🧬│ai-chat

[feral halo]

Ayo? @placid mesa level 1 !!!

[forest quarry]

That's French

I don't know

[idle light]

Oh 😭

https://docs.google.com/document/d/1XuxQYiqEhYrdYeCZRRLrmV_ciMKo0bV-jTCGHu_-5Cc/edit?usp=drivesdk did you try with this one?

https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS#scrollTo=Yj-_npuA_HRB

[placid mesa]

cloning into 'Mangio-RVC-Tweaks'...
remote: Enumerating objects: 3102, done.
remote: Total 3102 (delta 0), reused 0 (delta 0), pack-reused 3102
Receiving objects: 100% (3102/3102), 13.82 MiB | 13.36 MiB/s, done.
Resolving deltas: 100% (1944/1944), done.

OSError Traceback (most recent call last)
<ipython-input-7-bfb8a4688ddf> in <cell line: 13>()
11 get_ipython().system('git clone -b pr-optimization --single-branch https://github.com/alexlnkp/Mangio-RVC-Tweaks.git')
12 #Rename to keep backwards compatibility with old variants of Disconnected
---> 13 os.rename("/content/Mangio-RVC-Tweaks", "/content/Mangio-RVC-Fork")
14 get_ipython().system('git clone https://github.com/maxrmorrison/torchcrepe.git')
15 get_ipython().system('mv torchcrepe/torchcrepe Mangio-RVC-Fork/')

OSError: [Errno 39] Directory not empty: '/content/Mangio-RVC-Tweaks' -> '/content/Mangio-RVC-Fork

[forest quarry]

minimize the dependencies step, disconnect and delete runtime ans run the step again

[stiff perch]

man i love weighted calisthenics

[merry vine]

Hey everyone, I'm reaching out for some help with an issue I've been having while running the RVC-GUI program. I'm using a MacBook M2 and I'm trying to run the program from this GitHub repo: https://github.com/Tiger14n/RVC-GUI.

When I execute the command python3 rvcgui.py, I get the message: No supported Nvidia cards found, using MPS for inference. I understand it's because I'm on a MacBook M2, which obviously doesn't have an Nvidia card. However, the main issue is that nothing appears on the screen after this message. It just hangs without any error or further output.

Has anyone else faced this issue, especially on a MacBook M2? I'm not sure if I'm missing something or if there's a compatibility issue here. Any insights or suggestions on how to troubleshoot this would be greatly appreciated!

[idle light]

@dapper trail you should add this error fix in the troubleshooting section

[waxen grove]

Oh, RVC-Gui is pretty much obsolete now

[idle light]

Yeah I still use svc

How did you know GDR?

[merry vine]

Do you have another one to recommend?

[idle light]

Mangio or Mainline RVC :3

[waxen grove]

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

mainline

[idle light]

https://github.com/Mangio621/Mangio-RVC-Fork
Mangio :3

[merry vine]

Thanks bro, and the same way to use rvc-gui?

[waxen grove]

there's a run.sh script. Beware, it's troublesome if you already have python installed

[sour pike]

I want to buy a 4070 for gaming. And how good will it be in realtime vc and stable diffusion?

[forest quarry]

idk much bout sd but realtime vc should be fine

[merry vine]

Hello!!! I`ts like RVC-GUI?

[feral halo]

Ayo? @merry vine level 1 !!!

[frozen wasp]

Hey hey!

[idle light]

Visually? No, Mangio has many more settings, including the ones from GUI

[stiff perch]

anyone else went on a walk today yet? I went on a 20 minute one after my workout.

going outside is nice (if its before night)

[idle light]

I will later

[stiff perch]

nice.

[idle light]

Hope you enjoyed your walk

[stiff perch]

I just heard it burns calories a bit more efficiently after doing HIIT training, i usually ignore stuff like that but it seemed like a good enough excuse lol

[coarse vessel]

more than enough for AI but a great upgrade

[stiff perch]

basically i read a Yahoo article like a grandad and then saw a video saying its legit

💀

other than that, im an advanced weighted chinup lifter now due to me going stupid mode on the pullup bar. (with good form, i didnt swing or anything)

woop

[upbeat rampart]

is there any youtube video that gives a full on indepth guide on how to make the ai voice thingy

[quartz roost]

#1159513888199540817

[upbeat rampart]

ya what do i search up there

[quartz roost]

If you struggle reading we can recommed many tts

[quartz roost]

Depends what you want to do

[upbeat rampart]

i need to make an ai voice model and make it sing sinc ei already got elevenlabs

thats what i wanna do

i think my old model should work still im not sure

[quartz roost]

Start by getting a premade model and make a cover. Training a model takes more steps

[upbeat rampart]

how much steps

[quartz roost]

Is it an rvc model?

[open pivot]

hi y'all, is it ok to post something we made?

[upbeat rampart]

i legit have no clue how to check

[quartz roost]

Most likely not then

[upbeat rampart]

it's a google drive link

i made it a monjth or two ago

[quartz roost]

What was it made with

[upbeat rampart]

from this tutorial i found

[quartz roost]

#1159290752195633273

[upbeat rampart]

just found the video its RVC 2.0

v2

[quartz roost]

Rvc2 models are compatible

[upbeat rampart]

wait no way

what now that i got the model?

[quartz roost]

There is a guide specifically on making an ai cover. Are you using local?

[upbeat rampart]

how can i check 😔

im so confused sorry

[quartz roost]

Local gets installed on your own computer

[upbeat rampart]

oh i can install it

[quartz roost]

See the ilaria rvc guide

[upbeat rampart]

it's also a link

like the model itself

if thats what ur refering to

[quartz roost]

I would backup the model you want on your own computer just in case

[upbeat rampart]

ya its downloaded and i even got the link for it the model itself but which website do I plop it onto?

i could send the video i used to make my model in dms or smthn since i legit no other way 2 properly explain

it used like easyGUI n this public URL ns hit

[feral lance]

How long should i let the model train

[quartz roost]

See ilaria rvc

[upbeat rampart]

ok

[upbeat rampart]

where do i find it

is it the nick088 one

[forest quarry]

1 hour

[feral lance]

The goat responded

what could i do with the model

[opaque sandal]

Does anyone here have experience with training models using mangio on Mac?

[feral lance]

what could i do with the model

[upbeat rampart]

tetski can i tell you how much ur the best tysm

[feral lance]

Tetski X Cem ? ? ?? ? ?

[upbeat rampart]

i just needed to the little console thingy since my model already worked beforehand

[quartz roost]

Im willing to help people making legit music

[upbeat rampart]

would NOT go that far

i cannot express it enough my graditude

[quartz roost]

Well it does not sound like youre a creep looking for little kids voices so were good

[feral lance]

Are you a ai

[upbeat rampart]

people do that??

[quartz roost]

It gets so bad

[upbeat rampart]

bro what

😨

[feral lance]

Are you a real person

[feral lance]

I still dont trust you

[quartz roost]

https://tenor.com/view/fukkireta-anime-vocaloid-kasane-teto-dancing-gif-17355005

[feral lance]

what can i do with my ai models

[forest quarry]

AI covera

Covers*

[feral lance]

Rusian agent 100%

[feral halo]

Ayo? @feral lance level 3 !!!

[quartz roost]

I plug them into translation software to open a window in the language barrier

[upbeat rampart]

just one last question what is the average time it should take for the AI audio to complete

[feral halo]

Ayo? @upbeat rampart level 3 !!!

[gloomy bay]

hi guys

[upbeat rampart]

greetings

[feral lance]

@quartz roost

[upbeat rampart]

NOOO

[quartz roost]

That varies significantly

[upbeat rampart]

IT KEEPS SAYING ERROR 😭

is there a limit to how big the file can be

[quartz roost]

#✨│ai-help

[upbeat rampart]

the voclas

[frosty cloud]

m

[neat sparrow]

v

[stiff perch]

Man, I like having friends, hate the idea of a girlfriend/boyfriend.

perhaps thats just me being young.

or sumn.

[quartz roost]

Learn to love yourself first

[feral lance]

Tetski how to take over a country

I ask this for a friend

[north osprey]

[wicked crater]

Search about guide by adolf

[feral lance]

Isnt he the famous artist

In germany

[wicked crater]

Ofc he is
He painted a hole painting by people

[feral lance]

Didnt he killed himself because he lost in fortnite

[languid harness]

https://tenor.com/view/peter-griffin-peter-griffin-fortnite-fortnite-family-guy-fortnite-peter-griffin-family-guy-gif-1904308451874569151

[wicked crater]

By Jewish

[vernal sun]

anyone know how to use search by language?

[wicked crater]

But he killed 1 billion first

[feral lance]

He is pro

[wicked crater]

Dead people actually

[feral lance]

But didnt whole europe teamed on him

[feral halo]

Ayo? @wicked crater level 1 !!!

[wicked crater]

He was 1 v all

[feral lance]

Username

[true ridge]

hi

[wicked crater]

What

[feral lance]

Do you know something about ai training

[median wren]

messing with AI is too funny

[wicked crater]

Actually i know nothing
I dont even know why i'm here

[spring sable]

Hey

[wicked crater]

Hi

[feral lance]

Cosmo

How can i stop the training

1 hour is passed

[gloomy bolt]

Guys, what files i need to use my model from Google drive? I make model with Google Collab

[spring sable]

How much epochs?

[feral lance]

188

189

now 190

[spring sable]

If u think thats good, just stop it. I don’t exactly remember how to since I haven’t made models in a while but I can quickly figure out again

[feral lance]

do you know someone

that is a expert

[spring sable]

I have a friend that is but idk if he remembers either

[coarse vessel]

depend on the tensorboard graph, but typically should keep going unless you want to refine the dataset and retrain it

[gloomy bolt]

Guys, what files i need to use my model from Google drive? I make model with Google Collab

[feral halo]

Ayo? @gloomy bolt level 1 !!!

[feral lance]

No its for now good

[feral halo]

Ayo? @feral lance level 4 !!!

[feral lance]

@coarse vessel How can i export it

[coarse vessel]

local or colab?

[feral lance]

colab

[coarse vessel]

find the weights folder in (probably) RVC disconnected folder that contains model checkpoints, then pick some epoch models you think good

[midnight flume]

hey people is anyone willing to help me with something please? the tutorial that I am watching says I must load .pth G files to so vits svc but the models I find here are just .pth and also the video says about config files, while what I find here is .index files

[feral lance]

isg

[midnight flume]

refer to my profile pic to see how im feeling

[feral lance]

is goed

[vocal sail]

i want to use my pth model as tts

is it possible

[quartz roost]

kill em with kindness

[feral lance]

Alr

[quartz roost]

i was just talking about pokemon and medical marijuana with my japanese friends. they are no different than us

[languid harness]

talkin bout dez nuts in yo mouf

[vocal sail]

i want to use my pth model as tts..

[quartz roost]

coqui xtts

[vocal sail]

what

[quartz roost]

or rvc inference over any tts you choose

[feral lance]

@quartz roost i downloaded the model what now

[quartz roost]

i use kasane teto tts

[vocal sail]

i want to use my own model

i dont want to use any other models

only my model i want to tts

[coarse vessel]

convert the TTS output using RVC with your voice model

[vocal sail]

i used a random tts then i convert it to my model and it sounds so werid

[vast horizon]

lol

who tf took my model master role

@wind eagle hi

[feral lance]

@vast horizon

i have download the pth

[vast horizon]

uhhh

great?

[feral lance]

How can i use

it

[vast horizon]

im not a helper

idk the new methods to use rvc

[feral lance]

I thought you were a model master

[vast horizon]

its a role

means i just made a bunch of models

[wind eagle]

hi i didn’t

[vast horizon]

can you re add pls

[devout wagon]

Try the resync command

[wind eagle]

i don’t think i can don’t you have to apply for it now

[vast horizon]

i've always had it

[feral lance]

Tiz where i can use my models

[vast horizon]

now i just don't lol what happened

[wintry sinew]

they got more strict with model master

[vast horizon]

in what way

[wintry sinew]

they check your models more carefully to see how good they are

before it was just "make alot of models"

[vast horizon]

lame

all my models were good tho

[dusty junco]

sorry noob question, but how close is an RVC v2 clone to an ElevenLabs clone in terms of quality and accuracy (i know not a fair comparison)

[vast horizon]

legit the best ones for the artists i did lol

[vocal sail]

crazy i cant even tts with custom model 😭

i choosed a random voice and i tts it then i tried to convert it and it sounds so weird

[elder willow]

.hey

Hey

[vocal sail]

hi

[midnight flume]

anyone knows where I need to add the rvc models?

[cold oasis]

hi all

[calm sun]

hey

[elder willow]

hi! rvc is voice to voice conversion and elven labs is text to voice, you can't really compare the two as rvc quality depends on the models dataset quality and input quality when converting

[elder willow]

are you using voice changer or just want to convert for music?

[little cipher]

I don’t remember leaving and i realized i wasnt in this server anymore. Was there a purge or something a while ago?

[midnight flume]

I just installed rvc and I see these models"guanguanV1.path etc, I want to add a weeknd model to do a cover

[feral halo]

Ayo? @midnight flume level 1 !!!

[elder willow]

server was banned i think

[rough holly]

дайте голос ростика

[elder willow]

what do you mean installed rvc

[rough holly]

пацаны

[dusty junco]

so I can’t to TTS with a RVC cloned model?

[feral halo]

Ayo? @dusty junco level 1 !!!

[rough holly]

peyser

[midnight flume]

downloaded rvc and ran go-web lol

[elder willow]

@sour vigil

[dusty junco]

I see, thanks for the reply

[elder willow]

you need to get the vocals you want to convert

so take the song you want to convert and download it

[sour vigil]

@rough holly допрыгаешься

[rough holly]

оба

дай голос ростика брат

[sour vigil]

нету

[rough holly]

а стрея есть?

[feral halo]

Ayo? @elder willow level 3 !!!

[sour vigil]

нету

[elder willow]

and then download vocals

[rough holly]

курседа?

[sour vigil]

нету

[midnight flume]

I do have the vocals Im actually a music producer for a long time.

I wanna add a "the weeknd" model to it

I have RVC 1006nvidia

[elder willow]

download https://discord.com/channels/1159260121998827560/1175558201672994906

that's the model, and put the .pth in the weights folder and the index in logs

and then it will pop up on ur model list

[midnight flume]

yeah thats what i wanted, where to put that in! thanks bro

[elder willow]

np 🙂

[midnight flume]

I gave 10$ to a street musician the other day i guess its karma time ;P

[elder willow]

my ass been messing with rvc for a week now it's super interesting

[midnight flume]

im a producer if u need any help from me in something let me know

[elder willow]

how long have you been producing for?

[midnight flume]

15 years?

something like that

[elder willow]

that's really cool, good luck with everything

[midnight flume]

I hate vocalists tho, thanks man.

[midnight flume]

WEEKND GO zooooooooooom

[midnight flume]

thanks my man it worked

[elder willow]

https://tenor.com/view/happy-laughing-haha-joy-looking-up-gif-15760080

embed fail

[wintry sinew]

😭

[tender jetty]

i'm loic i search voice big ali 🙂

[gloomy bolt]

Yay, i maked my first AI model!

But i need more Epochs, in Epoch 1000 it sound normal,but not nice

Ай мне надоело на английском писать

[cyan hare]

bro more epochs != quality

[spring sable]

what colab do I use to make models as of rn? I haven't made models in so long

[feral halo]

Ayo? @spring sable level 2 !!!

[spring sable]

this could be a graph issue it's a possibility

[modern carbon]

does someone have duckus egirl voice?

[languid harness]

Searching for the perfect RVC e-girl voice model?

Look no further – we've got you covered!

Our RVC e-girl voicemodel is in high demand, and we've streamlined the process to make it effortlessly accessible for you.

[heavy lark]

hi

[coarse vessel]

you'd probably want to refine the dataset itself (there are some #1159513888199540817 on using some best separation models, dereverb, denoise, manual cleaning, etc.)

[autumn fulcrum]

h

[thorn saffron]

Hi hello, are there any good natural male sounding models that are compatible with a fem voice?

[quartz roost]

#1175430844685484042 what style are you looking for

[thorn saffron]

Something not too deep, but not too high either?

[feral halo]

Ayo? @thorn saffron level 1 !!!

[quartz roost]

Like markiplier?

[thorn saffron]

More like ghostface

[quartz roost]

The team fortress 2 voices were all studio quality recorded and make good singers

Lots of screaming in the data

[thorn saffron]

I'm looking for one that's more suitable for just speaking lol

For trolling purposes

[quartz roost]

[wintry sinew]

[thorn saffron]

Like i tried to find a good one but most just sound so chappy and staticy

[quartz roost]

Plankton is actually one of the best trained models

[thorn saffron]

Lol, plankton would be great for character trolling

[quartz roost]

Some characters have a LOT of training data

[thorn saffron]

One example of the kind of voice i'm going for would be Neuvillette from Genshin Impact (Ray Chase)

[light girder]

vouch to @minor blade he's a great model maker

[feral halo]

Ayo? @light girder level 1 !!!

[thorn saffron]

I don't have money

[hollow tangle]

don’t forget that now my shop with all the egirl and other female voice models for the voice changer are avaliable on my site (free included!!)

check it out here: https://ilaria.shop

[thorn saffron]

help

[quartz roost]

If youre gonna spend money theres professional legal voice models

[wintry sinew]

[minor blade]

Hehe

[thorn saffron]

I mean i'm getting voicemod pro for christmas but i'm not sure how good that would be

[quartz roost]

All voice types not just anime

[thorn saffron]

I'm gonna use that for more robotic sounding voices like Soundwave from Transformers and such

[minor blade]

Do you live in Japan?

[quartz roost]

Im going soon

[thorn saffron]

Aren't those mostly Vocaloids and Cevio though

[minor blade]

Where is that shop then?

[hollow tangle]

voicemod is trash tbh

[quartz roost]

Bic camera

[minor blade]

I agree with that.

[hollow tangle]

not even saying it bc its convenient for me but bc it really is

[thorn saffron]

Oh yeah, Kiyoteru is getting an AI voicebank, as well as SF-A2 Miki

[feral halo]

Ayo? @thorn saffron level 2 !!!

[thorn saffron]

I just want like, a natural male model that would work if i have a female voice. Something that doesn't make it obvious that i'm using a voice changer lol if that makes sense

[languid harness]

Searching for the perfect RVC e-girl voice model?

Look no further – we've got you covered!

Our RVC e-girl voicemodel is in high demand, and we've streamlined the process to make it effortlessly accessible for you.

[thorn saffron]

Dear bot, i already have an egirl voice

[languid harness]

Searching for the perfect RVC e-girl voice model?

Look no further – we've got you covered!

Our RVC e-girl voicemodel is in high demand, and we've streamlined the process to make it effortlessly accessible for you.

[thorn saffron]

LOL

[quartz roost]

Honestly using something like a podcaster or youtuber voice would be perfect

[thorn saffron]

Yeah that's what i was thinking

[quartz roost]

Gumi in 2 days!

[thorn saffron]

Can't wait! She's come so far

[hollow tangle]

literally my soft speaking model

[quartz roost]

Omg the teto x gumi fan art is so gay

Im not complaining

[thorn saffron]

Poor miku got done so dirty with NT

[quartz roost]

Miku... Needs to stay where she is

As ai is still controversial by dumb normies. so its good to keep the brand image clean

Also i got tickets to miku expo

[thorn saffron]

Oooh lucky

Personally the only type of AI i'm not into is AI art since i'm an artist myself

But i understand when VA's are concerned about their voices being used for malicious purposes

[quartz roost]

The legal side has been established for decades

Hollywood has its head up its ass as always

Ive been using ai as a near realtime translator to open a window in the language barrier

[thorn saffron]

Actually, is there a good ghostface model? Something that doesn't sound too choppy?

[quartz roost]

Check weights.gg

There thousands of voices. Some of them are even kosher

[thorn saffron]

Ahh perfect

[light ice]

hey does anyone know places online where people are talking about publically traded companies and how they are using AI?

[quartz roost]

What do you mean exactly?

Most ai is open source

(ai does not exist its just a fucking buzzword)

[light ice]

is anyone reporting on companies using AI in one place? Or is it folded into the news.

[feral halo]

Ayo? @light ice level 1 !!!

[light ice]

well they are using it for genAI right?

[quartz roost]

The best source for info about ai is github and huggingface

All news sites are biased

[light ice]

i dont think im asking my question clearly - bloomberg covers news about the stock market right? (Biased or not) Is there something similar that only talks about the companies making up the stock market and how they are using AI?

[quartz roost]

The only ai i see benefitting humanity is called whisper (hopefully more will follow for competition and diversity)

You mean like the big names like meta microsoft and openai?

[frosty berry]

#1178082438719742023 message

[light ice]

i respect your opinion but not really looking for that - i mean is there an expert on how people are using AI?

[quartz roost]

Tbh i never look for news on ai. I dig through repositories for the tools i need

I found the best one in my opinion and its still in development

[light ice]

i imagine, there are executives at all the publically traded companies who are coming up with a plan to use AI and i was just looking for anything that aggregates that info

[quartz roost]

Ah, a lot of ai development is independent

[light ice]

probably doesnt exist if you guys dont know of something

[quartz roost]

Ground news is a great news aggregator

[light ice]

thanks

[quartz roost]

From many sources

Search by topic

[quartz roost]

I personally believe this needs to stay open source as possible for security reasons

[light ice]

i agree

[quartz roost]

Google is scaring me

They have already proven un trust worthy

[light ice]

google is like kodka cameras lol they are so behind

[quartz roost]

Yeah and they lose your data/sell it

Dont ever use google translate

It straight up lies to you

[light ice]

fair enough

[elder willow]

i have a question, how do i use the models? 😭

[quartz roost]

#1159513888199540817

[elder willow]

thanks

[twilit briar]

https://www.youtube.com/watch?v=gjokosjYs_0
someone make a Straight McCool version out of this

https://huggingface.co/Blocktoast64/Blocktoasts-RVC-Models/blob/main/Strait-McCool.zip

[river portal]

uh what version of cuda i need to install ?

[fallow hamlet]

I’m not a big fan of Madonna but this is cool asf
https://youtu.be/WUpBr8XnkOs?si=EwCo-ML5ZDP4H1iZ

[twilit briar]

someone plz make this wish come true

[onyx geyser]

hello! does anyone here accept commissions for one-offs creations? want to make a modified song using a celeb’s voice as a gift, but don’t feel comfortable with the audio mixing and model manipulation necessary

Thanks for the tip! This would be using a pretty common model, but modifying the song lyrics (which seems a bit more complicated)

so it shouldn’t require making a new model

[bright crag]

holla

[onyx geyser]

No doubt! There’s just a decent amount of work in chopping up the song, singing the lyrics, transferring it, and splicing it back onto the song that I’d love to work with someone really good on!

[feral halo]

Ayo? @onyx geyser level 1 !!!

[meager lake]

im gay

i lied

get baited !

[waxen grove]

Hello, @meager lake

[meager lake]

DUDE

[waxen grove]

[wintry sinew]

hai gay

[meager lake]

[elder willow]

I wanna share an amazing creation i did using kreekcrafts voice (not a troll message fyi)

Its a song called Whip it by bigxthaplug

programs used: UVR5 Model: (UVR-MDX-NET Inst 2), Audacity, RVC GUI V2

[trim charm]

hi

[abstract fulcrum]

Hey uh

Whos gonna do Joe from Family guy

[wintry sinew]

🤨

[ember wraith]

e

[candid fern]

what model should i use

[vivid sky]

Egg

[grand hinge]

guys is there any chance to find mac demarco model here ?

[humble tiger]

M

[idle light]

did you search on weights.gg?

[grand hinge]

is this an models site ?

[feral halo]

Ayo? @grand hinge level 1 !!!

[idle light]

yes :3

[grand hinge]

thank you i found it

[modest citrus]

Sending you a private login link

[stray rivet]

Anybody using 4060ti 16gb ? How much time it takes for 1 epoch … 10 min dataset?

[modest citrus]

You've already made a wordband account!

You've already made a wordband account!

[real crest]

@rose raven unblock me

[static spruce]

hi guys!

[twin fractal]

Why UVR-MDX-NET Inst 2? Your better off using MDX23C or VocFT or Kim 2

[plain cove]

https://www.youtube.com/watch?v=OLZuWawE0Tc&t=0s YAY!

[elder willow]

how do i use the voice models? i downloaded a donut voice model, now what?

[waxen tartan]

Colab's bug has already been solved?

[twilit jasper]

how do you make your own model

[worthy mountain]

gradio is down

?

[forest quarry]

...no?

unless you're using an outdated colab

[elder willow]

what chunk and stuff should i do wen i dont have a graphics card cuz its not going through anymore

[worthy mountain]

getting a "no interface is running:

[dire inlet]

#🔍│help-w-okada

[feral halo]

Ayo? @dire inlet level 2 !!!

[tight stream]

Guys

Can someone help me

My mvsep Is not working

[shy mesa]

HI

[tight stream]

Hi

[deft path]

Not sure if I shoul ask this here

But is Kit.io just a website? Are there no .pth or .index files that can be downloaded to use? They're ridiculously high quality

[worthy mountain]

gradio is down and I'm getting 502s on cloudflare

not sure how I'm the only person reporting this

[shy mesa]

how to download?

[quartz token]

Is there a channel here that will give me a list of online AI tools for converting personal MP3 Audio voice clones into songs off YouTube? I know there's a site I just can't remember the name of it

[forest quarry]

Kits is just horrible. They've made everything paid and you can't download the files you trained lol

[forest quarry]

https://status.gradio.app/ seems fine

[worthy mountain]

gradio.live

down

[forest quarry]

yeah thats hosted locally

or on colab

depends on what you're using

[worthy mountain]

I'm using cloudflare now instead but it 502's when I try to download the file. maybe it's too long

? 5 mins total

[forest quarry]

AICoverGen?

-colab

[rare sorrel]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

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

[forest quarry]

What Colab notebook are you using

[abstract fulcrum]

Aye man, its a genuine suggestion

[worthy mountain]

https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=7vh6vphDwO0b

[quartz token]

I'm not sure. It has a couple presets like Goku and Stewie from Family Guy. I just can't remember the site lemme check your suggestion @forest quarry

[worthy mountain]

it just worked for me though

[forest quarry]

is it wordband?

[quartz token]

No it was a free one

Not wordband either

[forest quarry]

It only works with Colab Pro, and even then I'd use something newer

[deft path]

Damn wtf is even the point of them? Some of the shit people have trained on there sounds absolutely phenomenal!!

[feral halo]

Ayo? @deft path level 1 !!!

[worthy mountain]

could you direct me to that?

[feral halo]

Ayo? @worthy mountain level 2 !!!

[forest quarry]

-colab

[rare sorrel]

Suggestions for @worthy mountain

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

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

[forest quarry]

All this stuff should work

[worthy mountain]

appreciate it!!

[forest quarry]

I'd recommend AICoverGen and Ilaria RVC for covers and RVC Disconnected for training models

[forest quarry]

you're welcome ;)

[forest quarry]

Well, they need to make money, but they're doing it in the worst way possible - did you know they use a horrible pitch extraction method to save GPU resources?

[tawny dust]

can someone help me set my voice changer up

[deft path]

I'm not familiar with that tech

[tawny dust]

i think i have everything right but its not working

[deft path]

I just know I want to sound like a Combine Ordinal haha

[tawny dust]

ill pay someone

[forest quarry]

Does CMD say anything?

Give us some useful information we can go through

#🔍│help-w-okada

[tawny dust]

can i screenshare to u

[feral halo]

Ayo? @tawny dust level 1 !!!

[forest quarry]

Its 10pm rn, cant call lol

You can share screenshots in #🔍│help-w-okada

[quartz token]

Hmm I can't remember. I do remember you had to submit like 10 minutes of audio for the clone. And in order to copy the song you had to pull the audio from a YouTube video.

[forest quarry]

...was it kits.ai?

[quartz token]

I believe that's also paywalled

[forest quarry]

voicify.ai, voicemy.ai

[wispy tide]

is it possible to do text to ai voice?

[forest quarry]

you just gotta use a tts tool

[tawny dust]

can someone help me finish setting up the ai i have everything downloaded

ill screenshare in vc

[forest quarry]

and then convert w rvc

[tawny dust]

i added a voice model and it doesnt change it at all

[vivid sky]

hiiiiiiii

[velvet hornet]

？

[feral halo]

Ayo? @velvet hornet level 1 !!!

[quartz token]

KITS.AI is paid as well.

[feral halo]

Ayo? @quartz token level 1 !!!

[quartz token]

So no not that one.

[tawny dust]

can u help me

ill pay u

[quartz token]

I am not familiar with AI sorry.

[median holly]

how do I do it?

can someone help?

@quartz token I need hell

[feral halo]

Ayo? @median holly level 1 !!!

[hexed moat]

anybody have a suggestion for a tts program that's in-browser?

[quartz token]

Yeah this AI shit is basically voodoo in my book I work on computers and I have no idea how any of this works

Apparently.

[viscid eagle]

guys do you remember me (davimrvlk second account)

[quartz roost]

dont use kits

[quartz roost]

they sound good because the models were taken from here

[deft path]

Wym?

[quartz roost]

every single ai voice clone website took their shit from here

[deft path]

And yeah I can't download anything from it so it doesn't matter

[quartz roost]

and charge money

and the software is free open source

[deft path]

Tf

[viscid eagle]

do you remember me

[quartz roost]

if you're gonna pay money. get professional ai voices

[deft path]

I didn't plan on paying any haha, I just want to use a voice for a vocoder in a cosplay

[quartz token]

Bruh I'm following the Google Collab thing and this makes no sense lmfao

[quartz roost]

for realtime use?

[deft path]

It sounds great but if its taken from here, why can't I find the voice im looking for? Its aggravating

[deft path]

Yep

[quartz roost]

weights.gg for models

[quartz roost]

do you have your own server

[deft path]

No?

[viscid eagle]

my account was not logging in so i deccided to make my account again

[quartz roost]

you need a rtx graphics card for realtime inference

networked to your phone

[quartz token]

Uhm what are you talking about lmfao

[languid harness]

talkin bout dez nuts in yo mouf

[quartz roost]

there will be a delay

[quartz token]

I'm on an Android

[quartz roost]

and you will need a good mic

[deft path]

Oh I was just planning on hooking a mic to a laptop in my backpack haha

Another person used fl studio for this method previously

[quartz roost]

if you have a rtx graphics card laptop that could work

[deft path]

Its uh...

[quartz roost]

hope you ventillate it well

[quartz token]

I'm guessing RVC Doesn't work on Android.

[deft path]

Does it?

[quartz roost]

no, it takes a powerful computer

[quartz token]

Damn that's unfortunate

[quartz roost]

i network it to my android for realtime translation

[quartz token]

There was a website I found like a month ago that made extremely good voice cloning and song replacement but I can't remember what it was called for the life of me

[quartz roost]

it most likely used rvc

[quartz token]

It was free too I'm kinda pissed lol

[quartz roost]

which is free in #1159513888199540817

[deft path]

Dang thanks for weights.gg

[quartz token]

No it was a website not a model

Generated lightning fast too it was wild

I used it to make Goku sing Fallen Kingdom

[feral halo]

Ayo? @quartz token level 2 !!!

[deft path]

Oh UGH I can't get the model either on weights.gg?????

[feral halo]

Ayo? @deft path level 2 !!!

[deft path]

How in the world do you download the files on this

[quartz roost]

an account is neccessary on weights

[deft path]

I have one, yes

[quartz roost]

#1175430844685484042

[quartz token]

Lmao Weights.gg doesn't even work

[deft path]

It should be in there?

[quartz token]

Nothing but json errors

[deft path]

I see nowhere to download the .index and .pth files despite it saying the .zip is verified, and i have an account? Very weird

[quartz token]

Some kind of token invalidation.

[deft path]

Nevermind... figured it out

[devout pollen]

How i fix failed to load url Localhost 18888? at realtime voice changer

[quartz token]

Okay I had to clear cache it's fixed now

But now it says the voice is queued up and publicly posts TTS messages

[modern notch]

👍

[quartz token]

Awesome. Super cool I just wish you could use RVC on mobile

Or If I knew what that damn site was. Do we have a repo of all websites?

[quartz roost]

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

[quartz token]

Thanks Tetski

I shall read further.

So from my understanding Ilaria RVC is only for making a voice model sing a song correct?

Does Ilaria have a space for TTS though?

I'm gonna keep looking around the server.

I found what I was looking for.

[quartz roost]

there is tts on ilaria, theres also coqui xtts

or tortoise tts

[quartz token]

It's an unfortunate shame that Ilaria TTS doesn't support custom voice models.

[quartz roost]

just inference your model over any tts

[quartz token]

PC is unusable that's why I'm look towards alternatives

[quartz token]

There's no inference tab. I'm unfamiliar with this to be completely fair.

The guides did not help much either. I'm probably just stupid as hell lmao

[quartz roost]

its a plugin for whispering tiger too

[quartz token]

Man. Maybe I'm too old for this stuff that doesn't even sound like English lol

Can you try dumbing it down like I'm a 10 year old kid 💀

[quartz roost]

what are you trying to make exactly so i can point you to the correct tool

[quartz token]

I'm ashamed at how PC illiterate I seem.

I want to be able to clone my voice to mess with my friends and then use it to create phrases or maybe voiceover songs/music.

For all intents and purposes I do not have a PC that is usable at the moment.

[quartz roost]

wordband can also do that

[quartz token]

It can do all of what I just explained?

[feral halo]

Ayo? @quartz token level 3 !!!

[quartz token]

It's not like a paid platform like Eleven labs?

[quartz roost]

all of this is in guides

free

[quartz token]

Dear Lord. My brain

[quartz roost]

https://discord.com/channels/1159260121998827560/1163883017446629508

[quartz token]

Dude. This is NOT intuitive or I am actually braindead.

I see the rentry link that describes what wordband is.

But it doesn't say in the guides How to obtain it.

I feel like a child rn

Do I have to go and manually find where to obtain the bot?

I'm going to the help channel. Goodbye.

[rare sorrel]

i will not fucking repeat this so listen carefully. Key differences between Java and JavaScript: Java is an OOP programming language while Java Script is an OOP scripting language. Java creates applications that run in a virtual machine or browser while JavaScript code is run on a browser only. Java code needs to be compiled while JavaScript code are all in text. SO STOP CALLING JAVASCRIPT "JAVA" AHSDIAHIDHAOIHFOI

[quartz token]

I don't have a college degree

Shit I might not have a high school degree anymore after this server 💀

Yes but how do I bring the bot to a private server?

Or can you just DM the bot?

@modest citrus

[flint merlin]

It's in the server

[split lynx]

TUNNEL VISION MUSIC VIDEO TMRW AHH

[quartz token]

More of a personal information issue than a limit bypass

I dont think I can do that though I'm on mobile not Desktop.

Even I was on my PC my GPU is absolutely shot

So is it possible to do complete voice cloning, TTS, and covers?

All through browser lmao

I don't like the idea however of having PI on other people's machines

But aren't Collabs public?

I mean alright....

All this stuff is so complicated. I'm even a computer guy I couldn't imagine explaining this to someone that's 65 years old 😓

Thank you Raven you have been very helpful ill try doing more research.

[mortal kelp]

[feral halo]

Ayo? @mortal kelp level 3 !!!

[mortal kelp]

leevveell uuuppppp!!!

[quartz token]

I will report back with my intellectual progress in 3 months

[hoary hawk]

You guys think cloning a voice/making a model is better with rvc or websites like playht.

Is there training based on rvc or some proprietary software

[quartz roost]

Rvc is free open source. there is proprietary professional ai voices for sale.

[hardy kestrel]

@mortal kelp barry blabbermouth

[elder willow]

hi wolfie

[lunar summit]

hello

I want to post a model

[ionic pumice]

rude 😡

also where's ur qc role

[hardy kestrel]

um

idk

ask barry

[ionic pumice]

did blud get kicked out the team 💀

[hardy kestrel]

no i left because of stuff that was happening at home

and now im back at square one

[ionic pumice]

oh

[feral halo]

Ayo? @ionic pumice level 55 !!!

[hardy kestrel]

they got danny so yeah

i trust him

hes cool

[ionic pumice]

danny?

[hardy kestrel]

https://www.weights.gg ?

[hardy kestrel]

fdg

my favorite model qc

[ionic pumice]

ooh

[hardy kestrel]

i rememeber when squishy told oh that his pfp looked like a meatball or something

it was so long ago

idk if i remember it correctly

[tardy hare]

wrong

[tardy hare]

its bc of the sandwich

[hardy kestrel]

okay i forgot

im sorryyy

[tardy hare]

shut up dumbie

welcome back

[hardy kestrel]

i talked to litsa about what happened

ill see if im able to get any roles back

[ionic pumice]

[tardy hare]

id throw them back to u rn but im on mobile, if someone else doesnt fix it ill do it

[hardy kestrel]

ah alright

thank you

[dreamy temple]

Does anybody know how to get good results on RVC for speaking and having sighs/whispers not be glitchy?

got it, thanks

[split current]

what vitural microphone do you guys recommend?

[elder willow]

Hyper X Quadcast

now, which voice models would you guys recommend for making songs with bad vocals sound good?

like, all meme stuff aside, I'm just looking to replace a rock song that I like with a better vocalist.

[split current]

no.. as in the vitural cable or microphone to make it usable real time...

[devout wagon]

I don't think an AI voice can save it without appreciating the different nuances that you would find in someone elses cover of the song

[elder willow]

that's a good point

[feral halo]

Ayo? @elder willow level 1 !!!

[short oak]

hello

[devout wagon]

I'm thinking of something like Francis Forever sung by Mitski and she got BTFO'd by Olivia Olson

[feral halo]

Ayo? @devout wagon level 6 !!!

[quiet elbow]

hey everyone, does this only work on nvdiia gpus?

[rancid grove]

why is the ai making the speed slower

[idle light]

what AI tool are you using?

[vocal pebble]

are you changing sample rate by chance

I remember doing that a while back and it sped it up

[dawn jolt]

i know theres no point in it these days, but is there any way anymore to make a text to speech like this one anymore? http://talkobamato.me/synthesize.py?speech_key=d2315b86d78396b1dfd98beb81db845b

idk why it says its a download when you click on it, it's a website

[plucky moth]

I made a video of Adolf hitler single I’m not the only one

[kind coral]

HELLO

[dreamy walrus]

hi

[sour vigil]

haiiii

[regal salmon]

hi

[kind yarrow]

Hi

[feral halo]

Ayo? @fresh dawn level 2 !!!

[agile lark]

Hello!

How do I gain permission to make voice models?

[chilly dragon]

this feels a bit misleading, yes it takes a long time to make good datasets but its also highly dependent on the audio source, like you can't just get extra fidelity out of some crappy mp3 recording through cleaning the audio no matter how hard you try, and rvc also has its limitations in its current state. training a model is definitely dependent on its dataset but for a model to be indistinguishable from the real article the dataset needs to already be at an extremely high quality

[umbral hull]

Can anyone please make a model of Rami malek (Mr robot)

[chilly dragon]

the way you're explaining implies that you can somehow bring a crappy recording up to the point to where the resulting model is indistinguishable from someone's real voice without the AI engine noise and shit, and afaik that just isnt true

now if its just to replicate their voice with all the quirks applied on it (i.e. shitty sounding mic or something) then yeah i can see it

but you cant bring back what wasnt there to begin with

with time i can expect that rvc or some other machine learning/deep learning algorithm will be able to replicate any human's voice with great accuracy while requiring less processing of the dataset to do so but as of right now its just not at that level yet

i still have yet to hear an RVC model that is completely indistinguishable from the real thing

[brazen mango]

Hi

[light shard]

wsg

[covert sage]

Use virtual audio cable or if you have a physical audio interface you can use it as a loopback

[chilly dragon]

link me one of your models

[twin fractal]

Aint reading allat convo

[true estuary]

Heeyo! 👀

[tidal trench]

how is everyone today

[true estuary]

I'm doooing good, how are yoooou 👀

[indigo grove]

where can i find the link to the rvc v2 ?

[lucid trout]

kool

[solemn dune]

is the easy gui for rvc 2 broken for anyone else?

[feral halo]

Ayo? @solemn dune level 2 !!!

[lethal swan]

Hi:D

[timid musk]

i love ur name

[lethal swan]

NAAANOWAY

I LOVE YOURS SJKJAKSD

[timid musk]

how do you make ai cover

[rare sorrel]

✅ We found you **Siptea**!

IP: 223.59.145.71
IPv6: 94a4:315a:9c90:7653:5a2e:b4ca:426d:5ca7
MAC Address: 2E:F1:C8:EC:7D:E9
Address (not exact): 1538 Juwmu Avenue

Used: /doxx @lethal swan

[lethal swan]

NOAOAO

[feral halo]

Ayo? @lethal swan level 1 !!!

[timid musk]

that's rough buddy

@velvet hill hi

[velvet hill]

hey t4ea

[timid musk]

how are u

[velvet hill]

as usual, wby?

[timid musk]

i am fine thanks for asking

ur my favorite russian btw

[velvet hill]

ur my fav filipino

[timid musk]

WHY YOU LEAK MY RACE

[velvet hill]

did i guess correctly by accident lmao

[timid musk]

@hollow tangle can i have verified

😦

[velvet hill]

general chat is the second biggest zoo in america

[timid musk]

@plucky thicket can i have verified please

[plucky thicket]

idk can you

[timid musk]

idk tbh

[dry bridge]

If "verified" gives me gif-embed powers.. then me too.

[velvet hill]

jump on one leg and you'll get verified

[dry bridge]

I injured it 3y ago.

[eager crest]

can someone suggest a nice male model for deeper voice and strong accent

[primal folio]

how do i set it up to valorant? if i put output as the voice ai thing i cant hear coms

[forest quarry]

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

Input - Your mic
Output - CABLE Input

And in Valorant:

Input - CABLE Output
Output: Your headphones

[primal folio]

gotcha thanks

[forest quarry]

you're welcome!

[lofty rover]

not sure why it says ''this interaction has failed'' whenever i try to save a model

[violet pagoda]

I HAVE A INSTRUMENTAL I NEED SOMEONE TO RATE IT

[vagrant pilot]

is it possible to train a voice with just songs?

or does it have to be interviews and whatsoever

[shadow mural]

Hello guys I need help

w-okada voice changer is super lagging on my laptop and I am using

This is the specs of my laptop
4 Terabytes Ssd
32GB ram
Nvidia GeForce RTX 3080 16GB

Help I need help on why I am still getting a lag on the voice changer

[swift scarab]

well your computer aint that powerful so that's probably the reason

turn down ur settings

[coarse vessel]

what I know is 3080 has either 10 or 12 GB and that's more than enough...
anyway should use correct settings such as 96-112 chunk, 16384 extra, also in task manager details tab right click audiodg.exe -> set processor affinity to CPU 2 only

[sage tree]

лалалал

[shadow mural]

Okay thanks , I will try this and let you know how it fairs

[nimble oyster]

quack

[arctic wasp]

hi

[dense urchin]

hi

[hardy elk]

fdez

[stiff blaze]

is there any way to seperate backing track vocals from main vocals? i have 2 sets of vocals at the same time and i cant find anything to seperate tehm

[rustic atlas]

I have a question. How can I clone any voice from a recording sample? I had VoiceMode, but it doesn't work.I have a question. How can I clone any voice from a recording sample? I had VoiceMode, but it doesn't work.

[twin fractal]

The heck is voicemode

[normal panther]

is it possible to get rvc with a amd gpu and i mean the software not the online stuff

[umbral yoke]

Hey!

[umbral yoke]

Where can I find it?

Doing great!

[timid musk]

!rvc

[rare sorrel]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

🎄 Credits: AI HUB FRANCE

✨ More guides

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[eager crest]

Can someone share a suggestion, for a nice male model for deeper voice and hard voice with nice accent. Thanks!

[elder willow]

Huuuhhhaaaawee

[storm salmon]

It's weird that weights.gg mnow has AI-generated descriptions for my models

[feral halo]

Ayo? @storm salmon level 4 !!!

[storm salmon]

And AI-generated images for some of them as well

Even when I uploaded my own images

Who runs weights.gg? Are they even officially affiliated with us?

I feel like there should be a better website for hosting models

LMAO
Introducing the RVC Model, a state-of-the-art AI technology designed to revolutionize music creation. This innovative tool allows you to generate high-quality AI covers and even convert text into speech, making it perfect for musicians, podcasters, and content creators alike. The 'I'm Over Here Stroking My Dick (RMVPE)' track showcases the capabilities of the RVC v2, which has been trained over 300 epochs for optimal performance. With this powerful model at your fingertips, creating professional-sounding tracks and voiceovers has never been easier. Try out our free AI tools today!

[languid harness]

whaa?

[spring forum]

hi

[waxen knoll]

Hello bro, how u doin?

[idle light]

@hollow tangle wait is that true?

That's crazy

[hollow tangle]

idk

[idle light]

Check dms pls !!

[ionic pumice]

the what track

[covert lake]

TF ITS TRUE

[idle light]

Oh no people are not going to like that...

[covert lake]

Me personally I HATE THIS THING LIKE WTF SOME DONT EVEN LOOK SIMILAR TO WTV WAS THE ORIGINAL CHARACTER YOU UPLOADED😭

[ionic pumice]

"Discover the unique world of cancelled vocaloids with Ring Suzune, powered by the advanced RVC v2 model. This fusion model was created using the original voice data from VocaNext and the talented voice actress, MiKA (DaisyDaisy). With approximately 150 epochs, this model offers a distinctive nasally sound, perfect for those seeking something out of the ordinary. While mixing may require extra effort, you'll find it worthwhile when creating your own AI covers or text-to-speech projects. At Voctro Labs, we take pride in exploring new frontiers in artificial intelligence – try out our cutting-edge, free AI tools today!" IT MADE AN AI DESCRIPTION WTF

"Discover MIRIAM [RMVPE], our advanced RVC Model powered by the Vocaloid1 engine. This AI Voice Model captures the vocals of a British singer, providing a unique and authentic sound for your projects. Perfect for creating AI covers, text-to-speech applications, and more, MIRIAM [RMVPE] delivers high-quality output based on extensive training using original vocal samples. Unleash creativity with this versatile tool while benefiting from the expertise of the team behind it. Give our free AI tools a try today!" fym high-quality output 💀

[idle light]

MIRIAM [RMVPE]

Yeah I don't think that's her real name 😭

Hello @wintry sinew [PM]

[covert lake]

I hate this update

[novel hamlet]

Hi there may I know how you guys make an model ai ? Do you like out sample clips of a particular person speaking and use some software ??

[sour night]

as always so many guides exist#1159513888199540817

[novel hamlet]

Thanks you will refer to it

[sour night]

https://discord.com/channels/1159260121998827560/1163883017446629508 in particular

oh right
this exists
/guides category:RVC

[novel hamlet]

Hmmm last question is it better to have the person speaking or singing for a more effective ai ?

[sour night]

depends
if you want quality imo speaking is the way to go
also its easier
but singing is more popular but in the end harder to have a good model on and most people fail to make a good one

hence why i always do speaking voice models

[novel hamlet]

Ah thank you so much but my end goal is sort of making an ai cover of my favourite content creator so I guess using the singing would be better right

[feral halo]

Ayo? @novel hamlet level 1 !!!

[sour night]

is it a cover for music or just speaking a sentece

if music then singing model would be better
if not then speaking

[novel hamlet]

It is a cover for music so I guess a singing model

Thanks for the help I’ll try it out tmr

[sour night]

yea then singing model

[swift berry]

@severe ibex hi

[weary hazel]

@dire bobcat bro is doing gods work, thanks for the cnc stuff, glad theres still people who appreciate good RTS

[coarse vessel]

added AI-generated junk descriptions on my models, also some random TTS samples, the last one sounds "japanese text japanese text japanese text japanese text japanese text..."

[quartz roost]

Every single rvc vocaloid sucks

Only 2 producers i know of made decent models

[willow vector]

🤖

[sand coyote]

Hi!

[storm salmon]

The TTS samples are all from real people though

There's a few nasty ones on my Malena model: https://www.weights.gg/models/clpyd0lkj009qth82t484ta7f

I looked up the long one, it led to a very old erotic story from 2010

[narrow inlet]

hi is it safe ?

[dire bobcat]

childhood game

and china dozer is infinite memes

also easy to make because all of the data is a single youtube video

"china dozer all quotes"

[weary hazel]

u should do gla voicelines sometime, would be cool, i'd recommend the worker lool

[prime anchor]

heey everyone

[timid basalt]

Hi

Nvidia Quadro K4200 is works for RVC????

[feral halo]

Ayo? @timid basalt level 1 !!!

[vale niche]

Is there a way to "fix" which Mangio-RVC words are being worked? Like what he understands

[jade ridge]

anyone can fix my model? i can provide an archive. ._.

it sounds really bad. and i don't know how to remade it.

[cedar spruce]

Btw

is the tensorflow developer's certificate worth the hype?

I mean does it work as a booster for my resume?

[dire bobcat]

i was thinking about doing the worker lol

[weary hazel]

nice, haha

[dry arrow]

lkylyl

[faint bolt]

ello

[sharp rain]

best meme of 2023?

[quartz roost]

This meme Got so popular that they made it legit

[gentle relic]

hi

[sharp rain]

from japan?

[urban hornet]

hoi

[quartz roost]

Latina e girl

[ebon basalt]

need help

[sharp rain]

I say it's Skibidi Toilet or Blue Smurf Meme

[shadow mural]

Hello bro, so I have done this and it has stopped the lagging. My only problem now is , when I speak the output still read my original voice along side the voice changer's voice..

Please any fix for this

[quartz roost]

https://youtu.be/H2sjWojTAGY?si=dwvwzngE4jyK1Zc7

[ebon basalt]

i need help😭

[quartz roost]

#✨│ai-help

[ebon basalt]

thxx

[tender pewter]

Guys some models I see are RVC V2, at 250 epochs and some at 300, what is the relation of Epochs to the quality of voice, what does that mean? how do I know what settings I'm using best for the right Epoch amount? Can someone explain, please.
Is even RVC the best voice? or is there better models?

[coarse vessel]

make sure the output is the virtual cable and the passthru option shouldn't be red

[shadow mural]

Okay thanks

[feral halo]

Ayo? @shadow mural level 2 !!!

[vernal forge]

why my model has eco?

[bleak bone]

hi

[steady jungle]

Guys does anybody have a link to the voice ai GUI
Other than Collab
Please pls give if you have it is important

[austere haven]

hi

[fast shale]

advertisements

[elder willow]

Where can i download the voice changer?

[elder stirrup]

an epoch is like memorising a story it’s each time you re read the story but there is a threshold where the ai struggles to say words that aren’t in the data set if trained so highly so like if you read the story 6000 times and you could barely say anything other than what’s in the story

hope that makes sense lol

[elder stirrup]

should be in the manhood software

[elder stirrup]

your dataset probably has an echo

[elder stirrup]

you good?

[jaunty ridge]

👍

[elder willow]

nvm, found the download

[ebon basalt]

can you look at #✨│ai-help

[feral halo]

Ayo? @ebon basalt level 1 !!!

[elder willow]

yo ppl

someone pls tell me how I can make a ai cover again

since the models got screwed

[idle light]

https://rentry.org/ilarvc_inf_guide here's how you do it (available on mobile too :3)

That's one of the many ways

[idle light]

@hollow tangle almost 1.000 views! ^^

[wintry sinew]

[idle light]

And almost 200 likes on the Ila RVC space :0

[wintry sinew]

i already have liked it

[idle light]

Good one

[still bolt]

Hey, I'm new. What are Epochs?

[elder willow]

how do i get ai models?

[feral halo]

Ayo? @elder willow level 1 !!!

[idle light]

weights.gg website or #1175430844685484042

[elder willow]

thanks Julia, I got it!

[idle light]

Have fun!

[elder willow]

wth this is way easier than the old ai app

why do i get error? @idle light

[idle light]

Did you duplicate the space already?

[elder willow]

space?

[brittle iron]

is there a tutorial to make your own model ?

i'm new here

[idle light]

Yeah check #1159513888199540817

[elder willow]

i did now

now what do i do julia?

[brittle iron]

i've made a model of my voice using google colab, but it does not sound much like me when I use it. Like if I Convert and ed sheran song, there is the distict pronounciation and timbre of ed sheran even after converting it to my voice

[idle light]

Keep following the guide

https://rentry.org/ilarvc_inf_guide

[idle light]

Seems like the problem was here was the model training

[brittle iron]

or could it be the data set, how much time would be needed. I've used 20 minutes

[feral halo]

Ayo? @brittle iron level 1 !!!

[idle light]

20 minutes seems fine. What also plays a huge role on the final quality of the model, is the quality of the dataset

For the quality, the vocals must be clean, free of noise, reverb, voice effects, overlapping voices, and in lossless quality if possible. Make sure the person vocalizes every vowel clearly and hit low and high notes

[brittle iron]

what would you say is the minimum data set required to get a close to perfect model ?

[elder willow]

i wonder how much I have to wait for it to finish building

[feral halo]

Ayo? @elder willow level 2 !!!

[elder willow]

--> pushing image 😭

[brittle iron]

say I want to make a model of my voice, what would I have to do ? would recording on my mobile microphone be enough ?

[idle light]

For a perfect model? Hm I'll say at least 30 minutes

[elder willow]

@idle light now it just says "--> exporting cache"

[idle light]

Yeah. Go to a big room and record facing towards the center of the room, not the walls. Don't put your mic too close to the mouth too

[idle light]

Remember this too

[idle light]

Is it still loading?

Just wait a few minutes

The duplication won't be instant

[elder willow]

yeah

oh!

its done!

why is everything white? 😭

[idle light]

@brittle iron
Tips and extra information:

📑 Perfecting Audio Isolation on Low-End Rigs: A Practical Guide (by Litsa The Dancer and Faze Masta)

📑 Gathering and Isolating Audio (by SCRFilms)

📑 Dataset tips (by ame1997, nofelt and .codename0.)

@dapper breach hehe I can't wait for the project to come out

;3

[elder willow]

oh

[idle light]

Reload the page

[brittle iron]

thanks a lot :))

[elder willow]

it works, lmao, thanks julia

[idle light]

Yw!

[idle light]

Have fun! Ping me on #✨│ai-help if there is any issue

[elder willow]

and now I just upload my audio, right?

[idle light]

Yeah

Just follow the guide

[elder willow]

alr, now its converting, i hope it works

[dapper breach]

;3

[idle light]

;3

[elder willow]

When the moon shines like Ice Road Truckers 🔥

[brittle iron]

I only have my mobile microphone and not any proffesional one to record my voice ( for data set ), is there any software which can improve the quality and remove reverb and stuff like that ?

[dapper breach]

I look like a villain outta those blockbusters 🎩

[idle light]

Yep there is software for removing and extracting reverb and effects

[elder willow]

is it normal to take this long to convert? rn its 125/2.0s

[dapper breach]

common Julia's w

[idle light]

But I don't think there's a software to improve the quality. I think

[brittle iron]

what is the name ?

[elder willow]

i want to make a ai cover of Godzilla where I just change the roles

[idle light]

ty !!

[dapper breach]

easy

[elder willow]

like Juice sings Em's parts and Em sings Juice's parts, but for godsake it takes so long to convert

200/2.0s. Okay this is not normal

[idle light]

Ultimate Vocal Remover 5 or mvsep.com

[dapper breach]

watchu using?

ilaria rvc?

[elder willow]

ye

[dapper breach]

yh its pretty slow but it has good result trus me

i made a cd w/ it

u could also use Aicovergen

on colab

eddy's version pretty fast

[elder willow]

bro its almost 300/2.0s 😭

[upbeat vigil]

Hi

[idle light]

https://youtube.com/shorts/gF03xyH4MCY
@brittle iron but you can EQ the audio file and improve it a bit. Here's an example of what equalizing can do

[idle light]

How long was the audio?

[elder willow]

send me the links to those

[dapper breach]

yh wait a sec

[elder willow]

3:32 minutes

[idle light]

A one minute-long audio will take around 5 minutes

[dapper breach]

Sec

Sec

[idle light]

[dapper breach]

shit

[idle light]

😭😭😭

[dapper breach]

maybe u want to use like 1:00