🧬│ai-chat

[sly herald]

hey

[midnight aurora]

js put my new juice wrld ai in #1159290752195633273 🫡

[outer sierra]

How to put files into eleven labs?

It only gave me index and pth files

No audio files

[covert lake]

RVC models and 11labs are 2 different programs

U can't use RVC models in 11labs nor the other way around

RVC is Speech to Speech, 11labs is text to speech

[earnest dragon]

applio supports eleven labs as input for tts to rvc

[covert lake]

You sure? Applio TTS uses Microsoft Edge TTS API

https://docs.applio.org/getting-started/tts#disclaimer

I mean there's a plugin

But it's still not natively using them

It's just using an 11labs model to make a TTS file audio, then using it as an input in RVC, then converting that to the RVC model

[fresh aurora]

Hi, I generated a clean audio track in Audacity .flac I have a question how to extract .index and .pth from this file because I can't find it anywhere on the internet. Thanks for the answer

[viral sky]

Hi

[earnest dragon]

there is a addon that allows Elevenlabs support

[earnest dragon]

https://github.com/IAHispano/Applio-Plugins/releases/tag/1.0.3

[covert lake]

Yes I replied there's a plugin for that but it's still not native

And also seems like the guy wanted to use an RVC model in 11labs

[earnest dragon]

yes it is its linked on the applio docs under advanced and is made by applio creator

elevenlabs for applio tts

[earnest dragon]

either way still does the same thing

[magic hazel]

Hello

I'm trying to produce a song for a video I'm making.

[broken moth]

We got another one in the U.S just recently because of port people doing work strikes

[magic hazel]

I have two AI generated audio clips from Suno. I'm looking for a user who can help me copy the style of both music and vocals from one and create a special voice that I love with similiar yet different lyric.

Who can help me with this?

[vivid flicker]

If you manage to get me to 150 instagram followers before the end of the year ill help

[amber light]

dawg wants followers

lmao

[earnest dragon]

why not just do all with suno

[vivid flicker]

No shit lmao im getting desperate because my old account with like 300 got banned

[feral halo]

Ayo? @vivid flicker level 1 !!!

[magic hazel]

I'm trying to get the exact style of one clip

[earnest dragon]

no one is going to get you insta followers

[feral halo]

Ayo? @magic hazel level 1 !!!

[vernal abyss]

someone just ban this guy

[feral halo]

Ayo? @vernal abyss level 2 !!!

[earnest dragon]

then use suno to do it all

[vivid flicker]

Jkjk

Im sorry for using profanity

[magic hazel]

I tried and its not working and there's no way I'm buying premium on suno.

[amber light]

if you’re gonna promo do it in #1159290752195633273

[earnest dragon]

Nah I feel like he should get the Russian only chat punishment

[amber light]

not a bad idea

[vivid flicker]

Thats a good idea

[amber light]

https://tenor.com/view/ken-carson-goomba-ken-carson-goomba-gif-7954500978985618872

[vivid flicker]

I studied russian when i was in slovenia a couple years ago

I would be fine

[earnest dragon]

You saying that means you don't know how ai hub Russians are.

[vivid flicker]

Lmao its fine

[jagged wasp]

o-kada is just not opening

[amber light]

#🔍│help-w-okada

[earnest dragon]

which one the fork or the orginal

[vivid flicker]

Yall mfs get mad when i self promo lmao

[jagged wasp]

lol, my anti virus was on.

[vernal abyss]

you're fucking annoying and nobody cares about your instagram. what, do you think people just look at one channel forever and so you have to send that fucking garbage in every single channel?

how about zero?

[vivid flicker]

Damn ethan you got anger issues or something?

Discord mod looking ahh

[long jasper]

hello. does anyone have chat gbt plus and have multiple people using their account?

[earnest dragon]

what is the point of doing this here is the good question.

[long jasper]

and will i get banned if that happens

[earnest dragon]

why

[long jasper]

its in their rules that you cant do that

why does it matter

[vivid flicker]

The real question is if yall mfs were so ignorant and like yall say "dont care" why tf reply to me and keep this thing going?

[vernal abyss]

because it's a product for sale? and you'd be giving it to people who didn't pay for it? not like they're ever gonna be able to find out anyway

[long jasper]

i want to get chat gpt plus and have multiple people use my account

[vivid flicker]

Karen lmao, you the type of mf to not share your netflix password so you dont get banned

[long jasper]

but i dont want to get banned . i dont think they will ban me bc i will just call my bank and get my money back

[feral halo]

Ayo? @long jasper level 1 !!!

[nocturne basin]

how do i make money from AI songs?

[long jasper]

useless

[vernal abyss]

if you have to ask your product is shit

grifters not welcome

[amber light]

https://tenor.com/view/pffttt-cantholdlaugh-father-laugh-dad-laugh-gif-13166789396741539132

[jagged wasp]

what do i do if i can't hear myself talking with the voice changer?

[earnest dragon]

Because it's funny your responses

[earnest dragon]

Don't if it uses someone else's voice then you can't.

[earnest dragon]

I'm the person to use a service that has all movies and tvshows in highest quality for cheap sub and then share it with my whole family so it's 40 every 6 to 8 months instead of 20 or more a month

[earnest dragon]

You are the type of person that gets mad at people that pirate stuff.

[vernal abyss]

I know reading and context aren't your strong suit, but I was saying why it's against the terms of service, not defending it. I don't give two fucking shits what you do with your stuff

[long jasper]

bro u guys are sad. focus on making money or sum shit cuz u shouldnt let ppl waste ur time and emotiions like thia

cya

[earnest dragon]

who cares about terms of service if they really can't catch you.

[vernal abyss]

exactly, it's all fluff

[earnest dragon]

Yeah I do make money it's funny these people's responses

[vernal abyss]

it's like the AI companies trying to put conditions in their models to not train on their output - it's wishful thinking

[earnest dragon]

Why would you train a ai using ai that would just give a bad result.

[vernal abyss]

you can kind of supplement a large enough dataset with some synthetic data but yeah it's really pretty shit and not a good idea

[earnest dragon]

Not really

[vernal abyss]

https://arxiv.org/abs/2404.01413

doesn't embed, but there's been at least one paper on this subject that was p interesting

[earnest dragon]

Model collapse is possible like I see it happen with deepfakes and stuff when people train to long on not that good of data

[vernal abyss]

of course model collapse is possible, all i was talking about is that you can mix some synthetic data with a large enough real dataset before synthetic data becomes a fatal issue

[earnest dragon]

yeah

[polar flax]

do you mean mode collapse? in RVC context, it is indicated by a few downward spikes by about less than 30% than usual in loss/g graph. the possible causes are poor quality or improperly processed dataset.

[vernal abyss]

i'm not talking about mode collapse, no. just generally shitty results, and more so applied to generative models as a whole than specifically RVC

my results aren't bad either, was in relation to a mention of stupid ai company TOSes earlier

ty regardless tho

[polar flax]

mb if it sounds a bit ambiguous to me, anyway here's another thought: #1220364005034561628 message

the conclusion is to avoid including synthetic/augmented data, ideally

[earnest dragon]

No really ai is good for messing with images to make them better for training and audio also just not full ai currently.

[magic hazel]

Guys I have a really simple question, when editing music in external apps what's the best way to cut a large part of the song out without leaving a blantant split effect that everyone will notice.

[idle sandal]

Hello everyone, trying to learn w-okada here, does anyone have a microphone they suggest using. Preferably at a cheap price since I'm just getting started really. Any help would be appreciated, please and thank you.

[polar flax]

decent two-digit buck headsets are good enough if you don't care on audiophile grade quality

[idle sandal]

I mean... trying to learn the software here, and if possible download the resulted audio cause I wanted to use it for a story thing (Fanfiction really) so... does that work well enough (My headset doesn't have a microphone so)?

[polar flax]

I'm sure there are cheap headset mics you can find

[idle sandal]

welp...

delivery time is a month... but I did find so... thanks

[idle sandal]

Actually question, can you possible record yourself with the Ai voices and if so how to

[feral halo]

Ayo? @idle sandal level 1 !!!

[earnest dragon]

obs or audacity with w-okada

[polar flax]

convert the recording using non-realtime RVC

[idle sandal]

Mhmm, oki going to try Audacity, if I have questions can I ask it here?

[earnest dragon]

yeah that is the best way to do record voice then use non realtime.

[feral halo]

Ayo? @glass copper level 1 !!!

[marsh vessel]

so i clicked rvc and now i dont know what to do

am i cooked?

[polar flax]

Audacity or any audio editors (izotope RX, adobe audition, etc)

[south reef]

what was the website name with the old rvc models?

[ionic crystal]

Blavlavka

[ionic crystal]

Yes

[marsh vessel]

actually?

[placid pawn]

Asd

[earnest dragon]

what

[signal cairn]

-docs

[rare sorrel]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[signal cairn]

check the docs for tutorials everyone

[polar flax]

https://tenor.com/view/huh-cat-gif-26460616

[stable furnace]

Bas oui

Alias va faire kaka

[polar bobcat]

Hi

[stable furnace]

Ton pipi est tellement rouge qu’il est vert

[neat hare]

Hello dev bros. Having an interview in few days.

NEED HELP!

I have 3 years of experience in AI/ML. I have mostly worked with Classical Algorithms & Neural Networks (ViT's, CNN hybrids, EfficientNet, GNN, Pytorch, Capsule Networks, ONXX. etc). I will presently be interviewing for ML Research specialist role. The team is working on ML applications in drone camera's face detection and recognition.

[feral halo]

Ayo? @neat hare level 2 !!!

[covert lake]

Native means that it can actually use models instead of having to use an input file which is not the best as RVC is made for speech to speech not text to speech

[covert lake]

yea here its not about this, its about wokada & rvc, most people are are prolly underage lol

[covert lake]

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/ (login required)
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models
• https://voice-models.com/
• https://thevoicemodels.com/ (for Turkish Models, login required with discord and level 2 on their server)

if there isnt one, you can:

• #1159289738314919936
• #1191429836321849435
• make it yourself with our docs guides https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[earnest dragon]

Yeah but text to speech using Elevenlabs then that to the voice you want using rvc

[covert lake]

yea but not native as what the guy wanted to do

every promo out of #1159290752195633273 will get deleted.

im just saying its not the best in quality thats all compared to an actuall tts program

[earnest dragon]

yeah what the guy wanted would never happen you can use Elevenlabs to clone a voice but not rvc.

[covert lake]

Yea its 2 different programs

[earnest dragon]

Yeah I get that

[vernal sage]

👍

[icy pendant]

👍

[covert lake]

👍

[slow nexus]

Hi everyone,

I am working on the document classification((for the legal )for the large pdf having 100 pages or more .
in this large single pdf containing different -different types of legal document, .

How could classify the different -different documents from a single pdf?
Please suggest to me any approach you have.

[weak merlin]

sorry if this is a dumb question but what exactly does index do?

[covert lake]

its the file containing the accent of the voice model

[hidden vine]

I suppose OpenAI came out and crushed the competition for tts
I mean, it's exactly what Bark dreamt of being.
gpt-4o-realtime
a speech model that you can exactly prompt it the way you want it to say things

[mild ermine]

is this legit??

[covert lake]

i mean there are also good open source tts

[light storm]

hi

[unique mirage]

anyone know of a good deepfake server? If I'm not wrong deepfacelab is still produces the best results and I haven't found many tutorials on how to get it to work and the ones I have found I've had issues with.

[covert lake]

its all on their site btw https://www.deepfakevfx.com/downloads/deepfacelab/

if u want a more easier way theres facefusion which requires just an image and not actual training, but the results arent as good

[inner gorge]

e

[hidden vine]

I am sorry but none can come close to an actual voice actor that you can give directions to
which is where the new model is at at the moment

[earnest dragon]

My discord is the best deepfacelab discord

[earnest dragon]

Deepfakevfx is not good there facesets are not the best and there pretrained models are not that good and also there guide is bad.

[covert lake]

Thats the official deepfacelab site afaik

[covert lake]

i mean, gpt so vits is good quality tho

[earnest dragon]

it's not the official one that is just a website someone made for deepfacelab

[stray merlin]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

Don't forget to thank Ilaria if you find it useful! 💖

[flint yacht]

merhabalar

nasılsınız

fenerbahçe bu sene sampiyon olur mu

bence olur

sizce olur mu 2 level olmam lazım

hadi ama

evet

yok

[feral halo]

Ayo? @flint yacht level 1 !!!

[flint yacht]

olucaz 2 olucaz

az kaldı basaracam

hedeflerimi gerçeklestirmek icin buraya geldim

basaracam

hasan abi basaracam

annem basardı oglum diyecek

su an 2 level olmak icin tahminimce guney koreli cocuklarla konusuyorum

2 level olucam anne rahat ol

korkma basaracak cocugun

khontkar gibi sarkı yapıcam basardım ane

hadi artık bot lutfen 2 yap beni

isim var seni bekliyorum ayıp oluyor

5 saniye kuralı var bir de onu da bekliyorum

dostlar

sakanın sırası değil

arkadaslar 2 level olmak istiyorum artık lutfen

nolursunuz

arkadaslar el atın 2. olim

hadi artık gerçekten cok hastayım grip oldum

boğazlarım ağrıyor sizi bekliyorum yakısmıyor size

2 leveel olmak istiyorummmmmü

gel kalbime yatıa

[earnest dragon]

Stop doing wall of text not in english

[ionic crystal]

Ok

[real iron]

Guys what is wrong with my w okada voice changer it is not Woking properly so much noise even though there no sound

[dry rune]

.

[earnest dragon]

is your mic bad and picks up a bunch of noise

[real iron]

When i normally speak it works very good

[earnest dragon]

so when you are not talking it's bad but once you start you are saying it's fine

[polar flax]

enable Sup2 and raise threshold as needed

[real iron]

No like when i use mic normal way it sounds good but when i use okada it is going berserk

[cosmic geyser]

Selamlar

[earnest dragon]

yeah because background noise probably

[covert lake]

ah

[covert lake]

ask in help channels

ye just took care of it ty

[feral halo]

Ayo? @bold gazelle level 1 !!!

[idle urchin]

hey. i would like to know how i can train and make voice models.

[covert lake]

Train is making modela

What's ur PC gpu?

@hexed gust I deleted the promo post you did in #1192011222023950368 , use only #1159290752195633273

[idle urchin]

rtx 3060 12gb

[idle urchin]

oh okay

[covert lake]

Alr that's good for local (on ur PC)

[idle urchin]

how do i train a model?

[feral halo]

Ayo? @idle urchin level 1 !!!

[idle urchin]

i have no experience in training models.

[covert lake]

As you got a good PC, you can use RVC locally, you can choose between:

• Applio: A fork of RVC with some extra features like Applio TTS, same quality tho
• Mainline: The original RVC

[covert lake]

Check also https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[idle urchin]

alright

thanks!

5gb?!?!

what even is VoiceConversionWebUI

it will take me weeks to download that

[covert lake]

The program name

RVC = Retrieval-based-Voice-Conversion

[proven python]

-rl

[feral halo]

Ayo? @proven python level 2 !!!

[proven python]

-rt

[rare sorrel]

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

[old forge]

hello everyone

[silent condor]

does anyone know why that whenever i try to use the voice changer or just talk, i cannot hear anything come out?

[pale bison]

Why are my parents so on my ass for smoking, I'm 19 bro 😭

[mental vapor]

обезьяна

[wary olive]

Server de bosta

Fdp

Fuck

Fuckallamans

[tacit falcon]

do your parents also smoke

[pale bison]

My dad does but my mom quit

[tacit falcon]

idk where i was really going with it but sounds like they probably dont want you to be like them lol and also smoking isnt good for you but do whatever

[pale bison]

I'm not "starting" I used to smoke at 16, so... I'm just being open about it now, and not hiding it

[tacit falcon]

hope i dont ever smoke

cigarettes that is, i dont think i could care about something mild

[pale bison]

True. If you for some reason just want to try nicotine out, and are dead set on it, just use snus or chew tabacoo instead. It's thousand times healthier

Do not smoke or vape

I used to smoke like half a pack per day but now I'm back down to just 2. So im making progress at least

[tacit falcon]

my parents are chronic cigarette smokers

[feral halo]

Ayo? @tacit falcon level 3 !!!

[delicate vault]

LIke my comment if you are not gay

[analog cosmos]

[meager iron]

hello

[pale bison]

How did tyga ever live down "stimulated"

Rapper with 21 million monthly streamers making songs, bragging about having sex with underage girls while 25

[ionic pumice]

wgat

[pale bison]

It's as bad as it sounds

[ionic pumice]

[pale bison]

I feel like he should be investigated or something, idk

[ionic pumice]

wha

naw ts wild

[tepid basin]

Chief Keith

[pale bison]

Damn, shits kinda beautiful ngl

[analog cosmos]

real

[winter comet]

https://colab.research.google.com/github/jarredou/AudioSR-Colab-Fork/blob/main/AudioSR_Colab_Fork.ipynb

can anyone give a good explanation on how to use this

[feral halo]

Ayo? @winter comet level 1 !!!

[dense heart]

weed?

[fading star]

such a problem, the voicechanger processes the voices of everyone who speaks in the discord, how can this be fixed?

[elder willow]

oo

[pale bison]

Yes, it's resin afaik

https://youtu.be/8A0M2v_fEpI?si=sdyy2SCu6agIjjsZ

The Vicki leekx version is so much better than the one on matangi. Just hits much harder & is much more gritty

[ionic pumice]

why is miss chuyin weilai lowkey serving

[dapper drum]

Nah this one tho yes

[ionic pumice]

wgat

[agile plover]

Hello guys, I'm new to the server

Can someone help with this?

[pale bison]

<

[agile plover]

can someone me on this server?

help me

[earnest cypress]

anyone knows how to get amanda silvera model?

[ionic pumice]

AGSGSHQGAD

I HAVENT PLAYED 5HIS VN BEFORE

[glad nebula]

what vn is this?
asking for a friend

[ionic pumice]

https://echoesofarcane.itch.io/farbeyondtheworld

ive seen some of my friends play it lmao

[gentle canopy]

hi

[ionic pumice]

so far i've played like

3 vns rn

[analog cosmos]

:3

[pale bison]

Support the dev and get it from patreon

The itch io version is outdated by multiple years

[pale bison]

It's super, super, super great, play it

[bitter sapphire]

Recently, I built Paint Chatbot using RAG AI, next.js and Python
If anyone need my help, please ping me.

[pale bison]

And play yume 2kki too. It's not a furry vn but it's geniuenly the only game I enjoy playing

Game has no goal or objective or anxtgong, it's just about exploring and finding cool stuff

[ionic pumice]

i don't have a bank account to use on patreon yet lol

[pale bison]

Well, there's a certain site that I can't mention that's a archive of paywalled content

[left cedar]

Hey guys, does anyone here have experience with Replika AI?

[ionic pumice]

oh yea. "that"

[forest elbow]

how come the catbug voice model doesnt let you click on it to download?

[left cedar]

There's a catbug voice model now? Lol

[forest elbow]

well it wont work though

when i click it says invalid password or name

its old though so idk what happened

#1175588844054265858 message

[left cedar]

Ah gotcha. I'd assume either support for it ended or some code broke and the dev is working to fix it maybe? I'm unsure who the dev is/and the status of the app

[feral halo]

Ayo? @left cedar level 1 !!!

[forest elbow]

someone named fiction is the dev i believe

[feral halo]

Ayo? @forest elbow level 1 !!!

[left cedar]

They haven't been active here since early August this year so I'd assume the support probably ended for the bot

[forest elbow]

but then how am i supposed to become catbug

[left cedar]

You can't ;-;

[forest elbow]

thats not true and i wont believe it

[forest elbow]

if i make a model request, what are the odds someone who knows how to do it actually makes a catbug one?

[left cedar]

Probably a decent chance

I can't imagine it's hard to code

[analog cosmos]

if you want someone to do it for free there's a small chance someone might help (not guaranteed that someone will help tho)

[forest elbow]

i hope someone does the old one sounded so good

[analog cosmos]

that model is still on weights.gg if you want it

[cinder moat]

фуриебы ебанные

[forest elbow]

are you sure?

[feral halo]

Ayo? @forest elbow level 2 !!!

[analog cosmos]

yep
https://www.weights.gg/models/clml5dpbb003pws5a4rrqc8y7

[pale bison]

Pls make a ecco2k model

[analog cosmos]

no

[pale bison]

Ecco2k shook his ass at a concert onetime and everyone thought it was playboi carti for some reason

[heady zealot]

@cobalt oyster

Who are you?

[forest elbow]

thank you so much

[analog cosmos]

yw

iicr the old models were archived onto weights a while back, so some old models should still be on there (unless the creator removed them or it wasn't archived)

[left cedar]

Also you have a cool name :)

[forest elbow]

hey man thank you alot, yours is sick as well

[left cedar]

My actual name is Kafka, I just go by Kavade online :)

Would you like to be friends?

[forest elbow]

dude kafka is also a sick name, where is that name from?

[left cedar]

Czechia

[wise quiver]

KAFKA HAHAHA

im sorry bro but i find that hilarious

not because of the name itself

[left cedar]

What's wrong with it? Lol

[wise quiver]

there's nothing wrong with it per se

but there was a famous novelist/writer named kafka

[left cedar]

Franz Kafka

[wise quiver]

franz kafka

yes

[left cedar]

He was Austrian

I think

[wise quiver]

pretty sure he was born in prague actually

[left cedar]

Oh fr?

[feral halo]

Ayo? @left cedar level 2 !!!

[wise quiver]

yea

i only find it funny because in school, they had us analyse one of his most infamous novellas, called "metamorphosis" over and over

fever dream of a story

[left cedar]

I don't love that novel

[pale bison]

Omg carti reference

[left cedar]

It's..weird

[pale bison]

https://youtu.be/wGBsIelFe-E?si=wENC-bcwFnmaCUkz

[left cedar]

Carti is overrated

[pale bison]

Straight to my blocked list

[floral pasture]

[feral halo]

Ayo? @floral pasture level 1 !!!

[left cedar]

His only "decent" songs are Magnolia really

To each their own really tho

[floral pasture]

👋

[final rivet]

no? 😂

[left cedar]

That's my opinion lol

I just don't really like his style or voice all that much. He's definitely a talented rapper

I like NBA YB more tho

[final rivet]

okay thats fair

[feral halo]

Ayo? @final rivet level 1 !!!

[hot portal]

tell me the best female voice, so that it looks like a real one

[left cedar]

?

[left cedar]

Oh lord

[rare escarp]

omg, wut's wrong?

[rare escarp]

subjective af!

[lapis zodiac]

what is the best female model to download that sounds okay?

its for my mum

[vernal abyss]

you understand how ridiculous of a question this is right

[feral halo]

Ayo? @vernal abyss level 3 !!!

[rare cloak]

hi

[earnest pagoda]

How do you make the covers? I already got the model but I don't know how to do it

[left cedar]

Hi! How are you?

[hollow tangle]

my bio

[earnest dragon]

-hf ilaria zero

[rare sorrel]

Suggestions for @earnest pagoda

<:huggingface:1179800228946268270> Hugginface Spaces

• UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by thestingerx Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• 🆕 FaceFusion UI, by Nick088 Huggingface Spaces

More commands: -audio, - colabs, -kaggle, -local, -overtraining, -realtime, -rvc, -help

[earnest dragon]

and then weights.gg or this discord for models

[covert lake]

whats ur pc gpu ?

[arctic hill]

Server more fucked up than a whore's anus

[elder willow]

hi does anyone have any recs for websites with ai voices for yt voice overs? i've tried 11labs already, but yall know of other alternatives that dont require any payment

[covert lake]

There are different Text To Speech (TTS) AIs:

GPT So Vits: RVC isn't as good as GPT So Vits for tts, but gpt so vits (few shot tts, which means needs just a lil training for models) can't use rvc models (and viceversa), and its only limited to: english, chinese & japanese, if you wanna check gpt so vits instead, read https://docs.aihub.wtf/tts/gpt-sovits/

Freemium 11labs: A easy way to do TTS is https://elevenlabs.io/, you can't use RVC model on this but its a mostly premium easy way for good quality TTS

FishSpeech: FishSpeech is a 0 shot (no explicit training needed) TTS, if you got a good pc you can use it locally else use their site

With RVC Models:

RVC is natively for Speech To Speech, but forks such as ilaria rvc mainline & applio have built in tts (using Microsoft Edge TTS to make a generated tts audio, which i suggest you to choose a tts model that is the same gender and language of the rvc model you wanna use, and then convert it with rvc)

If you wanna do tts locally with RVC Voice Models (if you got a good pc):

• You can get Applio in our docs
• While Ilaria RVC Mainline here (no guide as of right now)

If you don't got a good pc you can do tts with RVC Voice Models on cloud:

• Ilaria RVC Zero (Running on A100 GPU, free fasted rvc on cloud) and the guide

• Use Applio UI Colab (with google colab T4 free daily limit gpu)

• if you don't wanna use edge tts, you could try another tts ai from our tts index and use the output as an input in rvc

[ionic heath]

-realtime

[rare sorrel]

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

[rigid moat]

No CUDA GPUs are available

[feral halo]

Ayo? @rigid moat level 1 !!!

[rigid moat]

on gradiowhat do I do to fix it?

[elder willow]

@gray rover sorry for the ping but are u porting your RVC fork to colab / lightning.ai someday? :>

[iron patrol]

Anyone have a good working colab link for rvc?

[feral halo]

Ayo? @iron patrol level 11 !!!

[covert lake]

training or inference?

[iron patrol]

Idk if my link is broken or not but it's for Illara

Just for making ai covers

[covert lake]

thats called inference

id suggest using ilaria rvc zero, a zerogpu huggingface space which is faster than google colab free gpu

[stray merlin]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

Don't forget to thank Ilaria if you find it useful! 💖

[covert lake]

if u really want to use google colab instead, use https://docs.aihub.wtf/rvc/cloud/applio-colab/

[iron patrol]

Is the huggingface one better then colab?

[covert lake]

yes, its a zerogpu (A100 gpu paid by ilaria, u wont pay anything and dont need to duplicate it) huggingface space, A100 is 11 times faster than google colab free gpu (T4)

[iron patrol]

Oh right on and is it regularly updated?

Cause the colabs iv used seem to stop working within a month or more

[covert lake]

yea its all working and updated

[iron patrol]

Right on thanks guys appreciate you both

[covert lake]

its normal, the rvc hype is kinda dead, and there can be somethings that stay out of date, also because google colab changes internal things sometimes and there are some old devs

[covert lake]

yw

[iron patrol]

Yeah and well I rarely do ai covers anymore just because my last youtube channel that was full of them got terminated for copyright so I'm doing ai covers rarely to avoid it

[covert lake]

yea unfortunately copyright doesnt allow ai covers

as u see there arent as many as it used to

[violet estuary]

: L'influence de la méritocratie sur la gouvernance institutionnelle en Tunisie : limites structurelles et réalités sociales

[left cedar]

Interesting

[analog cosmos]

very

[feral halo]

Ayo? @elder willow level 2 !!!

Ayo? @cunning ridge level 2 !!!

[gray rover]

don't have any plans

[elder willow]

okie dokie, i'll wait until someone does it :p

[ashen glacier]

Hello guys 👋

[crude tinsel]

I've been wondering if OpenAI does end up going bankrupt, will this have any major consequences or will that just leave a vacuum for shit like Grok & Gemini to fill?

[lean bough]

hi

[signal cairn]

asking for the impossible to happen

[scarlet spire]

Don't Click!|| if you are reading this, it is already too late. You have been infected by the curse of pee pee poo poo man. If you don't copy and paste this on 5 different servers, you will face the consequences. I was a victim like you, trying to be free.|| i warn you bro

[woven ether]

whats the best way to train song models and how do i make the data sets (they use alot of auto tone)

i usually use rvc

[midnight pasture]

are ckpts safe

[sterile cloak]

Hi everyone, I had an artist do a painting of me and my kids (single dad) as DC superheroes a few years ago.

I want to see if there's an ai tool I can use that can draw each of us, full body as DC heroes that I could then port into Photoshop and merge into one image? They have grown up now so id like a new family portrait of us as heroes

Any suggestions on how to tackle this without needing to subscribe to a service?

[amber light]

bro come on 😭 shits so corny

[tepid basin]

Don't act like you're not scared of pee pee poo poo man

[amber light]

https://tenor.com/view/spy-x-family-anya-anya-scared-scared-scary-gif-25559140

[hardy token]

ji

[sharp bridge]

Hola

[elfin ravine]

.

[smoky otter]

what's wrong?

im tired of staying this server

[humble minnow]

Do udio/Suno songs also get copystriked?

[iron patrol]

Iv made ai music songs with suno about random things and I haven't got a copyright strike

[earnest dragon]

No because they are allowed to be used I think for only the paid tier for suno but don't know about udio

[real spire]

what i need to setting my programm? rvmpe_onnx is best?

[latent ivy]

Hello so I don't know if you guys remember me like probably since 1 year I think

I was been hacked that year

Because it wasn't me I was doing something like this

[elder willow]

mommy vibes slayyyyy

[ionic pumice]

real

[cobalt temple]

Hello, new to the server. I just discovered the world of AI Generated Music, I normally use Suno and have created two EPs with my AI Band Zombie Dust, with lyrics written by me. Nice to meet you.

[fluid bobcat]

потные мужщины тягают железо

[sharp verge]

[pastel pawn]

So when training a voice model, what are some things to keep in mind? I've got about 15 mins of adr Inverview recordings from a specific actor. However if trained on lets say 500 epochs or 750 epochs, will the model be able to sound less flat or monotone which often is the case with many voice models?

[feral halo]

Ayo? @pastel pawn level 2 !!!

[covert lake]

More epochs dont mean more quality, epochs depends, use the tensorbaord

and be sure to have a good quality dataset

[pastel pawn]

hmm intersting i thought the more opochs the better the quality. Like i said i have about 15 mins of interview recordings, which is just the actor speaking into the same mic, same audio quality. All from one source which is how i prefer to work because i imagine that if you have a dataset which jumps around in quality its going to be very noticable in the final model

[covert lake]

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

[pastel pawn]

yeah just found the link myself, thanks for the guidance thought its really appreciated 😉

[covert lake]

yw

[elder willow]

https://tenor.com/view/whizzy-imposterfox-smash-anime-smash-anime-girl-gif-21682300

[polar flax]

https://tenor.com/view/anya-embed-fail-gif-25814121

[winter quest]

heloo

küfürbaz haydo sound vocal download

[covert lake]

Dont promo here, use #1159290752195633273

[inland jetty]

Hi, noob here what is good local ai image generator. looking for something working great with rtx 40 series gpu

[thick lake]

im curious about this too

[tacit falcon]

im about to play this shit

[covert lake]

Prolly automatic111 or comfyui

Not really a local user but that's what most ones use

[nimble garden]

do i need pytorch to install

im getting this permission denied on the cmd

oh wait i fixed it

i just moved it from program files to any other files

[tacit falcon]

the game in the picture you just sent

[rancid garnet]

hi this is haru nice to meet u all

[latent tundra]

I can finally hunt preds in vcs 🙏

[fresh aurora]

Hi everyone, do you know any AI tool that can process data from images or scans and extract specific information? OCR struggles with photos. It can handle scanned documents, but when it comes to photos, it can't read them.

[latent tundra]

Like a image to text?

[fresh aurora]

I'm specifically looking for a solution to recognize PDF documents that contain scans or images of ID cards, and I need to extract the PESEL number, date of birth, first name, and last name from them.

In my job, we have to do this manually, but such a bot would greatly streamline my work instead of searching through 20 pages of PDF for client data.

[willow harness]

do u guys know how to like remix songs with a ai voice of the original singer with different lyrics?

[covert lake]

You'd need to sing the song yourself, convert the voice, then apply the instrumentals

[quiet cobalt]

hello
I would like to know if it would be possible to use an AI to translate voices in games for free (in the form of an audio file or a YouTube link to translate the voices, not directly in the game) and recreate them in another language

[molten stone]

not accurate but a method that im thinking for that is grab the transcript and translate it, then using something like gpt sovits or elevenlabs maybe to make a realistic sounding translated audio, problem is, it wont match the mouth movement or length of talking

[ionic pumice]

i forget creators also put nsfw in there

[boreal badger]

hello

[ionic pumice]

mm true

[vagrant dawn]

ola

[covert lake]

Use #1159290752195633273

[vagrant dawn]

thanks

[icy breach]

Uati

[feral halo]

Ayo? @icy breach level 1 !!!

[halcyon meteor]

Hi, i'm new here nice to meet ya all

[lapis bobcat]

[ruby saddle]

hey

lots of members here!

[ionic pumice]

yee

not a whole lot talks though

[tacit falcon]

come for resources and go

all the people i remember talking with before on here left lol

or most of them more like

[distant isle]

hello

[covert lake]

There's #1159290752195633273

[vital root]

oh srry, didn't see it

[covert lake]

It's fine

@atomic vector is paid 69 cents an hour for that 🔥

[atomic vector]

lol that's not a bad idea realecco2k, but i think nick088 might be joking about the 69 cents an hour thing 😂 what do you guys think a fair rate would be for chatting on here?
-# AI-generated responses may be inaccurate; please verify important information.

[ionic pumice]

what the hell the applio bot can read from the replied message too

[covert lake]

Didn't know that either tbh

[covert lake]

Resources change tho, many colabs got deleted after time or either they are not updated anymore

[tacit falcon]

yeah but thats what i would think the people that join and dont talk much are here for; updates and new resources, but idk i havent talked in this server in a year i wouldnt know what they were doing

[covert lake]

Yea ig for models too

[ember dagger]

I remember there was alot of like diffrent voice models, where did they go?

[covert lake]

Server got deleted once for copyright so super old models may not be found here, they should be in weights tho

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/ (login required)
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models
• https://voice-models.com/
• https://thevoicemodels.com/ (for Turkish Models, login required with discord and level 2 on their server)

if there isnt one, you can:

• #1159289738314919936
• #1191429836321849435
• make it yourself with our docs guides https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[ember dagger]

Ah i see, thanks

[covert lake]

I sent u all links where u can find models

[ember dagger]

Yeah ill look into those, wanted to find few russian one, and supriced it was empty

[covert lake]

Yea

[gray bluff]

i forgot id joined this server tbh, lol

[covert lake]

😭

[gray bluff]

i checked back in recently tho, because im tryna make my own simple speech to speech voice converter from scratch

[viral jacinth]

are there any good tortoise-tts inference and training gradio web ui's available? everything i have seen has been abandoned, or is xtts better?

[gray bluff]

and part of me feels like im way in over my head

[covert lake]

so your RVC fork?

rvc is the program used for that which is the best

[gray bluff]

is that the fundamental algorithm behind voice conversion shit, like if someone wants to talk as snape from harry potter or something?

for the past month id been playing around with and learning audio signal processing and deep learning to pull this off

[covert lake]

xtts2 is better btw, its built on tortoise tts, but both of those are discontinued btw
If u care about quality, use gpt so vits (few shows, a bit of training) which needs actual training

If oyu care about speed (0 shot, no explicit trainng), use fishspeech or xtts2

[viral jacinth]

oh ok i have gpt-sovits installed, but i think it's a chinese version

[covert lake]

this is the program yea https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI

this is the original one

[gray bluff]

ahh ok

what i was gonna pull off is basically just implementing that from scratch

[covert lake]

maybe try https://docs.aihub.wtf/tts/gpt-sovits/ but idk of local things. Whats ur pc gpu btw?

[covert lake]

so like, re make it?

[viral jacinth]

4070 12GB

[feral halo]

Ayo? @viral jacinth level 1 !!!

[covert lake]

ig u mean a fork (modified version)

[covert lake]

ye good enough, u could try that guide

[gray bluff]

idk how exactly RVC works, like in terms of the step by step algorithm, so i was gonna reinvent the wheel for learning purposes essentially

and see if i wind up remaking something with the same fundamental functionality

but i have no clue how deep thats gonna go. all ik is that i need the software to take in vocal input, extract the audio features and feed them into a neural network that'll have been trained off of a lossless audio dataset from (insert fictional character here)

[viral jacinth]

i've stared at and tried so many voice cloninig prograsm the past few days, my poor hard drive is hating me right now

[covert lake]

you are looking for text to speech right?

just checking, cus if u want speech to speech its another program

[ionic crystal]

Thats not how it works

[viral jacinth]

want to do text to speech with models that i train

[covert lake]

then gpt so vits should be good

try to check the guide out

[viral jacinth]

what is the retrieval based voice conversion program for?

[ionic crystal]

@covert lake stop trying bro

[covert lake]

rvc, its for speech to speech

[covert lake]

?

[viral jacinth]

ah ok cool

Hey i've been doing this for 2 days, mainly did comfyui stuff. I appreciate any tips and advice

[gentle canopy]

hi

[mental vapor]

обезьяна

[sacred glade]

Anyone have knowledge about like to why w okada repeating itself when talking into applications like discord

[covert lake]

ask in help channels and be specific

[elder willow]

hi, can i ask you a question

[ancient pasture]

whats a good video tutorial to merge rvc models?

[dusty jacinth]

Guys who want to join owr clan

[elder willow]

Hi

I'm making a rvc model

I'm trying to clean the dataset as much as possible

[covert lake]

if its related to help, there's help channels for that

[regal crypt]

what do you guys think of my new model?

[magic dew]

hi guys

[novel nymph]

sounds great

[regal crypt]

does it sound believable

[novel nymph]

I think it does

[tacit falcon]

sounds good, im probably stupid for saying this but you'd probably have to change your speaking mannerisms a little for it to sound more convincing when converted

[pliant gust]

)

[light kiln]

Wtf is that bios 💀

[feral halo]

Ayo? @light kiln level 2 !!!

[tacit falcon]

alr

[regal crypt]

wdym

[brave badger]

.

[topaz jay]

yo

thanks for inviting me @ruby saddle

great server

[ruby saddle]

ye ik

[shrewd mountain]

hi im a new guy here

[heady lava]

ejjj mam prośbę bo kolega mi dał wyzwanie o suby pod jego kanałem sub nie musi być na długo ale bardzo by mi zależało tt: AnarchicznyBanan

[hoary geode]

Beam

[obtuse flax]

I tested out Just way you are by Bruno miles on 64kb mp3 and use AI to enhance the frequencies to make it 'lossless' it sounded really good

[digital heart]

👋🏽

[wind stratus]

omg which ai is this

[feral halo]

Ayo? @wind stratus level 9 !!!

[obtuse flax]

Mvsep AI enhancer

[tepid basin]

Bruno Miles sounds like a nightmare combination of Bruno Mars and Yuno Miles

Idk if that instrumental counts, but generally, no instrumentals in samples in chat

I think that's fine tho idk

[obtuse flax]

I like that instrumental because it's nostalgic

[obtuse flax]

There's some problems, But it removes the lossy effect

[obtuse flax]

Shit I got it mixed up

[graceful wren]

ale ty masz 2 iq bo pytasz sie ludzi co piszą po angielsku a ty piszesz po polsku jak mają to zrozumieć...

[cursive gazelle]

can ai do cold call in french

[unkempt bluff]

https://open.spotify.com/artist/5OKZ6HhHbUkyAgEUeIPWQa

[unkempt bluff]

ai?

[elder willow]

down talk here :_:

[fallen kraken]

sorry

i use my phone

[green panther]

guys

i cant search on find voice models

[elder willow]

Hello!
Web3 platform is expanding our team and looking for: developers, moderators, beta testers, analysts, and designers. The salary depends on the role.
You can write without experience. I'm waiting for you in DM.

[turbid yew]

Hey all, I have been working on a very intriguing and promising project for the past month, and I am looking for other AI-fellows to join and collaborate! The project is related to easy, use-case specific chatbots & virtual assistants using Retrieval-Augmented Generation (RAG). You can find more details on https://voiai.io .If you are interested, please DM me so I can explain the details of the project and how we can collaborate:)

[upbeat bone]

i need the best settings for W Okada newest release

[gentle perch]

hola 🙂

[timid mango]

-df

-hf

[rare sorrel]

<:huggingface:1179800228946268270> Hugginface Spaces

• UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by thestingerx Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• 🆕 FaceFusion UI, by Nick088 Huggingface Spaces

More commands: -audio, - colabs, -kaggle, -local, -overtraining, -realtime, -rvc, -help

[civic quarry]

Hello, I'm Anny, hope i'm welcome?

[idle sandal]

Ok general question, is an AMD gpu not compatible with W-Okada?

[timid musk]

you have to download the amd version

[idle sandal]

How do I know which version is the AMD version?

[timid musk]

Onyx version

[idle sandal]

Mhmm, oki thanks

[feral halo]

Ayo? @idle sandal level 2 !!!

[polar flax]

I'd suggest fork wokada that uses less cpu for your AMD gpu
https://rentry.co/forkvoicechangerguide#download-amd-intel-and-cpu

[spark agate]

anyone know how to fix this

[chilly lake]

it says how

[spark agate]

i already moved it to a dif folder

[hollow plume]

hi

[idle sandal]

Oki so, got this working, and honestly thank you very much for the help.

Issue now is just me talking too fast and my voice being a bit incompitble with Jp voices

[chilly lake]

directml is so bad performance-wise

[earnest dragon]

No it isn't what is the alternative there really isn't a alternative.

[idle sandal]

That and the voice I got was a bit choppy, so... should I talk slower? I assume that should help but honestly I'm unsure...

That and I wanna record audio with it which... I don't think it can do?

[chilly lake]

patch zluda into nvidia build

that's the alternative

[idle sandal]

I mean, I could try to redownload the AMD version, issue is well... I dunno which one that is,

Like my GPU is an AMD Radeon RX 6800M if that helps, but yea at least it's workin so it's better than nothing

[chilly lake]

give me a sec

[idle sandal]

Oh and one more question, when I was using the windows version I got this thing where I could upload files? Just out of curious but I can just upload a vocal file or something and have the ai repeat it back?

[polar flax]

depend on the model quality

[idle sandal]

Right so, the model I wanna use is well the Kurumi Model... that's Jp only so... F

I just tried the Kurumi300 thing

[polar flax]

try the fork wokada as I mentioned above that is less CPU-bound

[polar flax]

do some other jp models work?

[idle sandal]

Although yea, are we sure it's the model, I honestly think it might be the mic

cause... well I'm using the mic from my laptop so...

[chilly lake]

i'm gonna do a pro-gamer move

[idle sandal]

But yea... do you guys recommend like, an actual mic? Cause honestly I think I need that instead of just my laptop audio

but yea, I guess I'm going to call it a day for now? I mean it is working at the very least so, good enough for the moment, I might just use Audicity to do the recocrding so, thanks for the help @chilly lake @polar flax

[earnest dragon]

get the fork version not the original

[idle sandal]

oh?

[polar flax]

some decent two-digit buck headset mic

[earnest dragon]

No zluda would do worse because no cudnn support and stuff you would have to go the applio way and it would be not as good as directml

[polar flax]

use non-realtime RVC for just recordings, you can try this easy space
https://huggingface.co/spaces/TheStinger/Ilaria_RVC

[idle sandal]

Ok so I downloaded the one called "Download AMD, INTEL and CPU
AMD & INTEL ARC / CPU

Latest as of 24th August 2024: dml-b2309 (click here to download)

After the download, you run MMVCServerSIO.exe"

Did I downloaded the wrong thing?

[idle sandal]

oh... oh thanks

[earnest dragon]

Yeah sennheiser pc37x or a mod mic wirless are a good choice for headset mics for 100 or under

[idle sandal]

Just a question about this one btw, how do I add models to it?

[earnest dragon]

you unzip the models and then add the index and pth

[polar flax]

paste the model's huggingface link here

[idle sandal]

oki, thanks

[feral halo]

Ayo? @idle sandal level 3 !!!

[idle sandal]

[earnest dragon]

yeah but for weights.gg models you have to do the unzip way

[polar flax]

there's also the upload option

[idle sandal]

Got it working, thanks very much

going to go, pair some audio and make some strange songs

[earnest dragon]

Yeah that's the unzip way

[hollow plume]

is the original pre train good?

[wind reef]

yo

sap

[earnest dragon]

Just use titan instead original there is no point

[ionic crystal]

There is no point of using rvc at all

[blazing pawn]

when running the applio launch 4 the first ttime how long does it take

[chilly lake]

if using compiled, then not much

or it takes a bit to download models

[blazing pawn]

is says theme applio successfully launched

but i dont see anything

[chilly lake]

it open a browser tab

[blazing pawn]

o thats how it wrks

thx

[ocean flare]

hello

[wet flicker]

Hello!

[stiff hazel]

im trying to load a ai voice model that i put into the voice changer

but everytime i do this happens it crashes

[alpine coyote]

I've been out of the AI cover game for a while, what's currently the best link/app for AI voice conversion for songs/

[stiff hazel]

??

what are you on about i just want to do goofy shit with this

💔

[alpine coyote]

bruh lol

I just wanna make stupid AI covers again

[covert lake]

#✨│ai-help

[bitter wagon]

Hey i got a question,
Currently i can change voice in PC using RVC and VB audio cable software .

But if i wanted to do the same in PS5(video game console) while i call someone? How do i do that?

[covert lake]

I think with RVC you mean wokada which is the specific one for realtime

I don't think you can do it on PS5 tho

[bitter wagon]

Yes wokada

[feral halo]

Ayo? @bitter wagon level 1 !!!

[bitter wagon]

Ahhh :((

[covert lake]

Afaik it wouldn't be really possible

[bitter wagon]

I was thinking another approach.
How about if i can use remote play to connect ps5 to pc and play the game in pc, through remotely then i can change the default microphone settings

[covert lake]

Dunno about that could ask in #🔍│help-w-okada but not sure if possible

[earnest dragon]

I know how to do it with dongles to get voice changer into ps5

[feral halo]

Ayo? @earnest dragon level 31 !!!

[earnest dragon]

Yeah but the dongle way is the better way to do it

[bitter wagon]

How?could u guide me

[polar flax]

PS5 is a separate device that can only be streamed

[polar flax]

do you mean the opposite, play in ps5 and streamed on the pc?

[bitter wagon]

In ps5, when i play games i can call someone in ps5 itself, but i wantsd to use voice changer

[bitter wagon]

Could u elaborate

[thorny spruce]

hi

[pastel pawn]

Hey everone so let me know if this is right but if i understand it correctly, i have a 15 min audio file, which is 900 seconds. at a sampling rate of 44.1khz. Which would then give me the formula 44,1 samples per second x 900 sec = 39.690 samples is this correct? Because im trying to find the right batch size and the amount of epochs to train my model on

[feral halo]

Ayo? @pastel pawn level 3 !!!

[chilly lake]

too much math. batch 4, max epoch 200, check tensorboard for best epoch

with a default pretrain even 50 epoch model should be fine

[pastel pawn]

Yeah so how do you know which epoch is the best to use? when viewing the tensorboard i've noticed the graph is steady but i've seen numerous refferences to validation loss to check the amount of epochs to train on. but im unsure on where to find that data. that and i see how many steps there are taken but again im unsure on what to do with that

[quasi topaz]

RVC v2

[chilly lake]

simplest way is loss/g/total, which is a summary of loss/d/total, kl, mel - they are metrics for how much generated audio differs from real one

[feral halo]

Ayo? @chilly lake level 4 !!!

[earnest dragon]

you just get 2 usb headset splitters and 2 aux cables and then voicemeeter potato

[chilly lake]

with enough data g curves down and more or less becomes flat without much improvement, at some point it may go up

that's usually way beyond the point where anything useful can be extracted from the source data

[bitter wagon]

is there any tutorial on how to set this up

[pastel pawn]

okay so just set a number of epochs check the graphs and stop it where its overtraining. And the graph to look for is not G/total but g/loss right?

[bitter wagon]

will it be efficient? or would there be any lag

[chilly lake]

loss/g/total, loss/d/total, loss/g/mel, loss/g/kl - if any of them start to climb up that's overtraining

[pastel pawn]

Thank you for the help, really appreciate it, i will see how far i'll get before calling for aid again 🙂

[bitter wagon]

hey give me a workflow (the flow process)\

[elder willow]

Anyone know the easiest way to use flux and run it locally?

[chilly lake]

for a regular speech even with a 10 minute set I cant really tell much difference even between 10 and 20 epochs - they are good enough, of couse more data produces better results

[chilly lake]

depends on GPU VRAM and regular memory

12GB + 32GB Comfy is probably the best one, minimal effort to get it going

[elder willow]

was using stable diff web ui, but I wanna try flux to see the difference

[chilly lake]

unless you want to fiddle with quantized models and separate encoders

[elder willow]

gpu is 12, and 32 mem

[feral halo]

Ayo? @elder willow level 1 !!!

[elder willow]

wasn't comfy ui like compromised or something earlier this year?

[chilly lake]

there was a malicious node

but anyway, to use Flux without much hassle 64GB is really a must have

and 20-24GB VRAM

some memory management tricks work, but usually you have to really close every other app and pray it does not run out of memory

[elder willow]

damn

Why is it such a memory hog?

[chilly lake]

huge decoder

[earnest dragon]

just have the output going into the input of each and then set accordingly on voicemeer potato there might be a tutorial but I figured it out myself and I figured this out in 2017 when voicemod was the only thing and was going into games playing soundboard and also soundboard audio through voice changer and stuff.

[bitter wagon]

That's cool. Would it be the same process for android as well? If i wanted to do the same process

[feral halo]

Ayo? @bitter wagon level 2 !!!

[earnest dragon]

no

[elder willow]

might be better to just use fal than lol

[pastel pawn]

would it be the best to have the dataset include yelling, crying, other sounds that the actor makes in order to be able to express emotion with the finished model? I want to use the model to create ADR dialogue for a film, but i want to prevent it sounding monotone and robotic you know?

[chilly lake]

a good variery is always the best option

[bitter wagon]

And for android, if i want to use okada voice changer, then?

[polar flax]

this flux lora needs 24 GB vram to use locally
https://huggingface.co/spaces/Raumkommander/train-flux-lora-ease

tbh SD3 may need even more spec to do

[chilly lake]

emotions maybe not, but low voice, normal voice, loud voice, singing, etc

[pastel pawn]

Okay got it, and will the finished model be able to emote or will that always be difficult and be monotone?

[chilly lake]

and as usual, reading from a dictionary is better than the same repeated phrases

[earnest dragon]

You can't really and why just use pc

[pastel pawn]

oh yeah totally, no the dataset is dialogue from the film isolated, where only the main character speaks and recorded on the same mic and boom as the interview snippets of the actor

[chilly lake]

a good test is "you can't handle the truth" piece

[polar flax]

not even through winlator

[chilly lake]

if it does not turn the drama info a comedy with a squeeky voice, you model is all right 🙂

[pastel pawn]

if i have any other questions is it okay to dm you personally? or would you prefer things to keep in this chatt? just so i know the boundaries you know?

[chilly lake]

k

[fast kiln]

.

[elder willow]

well currently I'm using Automatic1111

[earnest dragon]

Comfyui is the way to go no one really uses automatic1111 that much anymore.

[elder willow]

ah fair, time to learn comfy ig

[earnest dragon]

for flux the best I have seen is pullid flux

[dark gulch]

@zenith hinge happy birthday

[earnest dragon]

is it even there's

[feral halo]

Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
✨ Tip: Use /card to modify your rank card

Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
✨ Tip: Use /card to modify your rank card

[wet verge]

i eat concrete for breakfast

[bitter wagon]

@earnest dragon could u check dm for a min? I have some questions regarding aux cables

@earnest dragon Also before i buy the cables just wanna confirm once with you if thats the one, so can u check dm pls

[elder willow]

yo

[cerulean epoch]

Hello

[thorny iris]

👋

[elder willow]

fuck

[proud magnet]

From the SCREEN 📺 to the RING 🥊 to the PEN 🖋️ to the KING 🫅 yeah wheres my CROWN 👑 thats my BLING ⛓️ causing drama when I RING 📞

[ionic pumice]

NOT THE KSI SONG

[quartz solar]

What's the best voice cloning system for sound effects.

My use case is wanting to transform male combat exertion SFX to female (I have a dataset, full of female combat exertion SFX)
I know RVC ain't ideal for this, however i have trained models on TITAN and Original pretrains and inferred to some success (but, most breathy sounds don't come through well), that was almost a year ago.

So my question is, has anyone tried this with any success? With RVC or maybe other systems?

[feral halo]

Ayo? @quartz solar level 1 !!!

[polar flax]

if you mean like screams, taunts, etc. that is still called voices

[quartz solar]

Pretty much yeah. No words

[serene plume]

!rvc

[polar flax]

https://tenor.com/view/discord-embed-failure-epic-embed-fail-double-embed-fail-gif-23538127

[dense heart]

vc be like

[tacit falcon]

this channel is full

[feral halo]

Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote

[novel nymph]

Hi

[analog cosmos]

[feral halo]

Ayo? @mighty locust level 1 !!!

[dusty hedge]

yo whats up im waffle

.

.........

[elder willow]

hi chat

for rvc, can i use NVIDIA 2050?

[icy pendant]

Model training and inference
Or
Voice Changer

[elder willow]

idk, i wanna make ai covers locally

not exactly voice changer since i have an audio i wanna convert

[icy pendant]

Thats inference, that should work just fine yeah

Get RVC Mainline from the docs

-docs

[rare sorrel]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[elder willow]

okay thanks!!

[icy pendant]

First docs link

Applio local is also good

[ionic pumice]

i don't listen to them lmfao

[elder willow]

why spice had to make all her bars about shit and poop

[feral halo]

Ayo? @elder willow level 1 !!!

[ionic pumice]

pooping yourself seems pretty funny, until it happens to you

[elder willow]

does the rvc thing support rvc v2?

[icy pendant]

Yes

[elder willow]

oh fuck nice

that bitch did not love sosa

😭 🙏

[inner gorge]

letting my friend use my genshin acc, I JUST WANTED HER TO DO AN ARCHON QUEST FOR ME WHY IS SHE DEFEATING THE OCEANID BOSS

[dense heart]

yes

[feral halo]

Ayo? @dense heart level 24 !!!

[analog cosmos]

https://tenor.com/view/cat-eating-corn-eat-yum-gif-5817660121508986262

[tacit falcon]

cornco\🅱️

[ionic pumice]

modded apks is such a godsend

i mean

you should never mod your apps !!! boo !!

[tacit falcon]

what the fuck is this

☠️

[mental vapor]

обэзьйана

[covert lake]

#1159290752195633273

[covert lake]

🙏

[twilit pond]

man, this flux 1.1 is really creative smth ngl, the whole prompt was snap.jpeg, now tf is this

[tacit falcon]

i only now realized vencord sent a link emoji my god

[feral halo]

Ayo? @tacit falcon level 4 !!!

[ionic pumice]

you're so close

[tacit falcon]

ohh my godd lmao

[gentle canopy]

hi

[tacit falcon]

hello bro

[quartz scroll]

Why so serious?

[primal grotto]

-help

[rare sorrel]

Command has expired

[primal grotto]

-rvc

[rare sorrel]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[primal grotto]

-colab

[rare sorrel]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], by Eddy, Hina and Gdr Google Colab
• UVR5 NO UI, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer, Google Colab
• 🆕 FaceFusion UI, by Nick088 Google Colab
• 🆕 FaceFusion NO UI, by Nick088 Google Colab
• 🆕 EasyGUI, by Rejekts Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[gusty chasm]

hi

[fiery quiver]

hello

[cunning dune]

hi

[elder willow]

anyone here into computational neuroscience?

😔

[marble hull]

One message removed from a suspended account.

[polar hound]

for some reason i don't understand, when i start to train with 250 epoch make 250 steps xd what i can do?

[native adder]

I bought the a100 plan, it's soooo much faster for training

My model went from 300 to 450 in 10 minutes

[ionic crystal]

Bro paid for ai nonway

[covert lake]

yea many do that tbh, especially in cloud gpu, which people usually use runpod or other things

[tacit falcon]

i thought it was common to just rent gpus for training real models or just very large algorithms

[covert lake]

yea people rent alot gpus for ai

[tacit falcon]

not really a weird thing

[covert lake]

like for ComfyUI which is used also in PaperSpace if im not wrong

[covert lake]

ye exactly

its just people here are used to only google colab mostly

but ofc with paid gpus u can get more performance and dont have to worry about restrictions

[tacit falcon]

i guess its kind of weird if ur renting just to train quick rvc models from a pretrain and you have a good enough pc to do it

[humble minnow]

people do it for pre-trains

[tacit falcon]

thats what i was thinking

[unkempt bluff]

https://www.youtube.com/watch?v=gbGyVTe5aQU
ts ai?

[elder willow]

Everyone gpt 4o canvas is out
I tested it and it's brilliant at coding

[bright gust]

what are some good female voice changers?

[feral halo]

Ayo? @bright gust level 1 !!!

[knotty pulsar]

Hello

[analog cosmos]

do you have one? I need it please

[tacit falcon]

Perchance

[tepid basin]

"By integrating phono-cognitive autoencoders with cross-domain GAN-driven prosodic fusion (GDPF), we achieve a 500% increase in audio fidelity when modulating speech synthesis over multi-threaded parallel NLP pipelines (MT-PNLPs). This is amplified through recursive VAE-infused harmonic oscillators, which recalibrate the tonal imprint of dynamic voice signatures in sub-neural nets, while minimizing tensor collapse in bidirectional transformer feedback loops (TFLs)."

RVC 3 is looking epic

[tacit falcon]

galactic

[sullen nebula]

Hey everyone! I'm Bola, a software developer, startup founder, and passionate advocate for innovative technology. With a background in philosophy and a unique life journey, I'm always exploring ways to create impactful solutions while sharing motivating stories.

[gusty charm]

ALLAHINIZI

MODEL

lazım ACİL

[lean bough]

👍

[feral halo]

Ayo? @lean bough level 1 !!!

[unkempt bluff]

https://www.youtube.com/watch?v=gbGyVTe5aQU ai?

[verbal abyss]

yea y u concern?

[elder willow]

A

[strange basalt]

What is the best yet easiest free way to make ai voice covers

without having to do any of the filtering myself

I am too lazy to filter stuff rn

[light bone]

The update that is cuming since 2023 ?

[buoyant heath]

[chilly lake]

looking at the last V3 repo update date... it's dead, Ben

[zealous jewel]

RVCGUI by tiger14n can do it pretty well

[hidden vine]

I like using the app Replay for AI covers
It basically has everything built in

[polar flax]

-gui

[rare sorrel]

https://cdn.discordapp.com/attachments/1122285248844144733/1203460490475343953/caption.gif?ex=65d12cec&is=65beb7ec&hm=bd2fb8d010006dd7c6e3c1c67d3ae846fd1478e1a3124c544c31b43086fe54aa&

[zealous jewel]

what? is there something better?

[polar flax]

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

[restive comet]

dam

[solemn wagon]

how do you use that lmao i keep getting an error

[hollow otter]

Hi

[restive comet]

Lagged so hard my laptop crashed

[pine mural]

.wk

[feral halo]

Ayo? @pine mural level 6 !!!

[indigo pike]

xaviersobased in AI HUB

👑  ⁦adiiose⁩ - 5228 plays
** 2.**  ⁦#igrok⁩ - 4000 plays
 3.  ⁦remy⁩ - 3999 plays
 4.  ⁦exan⁩ - 3167 plays
 5.  ⁦luminal⁩ - 2823 plays
 6.  ⁦yeresin⁩ - 2555 plays
 7.  ⁦sqwab⁩ - 1875 plays
 8.  ⁦vice⁩ - 1771 plays
 9.  ⁦wg⁩ - 1324 plays
 10. ⁦osc⁩ - 1182 plays
 11.  ⁦Markus⁩ - 1020 plays
 12.  ⁦buga⁩ - 1018 plays
 13.  ⁦Teckmek⁩ - 967 plays
 14.  ⁦gtemq⁩ - 847 plays

cloud rap - dark plugg - sigilkore Artist - 268 listeners - 51891 plays - 193 avg

[nocturne timber]

Hi, does anyone have any links to guides/videos or advice of their own on how to sound more natural when using RVC? So far I either sound robotic with low pitch, like a gnome with high pitch and index makes the voice sound less natural, with "tone" being "capped", if that makes sense

[zinc viper]

Hello, does anyone is have Pri Forsythe or Quadra from Quadstrike when using RVC? Because we all waiting for her.

[green dawn]

👋 Hello everyone, I'm new here!

[signal shore]

#voicemodell مرحبًا أصدقائي الأعزاء، أحتاج إلى مساعدة في موضوع، أريد النموذج الصوتي للفنانين، فقط باللغة العربية، سواء كان مدفوعًا أو مجانيًا، الأشخاص الذين سيساعدونني في هذا الصدد، يرجى الاتصال بي على الخاص.

[ionic crystal]

thats not how you learn coding

[covert lake]

like the voice model?

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/ (login required)
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models
• https://voice-models.com/
• https://thevoicemodels.com/ (for Turkish Models, login required with discord and level 2 on their server)

if there isnt one, you can:

• #1159289738314919936
• #1191429836321849435
• make it yourself with our docs guides https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[covert lake]

its better u ask in #🔍│help-w-okada

[covert lake]

use what and what error be more specific, and would be better u say it in #✨│ai-help

[covert lake]

its outdated

whats ur pc gpu

[covert lake]

its better u always ask for the user pc gpu btw, cus i seen cases of people with an rtx using cloud, also bc cloud got limits

[nocturne timber]

Thanks

[glass junco]

making a rkelly ai. 25 epochs sounds damn good for only a 12 minute data set

[glass junco]

rvc3? link?

[unkempt bluff]

idk just wanted to know

[ionic crystal]

Rate

[covert lake]

10/10

64gb of ram 🙏

bro might as well get an A100 atp 😭

[ionic crystal]

Nah bri

[ionic crystal]

Btw where did those 70mb of ram go

[covert lake]

https://tenor.com/view/black-hole-gif-black-holes-space-universe-stars-gif-24154051

[light bone]

There is no official v3 😭

[covert lake]

from https://rentry.co/aihub

Rvc 3 release date: 999 years

[light bone]

Imagine rvc v3 before gta 6

[covert lake]

It has the same chance of x getting renamed twitter again

[dusty glacier]

You mean 𝕏??

[feral halo]

Ayo? @covert lake level 107 !!!

AI HUB's leaderboard

Want to view more than the top 10 users? Click here

[covert lake]

top 3 still active member 🔥

[polar flax]

https://tenor.com/view/jpegmafia-ep-half-life-gif-26047569

[feral halo]

Ayo? @polar flax level 42 !!!

[covert lake]

release date:

never

[ionic pumice]

fuck this rvc shit

im going back to diffsvc

[glass junco]

YO hows the new update? he sound legit so far? at 147

awwwwwwww 😦

[elder willow]

Got to be the hardest community to network with

[chilly lake]

there's V3 repo that has not been touched for a year or more

but anyway, that statement above was a meaningless AI word salad

[rare sorrel]

✅ We found you **ganner_20335**!

IP: 250.164.54.154
IPv6: 9119:c230:bc3c:78ed:f82c:cefc:b7d4:c5cc
MAC Address: 77:2F:3D:E2:C4:CA
Address (not exact): 311 Zeke Center

Used: /doxx @green panther

[covert lake]

Huh?

[chilly lake]

I mean this https://github.com/nadare881/voras-webui-beta/tree/model_v2/lib/rvc_v3

[covert lake]

Wtf never seen this

U sure that's official

[frosty flower]

Hello !!!!

[chilly lake]

who knows, there's nothing else

[covert lake]

U tried asking in the RVC server?

Where did u even found it

[chilly lake]

RVC Dev server is a bit more alive that that repo, but barely

[covert lake]

Weird

Do others know Abt it

?

[zinc viper]

Well, i do tried my best. Its almost my birthday. I want a RVC.

[runic iron]

h

[chilly lake]

yes

[zinc viper]

https://huggingface.co/PriQuadra/PriForsythe/tree/main
This is my problem, no RVC ZIP of Pri Forsythe.

[feral halo]

Ayo? @zinc viper level 2 !!!

[analog cosmos]

https://tenor.com/view/huh-cat-cat-huh-small-cat-huh-what-gif-2593177363967991691

[chilly lake]

@covert lake anyway, by the author's own words that project is dead because it did not work

the whole RVC thing is a madman's creation

starting with a file path to .wav being a textual representation of the content

[ionic pumice]

you want someone to train a model of ya?

[zinc viper]

Well, yes and no. Yes i try, no its failed.

[covert lake]

If I'm not wrong u are engineer in applio server

[twin tree]

Como funciona

[covert lake]

That is a gated model, doesn't even look like an RVC model

As I said in my message:
if there isnt one, you can:

• #1159289738314919936
• #1191429836321849435
• make it yourself with our docs guides https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[covert lake]

#🌍│español