🧬│ai-chat

[covert lake]

Realtime for calls?

[sour gulch]

yea

[covert lake]

-rt

[rare sorrel]

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

[slow nexus]

earlier we were supposed to make kg on neo4j so there was not any hierarchical structure, we tried to incorporate levels attributes and classes(entities) as properties in our structure, and now we would like to make it to classes and subclasses structure.

[covert lake]

1st link, it's a wokada fork, modified version with better performance

Do NOT watch yt tuts

[solar merlin]

hello
can i use voicechanger in windows 11?

[covert lake]

That's a bot, and here it's mostly about RVC and wokada

[covert lake]

Yes, what's ur PC GPU? And do u mean realtime?

[polar flax]

no, in windows 3.1

[solar merlin]

i have laptop. with rtx 3060ti. and yes realtime

[solar merlin]

what?

[sour gulch]

im trying to follow it but what it says doesnt match whats happening

[tacit falcon]

me using rvc on windows xp:

i should try that, unironically

[cursive mirage]

Rtx 3060

[feral halo]

Ayo? @cursive mirage level 1 !!!

[real hazel]

How many guys here with a MacBook? I'm having trouble with my AI tools on my Mac.

[covert lake]

Yeah should be good

-rt

[rare sorrel]

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

[covert lake]

1st link, wokada, especially a fork which is a modified version that is optimized

[covert lake]

If u have issues ask in help channels

[covert lake]

It's a joke

[covert lake]

Yk people believe that right 😭

[covert lake]

Alright, explain what u are using and the issue better in #✨│ai-help pls

[covert lake]

Mac isn't the best for AI, are you using RVC or wokada?

You would be able to run wokada and rvc inference (using models) fine, but not training (making models)

[novel nymph]

Hi @covert lake

[slow nexus]

????????????????????????????????

[real hazel]

What's the best for AI you'd say?

wokanda

[ionic pumice]

peak

[novel nymph]

4ever

[real hazel]

For sure, for sure, always

[feral halo]

Ayo? @real hazel level 1 !!!

[abstract estuary]

ANANIZI SİKİYİM

BABANIZI SİKSİNLER

[zenith nova]

https://tenor.com/view/hello-chat-chat-gif-22869152

[abstract estuary]

SİKSİNLER

[analog cosmos]

true

[ionic pumice]

hello air

[feral halo]

Ayo? @echo bough level 2 !!!

[elder willow]

Is there a 48k Beatzforge PTH file?

[ionic pumice]

true

_ _

[zenith nova]

https://tenor.com/view/speak-in-gifs-blue-gif-14626908

[elder willow]

guys what do i do if i need .json format for voice mode

for config

[covert lake]

hey

[covert lake]

u don't need it, u need a .pth and added.index for rvc models

unless ur using another ai

[covert lake]

for what ?

there are LOTS of ai fields

[covert lake]

here its mostly about rvc and wokada

[novel nymph]

Do you know coqui xtts?

[covert lake]

yea, its discontinued btw

[novel nymph]

KIZ KARDEŞİNİ SİKİM

[novel nymph]

Whyyy

[covert lake]

https://coqui.ai/

its been like a year or some months

[novel nymph]

https://tenor.com/view/agonyman-gif-25333932

[elder willow]

-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

[elder willow]

-svc

[rare sorrel]

So-Vits-SVC

[novel nymph]

It was making better tts than elevenlabs

[covert lake]

in terms of quality gpt so vits is the best rn

[covert lake]

so vits is old, rvc is better

[abstract estuary]

yeah and ı am sorry my brother is... Freaky?

[elder willow]

is there a rvc inferrence from huggingface that is free!?

[novel nymph]

Oh ok
Np

[abstract estuary]

k

[covert lake]

yes, ilaria rvc zero, did you try what i told you?

[elder willow]

im going to try today

[elder willow]

wish me some luck

[covert lake]

reminder to show me the error on top right

i need to see the error to help u out

[elder willow]

ok

[covert lake]

goodluck lol

[chilly lake]

their product was, while good in quality, not hands off... so they lost to other competitors that were many

[novel nymph]

Oh
I see

[chilly lake]

you cant put hallucinating tts into production without any supervision

[novel nymph]

Is there a tut video for gpt so vits

[chilly lake]

because it either starts repeating things or speaks demonic voices, dont want some company scare a grandma with that

[elder willow]

why does my final model sound robotic!? @covert lake

[novel nymph]

But
I do want

[novel nymph]

Your dataset is noisy

[chilly lake]

you can still use it for personal projects

get a forked repo

github.com/idiap/coqui-ai-TTS

[junior pollen]

hi

[feral halo]

Ayo? @junior pollen level 1 !!!

[covert lake]

that depends on the model quality, is it a model u trained urself?

[elder willow]

yeah

[elder willow]

i trained it myself

2 min 250 epochs

[covert lake]

i dont really think so, u could check https://docs.ai-hub.wtf/tts/gpt-sovits/

[covert lake]

did u use tensorboard?

[elder willow]

no?

[covert lake]

u need a CLEAN dataset

[covert lake]

more epochs dont mean more quality, check https://docs.ai-hub.wtf/rvc/resources/epochs-tensorboard/

u need to use the tensorboard

[elder willow]

@covert lake What Index influence should i use?

[covert lake]

the index influece is the setting that specifies how much index is being used, index is the file that contains the model accent

id leave it default

u would need to retrain ur model with the tensorboard if u want an higher quality one, and also be sure the dataset is cleaned

[elder willow]

What is the best vocal remover on MVSEP? Any suggestions

[sturdy elbow]

BS Roformer

[polar flax]

tell me some of them who play ranked league, apex, valorant, etc. and what kind of issue it could be

[rare moss]

-_-

[icy pendant]

-_-

[covert lake]

-_-

I haven't really seen those types tbh

Unless u mean those ai who just play the game and take alot of time to finish a level like for super mario

yea didn't see those type lol

[raw path]

Hello

[flint wharf]

hi

[heady whale]

Yo

[untold abyss]

someone know how can i change the name of the .pth?

[narrow trail]

does anyone have a good 2010 era eminem model pth i can use (love the way you lie sounding)

[ionic pumice]

just change it lol

[untold abyss]

i cant

[tawny carbon]

How to use this?;)

[iron junco]

HI

[gray dagger]

h

[feral halo]

Ayo? @gray dagger level 1 !!!

[ionic pumice]

goodnight ai hub

see you soon, maybe, or never

[fathom zealot]

a

[plain cove]

https://tenor.com/view/goodnight-good-night-gn-nigh-night-night-gif-15135428212810379047

[drowsy minnow]

join me play this game

https://wolfy.net/game/HLLVC4GC48

[elder willow]

can i see a link to a guide on how to train an rvc mode

[feral halo]

Ayo? @elder willow level 2 !!!

[elder willow]

model**

i found it now

thx

[indigo idol]

Any chance someone can do one free ai voice if that not problem?

[covert lake]

for models request, ask in #1159289738314919936 or #1191429836321849435

[indigo idol]

Did it

[covert lake]

will reply in #✨│ai-help as i seen u talked there too

[novel nymph]

Can I ask you something about gpt so vits

[covert lake]

I can't really help you as I never used it, sorry

[novel nymph]

Okeyy

Np

[dreamy plover]

.fm

[indigo pike]

Hi Sebbe, welcome to .fmbot.
To use this bot you first need to add your Last.fm account.

Use the buttons below to sign up or connect your existing Last.fm account.

[glad forum]

hello where am i? it's RVC group here?

i'm new to RVC, came from youtube. how can i start up?

[timber mulch]

hey everybody

https://tenor.com/view/angel-galang-angelgalang-drag-race-drag-dragraceph-gif-6840389422617647714

why can i not send gifs

[covert lake]

RVC & Wokada yes

[covert lake]

youtube vids are old. WHat are u looking for and whats ur pc gpu?

[covert lake]

level 1: embed perms in help channels
5: embed perms in other channels

[glad forum]

I don't know whats my GPU (someone told me that this GPU can't work)

[feral halo]

Ayo? @glad forum level 1 !!!

[glad forum]

can I use my CPU?

I just want a voice transformer (Chinese), and I wanto learn training my models

I'm sorry my English is poor...

[covert lake]

You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU

[covert lake]

You can't like convert your English voice to Chinese, i mean u could but would sound bad

[glad forum]

Yeah I know, but my computer is not here...

[covert lake]

U mean like convert ur voice to another voice in realtime for calls? Or like translate

[glad forum]

Yeah, Chinese to Chinese (mandarin)

[covert lake]

CPU is slow, would be better to use cloud (remote good pc) than locally (on ur PC)

[covert lake]

Alright, could be possible if u get a Chinese model ofc

For Realtime Voice Changing for Calls on Cloud (remote good pc for those who don't have a good one, YOU CANT DO THIS ON MOBILE):

• Google Colabs (4 hours daily of free T4 gpu, easy to use, require only a google account) :
  • How to use Original W-Okada's Voice Changer Google Colab
  • Modified W-Okada's Voice Changer Google Colab
• Kaggles (30 hours weekly of better GPUs, T4x2 & P100, harder to use, requires an account and a phone number)
  • Original W-Okada's Voice Changer Kaggle
  • Modified W-Okada's Voice Changer Kaggle

[glad forum]

I remember, it's A2000

[covert lake]

A2000?

[glad forum]

I'm not sure, my computer is not here

[covert lake]

I haven't really heard off that one, it's better u come back when u have ur PC there

Also bc u can't use it even on mobile wokada

U need AT LEAST a pc

For wokada

[glad forum]

How can I get a sound model of a certain person (for example 郭德纲)

[glad forum]

gotit

now I have a 郭德纲's solo 相声 about 20min, can I train a model with it?

[feral halo]

Ayo? @glad forum level 2 !!!

[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/

[covert lake]

I don't know that language, and Wokada is just for using models in realtime for calls, for training you need RVC

It's another program, the same one used to make models for wokada

As you dont got a good PC, its better you use cloud (remote good pc) for training an RVC Voice Model:

• Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
  • Applio (ui)
  • Mainline (UI)
  • RVCDISCONNECTED (no ui)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
  • Mainline (UI)
  • Applio by Vidal (UI)
  • Applio by Shirou (UI, no guide as of right now)
• Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly):
  • Mainline (UI, No guide as of right now)
  • Applio (UI, No guide as of right now)

[glad forum]

is wokada something like RVC? I only know RVC yet

[covert lake]

Google colab = easy but low GPU time
Kaggle = hard but more GPU time

[covert lake]

Wokada is a program specified for using RVC models in realtime for calls

[glad forum]

Oh I see

Thanku very much I'll go to learn that

Sorry I'm new to discord too QwQ

[covert lake]

Alr yw

[fresh aurora]

Hey, is the guide on how to download TensorBoard on Applio from the AIHub website working for you? I’m getting a 404 error and can’t figure it out. Could someone send a link to the download page? On the official site, I can only install it through CMD, but that's not working for me.

[feral halo]

Ayo? @fresh aurora level 2 !!!

[covert lake]

#📰│dev-updates message

[fresh aurora]

solved problem thx

[feral halo]

Ayo? @main robin level 1 !!!

[glad forum]

@covert lake "Wokada" means the project "voice changer" by W-Okada, right?

and is called "日版RVC" (Japanese RVC) in China

[plain cove]

I'm not a nick but yes that's true

[glad forum]

I'm just using it but meeting some problem

[plain cove]

#🔍│help-w-okada ?

[glad forum]

Wao! I see 3q

[old sleet]

heyy

[wanton sparrow]

has anyone made an ai voice for the luigi from that one video where he says "hello mario"

[covert lake]

Yes

I deleted your yt video link, use #1159290752195633273

[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/

[fervent bone]

How long till I can use X-Minus again?

[glossy hill]

Hi all - Is there anything that'd prevent me in theory from training a YoloV8 model directly on an OpenCV Mat instead of NCHW blob data?

For the past 6-7 months I've been making a full body tracker for VR use and was thinking of ways to optimize inferencing process - currently around half of inferencing run time is just preparing the image and converting it to a blob

[covert lake]

#1159290752195633273

[glossy hill]

I'm looking at having 3-4 cameras all inferencing at 90FPS simultaneously while a game is running so performance is quite important - will likely be done on a companion compute card as not to hit game FPS too hard

I mean, I know it should work, but are there any special architectural considerations that were made that would mean NCHW works better?

[glossy hill]

hmph https://github.com/ultralytics/ultralytics/issues/13179

[feral halo]

Ayo? @glossy hill level 1 !!!

[glossy hill]

maybe there are hardware differences with memory access speed that make NCHW ideal

but then, wouldn't it still be quicker to access the memory slower and not full-copy the image when making a blob??

[livid nebula]

whats the best AI for music

[solar agate]

they not like us, they not like us, they not like us.

[chilly tartan]

hey, is it possilbe to remove background music from a video in order to hear a character's voice?

feel like that sort of thing should be possible in this day of age

[chilly lake]

if the music is faint enough, regular inference will remove it

or you can use UVR

[fleet tide]

does anyone have a google colab to put AI model on the apacella to make my AI sing the song?

[elder willow]

@covert lake Are there colabs for inference?

[feral halo]

Ayo? @elder willow level 3 !!!

[molten stone]

Yes

[chilly lake]

using colab for inference kinda not worth the hassle

either run local on cpu, it is not terribly bad

or use huggingspace

[glossy hill]

I wonder if you could just run it thru nvidia broadcast

[ionic flower]

yo

[fervent bone]

Hi

[ionic flower]

I had this idea of a ai cover album tbh

[feral halo]

Ayo? @ionic flower level 1 !!!

[earnest dragon]

yes you can if you have background noise when using w-okada instead of using the not as good w-okada noise suppression

[proven python]

-rt

[rare sorrel]

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

[main temple]

hi and hello, this is my into msg lol

[chilly tartan]

and that removes BGM?

[chilly lake]

it can separate background music out

[ionic flower]

and what will it sound like when you add ai to the vocal split??

[solar agate]

they not like us, they not like us, they not like us.

fein, fein, fein, fein!

eating my lunchly right now! #inthethickofit

[ionic flower]

what in the fuck?

[solar agate]

what in the hawk tuah!

i cant wait for the new episode of talk tuah. preordered the new prime drink and stocking up on feastables to go with my lunchly as we speak! #hawktuahtuesday! #spitonthatthang!

i dont like that they give you such a small size of feastables and prime so i ordered the party size!

[ionic flower]

https://tenor.com/view/way-tf2-demoman-gif-16565823

[solar agate]

who wants some?

[ionic flower]

no

[earnest dragon]

I feel like you should get mutted for this

[earnest dragon]

https://tenor.com/view/embed-fail-dissapointed-sad-teary-seal-gif-3391051642056876120

[ionic pumice]

from the hawk tuah to the lunchly to the

[polar flax]

https://tenor.com/view/tf2-epic-embed-fail-gif-26146453

[solar agate]

what in the hawk happened tuah freedom of speech?!

[earnest dragon]

Nah you sound under 13

[ionic flower]

Reversed Demoman Crying

[solar agate]

im 29

[ionic flower]

i kinda agree with cxsmo

[earnest dragon]

then that's even worse imagine being that old and still saying that stuff.

[solar agate]

im not under 13! that would be violating discords terms of service which i abide by dearly my friend

[ionic flower]

2 words. red flags.

[feral halo]

Ayo? @ionic flower level 2 !!!

[ionic pumice]

what yall know about the talk tuah podcast

[earnest dragon]

but a lot of people lie

[solar agate]

everything! i cant wait for her to reveal whom pookie is 👀

[earnest dragon]

Cringe

[weak scarab]

Hey guys

wkada is really laggy if I play games while using it

[earnest dragon]

yeah you're under 13

[weak scarab]

anny tips?

[ionic flower]

nah i got nothin

[earnest dragon]

yeah that will happen what gpu do you have

[weak scarab]

3070

ti

[solar agate]

youre so cool

[feral halo]

Ayo? @solar agate level 2 !!!

[solar agate]

ayo?! im level tuah?!?!

that is epic!

[weak scarab]

do I need a 4090 or something? lol?

[ionic pumice]

hawk 1: what's 4-2
hawk 2: uhh

[ionic flower]

how much did it cost for that?

[solar agate]

LMAO

[ionic flower]

bad joke detected

[ionic pumice]

cool

[earnest dragon]

yeah playing games and w-okada is not something you can do really on a 3070 while gaming unless you mess with the settings on but then you will have delay and also you would need to turn game down

[solar agate]

tell one right now mr funniest guy in the world

[weak scarab]

So I need a 4090 then basically?

[ionic flower]

no you brainrotted F A C K.

[earnest dragon]

yeah or you can just get a Intel ARC gpu as a secondary gpu just for gaming or the other way around.

[weak scarab]

Uhh which one is cheaper, how much does it cost?

[solar agate]

Why did the bird start a podcast about nature?

Because it wanted to share all its insights on the "Hawk Tuah Podcast!"

[polar flax]

u can still play games like valorant, not demanding ones

[weak scarab]

Im playing valorant but its still laggy

[ionic flower]

just to reopen your pc build and add it plus setting it up?

[weak scarab]

so sad

[earnest dragon]

260 for a a770

[polar flax]

better try fork wokada, and ur cpu wont suffer as well

[weak scarab]

I'm using fork wokada actually, but im lagging more using that

even at 1000 ms im laging, but I use 2.7s extra

[solar agate]

who wants to be my hawk-tuah the bird festival (im the pigeon)

[ionic flower]

get out.

[solar agate]

i am not in violation of the rules actually

[weak scarab]

Is it easy to run a 2 gpu setup?

I thought it was complicated

[polar flax]

lower game settings as needed, enable DLSS, and fps cap to 60 or according to monitor refresh rate

[solar agate]

intjs! i love those!

reminds me of when i went tuah the eagle-only party as a hawk! (hawk tuah) @polar flax

[weak scarab]

Is drowsy a bot?

[feral halo]

Ayo? @weak scarab level 3 !!!

[ionic flower]

im thinking the same thing or he might be a tos breaker.

[solar agate]

ayo?! erik passed up level tuah! congratulations!

[polar flax]

if u only trash talk and act like a kid who wastes ur parents' credit card, why are u here?

[solar agate]

trash talk?

when? ive only been nice and kind!

[weak scarab]

Damn I tried lowering to lowest, I think I just need a better PC

[ionic flower]

god damn! pure comedy!

[earnest dragon]

not that hard as long as your cpu has the pcie lans

[solar agate]

i am here, simply to share my interest in AI and download models that the kind members of this community has provided! and of course make friends along the way!

[earnest dragon]

and also you are 12

[ionic flower]

erik, you have some thermal paste to keep it cool?

[solar agate]

nope! i am 36

[weak scarab]

Yeah I applied new thermal paste a day ago actually

[ionic flower]

lies

[polar flax]

ok, feel free to share ur AI creations in #1159290752195633273

[weak scarab]

I have AMD cpu, but thats fine right

[ionic flower]

good.

[solar agate]

i dont have any AI creations

[weak scarab]

Should I use server instead of cleint?

[earnest dragon]

no one believes anything you say not even your parents or your cat or dog if you have one.

[polar flax]

ryzen 5600 should be fine

[ionic flower]

your choice

[earnest dragon]

then make one

[polar flax]

go make it

[solar agate]

i am not obligated to! that is not apart of the servers TOS! i will pass, thank you!

[ionic flower]

Loud Wheezing Intensifies

[earnest dragon]

imagine liking ai but never making anything with ai

[solar agate]

many people do, actually!

[earnest dragon]

they are the ones that don't have good pcs

[solar agate]

AI doesnt only exist on PC, it exists as a seperate entity i.e. Artificial intelligence is the science of making machines that can think like humans

[weak scarab]

I have

AMD Ryzen 9 5950x

[solar agate]

to correct you, @earnest dragon. cant argue with logic!

[earnest dragon]

but if you like ai and have a good pc then you should be at least making something with ai

[dull night]

@solar agate pop in vc and square up cuh

[solar agate]

that isn't true! In essence, while creating with AI can be fulfilling, it’s not a requirement for everyone who enjoys the technology. Each person's relationship with AI can be unique and valid, regardless of their creative output.

[feral halo]

Ayo? @solar agate level 3 !!!

[dull night]

drowzy pop in vc right now

[solar agate]

this response was made from AI, something i enjoy whilst having an android!

cant argue with logic @earnest dragon !

[dull night]

drowzy u suck

[ionic flower]

get in bro

[solar agate]

heyo! i leveled up from level tuah- tuah level 3!

[dull night]

drowzy are u mentally afk?

[feral halo]

Ayo? @dull night level 1 !!!

[solar agate]

heyo! congratulations on the level up @dull night ! next level is looking like youre going tuah level tuah!

[dull night]

fuck you

[weak scarab]

@earnest dragon which gpu do u recommend to use 2nd only for processing the voice?

[chilly tartan]

do you think its possible you can do a small requet? ^^

[feral halo]

Ayo? @chilly tartan level 1 !!!

[dull night]

get in vc rn

[ionic flower]

YO MATE WE GOT Lunchly in the chat

[feral halo]

Ayo? @ionic flower level 3 !!!

[dull night]

get in fucking vc right now before i get angry

[weak scarab]

im playing val bro

[solar agate]

ayo!! congratulations on the level up @chilly tartan next level is level tuah!

[dull night]

join vc

[solar agate]

ooooo! @ionic flower with the level 3 👀 big upgrades i see!

[dull night]

join vc

[solar agate]

who has tried the fiesta nacho luncly? its my favvvvvvorite!

[polar flax]

seems like ur system issue, try update nvidia driver and debloat ur windows/disable unneeded background processes

[dull night]

alisa

[solar agate]

@polar flax you wish to be a streamer someday? i know a thing or two about content!

[polar flax]

https://tenor.com/view/cat-no-nonono-noooo-cat-no-gif-13597132752790194338

[chilly lake]

what request?

[dull night]

Alisa I need your help

[solar agate]

basically to become a successful streamer you can:

1. make videos

2. alternatively, you can stream

[dull night]

pls stfu

drowzy is sexist cannonical

[ionic flower]

secksist

[polar flax]

I need to change my bio there

[chilly tartan]

im trying to get two voices out of these clips but the music is in the way

[polar flax]

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

[solar agate]

change of plans?

[dull night]

drowzy

[solar agate]

If you could have dinner with any fictional character, who would it be and what would you want to talk about?

[dull night]

too wimp to talk bud

join unlimited vc

square up wit me\

[ionic flower]

and me

[dull night]

we just wanna talk with you, you're so cool

[weak scarab]

I just updated my driver yday, how do I debloat? I have 4tb free space and no programs almost

[solar agate]

me personally, im more of a harry potter kinda dude! would love to hear thoughts and opinions!

[weak scarab]

I reset my entire pc yday

[solar agate]

if there are any AI harry potter voice models i would LOVE to try them!

[dull night]

drowzy, I wanna talk tuah you

[ionic flower]

i wanna hear your voice~

[solar agate]

thats it for tonight! im going tuah go through the motions and head to bed! hope to talk-tuah you guys tuah-morrow!

[dull night]

stfu you fucking basic bitch @solar agate

[proven python]

-rl

-rt

[rare sorrel]

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

[polar flax]

how is the valorant performance alone?

[weak scarab]

its good

[dull night]

@drowzy stfu dont even say a word tuah everyone in here

[feral halo]

Ayo? @dull night level 2 !!!

[dull night]

@solar agate

shut up

[weak scarab]

its weird because the regular w-okada worked better than the new one (forgot name)

maybe I have a setting that is wrong?

[solar agate]

heyo! congratulations on the level up @dull night you managed tuah get tuah level tuah! 3 tuahs in a row! holy cow!

[dull night]

shut up

[polar flax]

the regular one is smooth? well go use it

[weak scarab]

I guess but isnt the new one better

it sounds better to me

[solar agate]

If you could instantly become an expert in any skill or hobby, what would you choose and why?

[polar flax]

iirc 3080 users might have similar experience on it

[weak scarab]

alisa do you use voice changers urself? do other people notice you using it?

I feel like some people can tell

[solar agate]

@polar flax Fair enough! I’ll share my ‘expert’ tips after I wow you with my voice-changing skills. What character should I channel first?

[ionic flower]

scamming indian

[earnest dragon]

a arc a770 for the price

[polar flax]

I have tried the original one on a borrowed 4070 pc before, now I have no idea with just a potato igpu laptop

[solar agate]

ah, the classic scammer! Let me grab my best Indian accent and we’ll see who can pull off the best impression. ready?

[earnest dragon]

You have no tips or anything only people that have made stuff or are well knowledgeable with the stuff are the people that can give good tips and stuff

[weak scarab]

Ok thanks you two, appreciate it

[dull night]

are you ai

[solar agate]

hello! you have won a great prize today! just give me your bank details, and I will send you all the money! haha, just kidding! but seriously, let's talk about something fun!

[weak scarab]

I mean honestly its surprising how well voice changers sounds nowadays though

[earnest dragon]

Nah he's probably 12

[dull night]

ill send my bank details

[ionic flower]

or worse

[dull night]

but you have to call me with face time @solar agate

[weak scarab]

I think if you use voice changer for YouTube videos people wont be able to tell, if you use some post editing

[solar agate]

If you could travel anywhere in the world for free right now, where would you go and what’s the first thing you’d do there?

just trying to make friends! haha

[dull night]

your house

[polar flax]

gen chat moment

[earnest dragon]

yeah and one thing now one has done is surround sound rvc

[polar flax]

https://tenor.com/view/shirakami-fubuki-hololive-yabe-vtuber-gif-19227770

[weak scarab]

what do you mean?

[solar agate]

Ii you could swap lives with any anime character for a day, alisa, who would it be and what adventure would you want to have?

[ionic flower]

stop yapping DrowzyGPT

[polar flax]

drowzyGPT moment

[solar agate]

haha, I guess that makes me the ultimate sleepy sidekick! What’s your superhero name?

[feral halo]

Ayo? @solar agate level 4 !!!

[solar agate]

heyo! just reached level four! perfect time for the lunchly break!

[weak scarab]

surround sound rvc? what do you mean by that?

[solar agate]

man KSI's new song is so grrrrrrrrrreat says tony the tiger

[dull night]

drowzy whens the lobotomy update coming out for you @solar agate

drowzy pls talk

[fervent bone]

Hi

[polar flax]

typical social engineering using chatgpt

[dull night]

hes a real person

if u were in vc u would know

[polar flax]

unfortunately not this moment

[dull night]

ur not cool enough

[patent solstice]

lol i keep leaving this server just to rejoin
https://github.com/Tiger14n/RVC-GUI/
is there any newer versions of this? Its archived

[polar flax]

yea im too nerdy on my things

[patent solstice]

just dont wanna learn another tool if its out of date

ill keep doing research, ping me if anyone would like to respond (:

[polar flax]

check the recent rvc and guides in https://docs.ai-hub.wtf/

[minor raptor]

gptcore

[patent solstice]

Thanks!
About the vocals, is it possible to remove specific instruments aswell?

[minor raptor]

I don't update rvc cause im so cool 🥸

[tacit falcon]

i use mainline lol

[polar flax]

UVR (local), uvronline.app, mvsep.com

[patent solstice]

thanks!

[eager drum]

hi

[earnest dragon]

surround sound upmixing it so that it sound more spacious and like you're talking to them in there room and just like surround sound is normally

[weak scarab]

oh is that a good thing?

actually ima start a new YT channel and im planning on using a voice changer so any advice is good

[earnest dragon]

yes it would make it sound higher quality because of the positioning and Soundstage and stuff

[weak scarab]

interestinh

[feral halo]

Ayo? @weak scarab level 4 !!!

[weak scarab]

how do you do that? is it easy to do?

[hazy perch]

@dire summit que haces acá putita, quieres ser mujer?

[earnest dragon]

not really that easy you need to install a bunch of stuff

[weak scarab]

do you know how to do it? is there a tutorial?

I dont mind if its hard

[earnest dragon]

there is no tutorial because the setup is something I put together using some vsts a daw and vb matrix coconut.

[polar flax]

rvc datasets and inference outputs are always mono since dry voices/sounds like used in 3D games are so.

surround sound, reverb and shit are just post-processing

[weak scarab]

Ok makes sense

tbh im noob when it comes to sound

im just impressed by how quickly the tech is evolving, its crazy

[weak scarab]

do u have like an example of how it sounds, im kinda curious

[polar flax]

yea if u want to be a streamer and play games, ideally get a second gpu/laptop

[feral halo]

Ayo? @polar flax level 43 !!!

[weak scarab]

I dont think I will stream, I'll just release YT videos but I do play a lot of games and its fun to use different rvcs sometimes

[polar flax]

u dont need realtime voice changer for that, just use offline rvc to add on it

[weak scarab]

offline rvc? how does that work?

[polar flax]

https://tenor.com/view/po-facepalm-gif-27277915

[weak scarab]

I'm telling you im noob lol

a noob that wants to learn

[tired jasper]

is SVC the superset of RVC?

[polar flax]

it is regular rvc duh

[weak scarab]

Yeah but like, would I record a video with my voice and then add it in post editing?

I basically want to mix my real voice with a voice changer to have a balanced voice sort of

just to make it a bit 'better'

[tacit falcon]

if u mean so vits svc that was before rvc apparently??

[tired jasper]

really? seems svc has better quality, idk

[tacit falcon]

i literally dont know

dont take my word for it

[tired jasper]

while RVC model has only 50mb the SVC is like GB

[tacit falcon]

-svc

[rare sorrel]

https://cdn.discordapp.com/attachments/1159822429137412177/1194008385113301003/svc.gif?ex=65c13ef6&is=65aec9f6&hm=ce1f57bc316793f641c80f9b359ba9db7d1312af6918b1aed9c8468c74035bac&

[polar flax]

the mic voice (post process with denoise and stuff) and gameplay video with system sound

[tired jasper]

got it

[feral halo]

Ayo? @tired jasper level 1 !!!

[snow briar]

Hi

[weak scarab]

Damn I feel dumb af ngl, I think I need an explination that even a 10 year old understands

[snow briar]

So umm... I was searching for a Gura AI voice model on weights gg and found these comments. So does anybody know who Felt is and whether it's true he has a custom Gura model that he shares via DMs ? (looks like I can't attach screenshot here)

Screenshot on imgur bc I can't directly attach here https://imgur.com/a/So2hufn

[tacit falcon]

felts not in this server no more

[snow briar]

Sorry, I'm pretty new to all this, just looking for a good Gura model, so is Felt's Gura model publicly archived anywhere for download ?

[tacit falcon]

idk............

ask around methinks

[tired jasper]

useful info: RVC extrapolate autotune af. When original audio is very discrete, the output of RVC is like: it's obvious it's autotune

[vocal leaf]

I am the Globgogabgalab, The shwabble-dabble-wabble-gabble flibba blabba blab.
[11:26 PM]
I'm full of shwibbly liber-kind,
[11:27 PM]
I am the yeast of thoughts and mind,
[11:27 PM]
Ohohoh, Hahahaha.. Splendid! Simply delicious!

Im the gobalabagla

[tacit falcon]

dude what

[placid grove]

it's 3am for me rn and that's what i see

im dying rn

[tacit falcon]

[snow briar]

So what's the best Gura model out there rn ? The one with highest number on weights gg I tried didn't work well with a song (Legacy from DMC 5) because of cracked voice in a certain part

[feral halo]

Ayo? @snow briar level 1 !!!

[hasty elk]

dude rvc is awesome

[earnest dragon]

I have a example of what a song sounds using it but it's on the older settings so not as good because didn't have time to record the new setup. I can send a screenshot of what the setup looks like

[weak scarab]

Yee would be sick man

[polar flax]

https://tenor.com/view/vtuber-hololive-ホロライブ-hologra-ホロぐら-gif-15579180677117910141

[tired jasper]

another tip: more clean the input, better the output

not always because some models can't deal with all voice noises

i recorded the Grover laugh with my voice, no voice model can deal with better than grover model

[earnest dragon]

this is what the setup looks like

[snow briar]

Tried several different Gura models in the #1175430844685484042 so far haven't found a one that wouldn't voice-crack while singing this song yet...

[tacit falcon]

cubase

[tired jasper]

what RVC index file do exactly?

[tacit falcon]

.

[feral halo]

Ayo? @tacit falcon level 10 !!!

[snow briar]

is there any trick to generate output that is less likely to have voice cracks ? It looks like all the Gura models have the same issue with this one song. I don't know about the settings like Voice conversion options and Audio mixing options so I would really appreciate it if anyone can help me out in this part. Thank you

[tired jasper]

clear input

[feral halo]

Ayo? @tired jasper level 2 !!!

[tired jasper]

try put input in adobe podcast before RVC

[snow briar]

So I should get the acapella version of the song ?

[tired jasper]

for sure

the most cleanest possible acapella

any noise in your input will affect negatively the output

[snow briar]

Also which Gura model in your opinion is the best one rn, in an ocean of countless Gura models out there ?

[tired jasper]

idk who is Gura, let me search

[snow briar]

Sorry, I mean Gawr Gura (Hololive Virtual youtuber)

[tired jasper]

but one thing is true: bad input got usable outputs for me in good trained models, and very bad outputs in bad trained models. When i put good input, the bad trained models became usable

models with more epochs should be better if not overfit

[earnest dragon]

so a weeb

[tired jasper]

@snow briar what music are you trying?

[snow briar]

"Legacy" ( DMC 5 song)

[snow briar]

yeah

There is always a part near 1:05 that cracks inevitably

[polar flax]

inputs with background noise, sounds, polyphonic, reverb, and other distortions can ruin the output results

[snow briar]

In my case it's just the original song audio and not a cover or something

can't find an official acapella version of it anywhere on the internet though

[polar flax]

u need to separate from instruments, back vocals, reverb, noise. with good separation models in mvsep.com, UVR, etc can produce good enough outputs

also ideally avoid duet songs as there's no perfect model enough for that

[snow briar]

This requires strong GPU right ? If so rip for me because all I have is a weak laptop that can't even run AI art models

[polar flax]

do online in mvsep.com
there are also huggingface spaces for UVR, and also Flux space for AI art

[snow briar]

Thank you, I was under the impression that those online AI services that require GPU usage would be locked behind some sort of subscription paywall Lol

[solar agate]

hi skibidi rizzlers!

[polar flax]

huggingface spaces use shared ZeroGPU, if u hit usage limit, try again in another hour

[solar agate]

hi skibidi ohio gyatts!

[tired jasper]

@Zladin i'm doing for you, just show me the real voice tone of gawl

@snow briar

[snow briar]

Sorry, I just found out that it was an issue with the input, not of the AI models

[tired jasper]

sorry sending DM before asking

i can't send files here, and even in DM without compressing it and split

it's 16MB large

i'm impressing there's no channel to share audio

how should i post this?

solved

#🔊│ai-development message

[sharp karma]

Hii how do I use the AI links ( in this group) to make some people sing the song I want?

[covert lake]

so, make an ai cover right? Whats ur pc gpu?

[sharp karma]

I can't use pc so I'm using my phone

[covert lake]

alright, you could technically do it locally on ur phone (directly running on ur device) but would be very slow on CPU

Its way better you use cloud (remote good pc)

use ilaria rvc zero which is the fastest for inference (using models)

[sharp karma]

Okay,how do I use the links?

So*

[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]

follow the ilaria rvc zero guide

[sharp karma]

Ah Tysmm

[covert lake]

yw

[sharp karma]

Did I just got Rick rolled at the same time 😭💀

[covert lake]

Yes, I had to add that too

ofc its a real guide, i just added a rickroll to be silly, u can see below

[sharp karma]

Last question I think,how do I change the model?

[sharp karma]

Lol 😆

[covert lake]

you should download it in the model loader tab first, then in the inference tab click refresh models and select ur model

[sharp karma]

Ok

Tysmm

[rapid heath]

What's the difference between gemini vs gpt4 for coding, say a range from 1-10

if you were to rate them

[covert lake]

yw

[covert lake]

depends on which gemini, on the 1.5 pro id give it a 8, i use it mostly for coding as it got long text context, i use the google ai studio

[rapid heath]

Idk what im on but I just installed it's plugin on android studio, so that's prolly the free one

you used gpt4o? would you say gemini 1.5 pro is better? or they are around the same

[sharp karma]

Ah thank you so much I did it 🙏🏻I only couldn't find an AI voice for one singer that I needed

Tysmm

[covert lake]

yea i used that too, but the bad thing is the context lenght isn’t as long atleast in the free tier, but its good too

id say for short thing use gpt4o, for long ones use gemini pro

[covert lake]

oh i used the one in the google ai studio

[thick yew]

Has anyone noticed how easy it is to bypass chatgpt filters?

All you need is a thesaurus tbh

[earnest dragon]

which one

[thick yew]

4o and 4o mini,
Sometimes it just goes off the rail into AO3 level stuff

I don't know if it's just my account, but why is it so horned out. The trick is to give it an inch. It will take it 10 miles

[acoustic trench]

What’s the best video ai? Preferably an image to video sort.

[earnest dragon]

-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

[tropic carbon]

why is it that WITH ALL voice changer when i speak at the end it kinda twitches/echoes (the voice)

[covert lake]

ask in an help channel #🔍│help-w-okada

[sonic spindle]

hello

[thin gazelle]

hi

[reef ermine]

hi

[grand mountain]

.getkey

[feral halo]

Ayo? @grand mountain level 1 !!!

[royal laurel]

quick question whats the best ai tool to generate voices, i want engaging voices tha sound super realistic

[grand mountain]

7r mom

[covert lake]

You want to make voice models? What's ur PC GPU?

[royal laurel]

no i want online tools that convert text into voice

[feral halo]

Ayo? @royal laurel level 1 !!!

[royal laurel]

but i want realistic ones

[earnest dragon]

then GPT-SoVITS

[covert lake]

Online tools depend, it would be better you tell me ur PC gou

If ur PC is good enough u can do it locally so have to not worry about GPU time

[royal laurel]

is it easy to use? id prefer an online site thats quick i dont want to download anything

[royal laurel]

1650ti

[covert lake]

Yeah nvm

[covert lake]

Realistic and emotional?

[royal laurel]

any online tools yk?

[earnest dragon]

you to trash for ai

[earnest dragon]

paid there is elevenlabs

[royal laurel]

something that sounds like a real human and is engaging

[covert lake]

Gpt so vits is the best then, u want custom voices or just random tts?

[vapid oasis]

hi

[royal laurel]

depends on the voice

[covert lake]

Gpt is for voice models

[royal laurel]

eleven labs or hugging face?

[covert lake]

It's the best out there

[royal laurel]

chat gpt?

[earnest dragon]

no

[old seal]

Hey guys I am looking for someone who is interested in working on AI Models so we might work together.

[earnest dragon]

and there is this one paid zero shot tts I have that even works if the audio has background noise but the background noise sounds low quality but speech sound good.

[earnest dragon]

what ai models

[covert lake]

Got so vits

Lemme share u all ways to tts

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

[chilly lake]

@covert lake bark tts listed as best TTS is a big question

[royal laurel]

whats the difference between 11;abs and fishspeech?

[covert lake]

11labs is better but paid, fish speech is a open source 0 shot tts, easy but not as good as gpt so vits

[royal laurel]

i dont mind paying, Thanks

[covert lake]

It's good for emotions

[old seal]

Idk if I might post video of my work here?

[chilly lake]

but it is not good for any text

[covert lake]

Can't listen rn not home

[chilly lake]

1- 'to take a.... " then a sound turns into a propeller plane flying overhead for 5 seconds '... then the text continues

2 - random music cutting in

[old seal]

yeah I cant post my vid so..

[feral halo]

Ayo? @old seal level 1 !!!

[chilly lake]

if XTTS is simply hallucinating from time to time, then bark is an LSD trip from start to finish

CogX 2b, not even 5 - 06:48:47-256343 ERROR CogVideoX: CUDA out of memory. Tried to allocate 35.31 GiB. GPU

trying with less frames

[covert lake]

Wtf never had those type of issues

Did you use a random preset?

[chilly lake]

I'm not sure.. I've never managed to get anything useful from bark

[covert lake]

I remember it had issues with random presets

That's why it's better to use one

[chilly lake]

hm.. not sure about presets at all, I used it from coqui

[chilly lake]

and using a reference file

[covert lake]

I installed it using their package

I mean Google colab lol

[covert lake]

Uhh not sure about how coqui works, but it has presets voices

[tired jasper]

i need adobe podcast free alternative (opensource if possible), already tried Apollo and audiosr, both don't have same performance than Podcast

[covert lake]

Oh yeah u might mean 0 shot as in https://github.com/Nick088Official/bark-gui-fix/

I mostly used the presets voices which worked fine

[chilly lake]

lemme try again

from coqui that is

preset or 0-shot should not matter, it would just compute the speaker latents and embeddings

[covert lake]

Other than resemble enhance and hifigan + bwe (https://discord.com/channels/1159260121998827560/1205637385245958215) I don't think there's anything else to try

[polar flax]

performance? u mean speed or quality?

[tired jasper]

i had no luck with resemble enhance, it cropped all my audio

[tired jasper]

both if possible, but i'm focusing in quality

[polar flax]

note that Apollo (mp3 enhancer) was literally trained on original lossless dataset and their corresponding mp3-compressed one

[tired jasper]

isn't hifigan the base model for any other?

[polar flax]

well no perfect solutions so far, but u can denoise/clean the result audio from noise/artifacts (though still not ideal for making RVC datasets)

[tired jasper]

sure, all restoration model is trained like that, but the bad news is that it learns about a pattern of compression and become very bad in uncommon patterns that wasn't in the dataset

[covert lake]

Weird tho

[polar flax]

I hope the Apollo devs keep improving it, since it's still new

[tired jasper]

i'm trying to avoid denoise audio manually 😦

[covert lake]

Well rip then

[chilly lake]

Apollo trains on wavs it damages itself by compressing them into mp3

[covert lake]

There is no other open source upscale option

[chilly lake]

but I did not see any specific compression levels it uses

[tired jasper]

oh oh, synthetic dataset 💀

[chilly lake]

CogXVideo2b

[feral halo]

Ayo? @chilly lake level 19 !!!

[tired jasper]

that's what i'm talking, any other way of damage will be ignored

i guess Apollo can be finetuned to remove noise

the problem is to finetune by myself

lol

[covert lake]

I remember they said it was memory efficient

[polar flax]

seems more like it may need a hundred day length of dataset

[tired jasper]

maybe not too much

[covert lake]

Pirate adobe podcast 🔥

[tired jasper]

mission: find the most vary dataset of bad quality speech

with ground truth

💀

could be synthesized from adobe podcast itself if i had paid version

[feral halo]

Ayo? @tired jasper level 6 !!!

[tired jasper]

i will need to merge noises and normal speechs to create the LQ

someone want to join me?

[copper quiver]

hi
anyone can help me to get the right one of MMVCServerSIO to my rtx 3060 !!
Thx u

[chilly lake]

@covert lake ran another test with 8 different references. 1 - female voice instead of reference male, 2 - stuttering and repeated parts of the phrase, wrong voice, 3 - wrong voice, 4 - wrong voice, 5 - wrong voice, radio broadcast noise, 6 - female voice again, 7 - wrong, 8 - wrong

[gaunt sable]

yo guys. how do i make an ai voice on huggingface

[gaunt sable]

anyone

[covert lake]

You can't train voice models on HuggingFace spaces, the spaces are made for inference (using models)

[covert lake]

Weird, might be it's not good for 0 shot, I will try with presets in some mins when I get home

[gaunt sable]

how do i get the model

[covert lake]

Do you want to search for already pre made models or make one yourself?

[gaunt sable]

i wanna make one myself

[feral halo]

Ayo? @gaunt sable level 1 !!!

[covert lake]

What's your PC GPU?

[covert lake]

You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU

[gaunt sable]

ok

12th gen (R) core (TM) i5-12500H

[covert lake]

that's a cpu

be sure to check the GPU tab

[gaunt sable]

i dont see GPU

[covert lake]

you don't see any GPU 0 or GPU 1 in task manager?

could u send a screenshot in #✨│ai-help

[covert lake]

Hello, my friend! [laughs] how are you?
Voice_Preset:
v2/en_speaker_6

but yea added that 'umm'

[chilly lake]

so it hallucinates too

give me the plain banana business ```I want to take the first steps towards establishing my own banana business.
The plan is straightforward: find a suitable location along the coast with ample sunlight and fertile soil, invest in high-quality banana plants, and set up the necessary infrastructure for irrigation and maintenance.
Marketing will be key: I'll focus on spreading the word through various channels like social media, local markets, and forming strategic partnerships.
While there will undoubtedly be challenges along the way, such as pest management and navigating market fluctuations, I'm committed to facing them head-on with resilience and determination.
Above all, this venture is about pursuing something I'm truly passionate about and building a legacy that reflects that passion.

[covert lake]

yea just noticed

[covert lake]

u want me to inference that?

im inferencing another test rn

666

God, 'm so bored today
v2/en_speaker_7

[covert lake]

okay that hallucinated badly

will add an hallucinating warning to bark tts, thx for telling me

[chilly lake]

bark is the worst of them all

tortoise is slow and 0-shot does not match the voice at all

[covert lake]

nah the worst is gtts added just cus that index was of the tts i played around with

[covert lake]

oh ye i should deffo delete that, xtts is a better fork of it

that guide is like a year old tbh

gonna remove gtts too while at it

[copper quiver]

guys what is ONNX

ONNXモデルファイル

[drowsy shore]

Hi

[frank comet]

Hey ppl

[hazy storm]

can some one turn https://www.youtube.com/watch?v=lpmTabF4FVA&list=PLQU2xjVsgNWtlq7maEKmdR3RVi-1wBpdE&index=3 in to a mp3 please

[feral halo]

Ayo? @hazy storm level 1 !!!

[hazy storm]

can some one turn https://www.youtube.com/watch?v=lpmTabF4FVA&list=PLQU2xjVsgNWtlq7maEKmdR3RVi-1wBpdE&index=3 in to a mp3 please

[onyx mantle]

sup

[copper quiver]

here i can have ONNX

[feral halo]

Ayo? @copper quiver level 4 !!!

[copper quiver]

[covert lake]

don't talk about NSFW here.

[covert lake]

don't ask the same question multiple times, just use yt-dlp or any other youtube video downloading site

stop asking the same thing over and over 3 times

its not even an help channel

[hazy storm]

ok

[ionic pumice]

good evenin

[manic lodge]

no ??

[ionic pumice]

morning? afternoon? noon

[covert lake]

#1159290752195633273 , or else your promos will be deleted in every other channel

[elder willow]

what software do you guys use for ai covers?

[covert lake]

RVC,
What's ur pc GPU?

[vivid hound]

Hi guys, someone please help me with how to change webcam settings in facefusion. its stuck on my inbuit webcam but i have a better one already connected, works fine in other apps @covert lake

[elder willow]

rtx 4050

[feral halo]

Ayo? @elder willow level 1 !!!

[covert lake]

laptop gpu? and how much memory

[elder willow]

laptop 16 gb

[covert lake]

You should be able to do it locally (on ur pc), you can choose between:

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

[elder willow]

tysm

[covert lake]

yw

[vestal plaza]

gm

[zinc salmon]

someone experiencing input lag?

how to fix it?

[polar flax]

get better monitor, kb/mouse

[elder willow]

Is this worth it https://rentry.co/egirlmodel

25$ better be

[opaque coral]

supp

[solar torrent]

It's been a year since I last talk on an AI server.

[tacit falcon]

welcome back bro

[solar torrent]

I've done remastering the entire brainrot nursery rhymes EP by the Rizz Records since yesterday.

[solar torrent]

Thank you.

[ionic flower]

with what ai voices?

or did you used your voice ig

or...idk

[solar torrent]

Eh, what I meant is to remix the track stems to get louder. The AI cover is another project.

[feral halo]

Ayo? @solar torrent level 1 !!!

[ionic flower]

oh ok, well, i have some plans for my ai covers

[elder willow]

in training I don't understand the dataset step, it tells me to place the training folder but I don't have one, do I have to create it or use one from the rvc?

[solar torrent]

I've also done getting like 72 people (AI models) to sing each bar of BBL Drizzy by Metro Boomin on one single track.

[copper quiver]

how can i get ONNX voice

why no one answer !!

[solar torrent]

#✨│ai-help

[ionic flower]

fake scam ad

[echo meadow]

Hey may you help me with the accents?

[chilly lake]

how?

an accent is what happens when a voice model has no features that match the source feature exactly

index search=0 eliminates most of it

[solar torrent]

The second one sounds like some voice TTS from 20 years ago.

[chilly lake]

well, it is a tts inferred by a voice model

other way around

[echo meadow]

I am training a model on my voice but I have Indian accent, is there a way to remove an Indian accent and add american/british accent?

[chilly lake]

same way

use an audio with british accent, infer it with 0 index use

it will be your voice with british accent

there ya go

or...

alternatively get a british model index

[chilly lake]

here ya go

[elder willow]

guys is astra labs down

[elder willow]

@earnest dragon May I dm you?

[ashen notch]

I drive

[echo meadow]

hey may I send you my model so you can check it?

[chilly lake]

okay

[echo meadow]

I guess I need to explain the workflow I am looking for:
I have trained a model on my voice with an Indian accent.
now I am recording my own voice like I am narrating something or singing a song,
now I am trying speech-to-speech where if possible I am looking for a way to change the accent of my voice model, so I can use my voice but with an accent.

but I guess it is not possible

[chilly lake]

the accent comes from an index.. like in the example above I used an english text and an english speaker, but an index from a russian model with 1.0 index search, so the output has a russian accent even though neither original or voice model had an accent

[echo meadow]

for testing sake may you please give it a try if I send you my voice model?

[chilly lake]

okay

[dense pumice]

hi

[echo meadow]

this is awesome, the results are quite impressive

[chilly lake]

used https://huggingface.co/Coolwowsocoolwow/Elevenlabs_Stern_British_Man/resolve/main/Elevenlabs_Stern_British_Man_v2.zip https://huggingface.co/Coolwowsocoolwow/Random_Indian_Tutorial_Voice/resolve/main/Random_Indian_Tutorial_Voice.zip

[twilit apex]

lol

how am i level 1 again gang

and what happen to my server profile

[copper trellis]

hey guys

[sleek phoenix]

hello

[green pebble]

Anyone here make music or know of someone who does because I'm wanting to make videos I just can't with copyright ever video

Can someone teach me how to use the ai?

[tropic carbon]

why is it that WITH ALL voice changer when i speak at the end it kinda twitches/echoes (the voice)

[feral halo]

Ayo? @tropic carbon level 1 !!!

[halcyon yew]

I make a bit of music using Suno; you don't really need training per se you just need a good ear for what is good output and what isn't. Also don't just take the gen as is. Be picky. Extend it at the moment you stop liking the gen and keep it up until you find a good gen, then extend some more. You may look at the lyrics for each song in my public list for a good idea of what you can prompt engineer to produce many styles: https://suno.com/@chemicalor9

[cunning estuary]

anyone know an ai tool i can use that will take an mp3 and put an ai pth voice over it

[covert lake]

so, inference right (use models, in this case rvc models)?

or you mean making models (training)

[green pebble]

U using to many big words for my small brain to understand

[feral halo]

Ayo? @green pebble level 1 !!!

[gritty cipher]

for which program me say settings

[real thunder]

hello everybody

[covert lake]

what

for help ask in help channels

and be specific

[cunning estuary]

yes, using models, i already have the trained model i want to use, but i want to upload like a clip of me talking or something and get an mp3 back of the ai talking, i know theres live voice changers and i have one, but i want to know if what im describing exsists

[covert lake]

what's ur pc gpu?

and ye i understand, using models on pre-recorded audios

in case ur pc gpu is good enough, u could do it locally (on ur pc), else, u have to use cloud (remote good pc)

[ripe spade]

I've become interested in TTS, is there any software I can use to create custom TTS voice models (no RVC over TTS files)

[covert lake]

GPT So Vits is the best one in quality, its few-shots

while fishspeech is easier but its 0 shot, so its not as good

https://docs.ai-hub.wtf/tts/gpt-sovits

[ripe spade]

any colab links?

[covert lake]

afaik #1229627925658337360 message

@twin fractal are all the gpt so vits colabs updated ?

would be cool a kaggle or lightning.ai port

[ripe spade]

thank you

[covert lake]

i dont really do gpt so vits

im not sure if those colabs are updated, could give it a try tho

bc i dont see that there are any other colabs of gpu so vits 😭

last update seems 6 months ago

[cunning estuary]

i train my own ai models if that answers your question (mangio rvc)

[feral halo]

Ayo? @cunning estuary level 1 !!!

[cunning estuary]

@covert lake

[cunning estuary]

my goal is to do it locally

[earnest dragon]

mangio is outdated

[cunning estuary]

well it works

[covert lake]

You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU

also yea mangio is outdated, id suggest mainline or applio

[earnest dragon]

use applio instead which isn't outdated and has all the stuff you want and also use titan pretrain when making models

[real thunder]

im trying to create a text to speech program that uses real voices

and I really need help

[cunning estuary]

IAHispano applio?

[covert lake]

yes

could u check ur pc gpu tho?

[cunning estuary]

also i have an rtx 3050

[covert lake]

do u may remember the memory gpu? and if its a laptop gpu

[earnest dragon]

3050 not good enough

[earnest dragon]

yeah is his the 6gb or 8gb or what

[cunning estuary]

8gb

[covert lake]

mm, is it a desktop or laptop?

laptop version is shit

desktop could be good enough

[cunning estuary]

desktop

[covert lake]

yea technically should be good enough

[earnest dragon]

yeah but speed won't be the best

[cunning estuary]

thats fine

[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

[cunning estuary]

i usually do this stuff overnight

[covert lake]

he said his goal is to do it locally

[earnest dragon]

yeah I know

[covert lake]

btw don't ask my why but i have a feeling u use linux

[cunning estuary]

what do you suggest

[feral halo]

Ayo? @cunning estuary level 2 !!!

[covert lake]

honestly, they both are the same quality, the main difference is the UI (Interface), applio interface may seem easier

[cunning estuary]

alr thanks

[covert lake]

your welcome

[earnest dragon]

no because most of the ai stuff and audio stuff I use is windows only

[covert lake]

dunno why u gave me linux vibes then 😭

I use w11 too even if id give it a try to linux too

i mean used linux for cloud computing services anyways

[earnest dragon]

I use windows 10 enterprise ltsc and for cloud I still use windows

[covert lake]

oh lol

[mental burrow]

.

[lyric obsidian]

hello everybody

[earnest tree]

someone has a good german ai voice ? egirl?

[mighty atlas]

Hello everybody

[gloomy linden]

hi everyone

[real thunder]

@gloomy linden sup

[covert sequoia]

Does anyone know why when I try to make a ai cover on hugginface it gives me an error “'NoneType' object has no attribute 'setdefault'"” ?

[topaz orchid]

hello everyone

[lilac cloud]

hello , where can i find an ai service without all the "i cant help you with that" ?

[earnest dragon]

there really isn't any unless local or maybe freedom gpt but not the best and paid

[lilac cloud]

okay

[covert lake]

send the huggingface space link u are using

[elder willow]

guys wheres the website

[earnest dragon]

what website

[spark agate]

anyone know if udio uses rvc

or is there tech different

[little pelican]

sup guys

[earnest dragon]

it doesn't use rvc

[spark agate]

do you know what they use

[little pelican]

anyone can provide me some AI voice model generator?

[covert lake]

so, train (make voice models)? whats ur pc gpu

[covert lake]

what website

[little pelican]

yeah, I have an macbook. I already tried audiomodify but no luck

[feral halo]

Ayo? @little pelican level 1 !!!

[covert lake]

dont use audiomodify, all those scam sites just use a paywalled RVC

RVC is the open source program

however, macs arent powerful enough to train rvc models locally (on ur pc)

As you dont got a good PC, its better you use cloud (remote good pc) for training an RVC Voice Model:

• Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
  • Applio (ui)
  • Mainline (UI)
  • RVCDISCONNECTED (no ui)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
  • Mainline (UI)
  • Applio by Vidal (UI)
  • Applio by Shirou (UI, no guide as of right now)
• Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly):
  • Mainline (UI, No guide as of right now)
  • Applio (UI, No guide as of right now)

Google Colab = Easy but low gpu time

Kaggle = Hard but wayyy more gpu time, much more suggested

[little pelican]

thanks sir

[covert lake]

Yw

[icy steeple]

So I'm thinking of commissioning a new model of my voice, I have 2 questions:

How long should my sample audio be?

And should I use post editing to remove stuff like reverb and etc?

[covert lake]

So paying someone to make it for you? #1159289738314919936 or #1191429836321849435
The sample audio could be around 10 mins, and yea u should remove that stuff, unless the person ur paying does that for u

[covert sequoia]

Been using this one https://huggingface.co/spaces/r3gm/AICoverGen

[covert lake]

yea its broken

use ilaria rvc zero instead

[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! 💖

[real thunder]

good evening everybody

[covert sequoia]

Okay bet thanks bro

[covert lake]

yw

[earnest dragon]

isn't applio zero gpu version better

[covert lake]

why would it be better

all rvc versions have the same quality whatsoever

[real thunder]

I was wonding if someone can show how to get started with text to speech or whatever that gets done here

[earnest dragon]

I was saying split audio and also the mp3 enhance

[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

[earnest dragon]

and there is this other tts not listed that is free to use but not make voices but paid users like me can create the voice for you to then use on the website

[real thunder]

this link isnt working: https://docs.aihub.wtf/tts/gpt-sovits/

[covert lake]

yea the docs are temporarely down, its https://docs.ai-hub.wtf/tts/gpt-sovits/ for now

[real thunder]

is this on youtube?

[earnest dragon]

gpt sovits is a model you have to train like rvc

[solar compass]

is it normal for an Applio training model to eat up disk space until its full? (60gb logs folder)

[covert lake]

which can't work bc it slows the inference that much that makes it go over the 1 min max inference time (time of inferencing, not time in the input audio) set in the zerogpu duration

so u cant upscale anything

zerogpu spaces on default can only take 60s to inference smt, that value could be changed by the creator but will eat ur acc quota faster tho

[earnest dragon]

no only mp3 enhance breaks but split works perfectly fine

[covert lake]

split vs no split

i never really heard of chunk inference making the quality better tbh

i dont seem to hear any difference, i used some random model and audio that was already inputted

[earnest dragon]

I think the pitch and stuff works better when doing that

[covert lake]

you really sure? i'm not able to see any difference between those files

its even the same size

[polar flax]

it mostly depends on batch size and vram constraint, in this case up to batch 8 for 8 GB vram (assuming in default fp16), though 3060 12 GB is better value given his budget range

[earnest dragon]

batch size is not always equal to vram 8gb gpus can usually run 10 batch size

[polar flax]

yea im talking about amount of vram usage for particular batch size, though for instance using batch more than 8 on 8 GB vram capacity may bottleneck the performance, but not throw OOM error (similar to using ultra texture and other demanding game settings that may use more than 8 GB on a 4090 and cause stuttering/texture popping on 8 GB/less card)

[covert lake]

#1159290752195633273

[ionic flower]

I wonder if pink guy song covers are ok here (for the rules to be exact)

[earnest dragon]

yes but only allowed in #1159290752195633273 and most be uploaded to YouTube or whatever.

[ionic flower]

cuz im doing one song called "ramen king" and it has some questionable lyrics

[polar flax]

upload to youtube and add content warning

[ionic flower]

its still a wip

[arctic eagle]

who knows how to create video (easy) to show: power and it's going up with percentage from 0 to 100%, strength, stamina etc with some background?

[pale swift]

what