✨│ai-help

[brittle wing]

i was watching this youtube video and he said make extra max

[pastel oak]

I hate these youtube videos that dont know what theyre doing

holy shit 💀

[brittle wing]

i had it on 131k

😭

[pastel oak]

wokada is too badly optimized for anything over 32k to work

and youd need a higher end gpu and cpu to run 32k+

and for gtx itll lag out, bug out, kill your cpu

[brittle wing]

ok it sounds better but still choppy

[pastel oak]

do you use CABLE or Line 1

[brittle wing]

im with my friends testing it and they hear me but cant understand what im saying

Nevermind figured it out myself and managed to get the command working. Thanks for the advice

[brittle wing]

where do i see that?

[pastel oak]

your wokada Output

does it say cable or line 1

[brittle wing]

ohh

cable

[pastel oak]

cable is also pretty bad

first, do this:

open task manager hit details

right click audiodg.exe and set Priority to HIGH

right click it again and press set affinity and only have an EVEN number CPU core set, 2 is what I use. AKA everything unchecked except CPU 2

see if it fixes it

[brittle wing]

ok i did that

[brittle wing]

anything else?

[pastel oak]

i recommend the other virtual cable

https://software.muzychenko.net/freeware/vac470lite.zip

"setup64"

[proper shale]

o nice, good to know ya got it! glad i could help (even if it was just a lil bit)

[brittle wing]

ok i did that

[brittle wing]

as soon as I sent that message i get this error brah do i have to resintall python and get version 3.11 instead of 3.12.3

[brittle wing]

should be like this right

[pastel oak]

ye

[brittle wing]

it still doesnt sound good, i think its the voice changer im using so ima use a good one from here and try it

[rare goblet]

Ayo? @brittle wing level 2 !!!

[pastel oak]

you can try RVC realtime

you want links?

[violet sky]

yo im trying to use the ozzy osbourne voice model i got from the search here and its saying its not valid when i upload it

[brittle wing]

what is that?

[blissful scaffold]

not all voice models work

so try another and if it works it means that one doesnt

[violet sky]

ok thanks

[blissful scaffold]

its sad wish it wanst like that 😭

[violet sky]

its all good ill still rock on 🔥

[blissful scaffold]

good mentally 😊 🙏🏻

[rare goblet]

Ayo? @blissful scaffold level 4 !!!

[pastel oak]

..the other voice changer

[brittle wing]

sure

im trying the caseoh audio right now

[pastel oak]

https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/RVC1006Nvidia.7z

Download this

Then I recommend you download the (old) dev version as theres a quick trick to reduce some more delay. To do this, heres a quick step by step:
https://huggingface.co/Shadicti/rvc-old-dev/tree/main

Open the zip, open the first folder called Retrieval-based-Voice-Conversion-WebUI-dev and drag all the files into your RVC folder. If it asks you to replace any files, say yes. If it does not ask you to replace any files, you probably dragged the folder Retrieval-based-Voice-Conversion-WebUI-dev in it, but you should drag the 8 folders and 26 data files directly to the RVC1006 folder.

Then you run go-realtime-gui.bat to start the voice changer

On there, make sure to change the default index thats there, just upload any random custom index to prevent crashes

Select input and output of the same driver type (MME) at the end.

Sample length is delay (not the best gpu just try 0.40 and see if you can go up or down)
ignore harvest algorithm, do max fade length and extra you could do full probably aswell
Choose rmvpe

[brittle wing]

so can i just delete the one im using right now for this one?

[pastel oak]

ye

[brittle wing]

ok im doing it rn

this one was a sturggle anyway

[graceful obsidian]

not only that, amd version ( before they tinkered with the release ) was broken anyway

inferencing on amd release is just, well, it's not 100% that way, but inferences start dropping pitch, kinda f0 halucinating after a bit

I'd say, past 15-25 secs ( longer the inferenced audio is, worse it gets )

totally not worth it + memory leak issues

much better to just infer on cpu
( dml training isn't a thing in rvc anyway )

[errant mantle]

I'm getting this issue, what should I do? The rvc just freezes when I try to import the model, it freezes when I click ''Start Audio Conversion''
''sounddevice.PortAudioError: Error opening Stream: Illegal combination of I/O devices [PaErrorCode -9993]''

[graceful obsidian]

also, the reclist won't work for the dude as it ( cmd ) has to be run at the place the txt is present

[pastel oak]

wrong input and output driver type

make sure both are (MME) at the end

[errant mantle]

Noticed that, thank you!

[pastel oak]

off topic feel free to check out my guide on reducing more delay

https://rentry.co/Realtime-less-delay

[brittle wing]

can we go in a call so i can share my screen im kind of confused

[noble forge]

can anyone help me? after start pressing "Start_http" it does open CMD but after that the program close immediately can anyone explain what is the problem?

[brazen wadi]

which is better?

[void salmon]

Hello here ! Sorry to come like this but I have an issue with Ilaria : it's stuck on "fetching packages 0/2" ..

EDIT / MY BAD IT WORKED.

[thin stump]

Check the guide here https://docs.google.com/document/d/1j9J8A8Oop9bMOHmCs3jDXzPujuD6TQ0Q396rJ0MyuIc/edit

[brittle wing]

What does "RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory" mean??

[viscid moss]

#🔍│help-w-okada message
Try this

[lusty knoll]

help im getting json decode error on rvc v2 colab i can't fix it

[viscid moss]

Provide more context. What App or Colab are u using?

!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.

[viscid moss]

I'm pretty sure that u re using an outdated colab

Try the new ones

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

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

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

[lusty knoll]

ty

[rare goblet]

Ayo? @lusty knoll level 1 !!!

[viscid moss]

guides here:
https://docs.aihub.wtf/

[dusty lava]

How’s gradio?

[junior halo]

Gradio is finally in operation again. :)

[steel forge]

-realtime @mild fern

[azure marsh]

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

[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

[barren acorn]

idk

[rare goblet]

Ayo? @barren acorn level 4 !!!

[edgy fjord]

How i an make a model on phone? I tried with google colab but it takes a long time and if I leave the site it doesn't work

[proud elbow]

phones dont have gpu as good as nvidia desktop ones, so the cloud colab is the only option

[edgy fjord]

I know, but isn't there another method ?

[pure scroll]

[proud elbow]

prob some GPU renting services like vast.ai but u need to set up anything needed, which is beyond my knowledge (no, not setting up with your phone)

[pure scroll]

pick up the lowest point shown in the graph of 'loss/g/total'

between 4000 and 5000, i can see

try exporting .pth of that epoch count

[proven hill]

excuse me??

[pure scroll]

LOL

[proven hill]

god i love misinformations in the morning

[edgy fjord]

Ok

I have .pth file how i use it

[viscid osprey]

people are saying my mic is hissy when using the official RVC's realtime VC on an AMD gpu. when i check it through my headphones it's totally fine

[graceful obsidian]

just woke up, In bed rn but

approx, try to look there

use just 0.1 - 0.3 smoothing, then confirm best one at 0.0 smoothing

just don't pick collapse ones

kl shouldn't be referenced, just used when you have 2 or 3 extremely close in loss FMs or MELs to determine which is more suitable

also, show me the total G, fm and mel at 0.2 and 0.4 smoothing

no ignore outliers checked

oh, wrong section, uhhh

go to scalars, actually, just send me tensorboard file

then, gonna be back in few mins with results

model's folder, file called tfevents file

cause you either trained in 2 sessions or past certain point, the file splits in 2

but ye, will be back

gimme few mins

oh, nono I meant the exact same model in 2 sessions

overal, that's quite a flatline graph

oscilating anywhere from sub 37-38 to 42-43

barely any progression, however

til around 4k it's def better than any further

4 - 6k is a zone I'd be careful about

6k - the last I wouldn't advise picking ckpts from there

total G: Mel: FM>:

• 1.74k <-> BEST FM aligns with this one

• 2.12k <-> Decent Mel aligns too

• 3.52k

• 3.56k

• 3.64k <-> Decent Mel aligns too

• 3.84k <-> BEST Mel aligns too

there

the heck, formatting

tl;dr, 1.74k steps has the best FM
3.84 has the best MEL

clarity ish fidelity, artifacts, quality in audio attributes

it's in pinned

this channel

ye, pinned msg

all is there

you just gotta go there and scroll down

total G at 1.74k steps has the best FM but not the best overal stats ( treat it as FM ckpt )
Total G at 2.12k steps has much better overal stats ( treat it as overal decent ish ckpt )
at 3.84k has the best MEL so, treat it as mel ckpt

3.84k might be sussy but let's say positive

overal, I'd take and test 3.84k and 1.74k separately

then, if 3.84 seems okay ( clarity should be your focus, ignore model:dataset accuracy in reproduction ) and 1.74 is okay ( ignore fidelity, focus on accuracy )

fuse both of these

in 50 / 50 proportions

FM being the slot A model

but I'd do the opposite combo too, so FM in slot B

yea, that won't work

that's the issue with saving freq :^

so, from those I sent

which ones you have

if even-

any of those I sent, align with yours?

as in, you got em?

well, you are in damn luck

lmao

let's see alternative for FM

2.68k - " best " fm other than the one you don't have
4k - probs crap but worth testing if there's no other options left for choice
3.64k - not the best of choice either

these are for fm

sadly, your general fm stats aren't the best

they oscilate anywhere from 13-14 to 17 or so

typically, decent FM for a model is from 10/9 to 7/8

as for that

in pinned msg, there's explanation of fusing models

and ye, that's all I can really say about your training :L

kinda flatline - nasty one tbf, but not all's lost

just test stuff around

that is also a factor
but batch also plays a huge role

Can't say much more without batch range testing

so

for reference

I always do range finding for my prototypes

then I more or less strike between rather stable but descending graphs ( due to batch choice

and if you have storage issues, just

as you train, as you go

remove bad ckpts

to save space

you know

all

of those that are up ( total G wise )

and not down

those aren't wanted

but always worth rechecking for Mel and FM

that's what I used to do when space limited on saturn + training big big models ( requiring tons of epochs - cause I am limited to 5 gb there

best of luck! I gotta go for now

sure thing ~

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

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

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

[graceful obsidian]

oh, any should have it, tho if yours doesn't

https://github.com/codename0og/hifi_testing_new/tree/main

if something doesn't work, use my requirements list
( just so you know, as I dunno what you use, hardcoded port used is 8000 - cause saturn. Haven't tested it on other cloud platforms. Just local n saturn ) Gluck ~

[main falcon]

why does it says file is not a zip file

i used the link in #1175430844685484042

[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

[slate halo]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

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

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

[severe violet]

hello can someone help me with voicemeter banana

[foggy folio]

No I am not making a model I am just making covers

[graceful obsidian]

that's a part of it

mdx23c or bsroformer ( 3rd edition ) are the best isolators for vocals

then for lead / backing acapella isolation.. uhhhh might be 44 BVE model ( in uvr models section )
Sadly, that one sucks ass but it's the best I can think of ^

[foggy folio]

what about the hp5 lead vocals model

[proud elbow]

BS roformer 2024.04 is the best though exclusive in mvsep, and its instrumental stem is a bit muddy

[foggy folio]

that is ok

putting the vocals over it covers it up

right??

[proud elbow]

it is vocal separation, but BVE is better as first attempt

use it on isolated vocals

[pure scroll]

BS Roformer version 3 (latest)

VR UVR-DeEcho-DeReverb

VR UVR-BVE-4B_SN-44100-1 (Backing Vocals)

VR UVR-DeNoise

This is all you need for cleaning

[foggy folio]

ty!!

time to make some caseoh ai covers now

[pure scroll]

then the other cleaning goes by using iZotope RX 10 and Audacity

[foggy folio]

all of them were comingg out robotic last night lool

[foggy folio]

that seems like the old method

I do remember that being popular but I think that is old now

[pure scroll]

who said!!???

using iZotope can clean the dataset way to good

[foggy folio]

yeah maybe

[proud elbow]

prob the model itself or the index rate, have you also tried inferring on normal speaking audio/TTS?

[brittle wing]

Could've been the sample I inferred with

[foggy folio]

nah they're very good now

[brittle wing]

You speak Italian?

[proven hill]

yes im italian

[brittle wing]

Bene

[brittle wing]

If it's not the fork it must be the sample I inferred

[proven hill]

youre italian?

[proven hill]

yea most probably

the fork simply works like any other fork

[brittle wing]

No I'm not I just know the language

[proven hill]

really?

[brittle wing]

Yeah cause I used Uvr bve2, then de-reverb but didn't Denoise
Now I Denoise the samples and have no issue

[brittle wing]

Si

[proven hill]

denoise is absolutely needed most of the time

[proven hill]

come mai parli italiano? studiato a scuola o hai qualche parente?

[brittle wing]

Yeah

[brittle wing]

No sono solo vissuta come bambina

Ma qui non e il posto dove parlare su queste cose

Si ho anche studiato a scuola

[proven hill]

vero, se vuoi i miei dm sono aperti per te, anche se ti serve una mano per rvc

[brittle wing]

Se puoi poi evaluare il mio modello e dirmi se e buono o no.

[proven hill]

certo

[brittle wing]

I submitted in #1159514067187277865 but I know I'm not gonna probably pass for a reason;)

[brittle wing]

E dirmi quali sono i problemi

[proven hill]

non è male ma probabilmente potevi allenarlo di più

quanto minuti di dataset? quanti epoch?

[brittle wing]

@proven hill di più?

[proven hill]

dm

[brittle wing]

Guarda il tensorboard

hello discord cant detect my voice i cant use vc

any idea how to fix?

but when i iuse my mic its sounds like my own voice not ai voice

[analog current]

guys, can i train a model with old gpu?

[rotund prawn]

Which petrain mode should I use for Japanese?

SnowieV3.1 doesn't exist on rvcDisconnected

[graceful obsidian]

if it's nvidia and above 4gb, yea sure ( 4gb cards aren't really worth the hassle. Batch_size would be the limiting factor sadly

for instance, back when I tested borrowed gtx 1050ti 4gb ( as I am an amd user ) t was just fine, can't promise anything for older or weaker than that gpus

[past bay]

hey guys, copying from main chat. ive been using this guide for ai covers but cant make any voice model work anymore: https://docs.google.com/document/d/1J-JlFGglJY7SpLtC-_3p8lW8CL_WJYqbo5X6CA7OnKM/edit

using Corpse's models from ⁠🎧┋voice-models and archives, but none are running in the Colab - giving the* JSONDecodeError: Expecting value: line 1 column 1 (char 0)* error

anyone ran into this issue and knows how to fix, perhaps? also tried downloading the zip file and uploading on own drive to share a link with Colab - no use. has there been a change in ai cover process?

thanks a ton in advance, noob here 🥲

[proud elbow]

at least 6 gb VRAM, or 8 gb recommended esp for batch size 8

[maiden raft]

which algorithm should I use for training and inference?

[rotund prawn]

rmvpe

[brittle wing]

can anyone help me wit hthe realtime voice changer?

[maiden raft]

thanks

[rare goblet]

Ayo? @maiden raft level 2 !!!

[rotund prawn]

np

[timid ruin]

Hello together. If I want to use an AI voice, I have to upload an audio file with my voice where I say the text.

Is there another way to do this? That I let say "Google Translater" speak the voice? Or can I also "store" a text field somehow?

Thank you :;)

[foggy folio]

why do ai covers sound so robotic for me

I made this on mangio rvc. I remember other ai covers sounding way better than this

acapella is pretty clear aswell I'd say

[pastel oak]

Might be the voice model or some setting

[foggy folio]

I think this is the main acapella

oh wait that isn't it

here it is. this one has denoise and I made the cover with this one

well I mean what is the most balanced index rate to use?

just made another one with the mario model

also very robotic

[fast shadow]

Hi!
I made a Google drive model but when I upload it to the gui it says it isn't a zip file and apparently it does this with any Google drive model :/
But when I put a hugginface model it works! I don't know what can I do?

[violet heron]

What pitch extraction?

And what index rate

[foggy folio]

rmvpe

assuming that is the search feature ratio

for index rate

[odd shale]

Harsh S's can screw up the final model results.

Sometimes you must some RX10 editing to your dataset.

[foggy folio]

it isn't my model btw

the mario one I found on weights.gg

https://weights.gg/models/clm72khs9011rcctcd2pr8yyg

I gtg for now

[violet heron]

Yes

[waxen mica]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

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

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

[maiden raft]

how do i split 2 vocals?

[analog current]

if i use google colab for training, do i need a gpu?

[proud elbow]

nah it is cloud gpu, not yours

[tardy topaz]

Having trouble with RVC1006Nvidia
Was working fine before and when I launched it randomly the next day the audio it outputted sounds glitchy and distorted like this

[maiden raft]

how much faster would an rtx 4070 super compared to an rtx 2060 6gb?

[pastel oak]

Whats your gpu and send screenshot of rvc

[tardy topaz]

GPU is Nvidia GeForce RTX 3060

[rare goblet]

Ayo? @tardy topaz level 2 !!!

[pastel oak]

Your gpu cant handle the sample length

[tardy topaz]

But it was working fine before ?

[pastel oak]

change pitch detection to rmvpe

[merry eagle]

i am getting this error does anyone knows how to fix this shit. @acoustic scarab @robust needle @timid olive @tired kraken

[tardy topaz]

oh yeah, it was the sample length I had to get it around .15-.20

Thank you 🙏

[pastel oak]

dont forget the rmvpe too

also you can get rid of more delay with my guide

https://rentry.co/Realtime-less-delay

Wasapi steps only

[pastel oak]

I dont know what kind of answer youre expecting? If you look at benchmarks youll know the big difference

what application do you want to run "faster"

[maiden raft]

rvc

training

cuz im trying to do something with my friend that has a 4070 super and i dont know if it would be worth the effort to get rvc running on his pc to speed up training

[proud elbow]

vanilla 4070 is also good enough, what matters more is 12 gb vram

[foggy folio]

man what version of rvc is this

[proud elbow]

latest one

[graceful obsidian]

wait

who pinged

[graceful obsidian]

I mean, no, you don't

but I don't really provide colab support so, someone else has to take over ( too much changed since the last time I used colab.

[foggy folio]

that doesn't look like mangio or applio

[proud elbow]

realtime one

[foggy folio]

ah

[nocturne mural]

uh

gradio share down again

:c

wait

[merry eagle]

when i use this on 1:00 am that was working for me correctly

[tawny lily]

Hi there. Is anyone around to answer 2 very simple RVC 2 questions I can't find an answer to anywhere?

[nocturne mural]

press f5

and try again

[merry eagle]

wait i am starting this again

[nocturne mural]

you could have just executed the second cell, I only made a small change, since I had already foreseen this.

but well

[tawny lily]

During inference, harvest has an option for feature retrieval rate. What is this and what's an ideal setting.
Crepe has a hop length option. What's this for and what's an ideal setting?

[rare goblet]

Ayo? @tawny lily level 1 !!!

[merry eagle]

nope not working

[tawny lily]

🤔 😬 not sure what I'm doing wrong that he's yelling at me about 😦

[nocturne mural]

screenshot

[nocturne mural]

hello? Are you sure that this is how the colab is displayed now?

[merry eagle]

now this displayed

[rare goblet]

Ayo? @merry eagle level 2 !!!

[nocturne mural]

When you executed the second cell, didn't you get a link and a password?

[merry eagle]

last time i tried this option was't there

[merry eagle]

trying

[nocturne mural]

I have updated it now because it had already happened before. I thought it had already been fixed but since you mentioned that there was no link, I put it back.

when traductor spanish -> english

[merry eagle]

what to do next?

[nocturne mural]

uh

[merry eagle]

what happend?

[nocturne mural]

access this link and the password is the one below, but delete the image because if not they will access your link xd

[merry eagle]

oh ok

[nocturne mural]

there you enter the password and hit submit and that's it.

tell me if it works or not

A
I guess since you don't say anything, it will be a yes.

[nimble flax]

where is the link for applio fork, im on the github but i canty find the download file

[split frigate]

Can any1 send me the steps to create a model?

I forgot from where can i get it

[nocturne mural]

https://docs.applio.org/get-started/installation/

[nimble flax]

but thats normal applio or is fork the same thing ?

[nocturne mural]

it is only the normal applio ,what do you mean the fork?

[nimble flax]

he has fork ?

OOH

[nocturne mural]

xd

[nimble flax]

nvm that was 2023

IM SO RETARTED

sorry

[nocturne mural]

dw

[brittle wing]

whats the best app to use for rvc? Ive been using MMVC for a while

[nimble flax]

same but i feel like it sounds to unrealistic

[merry eagle]

it works

thank love you

[nocturne mural]

it was the ip that was below the link, not your ip xd

but well

[merry eagle]

thank you my love

love you brother

[rare goblet]

Ayo? @nocturne mural level 12 !!!

[brittle wing]

how can I hear myself with the realtime voice changer?

while im running it through discord

[merry eagle]

would you change back this in feature?

[pastel oak]

You can go to sound devices -> recording -> Line 1/Cable. Right click, properties, then on the 2nd tab enable this

[merry eagle]

i'll give you a advice. don't change this one. because this one is working perfectly. since i am searching from yesterday i was't found any rvc like that. this one is easy and also have no issues.

[brittle wing]

anyway to fix lag

[pastel oak]

Define lag

[merry eagle]

love those teams

[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]

sometimes it will reapeat the last thing i said

like 4 times

[pastel oak]

then your gpu cant handle the sample length, raise it

[desert niche]

is there a way to inference gpt sovits using audio longer than 10 seconds, I've seen ppl inference using 1 min audios, for example, the one in the image below, longer than 10 seconds, at 22 seconds

[brittle wing]

I think it's MMVCServerSIO

[quaint dock]

Yo guys whats the best real time voice changer?

Local

[pastel oak]

We use RVC realtime and wokada. To put it simply:

If you have AMD gpu, you should 100% use RVC
If you have nvidia, you can use rvc or wokada. I prefer RVC still

Differences:

• wokada has user friendly ui, its very simple and easy to understand
• rvc is minimal, but has better performance and more flexible with settings
• wokada is more bloated and more hardware usage,

wokada has more hardware usage (cpu especially) so playing a game next to it sometimes is more difficult if you have a lower end gpu

[worthy lantern]

public url still down?

[quaint dock]

Alright, thanks!

[quaint dock]

I'll definetly use RVC

[sand iris]

lmao

[worthy lantern]

im using Ilaria RVC but not getting public url? why?

[violet heron]

Bro is sending Roblox malware/hacks in a discord server about ai covers

[odd valve]

where do checkpoints save in google collab?

[rare goblet]

Ayo? @odd valve level 3 !!!

[worthy lantern]

bruh

[violet heron]

Bottom file file button in logs

[odd valve]

it says logs/name/d_23333 but i cant find that anywhere in my drive

[odd valve]

?

where? i cant find it in my drive anywhere

[violet heron]

It’s not on your drive

[odd valve]

where is it then?

rvc v2

[violet heron]

There is a file button on the bottom left

[violet heron]

(Or tools, show file explorer)

[rapid wharf]

Idk if this is normal but I've been waiting for this thing to be done converting for over 25 mins

[violet heron]

Your using RVC-GUI which is verryyyyyyy outdated

[odd valve]

yep i did that

[violet heron]

Use original RVC, Ilaria RVC, or applio instead

[rapid wharf]

I'm just trying to make a fnf chromatic scale for Dr. Don't 😭

And the last time I used the original one it wouldn't even open 💀

[violet heron]

Which file were you opening

I would try Ilaria RVC Mainline

[odd valve]

thats what i thought, but i cant really do much with the files in there as they arent on my drive

[violet heron]

https://github.com/TheStingerX/Ilaria-RVC-Mainline/releases/tag/beta-2.0

[odd valve]

if i try to download them it doesnt do anything

this?

i think this needs the training to finish right?

im trying to get my checkpoints downloaded just incase i get disconnected or somethinbg

which has happened before and ive had to restart

how do i get my checkpoints/saves downloaded

[proper shale]

yea get 3.11 or 3.10

[rare goblet]

Ayo? @brittle wing level 49 !!!

[odd valve]

this download button doesnt work

ill get the download to work

what i want to know though, if i get disconnected and want to reconnect from last checkpoint i need the d_233333 file and the g_2333333 files right?

@brittle wing

alright thanks alot

[worthy lantern]

how long has gradio been down? 😭

[rapid wharf]

When I click on the download link it says: "Entry not found"

[brittle mango]

How should I switch?

[violet heron]

https://huggingface.co/TheStinger/IlariaRVCBeta/resolve/main/Ilaria RVC Beta 2.7z?download=true

Direct download

[brittle wing]

whats a good gpu to use while playing and using the voice changer

like 3090?

2080

[violet heron]

But isn’t it precompiled?

[pastel oak]

high vram amount, so the Ti versions of each gpu are not recommended

[brittle wing]

what is a good one to get that can run it well

[rare goblet]

Ayo? @brittle wing level 3 !!!

[pastel oak]

literally any rtx XX60 +

so 2060 3060 3070 4070 etc

as long as its not the Ti versions

they should all have 8+ vram

I personally have 3060 Ti, but with RVC realtime i never run into any issues, its just enough with its 8gb vram

so id say look for one with 8+ vram, if you find one of the Ti's with 8+ its good

[brittle wing]

ok good im getting a 3090 in a little

[solar glade]

hey guys

what can i do when this happens?

it never loads in

Please open the following URL in your browser.
http://<IP>:<PORT>/
In many cases, it will launch when you access any of the following URLs.
http://127.0.0.1:18888/

[VCClient] Access http://127.0.0.1:18888/
[VCClient] wait web server...0 http://127.0.0.1:18888/
Booting PHASE :main
Booting PHASE :MMVCServerSIO
[Voice Changer] VoiceChangerManager initializing...
[Voice Changer] Web Server Launch Exception, Expecting value: line 1 column 1 (char 0)
[VCClient] wait web server...10 http://127.0.0.1:18888/
[VCClient] wait web server...20 http://127.0.0.1:18888/
[VCClient] wait web server...30 http://127.0.0.1:18888/
[VCClient] wait web server...40 http://127.0.0.1:18888/
[VCClient] wait web server...50 http://127.0.0.1:18888/
[VCClient] wait web server...60 http://127.0.0.1:18888/
[VCClient] wait web server...70 http://127.0.0.1:18888/
[VCClient] wait web server...80 http://127.0.0.1:18888/
[VCClient] wait web server...90 http://127.0.0.1:18888/
[VCClient] wait web server...100 http://127.0.0.1:18888/
[VCClient] wait web server...110 http://127.0.0.1:18888/
[VCClient] wait web server...120 http://127.0.0.1:18888/
[VCClient] wait web server...130 http://127.0.0.1:18888/
[VCClient] wait web server...140 http://127.0.0.1:18888/

it just loads and loads

[rapid wharf]

I tried applio but it says: "env/python.exe: can't open file 'C:\Users\49162\app.py': [Errno 2] No such file or directory"

[pastel oak]

This is caused by:

• You havent unzipped the MMVCServerSIO folder
• Your Antivirus might interfere
• Your MMVCServerSIO folder is inside OneDrive. Put it somewhere else like Desktop
• Your path to the MMVCServerSIO has spaces in its name. Remove all spaces, make it one word "AIVoiceChanger" instead of "AI voice Changer"

If those werent the case or didnt work, delete the file stored_settings.json and try again

If all failed, your installation messed up. Delete the folders pretrain and model_dir, launch start_http again and itll reinstall

[violet heron]

Tried the installer or precompiled?

[rapid wharf]

I ran the installer and that finished

[rare goblet]

Ayo? @rapid wharf level 3 !!!

[marble forum]

Bro can this be used in real time..
And how updated is this. I've been having so many problems with RVC

[violet heron]

Use original RVC or w-Okada for real time

[marble forum]

What's the original RVC....please link

[violet heron]

-local

[azure marsh]

Suggestions for @marble forum

🖥️ Local stuff

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

Need some help? 🤔

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

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

[violet heron]

I would prefer w-Okada for real time tho

[marble forum]

Any link to that bro. I don't know if I have the correct one. That's compatible with my pc. I encounter so much problems with it. I have an rtx 3050 4gb

[brave garnet]

Voicechanger Download (Okada)

⠀
Download for Nvidia GPUs
Version 18a cuda

Download for AMD GPUs
Version 18a directml

Download for Intel GPUs
Version 18a directml

Download for Mac
Version 17b Mac
⠀

Made with love by Ilaria 💖

[violet heron]

@marble forum

[marble forum]

By the way these work best for just normal voice cloning right?

[violet heron]

No those are only for real time

Original RVC can do training, covers and realtime

[marble forum]

One major problem I have. Please is there any way to reduce the generated voice duration after training. I find that after training a voice for hours for text to speech. It still takes to much time to generate sentences. Please any solution to that ?

[plucky echo]

Got this error on loading program:

[12768:0429/060320.505:ERROR:ssl_client_socket_impl.cc(978)] handshake failed; returned -1, SSL error code 1, net_error -101

Error
unhandledrejection
no error stack

TypeError: (0 , Zz.removeDB) is not a function
TypeError: (0 , Zz.removeDB) is not a function
at http://127.0.0.1:18888/index.js:2:3293292
at m (http://127.0.0.1:18888/index.js:2:1536451)
at Generator.<anonymous> (http://127.0.0.1:18888/index.js:2:1537797)
at Generator.next (http://127.0.0.1:18888/index.js:2:1536880)
at e (http://127.0.0.1:18888/index.js:2:1543652)
at s (http://127.0.0.1:18888/index.js:2:1543855)```

[scenic panther]

when I open app a white blank screen pops us

up

[violet heron]

More training doesn’t mean it will infer faster

[vagrant mortar]

Why when training a model with a different hubert than the one I have now in the local app, after making inference the result is bad? (He sounds like he's mumbling and he doesn't understand what he's talking about at all)

help me!

[rare goblet]

Ayo? @vagrant mortar level 2 !!!

[verbal vapor]

guys

is yalls rvc taking hella long to process

the song isnt even 3 minutes.

[graceful obsidian]

might be it froze for you

that is, if you use it locally

do you?

[verbal vapor]

oh

i dont

i use the oublic

public

[graceful obsidian]

oh, then I dunno why it's like that

just know scenarios it can freeze locally, then spamming enter few times into the cmd does the job

but ye, that's that

long dataset?

also what did you set for cores / threads

[verbal vapor]

i already like refreshed the page changed the ai and song and still

[graceful obsidian]

oh, you mean inferencing

then

processing is about the set

you're talking about inferencing

if gpu's not available / if it's cpu only? dunno - then ye, it'll be slow ish

[verbal vapor]

ph yeah

i ran out of gpu (?)

[graceful obsidian]

Might be

[verbal vapor]

bc they woukld be times where the page would say error and error constantly

and i had to refresh

[graceful obsidian]

but overal, it's typically at least 1.5 or 2.3 times the length of the input

for cpu inferences

I'd say, 2 times the input's length ( acapella / song )

[graceful obsidian]

ye, then that's most likely it

give it a bit of time

[verbal vapor]

alright

thx

[graceful obsidian]

sure thing

[verbal vapor]

it finally loaded

[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

[scenic panther]

I need help

I don't know where to change pitch

[verbal vapor]

brother

[verbal vapor]

u just like scroll down

''advanced options''

and like read unless you need smth else

[timid ruin]

Hello together. I’m training a voice and it’s taking awhile and my PC is very loud. I wanted to ask if rvc can „rip“ my GPU? Because I can’t do anything when it’s running and I’m scared that I’m destroying my gpu with that

[rare goblet]

Ayo? @timid ruin level 1 !!!

[graceful obsidian]

Nah, you'll be fine

it's comparable to running a really demanding game really

[verbal vapor]

damn..any ways i can make it run faster

[rare goblet]

Ayo? @verbal vapor level 3 !!!

[graceful obsidian]

and it is loud because it's all heating up right

it gotta cool itself

[timid ruin]

Okay thank you. And another question how can I check that my Prozess isn’t freezing? Like it’s at a point always at 88% and it shows „◻️◻️◻️◻️?“

[timid ruin]

Ahh I got it. Thank you a lot for your quick answer 🫶🏼

[graceful obsidian]

well, you'll know if epochs do advance in count

but if you want to be sure, get into the cmd

u know, that thingy ( not rvc on the ss tho, just for demonstration )

click it so the window's active

and press enter few times

then click somewhere else, outside the window

( sometimes it " freezes " like that )

[timid ruin]

yeah but I’m waiting and waiting and than it’s shows ◻️◻️◻️? and nothings happening all the time

[verbal vapor]

sigh

[timid ruin]

Okay I will test it <3

[graceful obsidian]

can you show me a screenshot of your cmd?

[timid ruin]

Sorry I’m now in my bed and stopped the process because I can’t sleep with my Airplane haha

[graceful obsidian]

oh, got ya

dw then

[timid ruin]

Is it okay when I’m @ you tomorrow with the Screenshot?

[graceful obsidian]

but ye, that's all I can really say to be " safe "

[timid ruin]

Okay screenshotted :) Thank you a lot for your help

[graceful obsidian]

nono, that's fine. Unless something's going on, you can send me a ss anytime

yw

[timid ruin]

Thank you have a good night 🫶🏼

[verbal vapor]

so does anyone know how to make it quicker..

[violet heron]

Local or HuggingFace

[verbal vapor]

hugging face

[violet heron]

Did you duplicate the space

[verbal vapor]

how do u do that

[violet heron]

3 dots, duplicate this space

You need a HuggingFace account (free)

[verbal vapor]

ok done

now like

how do i duplicate the space

[violet heron]

3 dots the space you were using, duplicate this space

If your on iPhone it has to be public

[verbal vapor]

but like i dont get it where are the 3 dots

[violet heron]

On the space you were using

[verbal vapor]

''clone this''

repository?

[violet heron]

“Duplicate This Space”

[verbal vapor]

for some reason

do i have to confirm email or smth

[violet heron]

Check your email and confirm it

[proud elbow]

which huggingface space? it uses cpu, you'd better use a colab one or a local fork

[verbal vapor]

it still looks the same

[verbal vapor]

im using a colab one actually

[violet heron]

Reload HuggingFace and try again

[violet heron]

No your using a HF space

[verbal vapor]

oh

[violet heron]

Now go back to your space and duplicate it

[verbal vapor]

im so lost

like wdym my space

like the rvc??

[violet heron]

Yes

Go back to it and duplicate it

[verbal vapor]

but like

[rare goblet]

Ayo? @verbal vapor level 4 !!!

[scenic panther]

@violet heron how do I change pitch?

[verbal vapor]

i cant find the

3 dots in the rvc

[violet heron]

Pitch

Or transpose

[scenic panther]

where

[scenic panther]

I don't see it says that

I don't see it

[verbal vapor]

wait

OH

i dont use like the

hugging face space

..

like i copy the link from the hugging face paste it on rvc change the blob thing to resolve and just

[proud elbow]

which rvc are u using?

[verbal vapor]

llaria

[proud elbow]

yea the old ilaria rvc in hf space, it's slow even slower than average laptop cpu I had on a local fork

[violet heron]

Wait

Are you trying to use Ilaria on colab or hf

[verbal vapor]

[scenic panther]

I need help to change tune pls

[rare goblet]

Ayo? @scenic panther level 3 !!!

[scenic panther]

the pitch

[violet heron]

What are you using

[proud elbow]

-colab

[azure marsh]

Suggestions for @verbal vapor

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

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

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

[violet heron]

How did you get there

[verbal vapor]

/joke

[violet heron]

Drag the tune slider

[scenic panther]

but like how do I put it to its specific like pitch

some voice models need pitch

[verbal vapor]

well

i entered this server like the discord one

and was like lost

so someone like helped me out through dms

[violet heron]

Was the link a google colab or HuggingFace link

[verbal vapor]

google colab

[violet heron]

Ok so your using the colab version of Ilaria RVC

You should only use HuggingFace in this case for model download links

[proud elbow]

I will redirect you to the colab link from the guide above
https://colab.research.google.com/drive/16LkwvFZeudTpUOsE_6bMjOq2qkxFo8Hr?usp=sharing

[queen atlas]

how to get it to stop cutting out the audio

[exotic vapor]

Once you download a voice model, how do you use it?

[queen atlas]

bro can someone explain why it cuts out so much

[verbal vapor]

so theres no way of making it take less time

[junior halo]

Looks like the code is getting directories confused? I'm trying to resume training.

Did I enter in something wrong? or

@steel forge Any ideas?

[steel forge]

check your name you put for the model

is it a typo?

[junior halo]

The experiment name?

[steel forge]

yeah

[junior halo]

Nope, the experiment names on the colab and the folder name on Google Drive are the same.

Would it help if I share you my screen?

I'll just start over. .-.

[thin stump]

Don't use special characters and spaces if that makes a difference

[soft tiger]

could someone help me with making this damn voicemodel

idk what im doing wrong

and its driving me insane

@thin stump

[thin stump]

Like on what part? The dataset, training or?

[soft tiger]

idek what to call it can i just show oyu in vc

cause im abotu to throw this damn pc

[thin stump]

you can post a screenshot

[soft tiger]

its like idk how to explain it

its more than just a ss

after i do everything the voice model dosent sound like anyone at all

this dosent want to get renamed

this dosent want to index train

like idk what im doing wrong

[thin stump]

Did you run the dependencies then set the training variable? It's covered in this guide too https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/#2-set-up-colab

I only had problems with the "Load Dataset" part as a newbie. Your zip file needs to have no spaces and you should ignore whatever pic attach says. Just zip up the audio file

[soft tiger]

yes

[soft tiger]

and i made sure to name it with nothing more than one name

[thin stump]

Could you make a brand new zip file of the audio file with a slightly different name and redo it? Idk, it should work

[soft tiger]

acutally i think this worked

but well see when i export

[thin stump]

It's more of a google drive issue, or in some cases you forgot to include .zip after the end of the name

[thin stump]

Which pretrain do you want to run?

[soft tiger]

uhh idk what that is

[thin stump]

https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/#1-training-variables right here and the list here describes all the pretrains

[thin stump]

scroll down

[soft tiger]

oopsies

i missed a step ig

[thin stump]

https://docs.google.com/document/d/1j9J8A8Oop9bMOHmCs3jDXzPujuD6TQ0Q396rJ0MyuIc

that describes it, if one of them doesn't work then I can troubleshoot for you

your mostly using original or Ov2 anyways

[soft tiger]

is there any way i can make this go faster?

[thin stump]

how long is your dataset?

[soft tiger]

before it was doing like 4 seconds and now its going 17 seconds

[soft tiger]

a little over 2 minutes

[thin stump]

You can use Ov2 with a batch size of 4. It will not take you that long

if you do get an list index error, I have a fix for that

[soft tiger]

how do i use Ov2

oh nvm i see

[thin stump]

by the way, is your dataset 32k or 40k?

[soft tiger]

uh i thin

idk

[thin stump]

sorry, i'll just make it simpler. Did you extract it from youtube?

if yes, then set your target sample rate to 32k. It'll make your model sound cleaner

[soft tiger]

no

i had to reupload

it didnt upload

hole on

[thin stump]

I can't read that . Up to you if you want to do 32k or leave it as 40k

[soft tiger]

igh

still taking 3000 years

[thin stump]

There's different google gpus. So if you're rich and can buy a lot of credits, then switch to A100

makes model making go brrr

[soft tiger]

mannnnnnnnnnnnnnnn

theres goota be a cracked version somewhere 👀

[thin stump]

I haven't looked at paperspace though that cost money for credits too

[rare goblet]

Ayo? @thin stump level 24 !!!

[soft tiger]

damn damn damn

[rare goblet]

Ayo? @soft tiger level 6 !!!

[thin stump]

Doesn't hurt to try A100 just for one model. But not as a free user, alright. Click on the T4 to switch

[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

[open axle]

Can you help me with the RVCv2 link?
please

[blissful minnow]

C:\Users\peura\Downloads\Retrieval-based-Voice-Conversion-WebUI-main>runtime\python.exe gui_v1.py --nocheck
The system cannot find the file specified

C:\Users\peura\Downloads\Retrieval-based-Voice-Conversion-WebUI-main>pause
Press any key to continue
What is wrong with my RVC?

Pytorch already installed

[rare goblet]

Ayo? @blissful minnow level 2 !!!

[blissful minnow]

poetry is already installed

WHYYYY DAMN System cannot find the file specified!

[queen kraken]

did you run go-web.bat?

[blissful minnow]

both of them.

Same happens!

[queen kraken]

where did you download the RVC? if it's the latest version, it should be RVC1006 or something

[blissful minnow]

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

[queen kraken]

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

Download the zip file that matches your graphics card. Since all the files are already in there, no need to download additional files

[blissful minnow]

I'll try and if it works, I'm gonna tell you ,eh

[queen kraken]

okay

[blissful minnow]

I really really appreciate!!!!

It works!

[blissful minnow]

um,, bro this one has only realtime not web, and realtime doesn't have a option to put origin music file and go-web doesn't work. They say "infer-web.py: error: unrecognized arguments: --nocheck"

[marble forum]

I understand that. But why does it still take so much time after training to just process a simple sentence. Since it's already trained isn't it supposed To process the input sentence faster?

[crimson prawn]

where do I find some german female models?

[deep prism]

rvc realtime can using with colab?

[radiant hinge]

why do I always get an error everytime when I'm trying to 'Clone repository and install dependencies' on google colab

[radiant hinge]

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

help

[split frigate]

what is the link for steps to create a model ?

i forgot it

[queen escarp]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

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

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

[blissful scaffold]

does anybody have advice to make it sound less robotic/autotune? every single voice sounds so robotic

[pastel oak]

For voice changer?

[blissful scaffold]

ye because my outcomes always get so robotic

even though what i put in is not

[pastel oak]

Assuming you have all settings optimized properly

It all depends on how the voice model was trained. If the model had a lot of variation, it sounds robotic on realtime.

So a general tip is to look for voices of streamers, vtubers, asmr, "mommy" etc. anyone who uses the same microphone and has the same proximity to their microphone

[carmine hearth]

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

GUIDE : https://docs.google.com/document/d/e/2PACX-1vTIceEcBfS6Zqolv_QEysrFfI_EJikPxozWptP_EjkpLVl-l-gdo-ijBonQMTviAHEYm5emmd9k9TdC/pub

[blissful scaffold]

interesting okay, what about changing voice for ai covers?

[pastel oak]

ai covers, also called Inference if you ever come across that term, any kind of properly trained voice model, even including variations and all, will sound pretty good

Its only realtime that is more tricky to make

[timid ruin]

Hello guys is rvc a Virus?

My Windows got detected

[naive violet]

It normally wouldn't be, but if you're worried you can check what type of threat it is

Then google it

[timid ruin]

Ohh okay google says it is a virus.

But is it from rvc?

@graceful obsidian :/ sorry for ping

[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]

thats not rvc...

[timid ruin]

ohh oaky what is it?

[fluid matrix]

Idk what you installed

Where did you get that file from

[timid ruin]

which file? I think i installed WInrar also for the unzip

[fluid matrix]

You say you got a virus

Im asking what you downloaded

Because it has nothing to do with rvc

If you wanna install rvc, check the docs

[timid ruin]

yeah and i dont know from where. And yeah i already installed rvc. So i asked?

[rare goblet]

Ayo? @timid ruin level 2 !!!

[fluid matrix]

How are we supposed to know 😭 it's in recycle bin anyways, just delete the file and scan your system again to make sure

[timid ruin]

i did i did bro and its still there im sorry yeah

sorry sorry i dont wanted to judge you

[fluid matrix]

Quarantine it

Go on aktionen

[timid ruin]

yeah i tried. Im trying it now which a tutorial

[fluid matrix]

What tutorial

What are you even trying to do

[timid ruin]

https://www.youtube.com/watch?v=ywhgYVBCr9w

[timid ruin]

to remove my virus? xd

haha are you trying to troll me bro

im sorry bro

[fluid matrix]

Ig you can try that but I already told you how to remove it 💀

Gl

[timid ruin]

bro i tried i tried but its still there

please trust me bro

im not that stupid

[fluid matrix]

Just know that rvc is bundled already so it wont need a executible

[timid ruin]

https://tenor.com/view/when-you-had-plans-but-avele-amosheke-nje-gif-17596278013398119965

[fluid matrix]

I didn't say that

Did you scan your system again tho?

[timid ruin]

yeah okay. My friends also said it isnt rvc also but i just wanted to ask dont want to annoying you

[timid ruin]

yeah but it seems like the tutorial helped

[fluid matrix]

Nice

[timid ruin]

thank you bro for you help

[fluid matrix]

Just run the scan again to make sure

You can also try adwcleaner

Np

[timid ruin]

ty

[minor pendant]

any tutorial how to download rvc sovits?

[rare goblet]

Ayo? @minor pendant level 1 !!!

[flat elbow]

what is sovits?

[minor pendant]

GPT-SoVITS

[flat elbow]

isn't it in readme of their github?

[minor pendant]

i mean i cant open it, when i open the go webui it automatically closes

[timid ruin]

How can I check if it’s still running and not freezing? And what does „◻️◻️◻️◻️(?)“ mean?

[pastel oak]

progress bar, its downloading it just shows the percentage as a "picture" if you wanna call it

[timid ruin]

Ich weiß. Habe das auch verstanden aber warum ist da dann ein Fragezeichen?

Und wie kann ich checken ob es immer noch läuft. Weil es schon länger bei den % ist

[pastel oak]

Was fürn program lädst du gerade runter

[timid ruin]

Ich trainiere eine Stimme

Musste auf so ein Feld klicken und dann dauert das

[pastel oak]

"auf so ein feld klicken" hilft mir jetzt nicht wirklich

es gibt verschiedene programme um stimmen zu trainieren, welches hast du

[timid ruin]

Sorry

[rare goblet]

Ayo? @timid ruin level 3 !!!

[timid ruin]

Rvc

https://www.youtube.com/watch?v=ixB9oalT3cQ&t=1129s&ab_channel=AISearch

Keine Ahnung wie das heißt ich dachte einfach „rvc“?

[pastel oak]

wenns nichts macht dann starte es halt von vorne

vllt ist deine internet verbindung zwischendrin flöten gegangen

[finite galleon]

https://docs.aihub.wtf/tts/gpt-sovits/

[peak osprey]

is anyone here, i have a problem?

It's "go-web" instead of "go-web.bat" and when i click on it it doesn't bring me the interface of the RVC

[proper shale]

that's because you don't have extensions enabled on explorer

[proper shale]

yeah u have to open the link it gives u

on command prompt

[peak osprey]

wait

[peak osprey]

That one?

[proper shale]

oh, god

move the mangio folder out of OneDrive

it wont work otherwise

[peak osprey]

Oh shit

[rare goblet]

Ayo? @peak osprey level 5 !!!

[peak osprey]

Oh shit

It worked!

Suddenly

Do i still have to move it from one drive?

Bro say something!

I don't wanna die for a peter griffin voice trolling

[proper shale]

nvm

i mean it worked

but like

idk i don't feel comfy with it shit breaks down

so if something breaks, move it out of OneDrive

[peak osprey]

I do i move it out of onedrive?

[proper shale]

your option

it'll work anyway apparently

[peak osprey]

No i mean how do i move it out of onedrive i don't know how

[versed fern]

hello

D:©RVC-beta0717>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897
runtime\python.exe' is not recognized as an internal or external command
or external command, an executable program or a command file.

why i have this error message ?

[proper shale]

folder name might have some special characters/spaces, if so remove said special characters

[proper shale]

just take the folder, go into ur c: drive and put it inside downloads

[nimble flax]

is there a better voice changer for mid end pcs then MMVC

[rare goblet]

Ayo? @nimble flax level 2 !!!

[proper shale]

like do ctrl + x when selecting the folder, then ctrl + v when u reach the c: downloads folder

[proper shale]

how mid range are we talking

3060?

6600xt or something

[nimble flax]

i have a gtx 1060 T_T

im saving up to buy a better sytem

[proper shale]

[nimble flax]

so is there or not ?

[proper shale]

I'd recommend using normal rvc realtime probably... for the time being

it comes alongside mainline RVC

[nimble flax]

wich one is that

is it this?

[proper shale]

this is W-Okada

[nimble flax]

is this bad ?

[proper shale]

also, if u wanna run it better... reduce extra

[nimble flax]

for my system

[proper shale]

no

[proper shale]

its mehhhh

id say u can still get a good experience from it

[nimble flax]

then what is the rvc ur talking abt

[proper shale]

its the one with horrible UI but has less delay

[nimble flax]

do u have a link ?

[proper shale]

https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/RVC1006Nvidia.7z

when u extract, just open go-realtime.bat

[nimble flax]

🫶

alright ill let you know how it goes, THANK YOU <3

[proper shale]

yw :)

[nimble flax]

so can i delete the okada i have or keep it

[marble forum]

Keep it. Run the other one. Compare both. See which is better for you

[rotund cairn]

bro can someone pls help im tryna make my own rvc but then it says this Enter the path of the training folder: C:\Users\ava\Downloads\Mangio-RVC-v23.7.0\Egirl1\ and i enter this and it says fail what am i doing wrong

[proper shale]

i mean... u can keep it
if u optimize it correctly yk

[proper shale]

remove the extra slash in Egirl1

maybe

[rotund cairn]

still dident work i dont kniow why

[brittle storm]

i have some issue using this, pls help

[rotund cairn]

what do yall use to train a model'

[junior halo]

It looks like TensorBoard is back up! Which graph should I be looking at?

[rotund cairn]

hey everybody its because im trying to train my own model and it says this my audio folder name is test1 so i tried to put "test1" instead of "dataset" and it says fauk what should i put in as the path

[round fog]

not supposed to be zipped

[rotund cairn]

so what do i do making it not zip and then writing test1 again and it works or do i making not zip and copy the folders path and put it in?

[proud elbow]

datasets\yourmodel that contains one or more wav files

[round fog]

your dataset folder is just a normal folder with sound files in it

[round fog]

we on the same wave length

https://tenor.com/view/theoffice-michael-scott-jim-halpert-gif-27622036

[rotund cairn]

wdym /yourmodel the name of the folder or

srry im a beginner

[odd shale]

Tea i need your affection

🐢

[round fog]

inside of dataset folder

name a folder whatever u want

that’s ur dataset folder

[rotund cairn]

i called my folder wawAudio but when i write it, it stille says fail

[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

[wispy burrow]

Try removing the last \

[thin stump]

@fringe shale I use a hot key with OBS and record only when necessary. Then I convert all my anime clips to wav/flac. Those audio files get combined into 4-5 minutes sections like "Sunjinwoo1, Sunjinwoo2" etc.

Because I run it all through Fine tuned BS ROFORMER and then UVR deecho-dereverb. You still have to denoise since I use Izotope rx10

[rotund cairn]

still not

will give 20 dollars paypal to the first 1 to solve the solution

[wispy burrow]

Since there's only 1 file in the folder C:\Users\ava\Downloads\Mangio-RVC-v23.7.0\wavAudio\wavAudio.wav should work too.

[rotund cairn]

still doesent work wtf

[rotund cairn]

do you know how to train models?

[wispy burrow]

Yes.
I even used Mangio.

[rotund cairn]

if i send you 2 sound files can you train it i will pay

or teach me how

[rotund cairn]

and theres 2 files in my folder i made 2 sound files

[nimble flax]

it keeps on not responding

@proper shale

[wispy burrow]

The image only showed 1 file so I assumed there was only one.

[rotund cairn]

oh np but how long would it take you to train a voice if i send you 3 min of database

[pastel oak]

You used the wrong inputs and outputs

See how it says Input device: Microphone (fifne microphone) (MME)

But the other is (Windows directsound)

Change the CABLE one to (MME) too

[nimble flax]

OOOOHH

omg thank you

[rotund cairn]

can you help me to

[pastel oak]

Training model?

are you training local or cloud

[pastel oak]

Theres a trick to reduce some more delay naturally, you wanna hear it? :p

[rotund cairn]

its because what should i write here then second picture is my audio file but when i change the dataset to wawAudio it says fail so what should i write for it to work

[pastel oak]

Is the "wavAudio" inside that folder?

or where is the folder

I give you example to make it easier

This is my dataset "Anni"

So the file path is: F:\RVCv2\RVC1006Nvidia\dataset\Anni

[nimble flax]

yeah im totally lost

this looks so diffrent

[pastel oak]

You need an explanation on what they all do?

[nimble flax]

if you can yes please, but if you dont want to thats also fine

[pastel oak]

I wrote a guide today

so ill copy paste lmao

[nimble flax]

omg thank you >3

[rotund cairn]

this is whats inside the "wawAudio" folder'

[nimble flax]

<3

[pastel oak]

Cant share the guide yet its still under works

[pastel oak]

I didnt ask whats inside it im just asking WHERE it is. Is it inside "C users ava downloads mangio\datasets" or where did you put it?

[rotund cairn]

its inside downloads

[pastel oak]

Send a screenshot of this part where audio clip 1 and audio clip 2 is located in

[pastel oak]

Ok then on that part you write:

C:\Users\ava\Downloads\wawAudio

this is the path

[timid ruin]

hello together. I trained a voice model and i want to continue to train it with new Input(Audios) but dont want to lose my old process. How can I do it?

One guy said with this option: But how?

[pastel oak]

Go into your "logs" folder

Then go into the folder of your voice model name

In there, copy the G and D paths for the "load pre-trained base model G/D path" each separately

Thats how you continue training

[rotund cairn]

did it work

[pastel oak]

Yes

[timid ruin]

yes!

[rotund cairn]

tysm bro can i do anything to support you

[timid ruin]

so?

Darf ich mit dir Deutsch schreiben?

[pastel oak]

No its all good thank you:)

Feel free to keep me updated on your models:D

[pastel oak]

Ne du musst den ganzen pfad angeben

d.h. bei dir ists dann irgendwas mit, lass mich raten, C:\User\JONA\Mangioordner\logs\modelname\G_2333...

also nach dem prinzip

[timid ruin]

guter guess wäre es auf C :D

also gebe ich das so ein?

[timid ruin]

ohh das ist schlau warte

[pastel oak]

du highlightest D part, dann pfad kopieren, und dann da rein

Du musst noch die anführungszeichen " entfernen, die kommen da immer mit dem pfad kopieren

und das gleiche für G

[rotund cairn]

alr tysm again