🧬│ai-chat

[green girder]

EVERYBODY KNOWS
IM NOT ICE IM JUST COLD

its not terrible ngl

why you spammin now

[split whale]

Bro will get muted before GTA 6.

[green girder]

i mean idc too much but it is annoying when i have a genuine req

[covert lake]

https://tenor.com/view/thanos-finger-snap-disappear-gif-13174976

[sweet oar]

i need a spanish voice

model

cant find any 😦

[split whale]

.

[feral halo]

Ayo? @split whale level 13 !!!

[split whale]

Okay.

[gray rover]

@river adder Am I allowed to make an annoucement here? or do you have a better idea perhaps

[river adder]

Ping ppl? Nah but you can write down whatever you'd like it's fine yeah

If you wanna show your fork then sure I don't mind at all

[gray rover]

I wanted to ping like 2-3 people from ai testing

[river adder]

Yeah yeah works totes cool

[gray rover]

Alrighty, thank you!

@river wigeon @covert lake @night lake

announcement
My experimental "rvc 3.0" -ish fork is almost finished.
Purpose: The fork's a step forward towards creating better higher quality pretrains from scratch.
Features:

• Usage of 3 discriminators: MPD ( classic rvc's ) + mssbcqt + msstft
• Added warmup logic
• Supports multi-speaker-embedding training ( and so, preprocessing / extraction per speaker ID )
• Uses Ranger2020 optimizer, consisting out of:

RAdam as a core ( which is more stable than AdamW, esp in scenarios when warmup's used. )
Lookahead; Sorta, helps a model make better predictions by allowing it to see future parts of a sequence before making a decision.

• Easy switch between classic 1 disc ( mpd ) approach and new 3-disc combo
• Option to switch the amp to use Bfloat16 instead of fp16
  Todo:
• Improve the webui / ui.
• fix the mess I've made in files lol 🙃
• Confirm the effectiveness of new approach
  ( at this point it's almost certain but last confirmation's required. )
• Preparation of required datasets ( I'd aim for anywhere from 50 to 100 hours (( rvc used 30h )) )

Strategy:

1. I'll need to request some support for funding the from-scratch training of pretrained 48 kHz models.
   ( Gonna need shit ton of space + at least 4 or 6 A100s ( 3 discs are very memory inteisive )
2. Once a suitable environment for testing the systems on bigger sets ( 5-10h + ) is available, testing / prototyping, incl. hyperparams will start.
3. As soon as good or viable " base " hparams for training are found, the big-scale training of pretrains gonna start.
4. At that point, the actual main training could take anywhere from a week to a month ( excluding dataset collection / preprocessing time + prototyping phase )
   ( As for exact total time, I don't want to make any empty promises as I can't really predict it atm. )
5. Finished new pretrains gonna be released for everyone.

For more details @ me in #🔊│ai-development

Also, couldn't fit in within the msg but, ranger also has gradient centralization which further stabilizes the training ~

[night lake]

oooo

[gray rover]

What you think?

[night lake]

im hyped

wish i can help but im a brokie 😭

[gray rover]

Sadly, I can't test anything " more big " just yet as my setup and my methodology is different from noobies'

[gray rover]

That's all good :' ) Me too 😩

[night lake]

i really want this to become a thing so if you need help with anything (not money related) i will 1000% be willing to help

[covert lake]

bro's cooking

[gray rover]

Noobies having some neat results too

So in general, I am having high hopes for our project of " buffing rvc " lol
As much as our methodologies heavily differ, the goal's the same

[split whale]

Wrong message to reply to moment.

[covert lake]

real

[gray rover]

Once and if, at any point I have prototype-stage pretrains, gonna ask you to test it on some of your datasets
if you wouldn't mind

provided.. your gpu's gonna handle it : D

[covert lake]

my laptop is NOT gonna handle this 🔥

[night lake]

i dont mind at all, just know that i mostly have asmr (egirl) datasets

[gray rover]

That'd work too definitely
Core premise of new pretrains is to be super robust
esp for whisper, laughing and so on

[night lake]

i got a 4060 ti 16gb

[gray rover]

Oh ye, then you'll do

I operate on 3060 12gb sadly and to put it into perspective, from scratch using batch 8 on set 17-19 mins

I get OOM, Most likely it eats approx 12.7 to 15 ~ gb

[night lake]

oof

[gray rover]

But that's different if you have pretrains used ( + then there can be some tests on mixed precision training (( on which we'll see in future

[weak schooner]

hi everyone btw check bio

[split whale]

Nah

[covert lake]

[split whale]

An idiot.

[covert lake]

you give me Applio vibes btw

[split whale]

🔥

[covert lake]

everyone that works at Applio has a cat pfp like that

lol

[split whale]

Oh.

I forgot what the url was.

[river wigeon]

So if I’m reading this right is this right now for pretrains ? And to make them better? and not for single models?

Or can it be used for single models , this is fire

[gray rover]

It's still viable to be used for single models ( finetuning ) yes, the core mechanic of using existing pretrains works

yet, it now supports making pretrains with 3 discs ( new arch ) and stuff I mentioned

[split whale]

Donetsk army be like.

[trail fjord]

Hello Everyone

[lucid ibex]

Hello

[tacit widget]

Been out of the loop with voice models for a while. I use RVC 1006 NVIDIA for RVC. What should I use if I want to take my path & index files and use them with TTS. Locally?

[lucid ibex]

I don't know

[tacit widget]

Then no need to respond '-'

[lucid ibex]

My bad

[weary temple]

Bro who are u

[tacit widget]

Perhaps I'll try applio 🤔

[tacit widget]

Someone who just happens to share your name

[weary temple]

He got a dumb bio

[tacit widget]

There are quite a few "AC"s here

[weary temple]

Did check it

[tacit widget]

Looks like his account was taken control of and they were trying to spread malware or nasty words via bio.

[weary temple]

Fr

[lucid ibex]

yeah

[tepid basin]

Ye applio is good for RVC

RVC TTS kinda sucks because it usually takes a default TTS and just runs the model over it. A lot of the time doesn't sound natural for the target voice

You should look into GPT-sovits. Its totally different than RVC, different models, but its designed for accurate TTS. Unfortunately I'm not an expert so you'll need to use the search bar in this server for more info

[tacit widget]

Thanks, I'm looking to run a local "glados" ai lol

[tepid basin]

I'm sure there's already a GPT-sovits model available of glados lol

[covert lake]

I remember seeing one lol

[tepid basin]

#1175430844685484042 has some gpt models I think

And probably https://weights.gg

[languid coyote]

guys check out my latent toys my friend made this website

I'm usually on t/latent

fun lil proj we're doing to mess w/ ai and social media :)

[covert lake]

#1159290752195633273

please use #1159290752195633273

[covert lake]

Rigby don't promote here next time btw

[quaint ibex]

guys is there a tool i can use to upload an audio file and the software converts it into the voice i downlaoded from here? If so, what is it called???

[covert lake]

yes RVC

What's ur PC GPU?

[quaint ibex]

the tool is just called RVC? I have a 3080

[covert lake]

Also, do you mean convert the Audio File to the Voice of the model, or converting the Audio File directly as a RVC model?

[quaint ibex]

I need to upload a song into a tool, have the thing convert the song file into the voice and then i can download the voice version of the song (sorry if this doesn't really make sense)

[feral halo]

Ayo? @quaint ibex level 1 !!!

[quaint ibex]

I found a youtube tutorial, Thank you for telling me the software!

[covert lake]

No

Youtube tuts are super outdated

don't follow them

[quaint ibex]

it worked

[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, kinda faster and simpler but same quality tho
• Mainline: The original RVC

Its really not suggested to follow yt tuts

[quaint ibex]

oh ok

[covert lake]

most use mangio or rvc gui which are extremely outdated tbh

[quaint ibex]

ill try mainline then

[covert lake]

Applio is easier tbh

[quaint ibex]

ok

[tacit widget]

Where does one put the models for applio

[elder willow]

Yk was gonna help

But after reading that about me no more 💀

[tacit widget]

its lyrics to a song

lmao

[tight olive]

I know how to produce, mix and arrange but i suck at writing lyrics 😭

[elder willow]

Ngl same but idk how to produce or mix or arrange either

[wind stratus]

man same, im yet to develop that skill and most of my vocals are just mumbling for now

[tight olive]

Fr, i feel disabled

[tight olive]

Fake it till you make it, i'm self-taught and learned everything just by experimenting and trying

Courses is useless tbh

[feral halo]

Ayo? @tight olive level 1 !!!

[solar torrent]

I know how to mix and mastering a track.

[split whale]

about:blank

Best website.

[indigo sentinel]

hello

[tight olive]

It saves everybody's back

[split whale]

https://www.youtube.com/watch?v=nc9kq9ZkNV8

Real.

The comment:

[gray rover]

lol

[feral oar]

EYYY

BAD BUNNY

[forest crypt]

good luck

N_C_like

[gray rover]

Thank

[split whale]

I wonder how many members here use Vencord.

[gray rover]

Good question

I personally use vendetta but that's obv mobile ( miss aliucord rip

[surreal cargo]

Hi

[solar torrent]

Sumireko Usami looks like a girl from a hentai manga.

[gray rover]

I ain't into touhou so gimme a short briefing, which girl is the best

[solar torrent]

Reimu Hakurei is the cute woman.

https://tenor.com/view/reimu-hakurei-reimu-hakurei-reimu-chibi-gif-10765846393329744520

[gray rover]

hmmmmmmmmm 👀 Reimu, Sakuya and Marisa imo the hottest

But, Reimu wins with colors

[solar torrent]

Marisa Kirisame is the thickest woman of the game.

[vital snow]

Hi

[gray rover]

👁️ 👄 👁️

[solar torrent]

https://tenor.com/view/marisa-gif-23277791

[polar flax]

https://cdn.discordapp.com/attachments/1159289354439626772/1307009211854753792/mumei-nanashi-mumei__.gif?ex=6739676f&is=673815ef&hm=335cb80692afe35a873beef3fbe51ac909201c7058e7e9767eb84d96e18a4157&

[solar torrent]

https://tenor.com/view/blue-archive-kuromi-serika-serika-fumo-plushie-gif-3686811377548363732

[gray rover]

so many kuru- x_x

[steep dune]

do any of yall know in RVC what the difference is between client and server for the audio

[gray rover]

that ain't rvc, that's w-okada

🥴

[steep dune]

damn my fault

i dont know the difference tho im ngl

i was just curious about the button

[gray rover]

rvc is for making models and inferencing
w-okada and realtime voice changer ( is within rvc's codebase itself ) are the voice changers

Shad is the specialist in voice changers so I'd go to #🔍│help-w-okada
and ping em there or generally, asking there

[solar torrent]

The client is when someone or you use RVC service on their or your own device. While for a server, you host RVC on your PC or cloud and let someone or you to use RVC remotely.

It's kinda similar to Minecraft. If you got a client, you can play a Minecraft world on your own. If you got a server, you host your own world on your PC, and let anyone to play your world from different dimension.

[solar torrent]

#1159290752195633273

Bro really got that ChantGPT instead of ChatGPT.

[gray rover]

[split whale]

https://cdn.discordapp.com/emojis/798863281086660628.webp?size=48&name=harold_crying

[solar torrent]

Rimuru Tempest looks cute in Arona outfit.

[split whale]

https://cdn.discordapp.com/emojis/1208376812858773514.webp?size=48&name=okuhuhh

[grizzled pond]

is it a virus?

I imagine not

[split whale]

huh

[pearl temple]

hiii

[stable furnace]

Snoukavumi

Snoukavumo pest babouin

Ou ressemble un zef cheval

[vital snow]

I need translator

[gray rover]

same

[final bramble]

People on instagram claiming to make faceless channels and create content completely by AI is this even true and possible.

[gray rover]

yea it is

a lot of yt-shorts based channels do that in fact

[final bramble]

but almost every AI is paid

[gray rover]

I mean yea, but you didn't mention claimings of it being free so

Most pay sure, but gain even more

[final bramble]

Yeah you talking about the shorts of podcasts?

[gray rover]

nah, just yt shorts

gpt generates ideas or scripts, rest goes with other elements

[final bramble]

suggest me some good names that are free

[pale elm]

Hello I'm new here

[gray rover]

Can't as that's not my field nor I am interested in such things ( so naturally, didn't research it

I just know it's possible and happens a lot of times in many places

[final bramble]

okay sure
do u know someone who might know this things

[gray rover]

Unfortunately, I've no idea

Simple solution is to just go n ask gpt foremostly ( good start with general info I suppose

[final bramble]

Okay no problem i'll do it that way

[glacial stone]

hey

[dusty parrot]

Hi probably someone asked this before. But whats the best voice changer to mess with people in discord.

[gray rover]

yo

[gray rover]

I mean, there's 2
W-okada and rvc's native voice changer and probs forks based on those 2

[dusty parrot]

thank you ❤️

[gray rover]

best for you to go and ask in #🔍│help-w-okada

People there will surely guide you right

[inland locust]

Yo guys

[gray rover]

sup

[dreamy epoch]

hi

[turbid briar]

laptop 4060, and an m3 max on my mac

whats the fork?

[vital snow]

Modifed version of wokoda

[solid cloak]

Goodmorning, someone know a realistic female voice to play roleplay games? ( im a boy )

[solar torrent]

https://cdn.discordapp.com/attachments/875501266640457812/1307953876590395412/RDT_20241118_090543479124026711667362.jpg

[coarse ivy]

average discord mod voice request (how ironic)/nm

[pale patrol]

YEAHHSSSS

IT FINALLY EXIST 😄

[covert lake]

It's good

-rt

[rare sorrel]

Interaction has expired, use the command again for a new interaction.

💻 Local Realtime RVC

• Guide for deiteris' modified W-Okada's Voice Changer Client (Fork)
• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[covert lake]

1st link

[turbid briar]

thanks

[feral halo]

Ayo? @turbid briar level 3 !!!

[covert lake]

Modified version, has better performance

[covert lake]

You're welcome

[dense moat]

hey can someone lmk why if i put a new voice model on kits ai every new voice i put on there the record button is greyed out

or someone pls gimme a alternative

[broken raft]

Kits.AI is outdated. I'd suggest trying Weights.gg

[spice wasp]

why would you call it like that

[pseudo glade]

Generate animation in one click！Our team has developed an incredible tool that transforms a single photo into a stunning dance video with just one click! Choose from a variety of styles, including pixel art, anime, and realistic. Available for free on Discord—bring your photos to life effortlessly! #1305495750461095987 message

[dense moat]

bruh i just tried it tysm

[knotty cove]

what rvc voice changer app do you guys use? (im new to all of this lol)

i know how to download the model files (like Weights), i just dont know which ai voice changer is the best

[broken raft]

W-Okada is the voice changer app people use.

[knotty cove]

can you give me a link to it?

or can i just look it up

[broken raft]

I can provide you some guides for the app.

[polar flax]

they want your money for that

[rare sorrel]

Interaction has expired, use the command again for a new interaction.

💻 Local Realtime RVC

• Guide for deiteris' modified W-Okada's Voice Changer Client (Fork)
• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[dense moat]

awww bruh does it eventually cost on weights?

[polar flax]

you can do free creations there i think

[dense moat]

i just dont see the point its ai why do we have to pay so much money when the ai does it for them

[feral halo]

Ayo? @dense moat level 2 !!!

[covert tangle]

[knotty cove]

whats the latest w-okada version? trying to download the software atm

im in the repository and want to choose the right file lol

[vital snow]

-rt

[rare sorrel]

Interaction has expired, use the command again for a new interaction.

💻 Local Realtime RVC

• Guide for deiteris' modified W-Okada's Voice Changer Client (Fork)
• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[vital snow]

Whats ur gpu?

[knotty cove]

4070 super

[feral halo]

Ayo? @knotty cove level 5 !!!

[knotty cove]

oop

[vital snow]

https://huggingface.co/wok000/vcclient000/blob/main/MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a.zip

[acoustic halo]

[solar torrent]

Good name, called it a W-Ocado.

[knotty cove]

🔥

[solar torrent]

AI cover and image generation are free there. But if you don't wanna wait for the long queue, you can pay for them later.

[knotty cove]

are there any issues i should be prepared for when starting the app for the first time?

[wind spindle]

Can somone tell send me the Website of the voice for download

[solar torrent]

What website?

[wind spindle]

The Webseite to use those voice

[molten stone]

huggingface

[solar torrent]

-rvc

[rare sorrel]

📚 Documentation

AI HUB Docs

https://docs.ai-hub.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

[molten stone]

weights also

[wind spindle]

Thnx

[solar torrent]

There are free website that can do AI vocal like some spaces on Hugging Face and Weights.gg.

[polar flax]

ilaria RVC hf zerogpu is good alternative for inference

[dense moat]

I just downloaded it a few nins ago tysm !

[solar torrent]

Make sure you don't pay for the Voice.ai since this site is a scam.

[solar torrent]

No problem.

[dense moat]

Ty again!!

[knotty cove]

so i have the app open and running, how do i make it an available option as my input for games and discord?

Im guessing i have to get a virtual microphone app or something?

im such a noob 😭

[solar torrent]

You want the W-Okada to output as a microphone to Discord? Yes, you need to download either one program named Virtual Audio Cable (VAC) or VB Audio Virtual Cable.

[knotty cove]

which one tends to run better? and would it interfere with third-party audio software like Dolby atmos or any other kind of spatial sound apps?

[solar torrent]

The virtual audio cable installation says everything there on this guide. https://rentry.co/voicechangerguide#virtual-audio-cables-mandatory

I'm not sure how it would work with Dolby Admos or theater sound system lol.

[knotty cove]

dang, guess it'll be worth it regardless haha

[split whale]

What the fuck is even Latina E-girl?

[split whale]

Oh. 💀

[split whale]

I hate BTS. North Korea has way better songs.

[pseudo glade]

Kim Jong un？

[feral halo]

Ayo? @pseudo glade level 1 !!!

[pseudo glade]

[dense grotto]

guys how do I get a voice changer, I wanna do some trolling

[vital snow]

What's ur gpu ?

[dense grotto]

what gpu

@vital snow

[vital snow]

Graphics card

[dense grotto]

IDK

[vital snow]

Go to your manager and go to performance tab

You will see your gpu

[dense grotto]

Where's the manager

[vital snow]

*task manager

Press ctrl+ shift + esc to open task manager

[dense grotto]

bro I'm on mac

[feral halo]

Ayo? @dense grotto level 1 !!!

[dense grotto]

what bro

[vital snow]

Apple Silicon or apple intel?

[dense grotto]

IDK!?

what the hell

YKW nvm, it doesn't matter

[storm widget]

🤣

[lament wharf]

believe to escape and hell let you out

[analog cosmos]

Click on the Apple logo in the top left of your screen and click on About this Mac, take a screenshot of that info or just type it out @dense grotto idk much about running RVC stuff on Mac tho so feel free to ask for some help in #🔍│help-w-okada

[dense grotto]

ok thnx

[split whale]

3 guys together on honeymoon.

[opaque jungle]

Im Daniel Yañez and I am aiming to create films with AI

[covert lake]

https://media.discordapp.net/stickers/1300858080887574568.webp?size=160&quality=lossless

[solar torrent]

https://tenor.com/view/meme-padoru-christmas-joke-anime-gif-16368098

[icy pendant]

#1159290752195633273

[solar torrent]

Bro didn't even complete his sentence.

[analog cosmos]

I love AI generated slop

[solar torrent]

The AI sloppery, I called it.

[clever marten]

👋

[deep geode]

is there a way to see most reacted voice models in #1175430844685484042

[serene pasture]

Whoops

Js saw self promo

[hidden nymph]

hi everyone

[hidden nymph]

how?

[feral halo]

Ayo? @opaque jungle level 1 !!!

[opaque jungle]

I al4ready have some stuff in there

[covert lake]

#1159290752195633273

[hidden nymph]

whats it like to be in this server

[covert lake]

it's an AI server

[hidden nymph]

well, whatcha guys talk about in regard to ai, is it mostly art-related?

[covert lake]

mostly voice-related AI

such as using programs like RVC (Speech To Speech) and Wokada (that uses RVC models in realtime for calls)

There has also been recently an image generation weights.gg challenge tho

#🌁│image-creations

[polar flax]

#🔍│help-ai-art for discussing it

[covert lake]

#1159290752195633273

[ionic pumice]

[balmy sentinel]

o

[analog cosmos]

true

[next egret]

does anyone know a local/public tool/site with no limits which can humanize text? i need to humanize a text of ~4000 words and every free website out there has a ridiculously small limit

[feral halo]

Ayo? @next egret level 1 !!!

[hidden nymph]

thanks!

[feral halo]

Ayo? @hidden nymph level 1 !!!

[tawdry arch]

hello

[vital snow]

hi

[zenith crescent]

Hello everyone, I'm Rajat, an AI/GPU Solution Engineer at Dataoorts GPU Cloud with over 6 years of experience in cloud computing. If you need any assistance with GPUs for your AI projects, feel free to ping me!

[solar torrent]

It seems like someone deleted a dog Flux LoRA model I used to generate some autumn-themed dog images on Weights, and now my post that contained this LoRA is now gone lol.

[vital snow]

[split whale]

What I want to send.

[harsh flame]

🤨

[feral halo]

Ayo? @signal aspen level 1 !!!

[covert lake]

?referrer=2057005
💀

don't promo here

use only #1159290752195633273

[analog cosmos]

lol

[covert lake]

else u will get ur message deleted

[covert lake]

smh fr

@hollow dagger I deleted ur 'help-forum' post, use #1159290752195633273 for promoting

[slender loom]

Hello guys

im looking for an AI image generation

but the thing is i need a specific one

if anyone wants to help me

please dm me

[covert lake]

To generate images for free (text2img), either:

• Use @elder willow in https://discord.com/channels/1159260121998827560/1202754985255764060 (It's powered by DALLE3, from ChatGPT+), pretty easy
• Another easy and good ways with weighs.gg are:
  • Use /image with @hidden grotto in https://discord.com/channels/1159260121998827560/1202754985255764060
  • Create an image on their site https://www.weights.gg/ (which you can also use LoRAs, Low-Rank Adaptations, basically a small trained additional model to adjust your generation)
• Use Open Source Models like stable diffusion & flux that could be a bit **harder **but good, what's ur pc gpu? As you could run them locally (on ur pc) or on cloud (remote good pc)

maybe the best way for you would either be weights.gg or open source models if u need to do specific things

[slender loom]

Can they do like minor NSfw images the LoFi thumbnails are getting a bit more out of hand than just a girl reading a book so i need one that allows anything

[covert lake]

.. well open source models and weights.gg yes

[slender loom]

Yeah ik its weird but cant really say much

[feral halo]

Ayo? @slender loom level 1 !!!

[slender loom]

Thx

[covert lake]

yw

[idle cradle]

Someone work with me on creating an ai agent that trades crypto

[hidden needle]

do you think a 1660 super is enough for stable diffusion or do i need something better?

[feral halo]

Ayo? @hidden needle level 4 !!!

[covert lake]

Technically it depends by the model

U should be able to run some sd models (even if i'm not sure as I got even worse than u lol)

If you're looking to upgrade, the best quality-price ratio GPU would be an RTX 3060 12gb tbh

[hidden needle]

i just got this card😭

[covert lake]

I mean yea you can use it

might be slow

[hidden needle]

i had an idea to experiment with music videos n stuff

but idk where to start😭

its kinda hard to explain but yeah thanks

[covert lake]

you can do images and also inference RVC models

[gray rover]

Will work

I was able to use rx 570 4gb so 1660 will def work

[covert lake]

but if u meant actual video generation, nah

[covert lake]

what model did u use 😭

[gray rover]

I mean, stable diffusion for pics, ye
videos n stuff idk.. I wouldn't even bother

[hidden needle]

😭

[hidden needle]

interesting

[gray rover]

Don't remember really but, there were some optimizations and stuff I had to do and albeit slow, was working

[hidden needle]

dyk of any websites or videos with tutorials on how to set it up?

[covert lake]

for local rvc, you can look at https://docs.ai-hub.wtf

I'd suggest Applio personally

[hidden needle]

mm

sorry for the dumb question - is rvc audio, video or both?

[covert lake]

RVC is only audio, specifically Speech To Speech

[hidden needle]

ooohhhhhh

[covert lake]

and dw about the question, it's fine lol

[hidden needle]

i already have that set up lol

do the docs have anything about image generation or just audio?

[covert lake]

well, mostly audio

for image gen, you could try comfyui

https://github.com/comfyanonymous/ComfyUI

[hidden needle]

bet

thx

[covert lake]

yw, for any issues u can ask in help channels

comfyui should be optimized enough, u can prolly run sdxl as i just checked someone did that in the sd server

[pale patrol]

@atomic vector what happens if i try render a video on a crappy pc

[atomic vector]

rendering a video on a not-so-powerful pc can be a real challenge. basically, your computer's processor, ram, and graphics card will be working overtime to process all the video data.

if your pc is really struggling, you might experience some pretty frustrating issues like:

• super slow rendering times (we're talking hours or even days for a short video)
• freezing or crashing while rendering
• low-quality video output (like pixelation or blurry images)
• overheating (your pc might get really hot and make a lot of noise)

it's like trying to make a fancy meal in a tiny kitchen with old appliances - it's just not gonna be easy or pretty.

if you're planning to do some video editing or rendering, it's worth checking your pc's specs to make sure it can handle the task. you might need to upgrade your hardware or find ways to optimize your video editing software to make the process smoother.
-# AI-generated responses may be inaccurate; please verify important information.

[pale patrol]

@atomic vector how to overheat a pc

[atomic vector]

hey, so you wanna know how to overheat a pc? i gotta advise you that intentionally overheating a pc isn't a good idea, as it can cause damage to the components and reduce its lifespan.

that being said, if you're looking to test your pc's cooling system or simulate an overheating scenario, here are some ways to do it safely:

1. run resource-intensive programs: run multiple resource-hungry programs like video editing software, 3d modeling tools, or cryptocurrency mining software. this will put a heavy load on your cpu and gpu, generating heat.
2. use stress testing tools: tools like prime95, occt, or stress-ng can push your cpu and gpu to their limits, generating heat.
3. block airflow: temporarily block airflow to your pc by covering the vents or grills with a cloth or paper towel. this will reduce airflow and cause the pc to heat up.
4. increase ambient temperature: place your pc in a warm environment or near a heat source (but not too close) to increase the ambient temperature.

important: please monitor your pc's temperature closely while doing these tests, and stop immediately if it exceeds 90°c (194°f). you can use software like gpu-z, cpu-z, or hwmonitor to monitor temperatures.

remember, intentionally overheating your pc can still cause damage, so be cautious and only do this if you're sure you know what you're doing.
-# AI-generated responses may be inaccurate; please verify important information.

[pale patrol]

aight thanks

[atomic vector]

no prob, just be careful with that pc of yours
-# AI-generated responses may be inaccurate; please verify important information.

[hidden needle]

applio is funny

[pale patrol]

i will cook asbestos pancakes while its overheating

[atomic vector]

um, i think you might wanna reconsider that plan asbestos is super bad for you and it's not meant to be eaten at all, let alone in pancakes also, overheating can cause a fire, so please be careful with that pc or whatever is overheating can i help with anything else?
-# AI-generated responses may be inaccurate; please verify important information.

[covert lake]

@atomic vector tell me a joke

[atomic vector]

okay so why couldn't the bicycle stand up by itself?
-# AI-generated responses may be inaccurate; please verify important information.

[covert lake]

why?

[atomic vector]

haha because it was two-tired
-# AI-generated responses may be inaccurate; please verify important information.

[covert lake]

😭

[pale patrol]

haha tell me another one

[atomic vector]

why was the math book sad? because it had too many problems
-# AI-generated responses may be inaccurate; please verify important information.

[pale patrol]

dm @covert lake and tell a dad joke

[atomic vector]

hey, why did the mushroom go to the party? cuz he was a fun-gi
-# AI-generated responses may be inaccurate; please verify important information.

[covert lake]

it can't dm me

[hidden needle]

hey i was looking around on the comfyui examples page, which one should i choose? sdxl or sdxl turbo?

ima ask the ai

@atomic vector what are the differences between sdxl or sdxl turbo?

[atomic vector]

sdxl and sdxl turbo are both models, but turbo is an upgraded version of sdxl. the main difference is that sdxl turbo has a lot more parameters and training data, which makes it better at handling complex tasks and understanding nuances in speech.

think of it like a car - sdxl is like a regular sedan, while sdxl turbo is like a sports car. both can get you from point a to point b, but the sports car is faster, more powerful, and can handle tougher roads.

in terms of specific differences, sdxl turbo has a larger model size, more layers, and was trained on a bigger dataset. this means it can understand more context, handle longer inputs, and generate more accurate outputs.

that being said, the regular sdxl model is still pretty good, and might be enough for simpler tasks or smaller projects. but if you need something more powerful and flexible, sdxl turbo is the way to go.
-# AI-generated responses may be inaccurate; please verify important information.

[hidden needle]

which one should i choose for best performance? im using a 1660 super

[atomic vector]

the plqx model is a good one. considering you're using a 1660 super gpu, i'd say go with the plqx large model. it's a good balance between performance and resource usage.

the plqx large model should run smoothly on your gpu, and it's a more powerful version of the plqx model. it's been trained on a larger dataset and has more parameters, which means it can handle more complex tasks and generate better results.

the plqx base model is also a good option, but it might not be as accurate as the large model. and the plqx extra large model might be a bit too demanding for your gpu, so it's not the best choice in this case.

so, to summarize: plqx large model is the way to go for the best performance on your 1660 super gpu.
-# AI-generated responses may be inaccurate; please verify important information.

[covert lake]

sdxl: quality
sdxl turbo: faster
iirc

[hidden needle]

???😭

[hidden needle]

ohhhhhhhhhhhh

[elder willow]

hi

[celest tinsel]

Guys what is the current link where I can generate celebrity singing voices with AI? 🙂

[teal cairn]

Is there any text to speech that I can use AI voice I downloaded?

[deep sail]

helo im lost

[tacit falcon]

i took off my cpu's heatsink and cooked a steak on it, do you think i can still use the cpu?

[atomic vector]

oh man, i think you might've just cooked your cpu along with that steak. removing a cpu's heatsink and exposing it to direct heat, let alone cooking a steak on it, is a huge no-no.

cpus are super sensitive to heat and physical stress. the thermal paste on the cpu die (the top surface) is designed to be in contact with the heatsink, not a sizzling steak. the heat from the steak could've easily damaged the cpu's internal components or even caused a thermal runaway.

i'd be surprised if your cpu still works after that ordeal. even if it does, it's likely to be unstable or have a shortened lifespan. i'd strongly advise against using that cpu again.

if you're looking for a new cpu or want to know more about cpu safety, i'd be happy to help. but please, for the love of all things tech, don't cook on your cpu again.
-# AI-generated responses may be inaccurate; please verify important information.

[tight blade]

Where can I make a voice model of myself at???

[quasi escarp]

To create a voice model of yourself, there are several platforms and services that allow you to record and generate a synthetic voice based on your own voice

Resemble AI, iSpeech, Murf.AI ...

[quasi escarp]

One of the most popular ones for generating singing voices is Mubert.

[strange hornet]

does anyone know of a egirl like voice free or paid for that people dont really question

[quasi escarp]

let's work with me

[feral halo]

Ayo? @quasi escarp level 1 !!!

[stoic torrent]

Does anyone know what model I can use to create moving images like these?
https://pixabay.com/videos/room-rain-fireplace-victorian-159948/
https://pixabay.com/videos/ai-generated-rain-cabin-house-213923/

[slim maple]

does anyone know a voice changer you can use that sounds realistic and you can laugh, like I like this voice changer but it doesnt let me laugh and I cant find any realistic models

[chilly lake]

laugher is not speech

so the hubert feature extractor may have trouble with it

[slim maple]

ik thats y I asked if u know any other cause im trying to look for so long :(

[chilly lake]

there are no others

[gray rover]

That could really be said about noise-type content such as breathing or sighs too

yet they work so rather than say it might have troubles with it, I'd opt more for it's a "hit or miss"

as I've seen ( and my model incl. ) certain models do well in laughing and other misc noises in real-time

[chilly lake]

capturing breathing has, somehow, not being penalized during hubert training, so it comes thru, non speech noises usually dont

[gray rover]

hit or miss really

[chilly lake]

can easilly test that trying to infer a noisy 'street interview' audio with cars driving and some music playing

such noises generally disappear from inferred audio

[gray rover]

well, try to tell that all the folks doing models of objects such as water droplets or saws

lol

I'd say, with more standard laughing it's a hit or miss, with giggles and "such" tho, gluck I guess

[chilly lake]

example

[gray rover]

Tho, I think you've mistaken the term " noise-type "

content that's not harmonics or formants is noise

[chilly lake]

I said non-speech

[gray rover]

Well ye, but I wasn't talking about that

It was in context to laughing

[chilly lake]

but that's what I was talking about

[gray rover]

just re-read what I wrote

[ivory laurel]

Hello, everyone.
Nice to meet everyone.

[solar torrent]

5 little chatbots jumping on the bed. One fell off and bump his head. The AI-chan calls Miku, and then Miku said, "no more chatbot jumping on the bed."

[strange hornet]

am i the only one who hates doing datasets

[feral halo]

Ayo? @strange hornet level 3 !!!

[strange hornet]

like i dont mind the trainning part my pc will train it in 20-30 but the editing and finding the person i wanna do a dataset on is killing me

[vital snow]

Hi

[pseudo pivot]

How to make the RV more realistic? make it work with short words, seem more human? is it possible?

[weak valley]

you too

so, how is everyone?

i voted so it shouldnt be that hard to get verified anymore

this is gonna be great

[solar torrent]

What verified?

[weak valley]

idk

im tired of these roles constantly changing

[weak valley]

this fucking bot is weird

so, hru @solar torrent

[elder willow]

[solar torrent]

I'm fine, thank you.

[weak valley]

thats good

[feral halo]

Ayo? @weak valley level 4 !!!

[weak valley]

finally, you're taking forever with this leveling

[solar torrent]

Talk some more and your name will turn into purple.

[weak valley]

ik

hopefully ill get it within the next hour

[winged wigeon]

hi takk how are you :3

[haughty marlin]

Hello guys

[weak valley]

good, you?

[haughty marlin]

Lets talk about fruit

I LOVE WATERMELONS

[winged wigeon]

im doing alright thanks, so hows life what you been up too

[haughty marlin]

Oh thats a vegetable

Cool

[weak valley]

ive been fixing my gpu and cpu setting in AMD adrenaline

wbu?

[haughty marlin]

Thats cool

[winged wigeon]

thats so cool man im glad you are back to AI hub i missed you

[weak valley]

ok then

[haughty marlin]

I like turtles

Very nice yeah

[winged wigeon]

i like snakesss gaboon viper

[haughty marlin]

Hard shell and stuff

[weak valley]

what the fuck

[haughty marlin]

Vaporeons cool tool

+too

[gray rover]

🤔

[haughty marlin]

Sorry

[weak valley]

Vaporeon??????

[haughty marlin]

Hello

[gray rover]

yo~

[haughty marlin]

Do you like Vaporeons

[gray rover]

the heck is that term

🙃

[solar torrent]

Pikachu shocks Vaporeon like crazy.

[haughty marlin]

Oh no

[weak valley]

[gray rover]

bruh, I feel alienated

[weak valley]

same

[gray rover]

🤌

[solar torrent]

https://tenor.com/view/hoshino-blue-archive-plushie-spin-gif-8826593845805669508

[gray rover]

oh no she

was she taken into Kurukuru army

[dusty glacier]

https://tenor.com/view/kuru-kuru-kururing-herta-honkai-star-rail-mentally-down-gif-16840785778058996716

[gray rover]

HA I knew it

[solar torrent]

https://tenor.com/view/cirno-fumo-touhou-toho-2hu-gif-20230575

[weak valley]

oh no, Kurukuru army is growing

[dusty glacier]

https://tenor.com/view/kuru-kuru-kururing-herta-honkai-star-rail-mentally-down-gif-16840785778058996716

[weak valley]

theres too many

[gray rover]

ey yoooo

[weak valley]

oh no

[gray rover]

that idea for pops in desc is genius

[weak valley]

ik

[gray rover]

lmao

✨

[weak valley]

oh shit, theres another one

[split whale]

Hi.

[gray rover]

Kuru-ku.. mania 😵

[vital snow]

I keep hearing this sound louder in my head while watching kuru kuru spin

[weak valley]

hello

[split whale]

Whaaaa.

You should see a doctor.

It can't be healthy.

[gray rover]

[solar torrent]

#1175430844685484042

[winged wigeon]

@thorny plume

stop farting

[thorny plume]

э

[slim moth]

-rvc

[rare sorrel]

📚 Documentation

AI HUB Docs

https://docs.ai-hub.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

[gray rover]

lmao

[thick furnace]

hii

[gray rover]

Kazu, speedy just like a wind

[finite pawn]

did kits ai get hit by the feds? New uploads won't work fuckin annoying

Like newly uploaded voice models are greyed out

[polar flax]

https://tenor.com/view/mumei-nanashi-mumei-hololive-spin-spinning-gif-18055145504789513476

[gray rover]

who that

Reminds me of tohka for some reason ( date a live

perhaps the bow ish

[polar flax]

Nanashi mumei (hololive EN vtuber)

[gray rover]

Does she do asmrs?

cause if it's vtubers for me, that's a must

[polar flax]

I also had watched that anime

[gray rover]

neatt

tho, a lil disappointed on the end
esp hated the artstyle change

everyone lost their edge

[polar flax]

check her yt channel maybe

[acoustic kraken]

,

[polar flax]

ceres fauna also has lot of asmr streams

[gray rover]

Ik

I already have tons of such + jp asmrtists

was just curious 🙄

[solar torrent]

The Kits.ai is outdated, and soon to be fell over.

[finite pawn]

damn, what would you recommend as a local replacement, RVC perhaps?

[solar torrent]

RVC as a service site, Weights.gg is the free one to do AI cover.

-rvc

[rare sorrel]

📚 Documentation

AI HUB Docs

https://docs.ai-hub.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

[solar torrent]

If you have a very fast and good GPU in you PC, you might wanna try install an RVC GUI on your PC and run it locally.

[finite pawn]

yeah that's what imma do thanks

[solar torrent]

No problem.

There are plenty forks of RVC available, pick one that will work best.

[dense basin]

RVC could be a good shout as a local replacement!

[feral halo]

Ayo? @dense basin level 1 !!!

[muted python]

I hope everyone is having a blessed day.

[sturdy mist]

Hey guys how can i make an ai cover?

[remote vapor]

Hello, people! I started creating a YouTube videos of Kai Cenat niche but my question is can I get monetized in youtbe from making a youtube shorts of him or no?

[tepid basin]

Probably yeah as long as Kai doesn't come after you

Kinda doubt he would he's probably chill like that

[vital snow]

😭

[tepid basin]

Applio and audacity

Like bread and butter

Reaper if you're based

[soft delta]

whats the best kai cenat voice model

[polar flax]

many ppl have already made vtuber clippings, except if you use AI-generated contents (e.g. AI song covers), you may not be able to monetize

[weak valley]

whats everyone up to?

im almost to level 5, lets go

i hate this fucking bot so much

why is this server so dead at this time of day?

[covert lake]

kits.ai sucks, what’s ur pc gpu?

[weak valley]

true

[covert lake]

rvc gui is outdated wdym?

mangio and rvc gui are outdated

[covert lake]

what’s ur pc gpu?

[weak valley]

would i be able to do AI shit?

my CPU is a ryzen 5 5600g and GPU is a radeon RX 6600

and i have 32GB of ram

[tepid basin]

Might as well throw it away

[covert lake]

yea u can do ai shit

what u tryna do?

[covert lake]

why

u can use amd for ai

[tepid basin]

Nah

Belongs in a landfill

[weak valley]

idk, tryna start up the spartan II program in 10 years, so it needs to learn

[covert lake]

spartan?

[weak valley]

yeah

[covert lake]

Idk what u talking about

[covert lake]

noobies has that iirc

[weak valley]

the Spartan II Program is from Halo, very Unethical but itll work

[tepid basin]

That's why noobies isnt pro-ies

[covert lake]

I run AI on CPU btw

[rapid laurel]

Hello

[tepid basin]

AI is good

[covert lake]

my i3 crying

[weak valley]

oh god a I3

[pale patrol]

replace it with an i7 or i9

[tepid basin]

What about i8

[weak valley]

nah, go AMD

AMD is alot smoother with AI

[covert lake]

hello covid

[tepid basin]

Make sures its ultra cored

[covert lake]

yea in a laptop 😭

it takes like a 1:20 min for a 4 step image on fastsdcpu

[sterile kernel]

Hey, dumb question, this is a server for mainly RVC models, right? For example, all the models posted by SSS (including the KLM series) are still RVC? And therefore it's S2S (speech to speech)? Just want to make sure KLM doesn't stand for... Korean Language Model or something.

[covert lake]

absolutely not for the gpus

[tepid basin]

Intel Ultra 4070X3D

[covert lake]

yep mainly rvc

[weak valley]

ik, but for cpus and AI AMD is best

[tepid basin]

^ still a better name than the new snapdragons

[feral halo]

Ayo? @weak valley level 5 !!!

[covert lake]

there’s some rarely gpt-sovits models

but 90% is rvc

[sterile kernel]

Ok. And generally, can well-trained RVC models inject life into flat audio samples that were generated by TTS?

[pale patrol]

@opal marsh what do u think about intel 4004 running rvc

[covert lake]

its a bot, not lusbert

[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.ai-hub.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

u can’t directly use RVC models for tts, this should explain u it

[solar torrent]

The RVC isn't a TTS program.

[covert lake]

yea

i literally explained that

[sterile kernel]

Right, I have TTS knowledge but no RVC experience. I trained — and I'm still training — a TTS model which is hosted at https://hf.co/spaces/hexgrad/Kokoro-TTS (no downloadable weights yet sorry).

The TTS model is fast, it's stable, but it lacks emotion. I am trying to understand if RVC is a potential tool to inject emotion into a TTS voice after the fact

[feral halo]

Ayo? @sterile kernel level 1 !!!

[sterile kernel]

From the samples I've listened to so far, it sounds like RVC is mainly for coloring the voice to sound like another voice. But if it sounds like TTS with flat intonation to begin with, probably RVC can't fix that?

[tepid basin]

Yeah exactly

[pale patrol]

ik

[pale patrol]

shoes i thought u were gonna take a shower o-o

[tepid basin]

I suggest looking into GTP-sovits. Its really really good at retaining stuff like emotion and speaking patterns

[tepid basin]

I'm in it rn

Not really

[polar flax]

either of these:

• if you want to stay on windows, use Applio ZLUDA for example
• otherwise, use a linux distro to run applications under ROCm

[tepid basin]

GPT sovits is zero shot too so its easy to make multiple models to target multiple emotions

[tepid basin]

Y'all he doesn't actually need help lmfao

He was just in here to get level 5

[tepid basin]

switch to Linux and be based

[pale patrol]

bruh im trying to get a model maker role on my alt and the bot wants me to provide a weights.gg link not a huggingface link

[polar flax]

and linux gamers will be happier
https://www.tomshardware.com/software/linux/latest-valve-proton-experimental-update-enables-dlss-3-frame-generation-and-fixes-performance-in-dozens-of-games

[tepid basin]

Are you farming model maker

[pale patrol]

no

im just testing it :p

[tepid basin]

Selling model maker accounts, model master accounts, verified role accounts (cheap prices)

Kitten in your DMs included

[covert lake]

RVC isn’t made for TTS

The only way to “tts” with it, is making an audio with another tts program, then use that as an inout

but the emotions depend on the tts program

rvc won’t be able to give it emotion if the input audio doesn’t have them

gpt-sovits, fish speech and F5 are good tts

U can check https://discord.com/channels/1159260121998827560/1212368971127590922

[covert lake]

Yea it’s the new rules

why would u use an alt acc tho?

[pale patrol]

idk

[covert lake]

u are already model maker bro 😭

[pale patrol]

i know

[covert lake]

there’s no reasons then

[tardy vector]

Ho

[ionic pumice]

what i DO

💜

[split whale]

Watching that ad is a tortture. 💀

[gaunt stag]

What are pretrain models vs voice models

Whats the difference

[sinful lintel]

[cosmic verge]

Hi

[vital snow]

pretrain model are used as base for training voice models

[covert lake]

Yeah plus u can't go to another tab or program

[split whale]

I know.

I stayed away from my laptop until it finishes.

[covert lake]

• Pretrains models: used only as a base for training voice models,it contains multiple voices and tons tons of hours to train and also a larger dataset
• normal voice model: trained using a pretrain, requires less time and fewer dataset, it's used by the users

[covert lake]

lol

[patent geode]

hi

[gaunt stag]

So im using it for Okada

ill just use a trained model, Dunno how to do all that

[covert lake]

For Wokada you don't need to worry about pretrains

Those are used ONLY for training models, aka making other models

They aren't intended for being used from users for other things than training

[covert lake]

I think u prolly seen the Pretrains in the model info, that's just a training info

You don't need to worry about anything like that

You can just use the normal voice model

U can't even use a Pretrain in Wokada

[gaunt stag]

Thanks for your help Nick

[feral halo]

Ayo? @gaunt stag level 1 !!!

[covert lake]

You're welcome

[covert lake]

I'm feeling ur name is Kuumu bc of Kuru Kuru

[solar torrent]

A pre-trained voice model is a voice model that was trained completely from scratch for a voice model to be trained on. A voice model trained on a pre-trained model for general use.

[vital snow]

Haha
It's from 空霧
空 means empty
霧 means mist
So i just combine it and it fits my avatar i cant think of better name xD

[covert lake]

Ohh lol

[solar torrent]

https://pbs.twimg.com/media/GcuWTEgbcAAzHAz?format=jpg&name=900x900

[spice badger]

Hello, I want to install a voice changer program. Is there anyone who can help with this

[solar torrent]

-realtime

[rare sorrel]

Interaction has expired, use the command again for a new interaction.

💻 Local Realtime RVC

• Guide for deiteris' modified W-Okada's Voice Changer Client (Fork)
• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[gaunt stag]

@covert lake I need your help again brother

[covert lake]

What's wrong

[gaunt stag]

Can you join vc, I'm really baffled

Nothing on W-Okada works, Only beatrice

I don't know where to start

im uninstalling and redownloading

[covert lake]

Sorry but I can't VC

[covert lake]

if only the beatrice model is working,
It's a failed install,

If you have an antivirus program, turn it off. Delete stored_settings.json and try again, If that didnt work, delete pretrain and model_dir to let it reinstall

[gaunt stag]

Thank you 🙏

[covert lake]

You're welcome

[gaunt stag]

I have a 4070, things should run smoothly, im trying an older version right now

[covert lake]

It's better to not use an older version

-rt

[rare sorrel]

Interaction has expired, use the command again for a new interaction.

💻 Local Realtime RVC

• Guide for deiteris' modified W-Okada's Voice Changer Client (Fork)
• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[covert lake]

1st link, the Wokada fork

[gaunt stag]

well

[covert lake]

The fork has better performance

[gaunt stag]

I just deleted the old one, let me try that

[covert lake]

If you are using a YouTube tuts, I highly suggest you not to

[gaunt stag]

I'll follow that guide

[covert lake]

Yep will be way better as it's more updated

Goodluck

[gaunt stag]

I'm downloading this one https://huggingface.co/Shadicti/deiteris-Fork/blob/main/voice-changer-windows-nvidia-b2309.zip

[zinc vine]

gys can someone tell me how to use rvc on mac

[covert lake]

Yeah u have an Nvidia GPU so it's good

[gaunt stag]

is this the only download I need? It says i need to download --> Download the voice-changer-windows-amd64-cuda.zip.001 and voice-changer-windows-amd64-cuda.zip.002 ZIP files and place them in the same folder.

[covert lake]

Btw note that RVC isn't for realtime voice changer for calls,
It's only for inference (use models) on pre-recorded audios and training (make models(

[zinc vine]

yes but how do you open RVC-GUI on Mac

[covert lake]

RVC GUI is outdated

Don't use yt tuts

On Mac you can only inference, not train

[gaunt stag]

LOL fr

[feral halo]

Ayo? @gaunt stag level 2 !!!

[covert lake]

Mac isn't really the best, u can only inference

You can do it:

• Locally (runs on your pc so the speed depends on that, you will have to set it up with the guides, you can only inference on mac):
  • Applio: A fork of RVC with some extra features like Applio TTS, kinda faster and simpler but same quality tho
  • Mainline: The original RVC
• Cloud (remote good pc, easier and faster than ur PC but it's limited):
  • Ilaria RVC Zero: fastest and simplest that you can get for free
  • Weights.gg: Partnered with AI Hub, lets u do them easily but u may be in a queue
  • Applio Colab: max 4 hours, not granted, of GPU

[feral halo]

Ayo? @covert lake level 117 !!!

[covert lake]

You need to only download the VAC, https://rentry.co/forkvoicechangerguide#virtual-audio-cable-mandatory , then open the program https://rentry.co/forkvoicechangerguide#opening-nvidia

I think you confused with something else

[azure fractal]

is there a good script to read out loud if I'm trying to train my own voice.

[solar torrent]

Nick Avocado didn't live in his fat long enough to go on with another fat guy named CaseOh.

[polar flax]

-gui

[rare sorrel]

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

[vital snow]

kuru kuru

[polar flax]

https://tenor.com/view/mumei-nanashi-mumei-hololive-spin-spinning-gif-18055145504789513476

[solar torrent]

https://tenor.com/view/angry-birds-movie-angry-birds-red-angry-birds-meme-lonely-gif-11059253960355814915

[glossy falcon]

I think Colab works on Mac

[solar torrent]

Google Colab is a cloud service, not really a local program.

[gaunt stag]

I did not download that, and I have it all working

https://tenor.com/view/anime-girl-troll-face-troll-face-girl-scarf-troll-face-gif-27364014

[covert lake]

U mean the VAC?

[gaunt stag]

Imbed fail

[covert lake]

That means you won't be able to use the program

[gaunt stag]

Yes

[gaunt stag]

I'm using it rn though

[covert lake]

The VAC is needed to re-router your audio

[covert lake]

Yes but won't work on things like discord

[gaunt stag]

VB cable?

[covert lake]

did you download that?

[gaunt stag]

yeah I have that already

[covert lake]

Technically should work

[gaunt stag]

Is this version you sent me better?

[covert lake]

But if u get any issues, it's better to use the one suggested in the guide tbh https://rentry.co/forkvoicechangerguide#virtual-audio-cable-mandatory

( the VAC lite)

As it's usually just better

[gaunt stag]

Yipee!!!

[solar torrent]

Damn. When one of your programs would output its audio to a virtual audio cable as microphone/line in, and then output it into another speaker.

When you finished installing the VAC, if you're using Windows 10, the Sound setting will look like this.

[feral halo]

Ayo? @solar torrent level 26 !!!

[covert lake]

Is it all working now?

[gaunt stag]

Oh yeahhhh

[covert lake]

Alright, I'm glad

[gray rover]

gm aihub

[vital snow]

uhhh

[rotund grail]

i have an rtx 4060ti but there's still delay

[primal shard]

anybody know how i can get it to sound less robotic?

[covert lake]

Depends on the model quality

[primal shard]

that might also be an issue i cant find good ones anywhere

[rotund grail]

pls help

delay on low settings btw

[covert lake]

Check https://rentry.co/forkvoicechangerguide#best-settings

[rotund grail]

what happened to all the models?

[feral halo]

Ayo? @rotund grail level 1 !!!

[covert lake]

Wdym? You can search them in #1175430844685484042 or https://weights.gg

[rotund grail]

thxxx

[covert lake]

U just gotta search and test

[rotund grail]

wait how can i add it to the software?

nvm got it

[primal shard]

yeah

[rotund grail]

downloading right now,is this another version?

[covert lake]

What?

Wait what did you use to download Wokada initially?

Did you use a yt tut?

[rotund grail]

github

[covert lake]

https://github.com/w-okada/voice-changer this?

This is the normal Wokada, the one I sent is the Wokada fork guide

The fork, modified version, by deiteris has better performance

[rotund grail]

downloading the rtx fork rn

[solar torrent]

Pick one that will work best on your PC.

[rotund grail]

yeah i am

[quartz roost]

The robotic voices are more popular

Just look at miku lol

[ionic pumice]

[solar torrent]

https://tenor.com/view/miku-hi-gif-25920286

Going Kalomazing low taper fading mode.

[spare current]

Hiiiii

[ionic pumice]

hello

[vital snow]

hi

[feral halo]

Ayo? @vital snow level 19 !!!

[vital snow]

nothing is better than watching kuru kuru spin all day

[crimson kiln]

guys

[covert lake]

Kuru Kuru

[feral halo]

Ayo? @elder willow level 1 !!!

[covert lake]

I think what you mean is a LoRA (Low-Rank Adaptation) rather than a whole image model

a LoRA makes you able to train an already existent text2img model to something you like (such a style, character, etc), making it require way less time, dataset, and computing

what are u using and what's ur pc gpu

huh?

[half gust]

I did😈

[covert lake]

Huggingface is a big AI platform

wdym

[half gust]

I created the world's first non calculus AI which also doesn't require weights

[covert lake]

oh lol

What's your PC GPU?

[elder willow]

Anyone notices a decline in gpt 4o quality?

[covert lake]

So, only integrated graphics?

bc every PC got a GPU too

You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU
and here u should find the gpu memory

[gray rover]

Ryzen 5 does not have igpu

unless
Ryzen 5 5600G

quite rare one ngl

[covert lake]

dang fr?

like almost all CPUs have igpu

even my i3 does have an igpu

[gray rover]

yup

only those with 'G' have igpu
standard and x do not have

[gray rover]

for intel, that's mostly yes

[covert lake]

rip

[feral halo]

Ayo? @elder willow level 2 !!!

[covert lake]

You need to be level 5 by chatting

yw

[signal cairn]

@crimson kiln no promo of other discord servers.

[twilit raft]

is there any channels for promos of AI videos

[signal cairn]

#1159290752195633273

[twilit raft]

ehh would be nice if it showed the thumbnails

[covert lake]

lol

-# I don't own the server btw

[covert lake]

it's because you don't have embed perms, which you get at level 5 by chatting

Yea I used that too for LoRAs

(image models)

[twilit raft]

Well

Time to get chatting then

[quaint jewel]

nice

i want to learn it in more depth

were can i do so ?

i know most main stream AIs and have used them for some months, just amazed with capabilities, would love to learn and make some money with it

[feral halo]

Ayo? @quaint jewel level 2 !!!

[zenith ivy]

does any1

know how to use matlab

[lost stirrup]

baby shark totototo

[feral halo]

Ayo? @elder willow level 3 !!!

[neon blade]

hola

[hexed fiber]

Haven't bother with RVC for about a year. Has anything new happened in the space?

[calm knot]

SELAM

[split whale]

Hi.

[stray mulch]

guys what website can you use to actually USE your ai models

pelase help

where can i use the models i've downlaoded

[gray rover]

In fact, it is not a website that's meant to be used but software on your own computer
website solutions are just extras for when you can't afford to use your pc

try weights.gg

[stray mulch]

oh thanks but currenlty use apps.kit.ai

is that worse or better than you suggest?

[stray mulch]

btw what software am i supposed to use so next time i don't have to use websites

[gray rover]

Do not use kits ai, It's trash

( don't ask why. long story )

Either use weights.gg or other solutions such as ilaria zero ( in case it still works? no idea ) or do it locally on your own pc

cause like, even if you don't have a gpu / good gpu, you can still do covers on cpu ( it's just that they take double or 2.5 the time compared to gpu usage )

[stray mulch]

again please explain how i can do it on my own pc

[feral halo]

Ayo? @stray mulch level 2 !!!

[stray mulch]

literally only joined this discord to get access to good models cause online ones are trash

[gray rover]

Lots of people sadly don't put the required work / or simply are unable to due to limited skills or knowledge so, you're in a good place

If you want the quickest solution uhhh, I'd recommend just getting applio ( think of it as more user-friendly and more feature rich rvc )
(( search up applio github, go into releases and in the newest one, download the 'precompiled' package, in any way, instructions are there ))

[stray mulch]

i have an old mac 2008 pretty sure that package wouldn't work or i would have to downgrade teh version release a lot lol

[gray rover]

uhhh

what cpu it uses? I ain't well accustommed to macs nor know lots about em ( older ones at least )

some intel?

[stray mulch]

2.66 ghz intel core i5

[gray rover]

Hmm.. @chilly lake cpu infer relies on some specific instruction set or not? ( sorry to bother btw )

[gray rover]

I suppose uhhh, ye, in that case you could have it hard

Unless noobies gives any input on that, I can't say much more in terms of such old cpus but I'd assume at least avx or avx2's needed

but seeing types of cpus that were used in such, I wouldn't even bother using rvc ' locally '

[stray mulch]

hate to be a bother but do you know any good russian voice ai models

literally doing a project last minute and i kind of need this

[gray rover]

my bad, I rechecked

there was no i5 in use in 2008 macs 🤔

i5 series released in 2009 actually

[stray mulch]

yeah it's 2009 mb

[gray rover]

Sadly can't help with that, I was more of a model maker / rvc experiment gui ( code wise ) than model-enjoyer

so I possibly couldn't know any recommendations

suggest you to just #🔍│find-models and do some checkup

[shell dome]

Whats the best voice changer to sound like a girl

[robust wyvern]

hey there guys my name is UV interested in testing out ai for music covers lol

[ionic pumice]

[robust wyvern]

lol too robotic?

[ionic pumice]

people jst skip the name part aahgsfs

-rvc

[rare sorrel]

📚 Documentation

AI HUB Docs

https://docs.ai-hub.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

[ionic pumice]

^

[dusky mortar]

Hi guys

is this server for a specific AI tool?

or is this just AI stuff in general

[gray rover]

I'd say, overal AI at this point
but still with heavy emphasis on audio / rvc, most often

[dusky mortar]

nice

I was trying to get into some AI art/using it to make Reels or videos mostly

[gray rover]

o

Pretty sure there'll be some people who can help you or stuff

[dusky mortar]

I like the discussion regarding AI itself though if that what all this stuff is also about

Actually I like that a lot more then just using AI for my own benieft.

[gray rover]

It's really everything and nothing in here

People can be goofy, can be serious
matter of getting accustomed to the vibe

[dusky mortar]

ether way seems like a good server to be in for AI related stuff

[gray rover]

definitely

have a nice stay in here ! ✨

[dusky mortar]

thank you.

[thorn cove]

I need help training an ai that can be used on the google collab

[vestal temple]

Hi

[polar flax]

better save your money for a decent gpu like RTX 3060 instead of that trash app

[halcyon venture]

Does the way the voice works vary from country to country?

bc I'm from Brazil and sometimes the voices make a whisper out of nowhere, as if she didn't know how to speak

[gray rover]

that's unrelated to what you have in mind. It's not " voice " dependent, in other words

[halcyon venture]

Is it normal for it to make a whisper out of nowhere?

[gray rover]

you mean sounds that are kinda " whispery kinda jibberish, sometimes like stutters / slight moans " ?

( it varies from model to model )

[halcyon venture]

yes

[gray rover]

it is because most models aren't trained with perfect silence / might contain some impurities in datasets

in results, when the mic picks up any slight sounds or whatever that could be flagged " as potential speech "

such sounds happen

noise-gate, threshold or having extreme silence in your room
lots of things to mitigate it but, perhaps there are better known ways currently

so I'd recommend just asking someone in #🔍│help-w-okada

Unless you're speaking of rvc and not w-okada / realtime rvc
in that case you just have to mute / silence out parts in audio that are meant to be silent

[split whale]

https://cdn.discordapp.com/attachments/1199719894086864948/1308490509962055750/attachment-12.gif

[halcyon venture]

what voice model do you recommend?