (0:01:20.837137)
INFO:hattavoice:Training is done. The program is closed.
INFO:hattavoice:saving final ckpt:Success.
/usr/lib/python3.10/multiprocessing/resource_tracker.py:224: UserWarning: resource_tracker: There appear to be 20 leaked semaphore objects to clean up at shutdown
warnings.warn('resource_tracker: There appear to be %d '

[steel forge]

colab realtime is like, really bad though. so you might want to try both to see what works best for you

[steel forge]

on the left sidebar, you can navitage through your folders. You're looking for the logs folder in there, your pth and index should be there

[wispy lodge]

colab with rvc realtime or wokada?

[steel forge]

wym

oh

no one uses mainline realtime except tea lmao

and tea told me to basically not bother with it

[wanton merlin]

i already looking for the logs folder but i can't find it. i think it should be on /content/RVC/logs/
but on the left sidebar theres only /content/sample_data , no RVC folder

[steel forge]

hmm. I'm not an expert with colab, have you tried navigating up in folders?

[steel forge]

oh, i think if it looks like this, your runtime was killed

if it wasnt backing up into your Google Drive, you might have lost the files

[manic sparrow]

hello guys did someone know which python i need?

[wanton merlin]

yeah it look like this

[wanton merlin]

i see, i think i lost the file haha. btw thank you!

[steel forge]

np!

[manic sparrow]

?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[lean orbit]

guys whenever i paste my model in collab it gives me erroe no matter what model im using

JSONDecodeError

File exists, proceed with creation of creds and client

these two messages is in red

can anyone help me ?

[low shard]

send colab link

[tacit tinsel]

i follow this to download the rvc mainline but after extracting I cant find "go-web.bat"

theres only go-realtime-gui-dml.bat

[lean orbit]

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

[proper shale]

old colab, use the ones in https://docs.aihub.wtf/ in the RVC > Cloud section

[low shard]

[lean orbit]

thanks guys

[rare goblet]

Ayo? @lean orbit level 1 !!!

[proper shale]

that happens with the DirectML (AMD and other GPU) version.... RVC doesn't work w it so it just doesn't include the go-web.bat

[tacit tinsel]

I dled the version before the newest and go-web was incluted

web ui works now

[proper shale]

ah

might be the prepackaged released version that's screwed up then..

[tacit tinsel]

def something because i dled the bat manually and it wouldnt start. 2023-08-13 version works (for now xD)

[tacit tinsel]

hmm it doesnt find my gpu. Even tho i dled the Amd version and i fullfill the requirments. Do i have to update my card or something?

[brittle wing]

In Mainline RVC, is it possible to pause mid training? To let my GPU chill down for a bit, and then resume training?

[proper shale]

yeah it... won't do it sadly

AMD and AI don't go well

[silk brook]

this is the most trained one

havent tested the others yet

[proper shale]

wow that sounds

very clean

Yeah the original audio is screwing with it. The model sounds great ngl

[silk brook]

yay

yeah these arent the best

just some voiclesses I found on yt

also this isnt the same language lmao

I tried this audio after the first one with the issues I noticed earlier and was like wow

this is really nice

thanks for all the help btw

oh

could you point me to where I can get educated about how to clean up my own audio? I wanna know the best resources so I can use better input

[cold tinsel]

hi, is this dynamic okay? still trying to wrap my head around the tensorboard stuff

^^also sounds good Flare

[silk brook]

thanks :)

its honestly probably cause the input is so good

I managed to find a libarary of all of her lines individually recorded and just took out the ones that were noise and not voice

sorry though idk what dynamic is either lol

[proper shale]

this is grads... they seem to be going alright i guess

but the ones you should focus on: loss/g/total, g/kl and g/mel

g/total is average, lower better

g/kl is how different model is from the dataset, lower better

g/mel is how the model is replicating the spectrogram of the original dataset, lower better

[proper shale]

hm, i suppose pins in here have em

or in #📑│making-models

[silk brook]

ok 👍

any tips from you?

[proper shale]

RX10 for a pro clean-up, use BSRoformer for vocal isolation whenever acapella isn't officially available, also use dereverbing tools... the two latter ones can be used in MVSep - https://mvsep.com

Create an account btw

So less queues

[silk brook]

what is RX10?

it just came up with a camera on google