✨│ai-help

[dense marsh]

Is it because it’s an mp3?

Is there a specific file type it has to be?

[half cove]

on colab run gui it will show the error

[mortal ibex]

I clicked the link from the gradio in the colab

and it brought me to the RVC thing

so i closed the colab tab

[half cove]

where you clicked the gradio it will show the error

[dense marsh]

Using this btw

[mortal ibex]

no I opened the link on the gradio and it brought me to the rvc page so i did the steps on the link u sent me and then when i converted it it gave me that error on screen

[half cove]

yes and if you get an error on that screen it will output it in the run cell

[half cove]

try using the colab

[mortal ibex]

so i had to keep the colab on

[half cove]

just go on colab and screenshot

the entire page

[mortal ibex]

k

[lime otter]

I'm using easy GUI

-colab

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[mortal ibex]

wait i got it

[half cove]

use a newer one

[lime otter]

I miss easy GUI 😢

[dense marsh]

Same thing, all of my mp3 audio files are just grey

[mortal ibex]

ok it sounds very choppy am i suppose to get rid of the reverbs and harmony

[half cove]

there is a newer easygui

[half cove]

yes

[half cove]

elaborate

[mortal ibex]

no instrumental?

just the voice right

[half cove]

yes

[dense marsh]

I can just send a ss

[mortal ibex]

alrighty ty

[half cove]

ok

[lime otter]

Fr?

[rare goblet]

Ayo? @lime otter level 2 !!!

[half cove]

-colab

[azure marsh]

Suggestions for @lime otter

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[dense marsh]

It’s my mp3 files, and it wont let me click them

[half cove]

its prob messing up cuz your on mobile

weird

[dense marsh]

I wish I could drag and drop

I mean ig I could try requesting desktop site

Didn’t work

I’ll try chrome

[brittle wing]

Help pls

[dense marsh]

Didn’t change it

Maybe I should try a different colab

[dense marsh]

I believe that’s an invalid huggingface link

[lime otter]

easy gui broke, use something else

[dense marsh]

Ok

I was wrong

[brittle wing]

i just copied out of the channel do you know the right way to put it?

[proper shale]

this colab doesnt work anymore

[dense marsh]

No it’s write

[proper shale]

use the ones in https://aihubdocs.github.io/en/

[dense marsh]

That was my mistake

[brittle wing]

allg

[brittle wing]

alr ty

[dense marsh]

I figured it out, how do I download it now

[rare goblet]

Ayo? @dense marsh level 4 !!!

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[low cliff]

send me the link

[fluid matrix]

that colab is very old, if you need a colab to train a model use the ones listed here instead

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

rvc disconnected for example

[brittle wing]

i got it fixed ty tho

[honest dune]

why when i export to onnx it trys to download a undifined.htm file

[obsidian marten]

same problem

[tawdry junco]

does anyone know if i can change the threshold settings in the app?
it picks up my mouse sounds even at max 🫠

[low pecan]

hey does anyone know a google colab that works now that this one doesn't work anymore

[astral sparrow]

How to fix it? Im using AMD graphic card

[thin stump]

I uploaded on #1175430844685484042 but the model isn't appearing on weights. Should I reupload again?

[willow lichen]

When did you upload? Want to send me a link? There's no need to upload again, we should automatically grab it.

There might be a small delay, but let me know if it's been a while

[thin stump]

It says 38 minutes ago

I had to delete the original submission since it lacked the title descriptions so weights gave it 32,000 epochs 💀

#1213697027318353940 message

[half cove]

Your using colab your gpu doesn't matter but make sure your using gpu runtime

[half cove]

-colab

[azure marsh]

Suggestions for @low pecan

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[astral sparrow]

how to tell if im using gpu runtime

[half cove]

Top right it will say t4

[astral sparrow]

Is that all?

[rare goblet]

Ayo? @astral sparrow level 2 !!!

[astral sparrow]

Im going to try after

[half cove]

Yes

[astral sparrow]

alr thanks

[brittle wing]

-colab

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[rugged stirrup]

Anyone know what hop length does in applio?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[astral sparrow]

Cannot connect to GPU backend Does that mean I have to wait for tmr

[deep phoenix]

I downloaded Pyfron and Pytorch and I think I downloaded pip but when I use the bat file the thing dosent pop up

[half cove]

Yes

[deep phoenix]

@peak urchin 3rd step isin't working

[peak urchin]

what graphic card are you using?

[deep phoenix]

Nividia

[peak urchin]

pip install -r requirements.txt

[deep phoenix]

it said no such file or directory

[peak urchin]

enter RVC folder with bunch of stuff in it

[deep phoenix]

the mmvc thing?

[peak urchin]

cd C:/Users/.../RVC1006Nvidia/

replace path with your

[deep phoenix]

It said it couldn't find it

do i gotta install something

[peak urchin]

oh, no

wait

MMVC?

[deep phoenix]

yeah the mmvc servio iso

[peak urchin]

you are installing w-okada voice changer?

[deep phoenix]

no

whats that

[peak urchin]

basically a voice changer for people to convert their voice with RVC and other architecture models

[deep phoenix]

I was just trying to get the thing from Ai tools search video

are they the same thing

[peak urchin]

can you mind posting where you download the file

[deep phoenix]

https://github.com/w-okada/voice-changer/blob/master/README_en.md

Oh is that the okada voice changer?

[peak urchin]

yeah

[deep phoenix]

oh

[peak urchin]

Help from that is another channel ( #🔍│help-w-okada ) but welp

[deep phoenix]

And you were trying to help me install the other one?

[peak urchin]

I thought you just say this

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

which to be fair can also realtime conversion.

but w-okada is also a general choice

[deep phoenix]

yeah my bad

[rare goblet]

Ayo? @deep phoenix level 3 !!!

[deep phoenix]

do you know what commands i should do then?

[peak urchin]

way simpler

you just run start_http.bat inside the folder

after unzipping the 3GB size full file

and you are pretty much on your way to go downloading models from #1175430844685484042 or default one

[deep phoenix]

i just need to finish donwloading pytorch

I still aint get it all the way

I have the 3.11.8 python model

[peak urchin]

w-okada doesn't need you to install Pytorch separately

it is inside the process of first starting start_http.bat

again misunderstanding

[deep phoenix]

it is???

they dont start when i drag them to voice changer

[peak urchin]

drag them, drag what?

Pytorch?

[deep phoenix]

the start bat

[peak urchin]

I mean just simply click the .bat

oh well

[deep phoenix]

its just when i do that it dosent pop up

i can use the broswer version tho

but it kinda sucks

[peak urchin]

browser version is https

it is for remote

I think your start_http.bat are in its original position?

because it is supposed to be

[deep phoenix]

I just got it to work

thank you

finally

4 hours

thank you so muchj

i dont got all the options though

[woeful sedge]

whats the list of rvc options

[woeful sedge]

ah here it is

[deep phoenix]

the ones I can use are "crepe_tiny" and rmvpe_onnx"

[woeful sedge]

ok

[deep phoenix]

how do i get the other ones?

[peak urchin]

you just use rmvpe for most case

[tacit grail]

so when making a model it is best to keep a single tone as best as you can right? Its not a good thing to diverse tones right?

[dawn nebula]

you need some diversity

[tacit grail]

so monotone voice?

[peak urchin]

depends what you wanted for your model

[tacit grail]

singing

[peak urchin]

I see some people individualize every single style of singing of the same actor.

[dawn nebula]

nope

you need to increase pitch range

[peak urchin]

which pro, you can have more space for testing and stuff and see if it suited you.

[dawn nebula]

so it has better performance

[tacit grail]

oh ok

[dawn nebula]

yeah would agree

[peak urchin]

cons is obviously you have to get more time to test and usually much harder to make dataset of

[dawn nebula]

fax

[peak urchin]

I think most people can just get away with one single model

with all funky singing or talking dataset

[tacit grail]

well im getting MrMOdZ to make a model of myself and i need to know how exactly i should be reading something for 10-15 minutes for best results

[peak urchin]

large tone-changing space for the actor? don't care just slap it and train it.

( some people's mindset )

[tacit grail]

Alright

ty

any recommendations on what i should read for best results?

@dawn nebula

[dawn nebula]

on dataset or training

[tacit grail]

for creating a dataset

sorry, data for training a model of myself

also, idk why but using crepe on Applio just messes up the output royally

[south vortex]

does anyone knows how to make clean vocal extraction to be infered in rvc?

i used UVR and it always have like vocal doubles and harmonies that messes the infer

[tacit grail]

I recommend using this instead. Yea theres a queue system but at least with this you can do stuff while it renders and all the stuff on UVR is on this. Its UVR but on a webpage https://mvsep.com/en/home

[south vortex]

ohh nice, any models that you recommend?

[tacit grail]

sorry that was the wrong link. Its updated now

[south vortex]

im checking it out now, and there's tons of models hahah

[tacit grail]

depends on the seperation type

[dawn nebula]

after the first step

using ensemble mode

[south vortex]

i have clean acapellas with harms and doubles in them, and i just want ro extract out a single main vocal without double or harms

[south vortex]

ohh nice, imma try this out, thanks alot

[dawn nebula]

5 hp karaoke uvr + mdx karaoke+ mdx karaoke 2

try

[south vortex]

👌

[dawn nebula]

all g

tell me if it works

[dawn nebula]

probs just pitch range

[tacit grail]

?

so i should try and sing sopmething not read something?

i was told talking normally gives best results

[south vortex]

the backing harmonies got removed, but the vocal double remains hahah

[rare goblet]

Ayo? @south vortex level 1 !!!

[south vortex]

i guess its a limitation i'll need to work around with

[tranquil fiber]

i'm using covergen no ui, but when i try to download the model from huggingface by using the "model download function" cell, it says the file is not a .zip file, even though it is. anybody know why's that?

[proud elbow]

have you tried each of MDX karaoke 1/2, 5/6HP, and UVR BVE via vocal splitter?

[south vortex]

i tried mdx karaoke 1/2 and 5/6 HP (all 4)
the lower harmonics is extracted, but the double vocal remains

[mossy rapids]

Yeah, me too. Does anyone know how to fix this?

[south vortex]

using MDX Instru HQ 3 extracted acapella as the source

[proud elbow]

there's pitch shift option in the MDX settings, have u tried changing it?

[south vortex]

pitch shift? what does that do

https://youtu.be/xpfRtFW6jGA?si=lxvMcLygFhPl0UqK
im trying to extract the chrous of this song if anyone is interested

[proud elbow]

what kind of vocal? m/f? and how high is the pitch?

[south vortex]

female. i guess quite high pitch? no idea whats the measuring standard

[proud elbow]

then try pitch shift -4 or maybe lower

[proud elbow]

^

[south vortex]

both vocals are of same pitch

she just sings twice and layer over it

[proud elbow]

same pitch but little different timing? not the opposite?

[south vortex]

yup

the lower pitch had been extracted

but the double same pitch hasnt

i'll upload what i have later hahah, if it helps

[proud elbow]

unfortunately no good models that can separate the kind of duet/unison yet. I heard there's MedleyVox but still flawed and its development seems abandoned.

[south vortex]

icic, thats unfortunate. thanks for the info

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[vocal sonnet]

HELP

[lost stone]

"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.
fastapi 0.97.0 requires pydantic!=1.8,!=1.8.1,<2.0.0,>=1.7.4, but you have pydantic 2.6.3 which is incompatible."

I'm getting this error message when using the Easy GUI colab, on the 1st step (Install)

[rare goblet]

Ayo? @lost stone level 1 !!!

[pliant sail]

is not overtrained right?

[indigo reef]

how can i get a download link directly from huggingface for a model zip?

or to be more specific how do i use the file directly if the author hasnt uploaded the model compressed in a zip

or does it no need to be zipped

[brittle wing]

What are the best settings for a mdx model?

[half cove]

check if loss/g/total goes up also turn smoothing up

[half cove]

wrong graph

loss/g/total

[half cove]

hasnt overtrained

[pliant sail]

oh okay thanks!!

[indigo reef]

you have to download them

[half cove]

send easy gui colab link your using

[half cove]

for model use mdx23c-instvoc hq

[brittle wing]

i have a nividia 3070 gpu

[half cove]

that works

[brittle wing]

are you sure

[half cove]

yes

[indigo reef]

How can i import a model from hugging face that is not zipped?

[half cove]

you could download it then put it into colab file manager (i assume your using colab)

[indigo reef]

yes

where is that?

i imported the link to the main.pth directly from huggingface and aparently it has ziped the model?

[rare goblet]

Ayo? @indigo reef level 2 !!!

[indigo reef]

(sorry if the model is inappropriate)

just kinda fooling around

[undone current]

that model

[rare goblet]

Ayo? @undone current level 1 !!!

[indigo reef]

where should the .pth file go?

[flint vapor]

the pth is inside weights folder

[indigo reef]

and then it will show up in the public ilaria page?

[flint vapor]

For use the model in ilaria rvc you can download from the interface itself, don' t need to upload the pth in weights

[indigo reef]

i know but the specific model isnt on a zip

[flint vapor]

You can upload the .pth in weights folder

and the .index you have to create a folder inside the folder logs

with the same name of .pth file

and you put the .index there

[indigo reef]

should i put the index too?

[flint vapor]

Yes

[indigo reef]

thanks

ill let you know if i have other issues

[flint vapor]

Ok 😄

[indigo reef]

illara keeps comming up with error when converting

[flint vapor]

what error you have?

[indigo reef]

Traceback (most recent call last):
File "/content/Ilaria-RVC/.venv/lib/python3.10/site-packages/gradio/routes.py", line 437, in run_predict
output = await app.get_blocks().process_api(
File "/content/Ilaria-RVC/.venv/lib/python3.10/site-packages/gradio/blocks.py", line 1349, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "/content/Ilaria-RVC/.venv/lib/python3.10/site-packages/gradio/blocks.py", line 1283, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "/content/Ilaria-RVC/.venv/lib/python3.10/site-packages/gradio/components.py", line 2586, in postprocess
file_path = self.audio_to_temp_file(
File "/content/Ilaria-RVC/.venv/lib/python3.10/site-packages/gradio/components.py", line 360, in audio_to_temp_file
temp_dir = Path(dir) / self.hash_bytes(data.tobytes())
AttributeError: 'NoneType' object has no attribute 'tobytes'

[flint vapor]

try to make the cover with mangio crepe and not rmvpe

see if works

[indigo reef]

how do i do that?

[flint vapor]

when you convert

you have options

[half cove]

it should work with rmvpe now

shouldnt have any problems

[indigo reef]

i dont

[half cove]

i think the audio wasnt selected properly

[indigo reef]

is it in advanced?

[flint vapor]

Which format is the sample you put?

[indigo reef]

wav

[half cove]

screenshot your ila rvc screen

[frosty lichen]

someone please help me

my ilaria is keep on loading

like whenever i convert

please someone how can i fix this

[indigo reef]

it did that to me before, just start over

[frosty lichen]

done that many many times

[indigo reef]

like do the colab again

[frosty lichen]

is it something wrong with my file

yes ive done that

[indigo reef]

purge it and start over

[frosty lichen]

over and over

this is the file im tryna use

is there a problem with it?

im seeing this on the start up thing

[half cove]

what does it say after loading rmvpe model

[frosty lichen]

it hasnt loaded

its still going and going

[half cove]

what runtime are you on?

[frosty lichen]

how do i check?

[half cove]

top right

[frosty lichen]

this obviously went up

[half cove]

doesnt answer the question

if your on cpu runtime it would take super long

[indigo reef]

they prob have a previous instance running

[indigo reef]

@frosty lichen

sellect dissocnect and delete

and start over

[half cove]

also make you change runtime type to t4 gpu

[half cove]

yes

[indigo reef]

@half cove so im trying it over again. I need to import the main.pth and main.index in the weights folder before loading the ilaria page?

[rare goblet]

Ayo? @indigo reef level 3 !!!

[half cove]

ran out of gpu runtime

[half cove]

index in logs pth in weights

or upload it to huggingface in a zip

[frosty lichen]

so what do i do now?

[half cove]

wait 12h or use an alt

[frosty lichen]

aight

[indigo reef]

index in mute or just in logs?

[half cove]

logs

or you can put em both in a folder in logs

iirc

[indigo reef]

ill do it in both for now to be safe

lets hope it works

[indigo reef]

it's still throwing up error

[half cove]

try another audio ig

[indigo reef]

is the audio the problem?

[indigo reef]

because if it helps it i can still compress them in a zip and upload them directly

[half cove]

that would just do the same thing

[indigo reef]

i edited the sound file in vegas pro. Could the render have messed it up?

i have a version not edited through it. ill try that

[half cove]

could try exporting it from audacity ig

[indigo reef]

i uploaded the zip on huggingface but now it turns up this when downloading it

username/authentication failed

[half cove]

is the model visibility private?

[indigo reef]

yes

i mean it's not my model

ok making it public worked

ill make it private again later

[civic laurel]

hi

my mangio rvc is not doing anything

can someone help

ok nvm its opened

[void pond]

does RVC have real time voice to ai stuff?

[half cove]

rvc can do realtime

[misty forge]

I need help. I keep having this white screen when I'm on tensorboard within Applio...

I asked before, but I got no response...

I... tried that too... but hold on.

Nothing.

What do I do now?

[half cove]

what does cmd say

[misty forge]

No errors, nothing.

[half cove]

try another browser maybe?

it seems correct

havent really seen this problem

[misty forge]

Edge and Chrome both don't work.

[sand iris]

I think we've seen it before

lemme just

[rain kiln]

Hi

[half cove]

gdr to the rescue

[rare goblet]

Ayo? @rain kiln level 1 !!!

[sand iris]

How the fuck does applio handle tensorboard ? I don't know

[rain kiln]

Why isn't this updated?

[sand iris]

But last time it was a simple case of isolating tensorboard because else it wouldn't behave

[sand iris]

two things :

[misty forge]

I think they're separate entities, where Tensorboard is just told where to find the logs folder.

[sand iris]

1. don't run as admin pls

2. RVC gui is dated by now

oh and third you can't use python 3.12

only 3.10, 3.9 or 3.8 will work

[sand iris]

Tempted to install tensorboard externally to be honest

[rain kiln]

Oh, okay. I installed 3.12 😅

Thankss ❤️

[sand iris]

last time this happened it was because tensorboard returned the browser javascript (responsible for all the UI logic) as text and not script

a reinstall simply worked, but I'm not sure how applio does it

it's one of those random wacky issues that shouldn't happen but here we are...

[sinful nexus]

I need a colab that works and is easy

[rare goblet]

Ayo? @sinful nexus level 1 !!!

[rain kiln]

Which of the 3.10.x versions do you recommend I download?

[sand iris]

Latest, from microsoft store if you can (they have 3.10 listed)

the python org doesn't provide the latest versions for the older branches

microsoft does though

[sinful nexus]

I need a colab that works and is easy

[sand iris]

model training ? or doing voice conversion ?

[sinful nexus]

Voice conversion

[rain kiln]

Thanks ❤️

[misty forge]

What python version works best for tensorboard?

[sand iris]

Any that's currently supported

[sand iris]

-colab

[azure marsh]

Suggestions for @sinful nexus

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[sand iris]

"Ilaria RVC"

[sinful nexus]

Thanks

[sinful nexus]

When I download everything it says it has to restart so when I let it restart it says it failed

[sand iris]

oh don't mind that

you don't have to restart it

just continue. It's kind of stubborn

[sinful nexus]

Oh ok

Would mega links work?

For model

?

[sand iris]

I think ? not sure

[sand iris]

blaise here

[clear lily]

any fix?

nvm

wait. yeah i still need help

[brittle wing]

Hello, let me ask, when you train a voice model, how much audio do you use to make it good? I know the more the merrier, but I'm going to ask because I'm a slow person

[tame mica]

basic models would probably be fine with around 5-10 mins. more than 1 hour is just too much

[edgy nova]

guys got a question
i got rtx 3060 ti but when i talk its a bit laggy why is that?
it was good at first but not now

[tame mica]

this would probably be best asked in #🔍│help-w-okada :3

[edgy nova]

alr thank you

[brittle wing]

Thank you, I was trying for an hour, I feel calmer.

[latent kettle]

how much Bach size is good for 5-10 minutes of dataset

[molten pecan]

8 is your safest pick

[latent kettle]

but i think training will be very fast.. nearly 10 seconds for pre epoch

[sinful nexus]

Why is it saying when I try to voice convert my audio it says sorry the content file doesn’t exist but it does

Im kinda confused of why its doing that

[sacred lichen]

can someone link me the rvc realtime turorial?

[molten pecan]

-realtime

[azure marsh]

Tag suggestion for @sacred lichen

Guide for W-Okada's RealTimeVoiceChangerClient

Description: This one has more options and is somewhat more convenient to use than Real-time RVC.
Link: Rentry
Credits: ravencutie21

How to use real time conversion with RVC (voice changer)

Description: More accurate than W-Okada's, but requires a decent GPU.
Link: Google Docs
Credits: teas.hi

RVC TTS

Description: Text to speech with RVC
Link: Rentry

[latent kettle]

https://rentry.co/VoiceChangerGuide

[sinful nexus]

Why is it saying when I try to voice convert my audio it says sorry the content file doesn’t exist but it does

[molten pecan]

in which RVC exactly?

[latent kettle]

but i think training will be very fast.. nearly 10 seconds for pre epoch

[molten pecan]

yeah

I used batch size of 8 in a 3 min dataset before

and it went well

[sinful nexus]

https://colab.research.google.com/drive/1u1brjK8IZt647UsbZuGYfW29oFM2I4tk?usp=sharing

This one

[molten pecan]

which step?

[latent kettle]

but I have a confusion. why people prepare 20-40 minutes of dataset and use batch size of 8? if batch size of 8 is good for 10 minutes of dataset and it train at very fast rate.. so why people prefer large dataset with 8 batch size and slow training?

[molten pecan]

is not mandatory using 8

[quick jungle]

https://cdn.discordapp.com/attachments/1160576561636966491/1165372018414342284/warning.gif

[molten pecan]

more advanced users use 10, 7, 5, 20, etc

but 8 it's advisable to use 8 as it's less prone to bugs

[latent kettle]

I want to know what this thing is actually and how it works

[molten pecan]

what batch size determines is the amount of information that'll be processed at once

[latent kettle]

like? any relevant examples

[molten pecan]

try asking @fluid matrix

[latent kettle]

@fluid matrix can you help me

[fluid matrix]

what are you using to train?

[latent kettle]

just want to know what the batch size is actually.

[sinful nexus]

Downloading the cover

[fluid matrix]

oh you mean what that is?

[latent kettle]

yes..

how it works and how to pick a correct batch size

[fluid matrix]

determines the steps per epoch ig

thats why i was asking what you are using to train

local or colab?

[latent kettle]

colab

[fluid matrix]

then just leave it at 8

no need to worry about it

only if your dataset is really short, you can try 4

[latent kettle]

I want to apply for #1159514067187277865 and I want that qc will not reject my model. That's why I want to get proper knowledge of everything

[fluid matrix]

yeah makes sense

the colabs all use the t4 gpu so i would just leave it at 8

a good way to determine the batch size is to look at your gpus vram

[pallid blaze]

need help: I try to use the hugging face link to a model like the Ilaria rvc says, but it says its invalid?

[fluid matrix]

but 8 is really safe

[molten pecan]

Hm well you don't really need to be knowledgeable of that much.
Check this out https://aihubdocs.github.io/en/extra/model-maker-role/

[latent kettle]

really short? in seconds or minutes. usually I train on 30-50 minutes of data

[molten pecan]

Just make sure you follow the rules and you'll be in =3

[molten pecan]

https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/ check the troubleshooting chapter

[fluid matrix]

short as in anything under 3 mins imo, then you could try 4

30-50min is a lot, make sure your dataset is clean tho, otherwise its just a waste of time

30min is great, 50 might be too much

never go above 2h

[mint musk]

https://tenor.com/view/shocked-shocked-cat-silly-cat-cat-kitten-gif-7414586676150300212

[naive violet]

.....how do i hear myself?

[fluid matrix]

same thing, why would it be different

[molten pecan]

The 30 hour long kanye dataset for SVC >>>

Or was it 16?

[fluid matrix]

[molten pecan]

Ask in #🔍│help-w-okada for voice changer help

[latent kettle]

and for model masters ?

[fluid matrix]

@latent kettle

[naive violet]

okay!

[molten pecan]

From what I heard is making plenty of HQ models & taking commissions

You'll have to be committed to do that

[fluid matrix]

thats not answering his question

but true

[latent kettle]

usually I extract vocals and de noise it by using FL Studio and then de echo reverb it

[fluid matrix]

dont denoise in fl

and how do you extract the vocals?

[latent kettle]

I use Vocalremover.org because I don't have a gpu

[fluid matrix]

the docs page julia sent is a good starting point for everything, read that first

yeah dont do that

use mvsep instead

[latent kettle]

Instead of that ?

[fluid matrix]

dont use denoise in fl at all, no vocalremover.org either

[latent kettle]

any alternatives?

[fluid matrix]

their algorithms are shit and arent enough for datasets anyways

just told you

please refer to the doc and read on mvsep

https://aihubdocs.github.io/en/vocal-isolation--datasets/datasets/

sry this link

https://aihubdocs.github.io/en/vocal-isolation--datasets/vocal-isolation/

you can ignore uvr, since you dont have a gpu

[latent kettle]

can I use uvr on colab?

[tranquil fiber]

When i try to connect to colab's runtime it displays a message saying "impossible to connect to gpu backend" (smth like that), anybody knows why and how to fix it?

[rare goblet]

Ayo? @tranquil fiber level 1 !!!

[latent kettle]

I think your GPU limit is over ...

[tranquil fiber]

meaning?

[naive violet]

how do i ping an helper?

[fluid matrix]

please just read the thing and use mvsep instead

it uses many of the same modes that are in uvr

there is a uvr option on colab

if you want to try that out instead

[tranquil fiber]

will the limit reset after some time?

[latent kettle]

I'm not native to English.. sometimes it is too hard for me to understand...

[latent kettle]

after 24-72 hrs

[fluid matrix]

you can try translating the site

[tranquil fiber]

damn, alright thx

[fluid matrix]

there is no translations for the new doc yet,sry

[pallid blaze]

still invalid

[latent kettle]

okay... thanks for your support 💐

[naive violet]

i need help in the other channel...

[rare goblet]

Ayo? @naive violet level 1 !!!

[odd shale]

Don't use general chat

[naive violet]

okay

[latent kettle]

using @ and the name of that helper you want to ping

[naive violet]

okay

[latent kettle]

@naive violet

like this

[naive violet]

i see

thanks

[latent kettle]

yw : )

[sinful nexus]

Why is it saying I need a NVDIA DRIVER

[rare goblet]

Ayo? @sinful nexus level 2 !!!

[odd shale]

Please, don't do that.

[sinful nexus]

Even tho it worked before

On my phine

Phone

[tired vessel]

Not sure if this is a thing or not, but is it possible to use RVC offline ? Incase a persons internet is down ?

[rare goblet]

Ayo? @tired vessel level 2 !!!

[tired vessel]

If so, how ?

[latent kettle]

@fluid matrix I have a doubt. is it necessary to convert audios into .wav from mspv or I can use FL Studio?

[rain kiln]

"Downlaod hubert_base.pt and place it in the root folder"

I couldn't understand the "root" folder mentioned here. Can anyone help?

[fluid matrix]

what audio? use fl studio for what?

generally you should choose hq quality wav rips, or lossless (flac for example if possible) for the source and then continue to work with flac files to preserve the quality

[latent kettle]

for converting mp3 into wav

[fluid matrix]

why would you want to do that

and avoid mp3

either use wav or flac

[latent kettle]

I want to use uvr

[fluid matrix]

on colab?

[latent kettle]

yes

[fluid matrix]

well then do that

you dont need to convert anything

and please avoid mp3

either use flac or wav

[latent kettle]

for dataset yes I do. but I downloaded songs from internet and they are in. mp3

[fluid matrix]

then dont download in mp3

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[fluid matrix]

and make sure they are high bitrate

[latent kettle]

but there is not any other option available

[fluid matrix]

youtube download for example is compressed by default but even in that case, make sure its hq

wdym

there is plenty

what are you using?

[rain kiln]

Where do we throw the "hubert_base.pt" file mentioned here? I don't see a folder called root

[sinful nexus]

Why is it saying I need a NVDIA DRIVER because it worked before on my phone

[sand iris]

root folder refers to the rvc gui folder

[latent kettle]

I just download songs from third party websites

[fluid matrix]

yes, what site

[sand iris]

the root being the top level folder, but in this care it's the rvc-gui folder

[fluid matrix]

you can try using yt-dlp if you really need to download using yt

but as i said yt audio is compressed by default

[latent kettle]

does it download with .wav

[rain kiln]

Just kidding, I'm looking for the root folder in the python folder or something, thanks

[fluid matrix]

yes ofc

if you really have to use that site, use this

[latent kettle]

the maximum bit rate they offer is 320kbps

[fluid matrix]

yes

[fluid matrix]

what are you trying to do

nvidia drivers on your phone? 😭

i guess it is

if thats the highest quality you can find

but try to avoid mp3

and make sure you use the highest bitrate available

[sinful nexus]

It worked before

Now it just doesn’t wanna work

[fluid matrix]

what worked. please be more specific

[latent kettle]

what should I do.. convert mp3 to .wav 32 bit float bit rate ?

[fluid matrix]

i dont know what you are talking about

[fluid matrix]

you can do that if you want

but it wont change anything

it wont magically increase the bitrate, just change format

the source itself needs to be hq

[tired vessel]

dam i got ignored entirly lol xD

[latent kettle]

so from where I can download .wav of songs ..

[fluid matrix]

idk what you want to download

spotify, rip it from elsewhere

youtube

idk

i told you some options

[fluid matrix]

what was your issue?

oh i see it now

[fluid matrix]

if you mean the colabs then no, you would have to use a local install of rvc on your pc but that requires a good pc

[tired vessel]

i currently have a i9 14th gen and a 4090

[fluid matrix]

i mean thats as good as it gets lol

[tired vessel]

not the collbs i dont think, just using a model

[fluid matrix]

i mean like i said, if you want a 24/7 offline option you have to use a local install

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

try this

[tired vessel]

idk about collabs lol i just use RVC and a model lol sry im new

okay, thanks !

[low shard]

@sinful nexus so what tool are you specifically using

[sinful nexus]

like what colab?

[low shard]

yea

[fluid matrix]

i asked you before, what are you trying to do? what were you using? you said it worked before, what did? you have to be specific

[latent kettle]

what if I download songs from YouTube using yt-dlp. is it good ?

[fluid matrix]

you dont need any nvidia drivers on your phone, i think you are confusing something

[low shard]

yea lol

!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.

[fluid matrix]

if you need audio from youtube, then yes

[sinful nexus]

https://colab.research.google.com/drive/1u1brjK8IZt647UsbZuGYfW29oFM2I4tk?usp=sharing

[fluid matrix]

if yt-dlp is confusing for you, you can try this too https://stacher.io/

same thing

[latent kettle]

Simply what I understand is that collect high quality data from any source (if possible) ryt

[sinful nexus]

And I was at this part @low shard

And it worked before but now it doesn’t

@low shard

Are you there?

[fluid matrix]

you need to run all the cells again

[low shard]

yea retry

[sinful nexus]

Okay

[ionic canopy]

How exactly do I stop rvc from adding .wav

To a just converted .wav twice

because it has the name of the file, .wav.wav

[hard frigate]

One message removed from a suspended account.

[cunning fox]

It says that i need a nvidia driver even though i downloaded it like 7 times

its confusing because sometimes it works and sometimes it says i need the nvidia driver

[fleet bobcat]

aanyone got the RVC link? i lost it and cant find it anymore

[molten pecan]

which version are you using?

[molten pecan]

https://aihubdocs.github.io/en/rvc/cloud/rvc-disconnected/ here's a guide for it =3

[molten pecan]

are you using Okada?

[molten pecan]

which RVC exactly? there are plenty of forks

[fleet bobcat]

found it in here

[ionic canopy]

Nvm. I fixed it.

[fleet bobcat]

its all good

[ionic canopy]

Had perplexity give me a puthon code and that dealt with it.

[cunning fox]

whats okada?

[molten pecan]

realtime voice conversion software. built on RVC

which RVC are you using?

mangio? applio? base version?

[cunning fox]

oh uhh rmvpe

is that even a thing i don’t know what im talking about tbh

[rare goblet]

Ayo? @cunning fox level 1 !!!

[edgy nova]

anyone free to help me with a prob?

[rare goblet]

Ayo? @pearl tulip level 2 !!!

[molten pecan]

RMVPE is a pitch extraction algorithm, not the app itself

where did you download the RVC from?

[pearl tulip]

Anyone know what this means for trying to run applio? I can't get the program to work no matter what I try

[brittle wing]

any sites for model searching

[molten pecan]

https://weights.gg - https://applio.org/models

[pearl tulip]

Anyone know if the is a tutorial because I can not get the program running at all

[cunning fox]

Ohhh i didn’t download it used colab

[strange reef]

when i start to install the llaria rvc it tells me i dont have gpu backend. everytime i try to connect without gpu it doesnt work

[rugged stirrup]

What's the difference between the new RVC WebUI VS the 23.7.0 one

[sinful nexus]

-colabV3

-colab

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[sinful nexus]

-colab

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[sinful nexus]

-colab

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[storm holly]

how much better is training on the new colab update?

[sinful nexus]

-train

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[small dragon]

Hi, I found some really well trained voices but there's no index files, what to do?😅

[fluid matrix]

can you link them? they shouldnt be uploaded if they dont have a index file

[small dragon]

https://huggingface.co/QuickWick/Music-AI-Voices/tree/main/NewJeans Hanni 22k & https://huggingface.co/QuickWick/Music-AI-Voices/tree/main/Huh Yunjin 5K

[fluid matrix]

those are really really old

i wouldnt use those

oh god

[fluid matrix]

please dont use those, where did you find these? 😭

those are svc models

not rvc

[small dragon]

😂 sorry, I just found them, I'm new here 😂

[fluid matrix]

yeah no worries

but how? 💀

either check the #1175430844685484042 channel or our backup, https://www.weights.gg

[small dragon]

To be fair some others work well from the same sight, just that those didn't have index files

[small dragon]

Some other sites linked that site to having models and so 🤷🏽‍♂️😄

[rare goblet]

Ayo? @small dragon level 1 !!!

[fluid matrix]

like i said, those models were trained using svc, really old tech. and huge file sizes. we are using rvc v2 already

what sites?

curious

pretty much every model there is comes from here

[lone vine]

What can i use to make ai covers using my voice model the colab i used to use doesnt work anymore

[fluid matrix]

wait what

they have that?????

[small dragon]

Weights.gg is great but not alot of Kpop idols

[fluid matrix]

you could try joining the kpop ai hub server, but if you cant find it on weights, then it probably doesnt exist yet

#1159289738314919936

who did you try to look up?

amazing

never knew that had that

im gonna delete the links, because of piracy and all that, but thanks for letting me know

they linked to old tech lol

[small dragon]

Rosé from blackpink

[lone vine]

anyone?

[fluid matrix]

search for just blackpink, there is so many

[small dragon]

Got it, thank you😄

[fluid matrix]

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

[small dragon]

I'll definitely do that when I find some time, I just thought a song would be nice in her voice but ended up going down a rabbit hole 😂

[fluid matrix]

if you just want to do covers, ilaria rvc can be a good choice

[fluid matrix]

https://www.weights.gg/models/clrx3rm48000e324b15q11kvm

[lone vine]

thx

[fluid matrix]

havent tested that one, but there is some just for rose

if you search blackpink

[small dragon]

Thank you, I'll definitely try that one next 🙏

[fluid matrix]

there are some other ones, i just picked out that one

like i said, havent tried it myself

[small dragon]

Yeah, I wasn't sure how to decipher which ones are good, like is the amount of downloads a sign or the higher amount of epochs🤔 I'm so new 😅

[sinful nexus]

How do you add your models to appilo

[fluid matrix]

best thing to do is to just try it out lol, but newer models with a large index will most likely be the better ones, epoch count doesnt matter in the grand scheme of things

more epochs doesnt mean better model

[small dragon]

Got it, thank you

[fluid matrix]

read this https://aihubdocs.github.io/en/rvc/local/applio/

either just download and drag & drop the file

[void pond]

Hey, is there a way to use RVC for realtime voice changing? As in, using a model so you can say something and it does its magic and plays as a microphone

[rare goblet]

Ayo? @void pond level 3 !!!

[small dragon]

Yeah, when I find an hour or 2, I'm just gonna download all voice models i see 😄

[fluid matrix]

or paste the link and let it do its thing

[sinful nexus]

Thanks

[fluid matrix]

yes, thats okada 👍 if you want to use it locally, you will need a decent pc but this will help you out https://rentry.co/VoiceChangerGuide

[void pond]

I'm guessing this has nothing to do with it then?

[fluid matrix]

no, thats not that great

[void pond]

I see

[fluid matrix]

it would still work, but why use that when you can use the other one

[void pond]

do I need anything besides the .pth file? I'm a bit stupid when it comes to RVC, I'm just a mere user

[fluid matrix]

for okada the pth is enough

[void pond]

nevermind it says on the website you sent, thanks!

[lone vine]

i was using google drive to upload my index and php any other websites where i can upload them?

[azure marsh]

RVC Guides (How to Make AI Cover)

[lime otter]

-colab

[rugged stirrup]

How do I use a model if I forgot to check the save a small weights file?

[lime otter]

Huh?

[rugged stirrup]

I forgot to check this setting

And the model did not save to the weights folder so I dont know how to use it it saved as a G and D model

[storm holly]

how come tensorboard only shows you results after 200 epochs

it actually reached it's lowest loss_gen all by 15 epochs

still sounds more like jackson than anything

[shut sorrel]

Does anyone know if this was done using RVC? If so, how? (Radiohead sings Britney Spears)
https://www.instagram.com/reel/C2ybxJru43b/

[storm holly]

for comparison, this is jack singing the other angel dust song

ah fuck

i think i used the same database

wait no

it's the rusty database

what the fuck

what the hell

the samples are all the same

yet it always sounds like jack

shit i was supposed to put these in a separate folder

no way

i submitted my friend's character set

it actually gave me an index

insert ctrl alt del did you know about video game meme

and get this

it actually gave me a higher step count

[fluid matrix]

made a pull request, that site is updated now 👍 no longer links to outdated stuff

[small dragon]

You're awesome 👍

[acoustic torrent]

yo i have a pth file

i trained a voice model

but idk how to host it on huggingface to get it on gradio

[storm holly]

and i finally got an index file

things are looking up for the both of us

[plush compass]

How do I upload models to the voice-models channel?

[acoustic torrent]

hell yeah

[fluid matrix]

https://aihubdocs.github.io/en/rvc/resources/uploading-models-to-hugging-face/

[fluid matrix]

if you made a model, you have to submit it in #1159514067187277865, get it accepted and get the model maker role once so you can post models. https://aihubdocs.github.io/en/extra/model-maker-role/

[plush compass]

thx

[storm holly]

should i download the .pth already or wait until it hits 300?

[fluid matrix]

your tensorboard is not really useable that way

set smoothing to 0.987

and center the view

[storm holly]

i did

[rare goblet]

Ayo? @storm holly level 6 !!!

[fluid matrix]

well thats how a graph is supposed to look (kinda)

yours is just flat

you probably didnt train it a lot yet

[storm holly]

now should i wait for it to hit 300?

so that's a no

[fluid matrix]

what is it at now?

[storm holly]

260-280 epochs

[fluid matrix]

yeah then just let it finish and use an older epoch state if needed

[thin stump]

look at kl, mel, d and see if one them is going up

[storm holly]

nope

all down

[solid holly]

Yo can someone help me plz. I have try so many yt vidz but none works i copy them 100% even the folder names still not working.

https://docs.google.com/document/d/1Uvf9txpkV3O6z25b6hoEab2yxDRKcYvcWR1cVG1HJDc/edit?usp=sharing (i don't know how send pic sorry (google doc))

[fluid matrix]

youtube vids are outdated and full of misinformation

follow this guide instead https://rentry.co/VoiceChangerGuide

make sure your pc is decent tho

and has a gpu

otherwise you will need to use the colab (guide at the top of the rentry)

[signal crater]

hi, which files should i save in order to retrain?

[fluid matrix]

retrain from scratch?

i mean you will just need your improved dataset and start the training process again ig

[rugged stirrup]

Anyone know what to do if you forgot to check the save small model button and it saves the G and D model instead for some reason

[signal crater]

i'm starting from scratch but i dont know which files to keep besides the model and the g and d files if i need to retrain later in case it takes a while

[rare goblet]

Ayo? @signal crater level 1 !!!

[signal crater]

cause last time i made one in one sitting with a smaller dataset

[solid holly]

@fluid matrix FIRST THANK U AND IS ADDING NEW VOICE THE SAME OR DIFFERENT?

[rare goblet]

Ayo? @solid holly level 1 !!!

[storm holly]

immediately shut off after 350 or so epochs

nearly 400

if it doesn't continue i'll just wing it at 300e

[brittle wing]

hey, im trying to use my voice changer, but my friends say they can only hear my normal voice but a bit more delayed than it used to be.
anyone sure how to fix this?

I trained it at 200 and then wxpoerted but now I can't resume

[molten pecan]

are you using RVC Disconnected?

[brittle wing]

How was resuming done I don't remember

Did the rec disconnected get any changes recently cause I last trained on December last year

[brittle wing]

Yes 😦

[molten pecan]

uhh not any big change iirc

[brittle wing]

I'm so angry

[molten pecan]

https://aihubdocs.github.io/en/rvc/cloud/rvc-disconnected/

this guide is oriented to resuming, from a different session or not

[molten pecan]

that's okay. you can start from scratch and prevent this mistake

[brittle wing]

Thx cause I'm losing my patience

[molten pecan]

yw. it only takes a few hours of waiting. good luck

[brittle wing]

I have already started from scratch

And it's at 200 epochs

I wanna increase them

To 400 then 500 it's a request of another person

Huh

[storm holly]

aaand rusty can't get his .index

whatever

[low cliff]

easy GUI is not working and disconnected isn't working ether

[brittle wing]

It worked when I first trained it yesterday at 200 epochs wdym

It worked like normal

Huh

[brittle wing]

Waste of life

[crimson stream]

my audio is very choppy and it keeps cutting how do i stop this

[brittle wing]

???????????

Not matched no such file or directory WTF

[crimson stream]

how do i stop the delay

[brittle wing]

But it's in the folder wtf is wrong now

[low cliff]

I run the same problem too.

[brittle wing]

This is angering

It's getting on my nerves

[rapid badge]

Applio 3.0.8 rvc local , works great in both Windows and Mac … 🔥

[brittle wing]

There's a Collab of applio where you CAN train models tho...

Maybe I fixed it but I ran every each cell as usual+import model from drive to notebook and it resumed!!!

The guide is misleading

[rapid badge]

I recommend applio local in future as i used it in both platforms without any issues..

[brittle wing]

Preprocessing and feature extraction should be ran too otherwise you get the error I got

Huh

[brittle wing]

I don't have a computer

Sadly

[rapid badge]

Might be the target destination , i mean path of the file format is misplaced..

[rare goblet]

Ayo? @rapid badge level 1 !!!

[brittle wing]

I fixed everything

Resumed

[rapid badge]

Cool

[humble shadow]

wait could you send me the link to it pls? 🙏

[brittle wing]

Hmm...if you want I can send you a whole Google doc

[rare goblet]

Ayo? @brittle wing level 4 !!!

[humble shadow]

sure

[brittle wing]

-colab

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[brittle wing]

There you go I summoned them all

[humble shadow]

thank you!

[brittle wing]

No problem but the doc I can send you in private message

[storm holly]

nope

still sounds like jack

[brittle wing]

Sent

-audio

[azure marsh]

📚 Audio Guides

• Perfecting Audio Isolation on Low-End Rigs: A Practical Guide, by Litsa The Dancer and Faze Masta Google Docs
• Gathering and Isolating Audio, by SCRFilms Google Docs
• Creating Datasets for RVC using iZotope RX, by LUSBERT Rentry
• Vocal Mixing Tutorial, by Roomie YouTube

🛠️ Tools

Audio Separation/Isolation

• https://mvsep.com/
• https://multimedia.easeus.com/vocal-remover/
• MDX-NET Colab

More commands: -colab, -uvr, -karafan, -overtrain, -help

[brittle wing]

Multimedia?uhuh

-uvr

[azure marsh]

Ultimate Vocal Remover

One of the best free and open source vocal and instrumental isolation tool.

Tip: Use /guides UVR to learn how to use it.

[molten pecan]

I made the guide. how is it misleading exactly?

[brittle wing]

I explained it already.

No offense buti followed it and it typed out error until I ran preprocessing & feature extraction cell too but the guide says to not run it

[molten pecan]

like mentioned before, the guide is oriented around retraining. so I explained it assuming preprocessing & feature extraction has already run

[storm holly]

this is how the OG voice sounds, for the record.

how do you get that from this?

[brittle wing]

But it typed out error when I I didn't run that cell

[molten pecan]

seems like the preprocessed files weren't on your google drive

[brittle wing]

They were

Everything's here

It didn't work til I ran this cell

The one the guide tells me not to when resuming

[molten pecan]

that's odd. I don't fully understand the context but I am 1000% sure it's not misleading. The RVC-D page is the top 5 most viewed page. If it included a big mistake like that I would have known at this point

either by users reporting the issue or by the multiple staff that has verified the guide

[brittle wing]

Idk but what I did worked

[brittle wing]

Idk

Seriously I just followed my own instincts

[storm holly]

probably because it's 15 epochs

and the dataset amounts to 50 seconds

[brittle wing]

I wanna get model maker & repost my models from this new account

Oh wait

I can reset the old one

I wanna get model maker role

[storm holly]

i want to give someone else the databases so they can make the models for me

[thin stump]

https://aihubdocs.github.io/en/extra/model-maker-role/

[brittle wing]

I'm the deleted user who posted jhope of BTS models

[thin stump]

you have to reapply. That's what everyone in the old server has been told

[brittle wing]

I know

[rare goblet]

Ayo? @brittle wing level 1 !!!

[brittle wing]

Can I claim the posts as my own or there's no such option

[thin stump]

The /claim command doesn't work

[brittle wing]

Oh

[brittle wing]

There's no such command

So I hv to reapply

I'll do

[dawn nebula]

you just reapply model maker

[brittle wing]

I did and I'm waiting

Also can someone delete my old posts I redid the AI covers and everything

Like the previous ones were too high pitched and sounded effeminate I redid them and re-uploaded them on my channel the samples are outdated I changed my handle on YouTube

-colab

[azure marsh]

☁️ Google Colabs

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

🤗 Hugginface Spaces

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

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

[tulip crypt]

i need help setting up

if anyone can help

[junior orbit]

@dawn nebula sorry to bother you but the voice program just stopped picking up my mic randomly

and wont work no matter what i do :\

[dawn nebula]

it might be your mic problem

[dawn nebula]

!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.

[junior orbit]

discord is picking it up just fine and its showing activity on the sound manager thing

[rare goblet]

Ayo? @junior orbit level 1 !!!

[dawn nebula]

so whats wrong then

[junior orbit]

its just not picking up my mic at all

[tulip crypt]

i need help setting everything up from what im seeing i have the right version i believe but i have no clue

[rare goblet]

Ayo? @tulip crypt level 1 !!!

[junior orbit]

[lilac gust]

is the google Collab glitched

[dawn nebula]

have you read the guide

[dawn nebula]

bro you need to press start

💀

[junior orbit]

yeah i know

even when i do that it doesnt work

[dawn nebula]

also too big chunk

set it to 112

[junior orbit]

ok i got it to work somehow

idk

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[brittle wing]

does this help get started?

[rugged stirrup]

Anyone know what this means and how to fix it?

[wary spoke]

guys, i know fuck all about coding. what does this mean? it keeps happening when i try import a model.

it won't let me send an image for some reason

[plush compass]

What do you think about ov2? Do you think it would be effective enough on datasets of less than 10 minutes or less?

[glad zealot]

Wdym

[edgy otter]

can someone help

i cant launch the start http bat

[glad zealot]

Yes

[sand iris]

you should be able to now

[sand iris]

give this a try