✨│ai-help

[crimson sand]

okay

now run

[grizzled blaze]

wait, its downloading packages

its updating conda

[crimson sand]

C:\Users\Admin\Miniconda3\condabin\conda.bat activate C:\Users\Admin\env

Sure but once that's done run that

wait

[grizzled blaze]

oh, edited

[crimson sand]

yeah i updated the command

[grizzled blaze]

ok

no answer

[crimson sand]

no thats good

now

cd X:\Games\RVC\Applio-RVC-Fork

[grizzled blaze]

k

[crimson sand]

pip install --upgrade setuptools
pip install -r X:\Games\RVC\Applio-RVC-Fork\assets\requirements\requirements.txt

uh wait

[crimson sand]

oh

its not changing

thats weird

okay edited the commands, run those

then once those are done run

pip install future==0.18.2
pip uninstall torch torchvision torchaudio -y
pip install torch==2.0.0 torchvision==0.15.1 torchaudio==2.0.1 --index-url https://download.pytorch.org/whl/cu117

[grizzled blaze]

k

its downloading

[crimson sand]

then you can run go-tensorboard.bat

no idea why the installer fucked up for you

i had to fix a bunch of stuff manually too, but i forgot what i did

[grizzled blaze]

does that matters?

[crimson sand]

it does, sec

OH

I know what happened

you didn't install VS build tools

show the full command line, whats the error coming from

[crimson sand]

yeah you dont have crt headers

You need to install the vscode build tools

winget install Microsoft.VisualStudio.2022.BuildTools --force --override "--wait --passive --add Microsoft.VisualStudio.Workload.ManagedDesktopBuildTools --add Microsoft.VisualStudio.Workload.VCTools --add Microsoft.VisualStudio.Component.Windows10SDK.19041 --add Microsoft.VisualStudio.Component.VC.Tools.x86.x64 --includeRecommended"

cause you dont have the w10 SDK

[grizzled blaze]

k

its installing

its gonna take a while

[crimson sand]

yeah

sorry about that

[grizzled blaze]

nah

np

you are helping me

thx for that

[crimson sand]

applio fork isgreat apart from the fucking tensorboard updating stuff

its so annoying

it only updates the core model board when you close applio

but it updates \eval every epoch -.- I BELIEVE that \eval should show different data, but never verified, maybe someone here knows

[grizzled blaze]

ready

[rare goblet]

Ayo? @grizzled blaze level 4 !!!

[grizzled blaze]

should i use the other commands now?

or is that enough?

[crimson sand]

yep once you're done that

pip install --upgrade setuptools
pip install -r X:\Games\RVC\Applio-RVC-Fork\assets\requirements\requirements.txt
pip install future==0.18.2
pip uninstall torch torchvision torchaudio -y
pip install torch==2.0.0 torchvision==0.15.1 torchaudio==2.0.1 --index-url https://download.pytorch.org/whl/cu117

[sand iris]

...Or just get the RVC-WebUI release archive at that point

unzip & run

No messing around

[crimson sand]

He used the installer but it didn't work shrug

[sand iris]

applio moment

[crimson sand]

but yeah applio has some issues for sure

[sand iris]

mainline RVC works

[crimson sand]

Any idea why applio doesn't update the tensorboard apart from the \eval one?

[sand iris]

¯_(ツ)_/¯

[crimson sand]

Do you recommend mainline RVC or Mangio?

[sand iris]

mangio hasn't been updated in a good while

mainline has all the goods now, there is little to no point in installing anything else tbh

[crimson sand]

Is it in english?

[sand iris]

Yeah

It has support for multiple languages

[crimson sand]

Nice

I'll grab that

[grizzled blaze]

whats the mainline

[sand iris]

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI

Releases are over at

https://huggingface.co/lj1995/VoiceConversionWebUI/tree/main

Pink the most recent .7z archive that corresponds to your system

[pseudo crow]

Hello I wanted to use it I have 4060 but it says

device detection error, fallback to cpu

@sand irisWhat I can do It has so long delay

[sand iris]

Not available atm

[pseudo crow]

wdym

[sand iris]

I'm not at my desk

[pseudo crow]

ah alright

[gaunt current]

about tensorboard:
i noticed in the guide that ROD/AOD has been removed from the guide, but in the steps following, i'm having trouble finding any of the files mentioned that i should open for running tensorboard. i have python installed as the guide stated, but can't find install-requirements.bat, config-editor.bat, or launch-tensorboard.bat in my applio folder. anyone know how to fix this?

[violet heron]

Ok so you are running it locally

Try restarting it

[crimson sand]

sounds like cuda isnt picking up the gpu

[pseudo crow]

I fixed it

thanks

[calm totem]

i did that earlier it didn't work

[violet heron]

Are you sure you put your model in weights and index in logs

[calm totem]

pth in the weights and index in logs

[robust marsh]

do you know what the link is to run rvc online?

[violet heron]

Did you select the inferring voice

[violet heron]

Inferencing or training

[calm totem]

what is that and how do i select it?

[violet heron]

Meant voice

The little box that says “Select your Inferencing voice:” or something like that

[robust marsh]

inferencing, the one that can i can use to make covers

[calm totem]

oh yes i did

[rare goblet]

Ayo? @calm totem level 2 !!!

[calm totem]

it also got stuck in the queue

[violet heron]

I don’t know sorry

Try a different model

[crimson sand]

Whats the deal with RVCv2 vs RVCv3? Has the pretrained model been released, is there any paper associated with it? Is it better than v2 measurably?

[calm totem]

there is this thing in the terminal tho

it said i could run it on a public link instead of a local link

how do i turn that on

[violet heron]

RVC 3 is supposedly faster with petter pretrains

[crimson sand]

Ah interesting, is the quality better?

[violet heron]

Should be

[tame mica]

petter pretrains?

[violet heron]

Peter pretrains 😭 oops

[violet heron]

Go to that link instead

It shows up with 2 links

[crimson sand]

Is there a link to RVCv3 or a paper @violet heron

[violet heron]

No

[violet heron]

“They are still working on it”

[calm totem]

idk where to find this

[crimson sand]

Ah okay cool

Is it people from a uni? Will there be a paper? or is this a community thing

[violet heron]

Have you tried a different model tho

[violet heron]

I hope there’s more documentation I would say it’s a community thing

[calm totem]

yea didn't work still

got stuck at loadin

[molten pecan]

https://rentry.org/losslessguide look what I did

[calm totem]

or queing

[molten pecan]

Mods did not approve 😭

So it can't be an official bot command

[robust marsh]

^

[rare goblet]

Ayo? @robust marsh level 1 !!!

[violet heron]

Sorry oops

Ima send you the link

[robust marsh]

ty!

[violet heron]

https://huggingface.co/spaces/TheStinger/Ilaria_RVC?duplicate=true create a HuggingFace account then go to this link

[proper shale]

isn't this technically piracu

piracy

😭

[molten pecan]

Yeah

But it's not affiliated with AI Hub

[violet heron]

Can be slow but it has lots of options and a GUI

[molten pecan]

So don't worry :b

[proper shale]

I'd recommend not posting that type of stuff, don't want any copyright complaints 💀

[molten pecan]

Oh Squishy told me it's fine

As long as it's not an actual command

[gaunt current]

anyone have a solution to this? anyone install tensorboard without AOD?

[proper shale]

there's the pins with two .bat files

and a little guide with non functional links, but they're the same files i uploaded

[crimson sand]

piracy is cool

[proper shale]

Installing TensorBoard Locally, for Mangio-RVC-Fork users

https://docs.google.com/document/d/1ihvqWnSyabvZ9jX6BsOxEY1FpplwjY3Jse3she_VO6I/edit?usp=drivesdk

[proper shale]

😭 wouldn't recommend it

[gaunt current]

thank you 🙏

[proper shale]

np

[molten pecan]

[mild cedar]

how do i use it on zoom,

[crimson sand]

all my homies love piracy

[molten pecan]

[molten pecan]

#🔍│help-w-okada

[proper shale]

download the vb audio cable

and use it as output in W-Okada and input in Zoom

[violet heron]

So your saying you go to the store, buy CDs, put them in a DVD player and rip the files from them and convert to .wav

[proper shale]

that's the right way to do it lol

or just listen to spotify

[violet heron]

But do you do it)

[proper shale]

well i currently don't have the money to spend on CDs but if i have it i will, + i just use tidal and im thinking to maybe buy qobuz albums

¯_(ツ)_/¯

once i have the money ofc

[violet heron]

Say qobuz without buz

[proper shale]

Qo

[violet heron]

🐀

[molten pecan]

I love piracy

[mild cedar]

where is the cable link

[violet heron]

https://google.it

[proper shale]

Just saying, it's a crime :)

[thorny grail]

-colab

[azure marsh]

All working colabs / spaces ☁️

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

Credits to Faze Masta and Antasma for compiling these links.

[molten pecan]

How mfs feel after they finish their sentence with a passive-aggressive " (: "

[upbeat ice]

so, i wanna create a voice model from my phone, is there any site or app that's compatible with android to do so in?

[finite galleon]

You will need to use google colab my friend

https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS

[pulsar tartan]

hello guys
how can i use rvc models
i need them to make voicelines

which app?

[upbeat ice]

i see, how do i make a dataset on my phone?

[finite galleon]

That my friend is gonna be a lil complicated

But i personally use the audiolab app

[upbeat ice]

audiolab? that's to make datasets?

[finite galleon]

More like to edit audios

[proper shale]

oh my god that was very passive aggressive 😭

[finite galleon]

To create a dataset you would need audios of your voice your trying to clone

[upbeat ice]

hmmm i can get those audios but then after I'm done getting the audio and possibly editing them, what should i do next?

[molten pecan]

There are different apps for it. First, do you own a gaming PC?

[finite galleon]

Go to rvc disconnected and start training

If you dont know how to train on rvc disconnected, read the guide

I forgot which one is it tho

Sooo search ig

[upbeat ice]

got it

[distant turtle]

does anyone know a free TTS online (colab or other sites), I tried tortoise tts on colab but I didn't get the model in Italian, I'm trying to create a tts model, does anyone know about other tts on colab that support Italian or more languages? Thank you

[finite galleon]

Edge tts

You can try it with ilaria tts HuggingFace space

[distant turtle]

thanks, I'll try it and see if it works, thanks

[distant turtle]

I tried, but I wanted to create a model from a dataset, to have an excellent result, Ilaria's site uses other voices to create yours, however I see the first site, Thanks

[distant turtle]

Thank you very much it seems that I have found a solution, I hope my PC is a bit slow which is why I couldn't install the local version, Thank you very much indeed!

[finite galleon]

No problem 👍

[molten pecan]

Boykissing (:

[pulsar tartan]

gaming pc no but i have an pc

[molten pecan]

In that case you should use Hugging Face or Colab

https://discord.com/channels/1159260121998827560/1182344731800387585

Here's my guide on HF :3

[pulsar tartan]

thank you

[molten pecan]

You are welcome

[crude siren]

hey i need some help setting up mangio on my computer

[candid jetty]

what should I change this to for local training? (i9-14900k)

[proud elbow]

2 cores for any CPU

[formal jungle]

https://gyazo.com/d71843e98a7ce735610551351379322f

I've no idea why some links just don't work with this covergen space

trying to figure out if I can just use the models I've downloaded on this as well

Would make things much less of a headache

[loud hazel]

there a webniar for rvc start now here : https://meet.google.com/zsh-wiwh-xta

[candid jetty]

cool ty

[rare goblet]

Ayo? @candid jetty level 9 !!!

[candid jetty]

do I need to do anything for this?

[crude siren]

i have it downloaded to my C: space how do i start it up

[candid jetty]

goweb.bat

run that

[crude siren]

that a file?

[candid jetty]

yeah should be in ur rvc folder

[crude siren]

this it?

[candid jetty]

ye

[crude siren]

what do i do?

[candid jetty]

oh wait have u not extracted ur rvc folder?

do that first

[crude siren]

ok

just click extract all oooooooooorrrr?

[lethal cradle]

Does anyone know how i can make my own ai?

[candid jetty]

yup

extract the whole folder

[crude siren]

guessing this is right

[candid jetty]

yeah should be fine

am I better to run with a 8 or 12 batch size?

[crude siren]

oh for-

wait nvm i think its working

[rare goblet]

Ayo? @crude siren level 7 !!!

[hallow wren]

hello there, I am using RVC colab version. Inspite of successfully downloading the model, it's not appearing when refreshed. And I don't see any error in Collab. Can someone please advise ?

[candid jetty]

tensorboard wont open, is there anything I need to do to set it up?

[molten pecan]

Did you try reloading the page?

[hallow wren]

Thanks for looking into this Julia. I tried multiple times, it didn't work. Now, I have switched to another AICover colab notebook and here the model is reflected however while generating the conversion, I am getting error even though I clearly see the model name folder is created within that directory. I even tried to manually add the model file .pth in that directory, then getting some pytorch error

[molten pecan]

That's weird

https://discord.com/channels/1159260121998827560/1182344731800387585

Use my guide

I'm sure it will work :3

If it doesn't, then the problem might be the model itself

[hallow wren]

Thanks Julia. I have worked with these notebooks multiple times in the past without issues ( I mean there were issues atleast it worked 😄 ) but yeah that's true that this is a new model I am working with. Just tried with your huggingfacespace and it worked successfully. So no issues with my model.But when I am trying it again, it says Error: None. Not sure why. On a separate note, I tried to duplicate the space to make it private to my space but now sure how to launch that private space ( sorry if this sounds stupid )

[pseudo crow]

does tune depends on my voice or there is good settings for general?

[molten pecan]

After you click "Duplicate this space" you should click "Duplicate" and it will automatically start loading

You just gotta wait

And after a few minutes all the text and settings will pop up

[candid jetty]

https://docs.google.com/document/d/1ihvqWnSyabvZ9jX6BsOxEY1FpplwjY3Jse3she_VO6I/edit has anyone got the 2 files that need to be put in RVC folder from here?

I think they were in the old server

[hallow wren]

yeah it;'s all loaded completely and it says, running locally but i don't see any other link. I don't think it's executing anymore step. May be i need to something to launch my private space

[molten pecan]

Uhh you don't need any link

If the URL of the space has your HF username, then you are in your duplicated space

[jaunty shale]

this easyrvc collab sometimes stops itself whenever I try to train 1000 epochs,
It stops usually around 700-800 epochs, why does it do that?

[candid jetty]

u might have run out of colab minutes

[pseudo crow]

Whenever I open start_http It opens a empty app that has nothing and on cmd It says wait web server

[W:onnxruntime:Default, onnxruntime_pybind_state.cc:1641 onnxruntime::python::CreateInferencePybindStateModule] Init provider bridge failed.

[rare goblet]

Ayo? @pseudo crow level 2 !!!

[jaunty shale]

I switched to another account for it, it still does that somehow

[hallow wren]

Got it. Thanks a ton Julia. I see the url has my HF username which means it's private. You are AWESOME !!! Btw, I am currently using this as voice changer on existing script with my model and the existing audio file is very expressive however with my model, the audio file is breaking. Do you suggest any specific setting to ensure the expression is captured or can be enhanced ?

[pseudo crow]

can someone help

[jaunty shale]

how do I prevent that from happening?

[candid jetty]

u cant rlly

[molten pecan]

Lower the search feature ratio value

Using the mangio-crepe algorithm might work too. It can also help with captivating the emotion of your model, since that algorithm is more pitch-accurate and detailed

[slate pawn]

can anyone help in this, the index file isn't working even after training

[rare goblet]

Ayo? @slate pawn level 5 !!!

[molten pecan]

You can play with the hop-length setting too. The hop length value determines the amount of time it takes for it to reach a certain pitch. I recommend having it between 10-150, but see which one sounds best. If you put it really high it will sound like it's blabbering and plain

[hallow wren]

I don't know why, after generating the output once, it's not working again with the same input and model. Just throwing this error

[rare goblet]

Ayo? @hallow wren level 3 !!!

[crude siren]

wheres the Any Key? /J

[molten pecan]

Are the numbers on the bottom still running? If so just ignore that error

Otherwise reload the page

If you still get an error, go to the three dots, Settings, and scroll until you find Restart this Space, and restart it

[crude siren]

ok to be serious a moment i did what it said then the program just ended

it just crashes when i try to open

[chrome cipher]

wont let me click it

[molten pecan]

Click the text

[chrome cipher]

i did

it wont let me

[molten pecan]

Click the YouTube tutorial on the bottom

Of the rentry.org guide

[chrome cipher]

i opened it in a new tab

hmmmmmmmmm

[crude siren]

i need some aid with this bug

[hallow wren]

Okay. Refreshing the site is doing the trick. For somereason, it's not working after few attempts but that's least of my problem. I tried lowering that search feature value and even selecting mango-crepe , didn't help. After playing around with setting, I feel the first generated one was better compared to tweaking advanced settings now I forgot how exactly that particular audio was generated. Is there a way on uploading a generated audio to check on what exact settings was that generated ?

[molten pecan]

When you reload the page the settings will reset to default, don't worry

[hallow wren]

Yeah but seems like that particular one was not with default as I am not getting same result with default again 😦

[molten pecan]

Hm that's weird

I don't know how to help you with that, sorry

[hallow wren]

No worries. You have been of great help already 🙏🏻. So for this HF space that I have created, I can simply keep using it right and all the files are private to my space ? Probably removing / del some models could help save the space

[molten pecan]

Yeah. No one else will see it except you

You are welcome. Have fun!

[reef steeple]

hello, I am complete noob in all this but really want to try to train my own model, as I understand this guide is exactly about this:
https://docs.google.com/document/d/1XuxQYiqEhYrdYeCZRRLrmV_ciMKo0bV-jTCGHu_-5Cc/edit#heading=h.hg5mjyg07cdn

what I do not see in guide is where to get this colb notebook and so on

can someone help me with basics

[molten pecan]

-colab

[azure marsh]

All working colabs / spaces ☁️

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

Credits to Faze Masta and Antasma for compiling these links.

[molten pecan]

Use RVC Disconnected

[candid jetty]

how can I install tensorboard for local 😦

[crude siren]

i installed python and it still crashes

[pearl wolf]

Hi, I'm following the tutorial at https://docs.aihub.wtf/guide-to-create-a-model/model-training-colab and I reached the end. The problem is that I can't find the .pth file in my drive, as it's not appearing. The index file is there though. Also, what is the "Inference" section about? The tutorial doesn't mention any of that section

[crude siren]

why do i keep getting this when i try to extract the files?

[nova holly]

bruh I just edited my message and it got deleted

[crude siren]

it gives me this message but keeps downloading

[violet heron]

Messages that get edited after 2 minutes get deleted

[nova holly]

weird

[rare goblet]

Ayo? @nova holly level 1 !!!

[grizzled blaze]

where is the tensorboard for mangio?

[molten pecan]

Wait what why 😭

[molten pecan]

Is there a guide for it on #1159513888199540817

[timid prism]

hey there, how do i open AICoverGen again after closing the window? Pasting the link only tells me that the website doesnt exist, and trying to run the command to get a link doesnt work either.

[molten pecan]

-colab

[azure marsh]

Suggestion for @timid prism

All working colabs / spaces ☁️

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

Credits to Faze Masta and Antasma for compiling these links.

[molten pecan]

Click it again

[candid jetty]

I dont think so

ive been trying to do it all day lol

[timid prism]

imma be honest i dont understand anything on each of the colabs, but thank you nonetheless

[molten pecan]

https://discord.com/channels/1159260121998827560/1182344731800387585 here's a guide I made :b

For HF

[timid prism]

thank you, ill check it out

[rare goblet]

Ayo? @timid prism level 1 !!!

[molten pecan]

Yw ping me if you need help !!!

[timid prism]

will do! appreciate the quick response and patience <3<3

[grizzled blaze]

i couldnt find the tensorboard though

[molten pecan]

Dang

@violet heron do you know how?

[half cove]

https://docs.aihub.wtf/guide-to-create-a-model/tensorboard-rvc

[timid prism]

@molten pecan Hey there, I got an error when I tried converting now... The audio is an mp3 and should be fine?

[half cove]

mp3 should be fine

[timid prism]

how come i get an error then?

[violet heron]

Local or HF Space?

[timid prism]

i think local

not sure

sorry, im very new to this

!

[violet heron]

Did you download something and open it to get there?

Or did you go on a website

[timid prism]

no i only followed the guide by julia

website

[violet heron]

Ok I know why

Did you duplicate the space?

[timid prism]

oh?

yes

i followed every step

[violet heron]

Send a full screenshot of the tab

[molten pecan]

Reload the page

[violet heron]

I’ll tell you if you did

[molten pecan]

That should do the trick

[violet heron]

Not the full tab

[timid prism]

oh

[molten pecan]

Reload the page. That will prob fix it

[timid prism]

alright thank you

ill give it one more try

if not ill see if i can fix the other version

thank you

[molten pecan]

If not I have another solution :b

Also check this out https://rentry.org/losslessguide

I highly recommend reading it

Since you are using mp3

[timid prism]

same issue

[timid prism]

ill save this thank you

[molten pecan]

When do you get that error?

[timid prism]

for now ill just see if i get the other site working

[molten pecan]

When you download the model?

[timid prism]

when i press convert

[violet heron]

Show a FULL screenshot

[molten pecan]

Do the numbers on the bottom run or do they only say Error?

[violet heron]

Like this top bar too

[violet heron]

Scroll up

That’s not the full screenshot

[timid prism]

there's literally nothing more to scroll up to

the thing over is just the hugging face banner

[molten pecan]

Timbo is telling you to show the URL

[violet heron]

Yeah that’s what I need

[timid prism]

like this?

[rare goblet]

Ayo? @timid prism level 2 !!!

[molten pecan]

Click the three dots, Settings, and scroll until you find Restart this space

And restart it

I hope that fixes it

[timid prism]

alright, ill give it a try thank you

@molten pecan it didnt work, but its alright

[molten pecan]

Dang

[timid prism]

ill see if i get the other thing working again

[molten pecan]

Alright good luck!

[timid prism]

thank you!

[molten pecan]

@violet heron the r3gm space has been giving that error a lot lately. I think I should instead revamp the Ila RVC guide instead

[violet heron]

Have you tried Ilaria RVC yet?

[proven hill]

thats because i update Ilaria RVC costantly :3

[molten pecan]

I think the other space hasn't been touched in months and it's still on beta 😭😭

[proven hill]

also Julia, i made a simple ReadMe file, if it can help you for the guide or to explain to people since it snot the first time i see you doing that

[proven hill]

yyyeaaaaa

we just implemented the best tts on mine sooo

[molten pecan]

Oh yeah I'm aware. I wanna revamp the readme so it's prettier and less cluttered

[proven hill]

nono, i completely remade it

[molten pecan]

Wait what when how

[molten pecan]

https://docs.google.com/document/d/186_a8d32iYxr1UAEm1LAFQvLu2w3huVIg-KTlTHshg8/edit?usp=sharing this one?

[timid prism]

Ill give it a try if this one doesnt work!

[grizzled blaze]

there is no installer

[weary fable]

for my question.. How did the voices get separated easily like this?

https://www.youtube.com/watch?v=3YVwhLIAs50

[rare goblet]

Ayo? @peak barn level 1 !!!

[hoary estuary]

hello

with amd gpu wich version should i use ?

for voice rvc

[marsh goblet]

Can anyone give me a rundown on setting up tensorboard?

[violet heron]

RVC or the real time voice changer

[hoary estuary]

I found that it was the DirectML version but I'm quite disappointed, is it because it only really works well on nvidia or is it something else?

[rare goblet]

Ayo? @hoary estuary level 1 !!!

[charred shell]

Ilaria rvc takes hundreds of seconds to process before giving me an error is this normal i dont think im messing up on any parts

[molten pecan]

Did you duplicate the space?

[charred shell]

Yes

[molten pecan]

Are you having a stable internet connection?

Are you exiting the tab while it processes?

[charred shell]

Yes and no

Oh shit i got my first inference with it

It did take 500 seconds tho

[violet heron]

DirectML is for AMD

Is your GPU selected

[hoary estuary]

I have to choice between gpu0 gpu1 gpu2 gpu4

I have a rx 6750xt

[violet heron]

GPU 0

[hoary estuary]

Yes

I touched all the buttons and watched config videos but it's still not very fluid

[flat jackal]

whats the best egirl voice

[violet heron]

#1181319399936893029

[flat jackal]

does it work for deep voices

[violet heron]

If you adjust the settings yes

[flat jackal]

what setting sshould i put

[rare goblet]

Ayo? @flat jackal level 1 !!!

[violet heron]

Depends

[flat jackal]

can you help me

[violet heron]

Can’t right now

I would just mess around with certain settings

[obsidian tree]

I just trained a custom model with rvc-studio, and I see a new '.pth' file in my models folder, but whenever i refresh the model listing on any other page it always comes up empty. Anyone know what could be going on here?

I can also see the model on tensorboard...

[rare goblet]

Ayo? @obsidian tree level 1 !!!

[kindred pike]

how can i connect the voice changer to my discord

[violet heron]

Download a virtual audio cable

Set it to your output on Okada and your input on discord

[atomic mulch]

what is this file that comes with a lot of models?

What is "metadata"

[violet heron]

What

A model should only have a PTH and Index

Metadeta is not needed and shouldn’t be there

[atomic mulch]

thats weird

came with two models

Is it okay to rename PTH and index files?

[past jungle]

What is the best configuration to listen to the audios after conversion in Hugging

please help me

rmvpe? mangio-crepe?

[vernal forge]

Rvc

[vapid gust]

hello

i need help on my dataset

i have some room echo on one of my dataset files and i wanted to know how to remove it

[violet heron]

RVMPE

-howtoask

[azure marsh]

How to troubleshoot

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

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

[pure pond]

what do i put on "Link"?

[peak barn]

Oops…

[pure pond]

it said hugging face, but i don't understand what exact link it need to use #1159514067187277865

[ornate gulch]

why i get a none error when i try upload a file to hugging face ?

[peak barn]

Sorry…

[violet heron]

A HuggingFace link

-upload

[azure marsh]

huggingface.co upload Guides

English guide

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

Spanish Guide

Link: https://rentry.org/AI_Hub_guide_esp
Credits: Julia (ailen2091)

Hindi guide

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

Turkish guide

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

[pure pond]

and i thought it could be easier..

[rare goblet]

Ayo? @pure pond level 1 !!!

[vapid gust]

Could someone help me?

[ornate gulch]

how do i upload files to rvc_hf ? hugging face

[viscid moss]

tell me

[viscid moss]

What space of HF exactly?

[ornate gulch]

when i upload a file and click convert it gives me a none error

RVC-HF, by r3gf

[viscid moss]

Have you duplicated the space in your HF account?

[ornate gulch]

no , i just used it without any account

[viscid moss]

When you do not duplicate it, a queue is created that can cause errors

Create an account, it's free

[ornate gulch]

is there are any tutorial on youtube that can help to do it ?

[viscid moss]

I've one but is in Spanish. do you want to see it?

[ornate gulch]

ok

[viscid moss]

https://youtu.be/s-aR-Y-ub9Q?si=MbS7DVoz-YuuAvJW

[ornate gulch]

thank u

[viscid moss]

If u have a problem ping me

First step is create an account

[tame mica]

then go to the space, click the button with 3 dots (looks like ":") and click duplicate space

[covert jetty]

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for fairseq
Failed to build fairseq
ERROR: Could not build wheels for fairseq, which is required to install pyproject.toml-based projects

help

tying to pip install error on installing fairseq

[rare goblet]

Ayo? @covert jetty level 1 !!!

[frosty jasper]

I'm trying to use RVC-GUI but it just says "running with the runtime python please wait" in the console and never launches the application

using windows 11

[little bobcat]

i need some help with creating an AI voice model

[distant turtle]

Does anyone know where I should put the pth and index file? Thanks

[slate pawn]

daily search for solution to the index file that is not working on the rvc despite being trained

[old radish]

why am i getting download failed for Ilaria_rvc space?

--2023-12-08 10:35:40--  https://huggingface.co/darthanonymous1/DarthRVC/blob/main/SquidwardTentacles.zip
Resolving huggingface.co (huggingface.co)... 10.0.140.123, 10.0.151.27, 10.0.117.141, ...
Connecting to huggingface.co (huggingface.co)|10.0.140.123|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 39987 (39K) [text/html]
Saving to: ‘./zips/squidwardtentacles.zip’

     0K .......... .......... .......... .........            100%  330M=0s

2023-12-08 10:35:40 (330 MB/s) - ‘./zips/squidwardtentacles.zip’ saved [39987/39987]

it obviously didnt fail

it downloaded fine

[brittle wing]

anyone know how to fix no module named "torch"?

[dense ore]

HELP everyone what rvc program do i download and how do i set it up cuz im using okada rn and its NOT good

[rare goblet]

Ayo? @dense ore level 1 !!!

[dense ore]

stfu bot

[hard gust]

Can we train with AMD GPU DML now???

[kindred pike]

what is the best virtual audio cable

[kindred pike]

how do i link my vertual cable to my discord and the voice changer and my mic
it seems that theres no voice

[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

[kindred pike]

is it casually laggy ?

it doesn't read all my worlds

[sand iris]

Nahh

[proper shale]

depends on your GPU and your discord mic threshold/volume of the ai output

[distant turtle]

where should I put the index and pth file?

[brittle wing]

what should i do whenever i open start_http.bat it just keeps quitting and not letting it load the voice changer client?

[sand iris]

Try this

#🔍│help-w-okada message

[brittle wing]

thank you!

[proper shale]

pth in weights

make a folder in logs with the model's name for index

[karmic heron]

someone got a RVC colab? i cant use hinas

[proper shale]

only Huggingface spaces

[karmic heron]

shitt

can you give me the link please?

ok i will just use illaria`s

[proper shale]

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

[karmic heron]

oh lol

[brittle wing]

can anyone just help me on how the fuck am i supposed to do in this

asked multiple time and noone helping at all

[proper shale]

maybe didn't install the requirements successfully?

its missing modules so

[brittle wing]

wdym

how do i insstall that specific modue

module

[proper shale]

you do a pip install

idk what pip install exactly bcuz theres torch audio n shi

[proper shale]

certified

[brittle wing]

why

im ussing the 2.20

2.2.0*

[karmic heron]

how do i train a AI on this?

[rare goblet]

Ayo? @karmic heron level 4 !!!

[brittle wing]

i dont have nvidia gpu

[proper shale]

you'd have to use RVC Disconnected for that

-colab

[azure marsh]

All working colabs / spaces ☁️

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

Credits to Faze Masta and Antasma for compiling these links.

[brittle wing]

does it work for you

ok

[crude siren]

so i cant get local programs to work and kits ai is crap

any alternitives to making models/covers?

[proper shale]

colab

AICoverGen

[crude siren]

how good is it?

and more importantly is it free

[karmic heron]

tsym @proper shale

now im gonna do something while its training

[alpine roost]

this happened to me sometimes then it stopped happening and now its happening again in gradio it shows the final model that was created but in gdrive it dosent show up? only these extra models show up https://cdn.discordapp.com/attachments/1155555886102749295/1182741773697687613/image.png?ex=6585cd1d&is=6573581d&hm=a04f9e01fcc878daba1f1ccc215ac43d6af7fb7314ae8c05a806564871655bc9& https://cdn.discordapp.com/attachments/1155555886102749295/1182741721918996641/image.png?ex=6585cd10&is=65735810&hm=7d9b0a9e664ca38fda16bcb03db12853104dfe917e6b7505859172dad99a2b6f&

[proper shale]

np

[proper shale]

free and most likely better than kits

[crude siren]

nice! can i have a link

also being better then kits itsnt hard to do

[proper shale]

https://colab.research.google.com/github/hinabl/AICoverGen-Colab/blob/main/Hina_Mod_AICoverGen_colab.ipynb

[crude siren]

ty

[proper shale]

np :)

[peak barn]

-howtoask who is recommended to ping?

[azure marsh]

How to troubleshoot

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

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

[alpine roost]

nvm i fixed it

[karmic heron]

i finished training my model but i just got my index file

INFO:idk the marcher guy:Training is done. The program is closed.
INFO:idk the marcher guy: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 '

[zealous shuttle]

if the sample rate had this much, then which sample rate is recommended for training?

[rotund zephyr]

guys how do you run with local tunell

[onyx scroll]

I need a little help, might be a silly question but how do i activate voice changer to make it work on things like discord etc and whenever i activate a voice i can always hear myself talking in set voice, would like to turn it off so i can focus on what i am saying

[karmic heron]

just put the monitor off

like

put the monitor as none

[onyx scroll]

it was by default

[karmic heron]

just put as none

it will work

[onyx scroll]

nothing changed sadly

[hasty mesa]

what does index do exactly

[full imp]

what's the most noob friendly way to train a model rn?

[alpine badger]

i need help

i activated it and it works but they cant hear me

[proper shale]

u can do 40khz easily

[merry ruin]

What does this setting do?

[kindred pike]

any idea on how to remove those weird electronic voice when i talk with the voice changer ?

[rare goblet]

Ayo? @kindred pike level 2 !!!

[karmic heron]

which one i should get?

[violet heron]

Download a virtual audio cable

You need to have it to use on other apps

[lethal crown]

well I would test them all, but the finished model is Caine.pth but idk how much audio you used to train, so you may have over trianed, so if you hear background sounds in your model or it sounds worst you should take one of the earlier models. you can tell how many epochs it trained for by the e in front of the number. so check your finished model first if sounds good you can delete the rest. if it sounds bad start checking the lower models, go from e1000 to e900 e e800 and so on till you like what you hear

IDK what was wrong with Mango RVC version 23.7.0. but both my models sounded the same after 10 or more try's of training. but I downloaded the version of RVC by it's self. and using harvest setting, it worked on first try. I'm still training but I checked the 50 epochs and the 100 epoch models I set it to 400 epochs so be done tonight anyway, both the 50 and 100 epochs models sound great.. so if you having problems with Mango use this version of RVC without the extra Mango. I am training Japanese Beatrice from Rezero. can't wait to try Alucard next from Castlevania SOTN PSX. very happy it's working so far now. 😁 just wanted to post this incase anyone else was having this problem with Mango.

[rare goblet]

Ayo? @lethal crown level 2 !!!

[onyx forum]

how many epochs should i use to train with a 13 minute dataset

[rare goblet]

Ayo? @onyx forum level 3 !!!

[violet heron]

Depends how clear and HQ it is

[onyx forum]

i had to extract it from the song with uvr, but i used 0.95 overlap so it sounds pretty high quality. i also removed noise/reverb and everything like that

[violet heron]

Try 300 and see if that works

[onyx forum]

but the singer also has lots of rasp so i dont know if that will affect it

but ill try 300

[rocky karma]

Got a really weird error

It literally just says 'None'

Tf happened?

[dreamy moon]

So i installed the AI Voice Changer but i can't hear anything when i talk

[violet heron]

Did you click start

Did you set your input and output

Did you select your GPU

[dreamy moon]

yep

[violet heron]

AMD or Nvidia?

[dreamy moon]

nvidia

not sure if that helps but console says "RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory"

[rare goblet]

Ayo? @dreamy moon level 1 !!!

[dreamy moon]

i mean, if i press passthru i can hear my normal voice

[sand iris]

Corrupted files

In the MMVCServerSIO folder, delete pretrain and model_dir. You'll have to reimport your models

But on the next launch, it'll redownload stuff that it needs to work and hopefully it won't corrupt that time

[raven bloom]

anyone know why all the female voice models sound the same? I'm trying to make a girl sing a guy's song but every female voice sounds the same when i convert

[violet heron]

Turn up the transpose

[raven bloom]

I did, every time they sound the same though

[rare goblet]

Ayo? @raven bloom level 1 !!!

[violet heron]

By how much?

[raven bloom]

I've tried 0, 4, 8, and 12 and everytime each conversion of different voice models have sounded the same

[violet heron]

The exact same?

[raven bloom]

the exact same

[violet heron]

Are you sure your converting it

[raven bloom]

yea, it works with the male models but the female models wont work

[violet heron]

Try a different female model

[raven bloom]

so you think all 4 voice models I've used were bad models?

[dreamy moon]

it worked 🙏

[raven bloom]

I just got 2 new models and they still sound the same

anyone else think they can help?

[violet heron]

It shouldn’t

[raven bloom]

maybe a bad download of the Mangio?

[raven bloom]

^

[sweet cradle]

I used Mangio the other night to great effect in all of its features. But today, when trying to train another model, the data set is unable to be processed. This is my assumption, as after pressing the button to process the dataset of a single 4-minute audio file, it has been running for over half an hour with no output. Even the console window is showing nothing past launching the web app.

[old radish]

Bump

[sweet cradle]

Update: I have received an error message from the console window.

future: <Task finished name='etidfh14ar_14' coro=<Queue.process_events() done, defined at C:\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py:343> exception=1 validation error for PredictBody
event_id
  Field required [type=missing, input_value={'data': ['C:\\Mangio-RVC...ion_hash': 'etidfh14ar'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.5/v/missing>
Traceback (most recent call last):
  File "C:\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 347, in process_events
    client_awake = await self.gather_event_data(event)
  File "C:\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 220, in gather_event_data
    data, client_awake = await self.get_message(event, timeout=receive_timeout)
  File "C:\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 456, in get_message
    return PredictBody(**data), True
  File "C:\Mangio-RVC-v23.7.0\runtime\lib\site-packages\pydantic\main.py", line 164, in __init__
    __pydantic_self__.__pydantic_validator__.validate_python(data, self_instance=__pydantic_self__)
pydantic_core._pydantic_core.ValidationError: 1 validation error for PredictBody
event_id
  Field required [type=missing, input_value={'data': ['C:\\Mangio-RVC...ion_hash': 'etidfh14ar'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.5/v/missing```

[peak barn]

Hi, I finished installing Applio and I'm trying to launch it using the default Windows command prompt. However, when I try to launch Applio it immediately gives me this error:
Traceback (most recent call last):
File "infer-web.py", line 14, in <module>
import lib.globals.globals as rvc_globals
ModuleNotFoundError: No module named 'lib.globals'
This is a python module within a two directories, but it's not importing correctly. Is this a code error? If so, can this be fixed, please? Otherwise, what can I do to fix this? I have Python 3.9.8 installed managed by PyEnv.

[rare goblet]

Ayo? @peak barn level 2 !!!

[raven bloom]

did you run the .bat file?

[rare goblet]

Ayo? @raven bloom level 2 !!!

[peak barn]

Go-applio.bat? If so, yes

[peak barn]

It's been fixed in the latest version.

[rotund zephyr]

guys why does res keep rising up when i speak?

[rare goblet]

Ayo? @rotund zephyr level 1 !!!

[peak barn]

Hi, I finished installing Applio and I'm training a model, but I want to train with my CPU because my GPU has 2GB of memory. (I chose option one to install and launch Applio) I put "1" for the GPU indexes but instead of stopping, it hangs with this message until I manually stop.
\Applio-RVC-Fork\env\lib\site-packages\torch\amp\autocast_mode.py:204: UserWarning: User provided device_type of 'cuda', but CUDA is not available. Disabling
warnings.warn('User provided device_type of 'cuda', but CUDA is not available. Disabling')
I want to train using my CPU in Applio but I'm not sure how else to do it.

[fossil glen]

you need vga gtx or rtx included cuda

[peak barn]

However I want to train using my CPU instead of my GPU. How can I do that?

[fossil glen]

train with cpu? are you want explode a pc

[peak barn]

Yes.

[fossil glen]

only inference if you want run with cpu

[fossil glen]

use

[fossil glen]

-colab

[azure marsh]

Suggestion for @peak barn

All working colabs / spaces ☁️

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

Credits to Faze Masta and Antasma for compiling these links.

[peak barn]

Ugh, I guess that's it for Applio.

[fossil glen]

its same base

recomend easyGUI, by rejects

[peak barn]

Thanks! There were a lot to choose from...

I could have sworn Applio allowed to train via CPU once...

[fossil glen]

never train with cpu

[rare goblet]

Ayo? @fossil glen level 12 !!!

[peak barn]

I never had any problems with that, but if you say so.

[peak barn]

easyGUI doesn't allow me to change sample rate. Imma stick with RVC disconnected for now

[proud elbow]

6 GB VRAM is the possible bare minimum, and trying to do using CPU is like trying to run some modern games in 0.1 fps with CPU software rendering (aka. Crysis meme)

[peak barn]

Jeez

[lethal crown]

like I said in my post before. that the same thing was happening with me every model I trained sounded the same. so I stopped using Mango RVC and just went with RVC and used harvest to train with. and it worked. idk what you are using to train but if it's Mango might be the problem. as I said IDK why. maybe it works for some gpu's and not for others? just a guess

[dusk tendon]

anyone know why rvc2disconnected just didn't create a "weights" directory?

[proud elbow]

did you train using mangio and have NaN loss values issue?

[pseudo steeple]

Hello, I didn't notice that AI Hub got blasted so can somebody give me a run down on what the new developments are?

[lethal crown]

did you look in assets? my weights folder in RVC is there where in Mango it was the main Mango folder idk why just a bit different.

[lethal crown]

sorry I'm new so only thing I know is I followed all the steps I see guy do in a video. and every model sounded the same. my audio was just voice audio took from game files

[rotund zephyr]

guys when i switched to server client, nothing comes out like zero on every 0 res 0 buf, nothin

[frosty jasper]

I keep getting "Error: None" on RVC HF

[rare goblet]

Ayo? @frosty jasper level 1 !!!

[proud elbow]

uh, the issue is more of the inference using Mangio for any models, right? yeah I guess you gotta try RVC or some others to see

[lethal crown]

well I already switched to just RVC and works great Beatrice sounded like Beatrice with 50 epochs so that's why I think it was Mango

still training and I'll keep the best version

funny thing was it worked just fine with downloaded models, but it would not train werid

well it trained fine it just didn't sound anything like the dataset, and even the girl model sounded like a genaric guys voice.

[raven bloom]

where can I find RVC?

[lethal crown]

I have a question. is there a way to have it run on the pc without opening a webpage? my worry is the program may stop working one day, would love to just have it instailed and don't have to worry about them updating something and one day when I hit go web.bat it says this website cannot load.

[lethal crown]

hold on a minute

[lethal crown]

watch this video

https://youtu.be/hB7zFyP99CY?si=J40vQEx-qgqHec0O

[violet heron]

#1159949278270193734

[hollow vigil]

is there a file that i can move that transfer all my settings for each voice onto a new updated version

instead of moving it manually

everytime

like i want it to save my settings for every voice but on a newer updated version

[pseudo steeple]

yo what's "Truncate silence," I've been reading "RVC_making-models" from the #1159513888199540817

[finite galleon]

Basically an automatic way to remove silence from audios

[pseudo steeple]

👍

[proud elbow]

basically shortening signals below set threshold to very short time, so it won't be detected as silence (RVC detects "silence" as signals below -42 dB in 0.4 sec)

[pseudo steeple]

how does one download and set up tensorboard again locally? The thread for that channel got deleted along with old RVC

And I don't want to go through the nightmare of waiting 4 days of trouble shooting and endless nights to go train...

[strong belfry]

I just installed it with pip install tensorboard

and you just run tensorboard and point it at the logs folder or run it in the logs folder directory

[pseudo steeple]

I downloaded RVC v2 1006, no sight of it

Where do I go get pip install tensorboard?

[strong belfry]

you open a cmd window and install tensorboard using pip

you can open a cmd window in your current folder by typing cmd in the file explorer address bar

[pseudo steeple]

cnd prompt is open as administrator of course

But what is pip?

I haven't heard of it before

[strong belfry]

google pip install

[pseudo steeple]

I've already downed python

[strong belfry]

its bundled with python

[pseudo steeple]

just as a heads up

...

[strong belfry]

pip install is the command to download python packages

so

pip install tensorboard

[pseudo steeple]

update?

[rare goblet]

Ayo? @pseudo steeple level 3 !!!

[strong belfry]

you can kind of just ignore the pip upgrade

no harm in doing it tho

[pseudo steeple]

:thumbs:

[strong belfry]

but now run 'tensorboard' in the cmd window

it will probably not work

[pseudo steeple]

you figured

[strong belfry]

cause you should run it in the rvc folder

[pseudo steeple]

alr

[strong belfry]

where the logs folder is

so

cd 'directory'

to nagivate the cmd window to the rvc folder

[strong belfry]

cd "path to the root rvc folder"

replace the text obvs

no quotes needed i think

[pseudo steeple]

oh shit

I took it too literally

huh

[strong belfry]

lol

pause

if you create a .bat file in the rvc folder

this should just work

if you know how to do that

[pseudo steeple]

probably not

it's been 4 months

since I worked with RVC

[strong belfry]

create a blank txt file

paste in the above

now you want to 'save as' the txt file, call it something like starttensorboard.bat

run the .bat file, which is basically a shortcut to typing in cmd commands

[pseudo steeple]

I'm in baby

[strong belfry]

yesyes

now open a web browser

[pseudo steeple]

I got it

dont worry

[strong belfry]

kkk

should be all good now

[pseudo steeple]

thanks

[raven bloom]

any tips to remove the popping out of the vocals?

[strong belfry]

what popping sounds?

[pseudo steeple]

wrong link?

[raven bloom]

when i converted my audio file, theres like a cracking sound when I play my audio file

[pseudo steeple]

wrong link...

[strong belfry]

i would need to hear it cause there are many different causes and kinds of popping sounds

[raven bloom]

I'm sending you the file in DM

[strong belfry]

voice model sounds kind of robotic to begin with, seems like there is also backgound noise/vocals that is not just the main vocals leaking though the inference. Not sure about the 'popping' sounds tho

and half way though seems you've gone completely out of the vocal range of the model

[raven bloom]

let me try with a different audio file in case its the file thats the problem

[strong belfry]

or that the audio was just bad

[raven bloom]

I did UVR Kim Vocal 2, the De echo/reverb, then HP karaoke

[strong belfry]

I couldnt tell you unless i also hear the original vocal

overlapping voices, screaming, yelling etc will be impossible to make a perfect inference out of

[raven bloom]

there was popping on the original, that was the issue

[pseudo steeple]

What's the general rule for batch size-vram ratio?

[strong belfry]

batch size < vram GB

not always the case

but on the safe side

If you tell me how many gb you have then we could just recommend a batch size

[proud elbow]

you can go 8 for 8 GB/more VRAM (the original RVC fits just below 8 GB but Applio may use slightly more than 8 GB)

[strong belfry]

Applio used less vram for me huh

[proud elbow]

probably not the latest applio version I have

[pseudo steeple]

got 8 gigs so yeah

figured I'd use 8 then

👍

although, I'll see if I can squeeze a 9 in there

Recommended dataset size?

There's a possibility of overfeeding know now (whoops)

so I'm a bit afraid of using uh

large datasets

[proud elbow]

5 mins-1 h, being too short may be more unstable (tends to have more mode collapses), unless you want to make short "meme" models, but quality matters more

[pseudo steeple]

1 hour is cutting it or it's within the goldilock zone?

[cobalt seal]

I've a runtime problem in RVC. What should I do? Please tell me

[pseudo steeple]

describe the proble,m

[cobalt seal]

Whenever I use it, runtime disconnected after every 4 min or 5 min

[rare goblet]

Ayo? @cobalt seal level 1 !!!

[cobalt seal]

Yes

[pseudo steeple]

collab?

[cobalt seal]

Yes

[pseudo steeple]

Your connection is stable, no?

And your storage isn't full?

[cobalt seal]

My storage isn't full.

[cobalt seal]

Internet connection is sometimes stable sometimes not

[pseudo steeple]

probably not the internet then

you run everything through step by step?

[cobalt seal]

Yes I do, but after sometime runtime disconnected

[pseudo steeple]

well, I have no clue, sorry

[rare goblet]

Ayo? @pseudo steeple level 4 !!!

[cobalt seal]

Okay thanks. If you get solution please tell me

[potent root]

should i manually do speaker diarization on a 1hr file

or is there an efficient algorithm for this

[proud elbow]

with silences truncated

[neon cosmos]

I’m trying to sound like a little kid for funny videos what voice model should I use?

[acoustic garnet]

find a kid like character and search up their voice model?

[neon cosmos]

Any recs?

[rare goblet]

Ayo? @neon cosmos level 1 !!!

[neon cosmos]

Can’t seem to find any

Tryna sound like a boy

[hasty sapphire]

does anyone know how to fix voice.ai when it says "you can't upload files of this type"? Idk what im doing wrong, its just a zip file i downloaded from happy face

[sudden tartan]

Why does this keep happening to me

[granite lynx]

where can i download rvc 2

[sudden tartan]

What does red text mean on rvc

[proper shale]

that's an error

have you tried using other forks?

[tame mica]

what the hell is that fork

also no rmvpe

strange

[proper shale]

that's the RVC-GUI by that Tiger guy

hasn't been updated in months

[sudden tartan]

Idk what that is

[proper shale]

forks are basically slightly different versions of an app with new GUI or features. RVC has other local forks, like Mangio-RVC-Fork and Applio.

[sudden tartan]

Whats the lasted rvc?

[proper shale]

there's mangio, applio and the mainline

[pale spindle]

it keeps saying 'bad cv2 for a model'

[proper shale]

-loc

-local

[azure marsh]

All Working Local Forks 🖥️

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

Need some help? 🤔

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

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

[proper shale]

Original RVC = Mainline

[pale spindle]

help

it says

"Bad CRC-32 for file 'insertmodelnamehere.pth'"

for every model

[rare goblet]

Ayo? @pale spindle level 1 !!!

[pale spindle]

i upload

[proper shale]

what are you using?

[pale spindle]

the google colab AI cover generator

[proper shale]

is it the one with no webUI?

[pale spindle]

no, its the one with web ui

[proper shale]

Can you send a screenshot of how the page looks like on the Download Model section?

[pale spindle]

ok

wait it's working again for some weird reason

[pale spindle]

help again, the "Bad CRC-32 for file 'insertmodelherepth'" error continues
i am on the google colab webui ai cover generator

here is a screenshot

[sand iris]

What's the model you are using?

[pale spindle]

sorry, im a bit confused on what you mean by 'model'

[sand iris]

To change the voice, you have to give it a voice model

Where did you get that voice ?

[pale spindle]

i got it from the voice model channel

[finite galleon]

Why dont u just use the link?

[pale spindle]

i did

[finite galleon]

Why download and upload

[pale spindle]

?

[rare goblet]

Ayo? @pale spindle level 2 !!!

[finite galleon]

Well it looks on the screenshot that you are uploading the file itself

[pale spindle]

i would

but i don't know

exactly how to insert a link

[finite galleon]

Go to the download model tab

And put in the link

[pale spindle]

ok

ok i'll try that?

[finite galleon]

Ye

[pale spindle]

doesn't work

[sand iris]

can you send the link ?

Model might be bugged somehow

[pale spindle]

ok

https://huggingface.co/yaniroo/YanAI/blob/main/MikuAI.zip

it basically says that error

for every model i upload after the first one is sucessful

so i don't think it's a model problem

and i've tested it on hugging face and colab

neither work

[sand iris]

I mean if the same voice model doesn't work on multiple platforms, I think it's fair to say it's the issue

[pale spindle]

true, but its the issue for every model

[sand iris]

ah

You tried other colabs ?

[pale spindle]

yep

either my pc is incompetent or idk what's wrong at all

[glad zealot]

coven gen acting weird?

[sand iris]

idk

[rare goblet]

Ayo? @sand iris level 47 !!!

[sand iris]

I'm thinking corrupted zip, but unsure

[glad zealot]

lemme try with my models

i mean if it fully breaks other than google breaking it, i have no idea how to fix it since i just made the bypass

[pale spindle]

if it helps i have the error code?

gradio.exceptions.Error: "Bad CRC-32 for file 'bfdiapinv1.pth'"
2023-12-09 12:04:37 |INFO|httpx|HTTP Request: Post: http://127.0.0.1:7860/api/predict

[glad zealot]

nope

no idea

[pale spindle]

it says its an internal server erro

r

[glad zealot]

the original creator of it isnt in the server so idk

[pale spindle]

but i have no idea what in the heck that means

[glad zealot]

what does the zip file look like inside?

yup its the zip file

i tried it on mine

[lament eagle]

Should I train this 32k or 40k

[glad zealot]

try using a different model

[pale spindle]

ok

i'll try

maybe all of the models i downloaded are outdated haha

[modern nest]

what does this mean

[brittle wing]

what should i open

[unkempt mortar]

rvc-gui.bat

[brittle wing]

ok

[copper pebble]

a bit outdated

ya wanna do covers dont you?

[brittle wing]

yeah

i am trying anime 💀

[copper pebble]

dont use that bro thats pretty outdated

https://discord.com/channels/1159260121998827560/1175615544037212200

https://discord.com/channels/1159260121998827560/1167844377046040769

use one of these two

way better @brittle wing

[brittle wing]

ok

thanks

[modern nest]

whta do i do

[pale spindle]

ok i know it's been 30 mins or so

but i've tried legit all models

and it doesn't work

on any ui

my pc is sucky

[pale spindle]

the ones on the front page, rvc v2, et

can anyone help

please

i can't upload ANY models at all

and it's rly annoying

[copper pebble]

watchu using?

[pale spindle]

Hina_Mod_AICoverGen_colab.ipynb

or the one with the web ui

i've tried EVERYTHING im desperate

[glad zealot]

i just used it and works for me

[pale spindle]

is it just me then?

[copper pebble]

https://discord.com/channels/1159260121998827560/1175615544037212200 this one?

[pale spindle]

nope

[copper pebble]

did u try ilaria rvc

[proper shale]

maybe? let's try to rule out some variables here

[pale spindle]

nope

[pale spindle]

ok

[proper shale]

what did you insert as the models nae

name*

[glad zealot]

can you give the link to the model?

[pale spindle]

there's multiple

[glad zealot]

just one of those models you tried

[pale spindle]

their name like if the character is called 'character' i name it 'character'

[pale spindle]

ok

[proper shale]

do you happen to be using any japanese characters in the name?

[pale spindle]

https://huggingface.co/AIMan2001/Deku/resolve/main/Midoriya.zip

[pale spindle]

nope

just english

[proper shale]

hm

any spaces?

[pale spindle]

nope, one word

the error is this

[finite galleon]

Check the colab to see which script and which line the error accoured on

[proper shale]

I'd suggest just using the link for the model

Or uploading it somewhere and copying the link

[pale spindle]

sure, but how to do that?

[proper shale]

Just use the links in #1175430844685484042

[finite galleon]

Just copy and paste it to the download tab...

[pale spindle]

IK BUT LIKE

WHICH ONE

[proper shale]

Or you could upload the models in Huggingface

[pale spindle]

ok

[proper shale]

In the Download model section specifically

-huggingface

[azure marsh]

huggingface.co upload Guides

English guide

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

Spanish Guide

Link: https://rentry.org/AI_Hub_guide_esp
Credits: Julia (ailen2091)

Hindi guide

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

Turkish guide

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

[finite galleon]

Should be lile https://huggingface.co/user/bla bla bla/something.zip

[pale spindle]

i know, but where to paste it

i know the links

[finite galleon]

On the download models tab

[glad zealot]

imma try the same model as then

[finite galleon]

On the models tab theres an input for a url

Paste it there

On the name input you put in the desired model name

[pale spindle]

ok

[glad zealot]

yup puting the model they sent works just fine

[pale spindle]

ah