#🧬│ai-chat
1 messages · Page 342 of 1
whut
wrong stiker
bro left
hi
0.0
Hey guys, are there a lot of beatrice models out now?
@icy pendant Hey man you online?
Foot fetish
Ayo? @elder willow level 3 !!! 
we
Hello
???
I can’t cure it
I want to but I can’t
I hope you're happy
Now that you're choosing this
- Mãe 2024
so true
Hi
hello
is chatgpt + worth it
Hi
Yes
is there an ai that can make a video that can also overlay text in a specific way
is this possible
/
does anyone know a ai like this
yes its possible
im pretty sure i saw an model like that somewjhere
not sure where though
hey guys, i was making an instagram account that teaches topics with homer simpsons voice. Is there any place i can share it?
np
@gray rover hi my w
hey taewoo
hii
dude how tf do you search for ai voices
many ways
they changed the system since last time i used it and idk how tf to do it
you mean in this server?
yeah
used to be as simple as /search lego man
or go to weights.gg
or whatnot
#1175430844685484042 this one
does anyone know why Okada doesnt work when EAC is in instance
/find
No query provided
yeag
do u know a lot about okada
i don't use realtime sorry
ohh yooo
YOOO TaewooOO 
thougt you gone
btw y'all
v3 prototype pretrains training has begun
my old acc was blown up
oh, shit
someone reported my age was 5 and they believed it 😭
bruh
that's such a d move
your typings are still fast af
ikr
arent you mad at me n shit from what i remember

thats nice
anyway, u excited for v3 pretrains? lol ( prototype ones but ye )
v3 is already out??
Training rn
o lmao
i think i havent got on discord since they announced v3 is going to be built
excited for better model results once they release it
me n noobies are working on it
as for my part, it's already started ( after all the darn frustrations erhm
tho as a sumup, improved architecture on the way
#🧬│ai-chat message if you want to take a look at everything added
also:
- Silu activation instead of leaky relu ( at least for prototypes that is
- tensorfloat32 training cause bfloat caused awful issues sadly
welcome back, haven't seen you for a while
thats a lot
yuh
so pretrain v3 releases are soon, and then v3 trainings?
speak english 😭
Basically, leaky relu is an activation funcion, responsible for activation of neurons in ML network, kinda like that
silu is essentially better, most often. Noobies proposed it so, why not
oh
you can technically start training now if you can get it to work #🔊│ai-development message
yea but I still wouldn't advise it yet, without pretrains one can effectively do nothing as of yet + it's a test repo so I can't really debug everything yet
proper one will be released along with pretrains, for testing
m1 mac will hate that 🥶
let me get back to my hibernation until v3 is released
wow
anyone will try porting your rvc to kaggle once the pretrain is finalized?
I may plan to try it by adopting hina's notebook..
So far nobody mentioned porting it so, probs there's no plans
so if you'd want to ( once it's done ) yea, ofc, I don't see any problems
@solar torrent how do I save drowning fish
Why did you ask me like that?
Kendrick Lamar's the Heart Part 6 and his latest album got released today. 

It is the General chat ?
hey lets get rid of verified
🔥
i think we should do it
yooo wait
someone already did
thats great actually, couldnt stand verified
why noooot
sorry, my bad
what
He
Hey wsg
yea verified totally wasnt useful at all so it makes sense to get rid of it

look at chat name
Hello
What
Hello.

shilling! :) i know damn well some of yg are from the bay area
No, thanks.
aint clicking a link on discord 😭
lol
ask for some help in #✨│ai-help , I'm sure someone will be more than happy to help you out
there aren't really much
if only the beatrice model are working, it means 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
You can search rvc ai voice models at:
- #1175430844685484042
- In #🔍│find-models :
- Send " @atomic vector search (name of the model)", without the ()
- Do /find with @hidden grotto
- 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.ai-hub.wtf/essentials/how-to-make-voice-models/
Unknown Command
aren't u the guy who also promoted twice latent.toys here
use #1159290752195633273
WHAT
hi bal
why is verified called ai chat 
its actually #✦│chat 
verified chat -> the lounge
general chat -> ai chat
okay 👍👍
now they should rename the verified role to lounge
eh
maybe
i mean verified role allows people for embed in AI HUB by Weights ig?
don't really think anyone has embed perms now
also bea the weights admin is a filipino women right?
i don't remember that
are u filipino?
🤚🏻hello
HI ALL
Ayo? @signal socket level 1 !!! 
how do i see a list of avail ai voices, i click voice models but nothing there ty xxooxoxxoox
have a look at the message Nick sent above, should be able to help you find the model/s you are looking for
#🧬│ai-chat message
jusnt sing and infer that
by removing the copyrighted content
hi guys
Yk the whole song
Ayo? @ocean summit level 1 !!! 
remake the song
Ayo? @copper sun level 2 !!! 
I tried it but it didn't worj
Looks like i need to write a personal mail to kanye west
just go to his house and ask him 
Guys, what is level 1? It's just a bot, periodically resets level 1
!!!! and I just don't know what it is
why UMG always do partailly block my ai video on Russia? because of the conflict??
maybe, and what was your video about?
AI Cover of Mephisto
its just the level you are in the sever, the more you talk and send messages the higher your level will become, you also get specific roles when hitting certain levels
bot just announces when you level up
mephisto, this is the blogger who does collaborations with the channel Chuck Review, you mean this mephisto?
no, i mean mephisto ed oshi no ko
Then you must be a really good producer if you made it almost 1:1 that you got a copyright strike for it. lol
Thank you! It's just not very pleasant when you are sent for several hours to a state where you cannot send messages to the chat, there was a feeling that I violated something and I was sent to a temporary ban
then it's strange, although in Russia it's very difficult to accept trends, like the one that proposed an initiative to deprive parents of parental rights because their children are quadrobists, but most Russians laugh at this and make memes
Ayo? @lapis raptor level 2 !!! 
Thanks bot for the notification
You separate instrumental and vocals. Keep the instrumental. Sing a new lyrics with your voice and convert it with some model. Place the instrumental and new converted vocals into audacity or some other audio editor and render/export it.
👋
Ayo? @elder willow level 1 !!! 
mvsep . com
weights gg
Audacity
-colab
- Applio, by IA Hispano Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- RVC Mainline, by Hina Google Colab
- AICoverGen-WebUI, by Hina Google Colab
- AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
- AICoverGen-NoWebUI [Spanish], by Eddy, Hina and Gdr Google Colab
- UVR5 NO UI, by Eddy Google Colab
- UVR5 UI, by Eddy Google Colab
- Modified W-Okada's Voice Changer, Google Colab
- 🆕 FaceFusion UI, by Nick088 Google Colab
- 🆕 FaceFusion NO UI, by Nick088 Google Colab
- 🆕 EasyGUI, by Rejekts Google Colab
While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.
Wth happened here?
There was a server exchange going on. 
Does verified chat still exist?
is there an AI deepfake that works well livetime and can use a cloud gpu ?
any help would be appreciated
realtime is what i mean
cloud gpu isnt a must
with livetime you mean webcam?
yes webcam
Ayo? @elder willow level 2 !!! 
that u can route through omegle or smth easily
you can try facefusion but you can use webcam only for local
I never really used the webcam thing, nor done it locally
what's ur pc gpu tho
3070
ye but surely you can use OBS to capture the local screen
and route through OBS into whatever else no? through OBS virtual cam
your pc gpu is good enough, you don't even need cloud gpu
technically as explained in https://docs.facefusion.io/usage/deepfake-webcam
ohh okay thank u
is "deep live cam" a rat
or is it legit, i see it on github and appears legit but not sure just wondering
never used that
oki

THIS GOOBER O;G
I'm not joining that.
Source: trust me bro.
ilaria rvc
does anyone have Latina E-Girl model? please will die if I don't get it
-rt
Interaction has expired, use the command again for a new interaction.
do you have?
I usually use anime woman voice models, not Latinos. 

hey I haven't been using this for a while, where do I download it?
Ayo? @narrow kiln level 1 !!! 
I have a question, What is the best vocal remover for Applio?
wow what is this cool server
- How to use RVC Mainline Colab by Cauthess
- AICoverGen Colab Guide by Eddy (Spanish Helper)
- Create a model with RVC disconnected (colab) by Angetyde
download what?
Me again with question cause i dumb, was here not long ago and had to get new account for dcord i rejoin today, before voive models had ton in it now not and takes me to weights place and before was a diff place i forget name any idea where i went wrong here lol
I think i already replied to u here #✦│chat message
Guys, I found an AI that you can play Minecraft with.
In Turkish conversations, it looks good if we don't count the artificiality and the waiting time.
@meager estuary duuuuude we need this
We need labor
where did u find that
if its the same I believe its this
https://github.com/kolbytn/mindcraft
@wooden jolt Excuse me for the ping. I need to talk to you about something.
nope
i prefer a real person over some NPC ai 😭
Umm alright
yes
what is it
I'm gonna try this
Update on training:
Sadly, prototyping failed due to an issue of which patterns are completely unknown to me.
I suspect the core handling of multi-gpu or the way ddp is configured.
Decided to adopt accelerate and potentially few other things from rvc-trainer repo
It's fine buddy, take your time.
Yeah, I'm still kinda sad cause getting it to run n properly unpack the set + preprocess it all took quite a bit of money away
I suspected issues but not on that scale :' )
Tho, thanks man 🥬
is there much difference between an rtx 4060 laptop and a normal rtx 4060
like ik it's worse but is it significately worse
or just a lil
.
does anyone knows an alternative to this? https://huggingface.co/spaces/Nick088/Fast-Subtitle-Maker , it time from time its giving me errors and i need this to work so please does anyone know an alternative?
in my experience everything in laptop is significantly worse and im never ever getting laptops again anymore
people says that is +or- 20% loss i think its more
oh that's my space lol
If u want I can help u out
bc i just tested it myself and it's fine
btw, an alternative would be using whisper locally, if u got a good pc
thank you, i dont know whats going on but today its giving me error all the time
Ayo? @rocky sage level 4 !!! 
it happens randomly some days
could u send here the input you're giving it ?
i heard people say it was just 10% worse
the one I tried was not open source
ahh, didn't even know of this one, seems about the same experience, open source one just doesn't have tts and you can feed a faster api into it
think Altera might be a little smarter tho, but I havent been able to test it with a better AI
is it free tho
oh lol
nope
Ayo? @analog cosmos level 30 !!! 
well limited free use
damn
open source one is better
Which did it better
you are pretty limited on free api's you can use, unless you want to run Ollama locally
groq gives 15k requests daily for free
gemini gives 1,5k requests daily only for the flash models for free
u surely have more than 30 mins with this
🐢: eats the lettuce
meow
good turtle
🐢 ❤️
10 minutes free daily renewed after 5 hours
5 hours is not exact, I may have exaggerated, I have 3 hours left.
There are some bugs. For example, I say we should sleep.
and when night comes, he says we should go to sleep.
different
Can someone give me a good configuration for the wokada? I use a rx 6600xt 8gb
Ayo? @elder willow level 1 !!! 
you can understand if you experience it
it also loves getting stuck in a loop from what Ive seen
Hello the bot told me to introduce myself what do I do
Dumbass bot
guys i cant find the link to the google colab to make ai covers anywhere 😭😭
what's ur pc gpu
google colab is only for people who don't got a good pc
KSI sings his new song to Ishowspeed on a facetime phone call and It's hillarious.
Thanks for watching my video, please like, share and subscribe.
#KSI #speed #ishowspeed
ksi is crazy as fuck
he's such a crybaby, he got mad just because speed "laughed" at his song
thick of it would be better if the song came out in 2016
o
Ayo? @jade hamlet level 1 !!! 
can i make an ai model cover a song without me having to sing it
like apply it as a "filter" ig to exisint audio
i have separate audios for the beat and the actual vocals
@hidden grotto
Unknown Command
Ayo? @alpine hearth level 1 !!! 
for modifying the lyrics of the song,
No
you could try with TTS (Text To Speech), but it wouldn't sound good as TTS isn't good for that
i worded it in a bad way
cuz i had the file with the vocals, yet i said i dont wanna sing it
either way i have the vocals, so doing it with /create works fire
either way, thanks for thehelp
oh lol
yea
anyone got any rvc gui
RVC GUI is outdated
don't watch old yt tuts
what's ur pc gpu?
4060 ti
is mainline better than applio?
mainline is just the original version
id suggest applio as it got an easier UI and is kind of faster
alr
Ayo? @lime basalt level 1 !!! 
thx
yw
I got banned for 6 days for posting rage bait on suno's discord server lmao
-gui
-svc
i never knew what rvc gui was
it was original rvc or what
Mangio as a forked version of RVC is outdated. I wonder how many people still using it for training models. 
rvc disconnected colab
simple as that.
Was I talking about Colab before?
no but it could be a reason why people still use mangio rvc
rvc disconnected uses a mangio rvc fork, for all the coding such as preprocess, training, etc.
so technically users still use mangio rvc in rvc disconnected
I think like people have ios and can’t figure out wtf kaggle is
My guess tho
ipad users should tho
yeah its more comfortable to train (on cloud) on tablets than a phone
can anyone help me , my mic is detecting audios from desktop
i am having this issue for so long
dont use Stereo Mix device as input, and use headphones that dont bleed to the mic
If you're using W-Okada, make sure you have Virtual Audio Cable installed.
i have installed it buddy , its not the issue with that , its just my pc i sending desktop audio as mic output and its so annoying
Ayo? @elder willow level 5 !!! 
i have it disabled
Someone have a website or anything else to make a text to speech with ai model ?
Weights is a free site that can do TTS with RVC voice model, but TTS there is kinda generic.
remove the monitor option and keep passthru option green in voice changer
its not the issue with the voice changerrrr
i even check it with discord
when i use hear myself and play something on spotify
Damn, so what else then?
it catches its audio
But i want something without number caracters restriction
Uh, there doesn't seem to be an AI TTS site with no character count limitation without paying for it.
then uncheck listen to device in the mic settings
https://cdn.discordapp.com/attachments/1159290161683767298/1308783304312623136/image.png?ex=6743d02f&is=67427eaf&hm=4d080ba453e6d09c7d6e867f968ca39cea5def682aaac533f32eb2ea4db68c84&
yeah

Do one of your programs just output to a virtual audio cable? Make sure you check every program that is outputing its sound, and make it output to your speak/headphones instead of a virtual one.
I'm asking this because I've been trying to make one for over 6 hours non-stop and I still haven't finished, I'm going crazy
Ayo? @topaz osprey level 1 !!! 
i am not a dummy cmon man 😭
i am not even using audio cable , its just issue with my main mic even 😭
Sigh.
if it was a normal issue why would i disturb you guys
i tried myself , i cant figure it out thats why i came to professionals
i have my mic turned off from my headset physical button
but its catching spotify
If it's not fixing, then I'm sure what the hell to tell you anymore. Try disable these virtual audio cables in both Playback and Recording, and see what happens.
its issue with my headhpone
when i disable it and use the speaker of my monitor
it works good
now what to do ?
man I just do brute force the possibilities, and if it is from unknown 3rd party program you might use, I cant help, but why ur reply pisses me off
Does your setting look like this? Because I couldn't identify the problem from cut screenshots you sent.
yesss sir
like i know people are dumb sometimes but i value your precious time so i wont be bothering you for shitty things that are easily available on google , i am sorry if you felt bad but my way of saying was not what you understood
i have reinstalled windows completely yesterday so its not much i have in my pc
Do your settings look like these?
Yes sir
If you microphone picks up your speaker audio, then I think it might be your system or PC issue.
If the mic is off with the physical button still how can it pick audio
Its the software issue for sure isnt it
YES, that's what I'm saying.
So whats the solution buddy
🥲
Please don't ping me again, leave me alone. I'm gonna do my own thing.
where did general channel go?
real
this just, it feels too wrong
What TTS to use to be able to use the downloaded models and not have to use 15 sec of audio?
i miss verified chat 😢
#1291523301843140678 message for more info on the merge or whatever
alright that makes sense now
but despite it all, it still kinda feels wrong
yea a bit
especially considering some earlier talks of channel name changing etc
it is what it is tho
the idea was declined
and now seeing the changes idk, but yeah, it is what it is
at least that's sort of a way to boost overal activity

yuh
and me
nerfing fast typers 😭
Yall i don understand nothing abt spectral denoise
it's easy
Glad I was the one to spread on rx processing 
awesome seeing people finally using right tools
that was you?
wow
i thought it was like a bunch of people started doing it
and then it spread from there
ofc youve been here forever
but I paid a high price
sanity?
all that speedy af learning on ml, rvc and python did cost me sanity and 5 burnouts
well I wouldn't go that far obviously, we've more talented people now than before ( first edition of ai hub )
back then it was mostly me, Hispano, Kalo, mangio and that one Russian dude
iirc
'that one russian dude' lol
don't remember him
simplcup?
nope
hmm
one that left due to dramas or something
oh
they did the formant shift in mangio rvc
i think i know him
tf keyboard correction
Tho, I still remember I once fought with Kalo on why tensorboard syncing was important
back then it was considered as not important
nor saving every epoch for accuracy
what was kalo argument?
yea surely not important if epoch can have 23 steps but you log every 200
lol
that was his ideology
interesting
yet I'm still grateful for his job and lots of indirect teaching when I first started
a good guy
ofc he helped a ton
btw, I'll digress but, I've replaced rvc's way of handling stuff in training with hugging face's accelerate
was actually a good call on noobies, too bad I took it in too late
such a helpful gimmick, automating a lot of head wrecking stuff
what does it do exactly? It is just a debugger? ive never looked into hugging faces accelerate
in a short, a nice wrapper ish for DDP
higher level library that simplified multi gpu training
data distribution and so on
i see
now I just need to figure out all speaker processing at once
im looking at it and looks nice and pain free
both preprocessing and feature extraction wise
to not sit for half an hour doing all speakers manually oof
oh thats rough
yeah, wasted quite a bit of compute on that as I expected for things to go faster, was wrong.
gonna aid it tomorrow and go back to training
anyhow, going sleep rn so, gn. see you around 
gn and gl on training
hi
hi
hello
Ayo? @compact plume level 1 !!! 
hey there
I'm new
I need a tiny favor
do you know any free platform to change voices?
Ayo? @light hound level 1 !!! 
Hello, can anyone tell me! Where can i find ai voices of everglow?
Have a look and search for it in #1175430844685484042, #🔍│find-models or look for it on Weights.gg
if you can't find one then nobody has trained one yet
Scroll a little bit at channels section and you'll find #1175430844685484042. Chat here is not where you find models.
is weights.gg for free , cz i i have visited tons of webs and all they ask is MOny
Ayo? @light hound level 2 !!! 
So i need to subscribe?
to just download models no, you might just need to make an account
An account needed for doing AI cover there, a premium subscription is optional.
pao
Amd Radeon graphics (?) Idk what that means but that's spperently my gpu
Integrated or dedicated? If your PC has Intel CPU, then your AMD Radeon GPU should be dedicated.
AMD Radeon(TM) Graphics is IGPU
Has w-okada been updated at all in the last 4 months?
-realtime
Interaction has expired, use the command again for a new interaction.
W-Okada has been implemented recently.
Flux and XL use more resources than Stable Diffusion 1.5, so no.
oh nooooooo
what's the best model that could used on gtx1060
😭 do you know? please tell me
Stable Diffusion 1.5, as what I said.
This version of Stable should be able to run fine like RVC, but it won't be that fast.
ok
kick this guy
omg really?
I have a gambling addiction from a minecraft server and this sounds like a great way for me to make some money
I mean with 2 people referring it I think it's safe
must gamba
GAMBLE GABLE BGAMBLE 🤑🤑🤑
Why is this server named "AI HUB by Weights"?
Hello guys , im new to the server
There was a server owner exchange between AI Hub and the Weights going on.

NOOOO
I gambled all my money away 
Welcome, new user. 
Ayo? @pearl relic level 1 !!! 
ais that doesnt completely "take over" human jobs\
ai that takes over top-level/CEO position
If...
merhaba. cengiz kurtoğlu ses modeli var mı acaba?
u know where i can upload my own song and give it lyrics to sing or no
is that not possible?
its not song server
so you should do a backflip inside your home so people can taste it
then you would be able to drink uranium
im not trippin, im just chilling
Ayo? @worthy coyote level 11 !!! 
with a cup of tea; inside of it, nothing.
huh
anyways, don't worry sir. someday you will miss these good days and sing a song like an angel with a shotgun.

anyways; Have A Unbad Night.
okay bro 😭
Integrated
But I jusr want to know where the colab link is mentioned, I can't find it anywhere 😭
yea your pc isn't really good enough
u could do it locally if u got 8gb ram
on cpu, but it would be slow
Oh, I see
Ayo? @ebon pine level 1 !!! 
You can do it on 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
you can use any of those
THANK YOUUUU
Oh so Ilara one is only text to speech?
All of those have TTS (Text To Speech) too btw
I do want to remind you that RVC is made for STS (Speech To Speech) tho, all that those programs do is just make an TTS audio with another program, and give it as an input for RVC
So might be they don't have the best emotions like 11labs (paid tts program)
Oh.. but are there ones that can do like singing ? (Sorry I'm a bit of a beginner) 😭
Not sure about weights.gg tho, haven't tried their TTS
like, sing just from your text?
Nooo, from an audio input
oh yes, all of them can, they all use RVC
oh, thank you sir
the first Iliana one keeps having errors but the second weights one seems pretty easy to use so far, so thanks for the help!
btw the 1st is named Ilaria RVC Zero
It's better you check out the trouble shooting part for that
Table Of Contents Introduction (with website link) Model Loader (Download & Upload) Inference (use RVC AI Voice Models) Ilaria TTS Settings (Inference) Vocal Separator (UVR) Troubleshooting “No GPU is currently available for you after 60 seconds” “GPU task aborted” “You have exceeded your GPU ...
but yea weights.gg works too
yo
Why does every voice model sound so "robotic" in the Voice models channel?
Can anyone recommend something that does not sound like wall-E?
quality varies per model maker 🤷♀️
I know, but the problem is we can*t filter for like "most likes" or "best" in this discord
Ayo? @spare gulch level 1 !!! 
you can try looking at models from people with either the "Model Master" role or "QC" role
or is this possible?
I will try that, but do you have any suggetions? Don´t care if it is male or female
i don't use much other models than mine tbh and mine's just trash too so I don't think I can give you suggestions 🤷♀️
Ok, but thanks for the thing with model master, Ill try that one
should atleast improve my search experience on here
who can help me
i agree. There was a suggestion to make a hq models channel https://discord.com/channels/1159260121998827560/1216988150623375390 but that never got anywhere sadly

Why are you even mentioning them in the first place no one's talking about them...
Yes I hope your new RVC version will be able to handle laughs
The name is shiroyasha
any voice changer tutorial
i love lightning.ai
Ayo? @elder willow level 46 !!! 
activate windows yo
license is expensive on my country
m
send a dm
ok
Quick question, what good de-reverb or de-echo models are around these days?
kind of want to get back into making models lol
bs roformer or mel de-reverb, and uvr de-echo normal. there's a wip mel deecho model which will be better than uvr de echo normal, so at the moment use uvr de-echo normal as de-echo model
Its also worth trying RX 11's dereverb. I hear its awesome
oh might give that a try, haven't touched my **legit **copy of rx 11 in a long time
UVR de reverbs are too destructive for datasets i think. Unless you already split it with something lol
I just throw Aggressive at any vocals I'm inferencing tho
every uvr de reverb are too destructive, the less destructive de-reverb is clarity vx dereverb pro
you can adjust how much the de-reverb is applied with an eq
Where is that?
Oooh Waves
waves yep
every automated de-reverb (uvr, rx) damages the audio in the same way
waves allows you to control that, being able to destroy less than the automated options
never noticed that uvr dereverbs destroyed my dataset. just that in some cases they're too aggresive
i remember that razer told me about clarity vx dereverb, so i may ask him about that other day
@night lake is clarity vx dereverb actually worth it?
yuh
hai
cool then
damn didn't know u pirate
REAL
might wanna send me in dms in case i ever need it?
I don't really need it now, but in case I get a new pc oh well
Wait
nah u are going to get ddosed if u run his command
nuh uh
nuh uh
domo
Where the best places to make images for ai
no.
its something different, let me send you the command
please
i know silly
its a magnet
oh ok then
before i dive further into the server are any of yall furrys just need to know(ignore the pfp)
wkwl'
what
hi
🥹
Hi guys! What's the best platform/app/whatever do you use to make AI song covers? The one I have right now is RVC but the songs come out kinda weird, not how it sounds on some of the samples
It also depends on how clean your audio is, plus you can play with the settings
any tips?
You could play with the index rate or change the timbre of the audio
explain what's wrong on? did you have trouble on separating vocals or something?
It has some sort of subtle distortions
Ayo? @maiden cedar level 1 !!! 
what kind distortion? could you post a sample audio through https://pillowcase.su/ ?
(please not as discord attachment regardless if copyrighted or not)
hiii
If by distortions you mean bitcrush like effect or "artifacts" in audio, that can't be helped as it's due to model's quality and no settings will fix it.
If, on the other hand, by distortions you mean " unstable pitch " or cracking and such, that's due to the input ( " vocals " ) you use being dirty ( contaminated with too much reverb, or with delay, having harmonies / layered vocals etc )
A distortion on audio could be either that a waveform for audio has been distorted and stretched like crazy, so it makes it to sound crunchy and loud, or a vibrato effect that makes audio to pitch up and down on repeat simultaneously.
Hi! I'm a new user! Used to be really into making AI covers last year but I got away from it for a while and am excited to get back into it!
Welcome, new user. That's good to know that you have used to make AI cover. 
the audio sample and spectrogram could be more evident to show lmao
That's real. 
hii! I'm nutst. I like to play with some voice models. it's interesting. nice to meet u guys~
Nice to meet you too. 
heyy
hey
Hey peeps! I am new to discord and in motion to build build build ai agents, I'm a fullstack dev that is switching to the world of ai to build some pretty amazing stuff. anywho, im excited to be apart and hope i can add to the discord in any way 🙂
Good morning AI HUB by Weights
is there anything better than audiosr
what are good input and output settings for okada
Goodmorning AI HUB by Weights
by Weights*
u can try apollo for mp3 only, or resemble enhance for speech only
oh yeah i already saw that
sounds nice but i don't like how it automatically finds the cutoff point
don't like this
hi
Ayo? @gritty umbra level 1 !!! 
new account brother
Hello
Hi
hlo
Hello
Hola
Greetings
setsuna
I know who I’m making a model of next 
uh
I've been wanting to make a Yena Youngblood model for some time lmao.
I've only understood "we request the takedown of misleading and offensive AI stuff made off our talents"
They didn't say anything about AI Covers or AI Music so i don't care 
As long as it doesn’t use a talents voice it’s chill to them
Hmm well they don’t really say that but I would assume so
Imma make a Yena Youngblood model on the future and make her sing any song along any holo talent and post it on the FWMC-AIR site haha
Lol
they’re probably mad at gooners making loras of their characters lol
Yeah
Yep, that's more likely than ai covers being the reason.
trvthnvke
i'm screwed
i have made 3 models of the idol en vtubers
except poko allows ai models so...
might take 2 of them down (by two of them i mean enya and roca once i'll get back to my pc)
Never take it down unless they personally told to
oh
Leave ‘em up you have a none zero chance of become famous now
okok i will
Nah you're not.
Is this for Hololive?
Nope.
Idol
Idolcorp
Ah
En es and jp idol
so i'm like still "allowed" to make idol vtubers' ai models right?
Yeah, i've seen little to no AI cover content from Idol talents on Youtube anyway.
Yep.
You're technically doing nothing wrong
Everything is allowed if you don’t get caught
one of the idol talents named poko allows ai models of her iirc
People might find it poor taste but that's their opinion
If I were you I kept continue making it
Tell it to me, i've done a lot of latin artist models that got used on viral remixes and AI songs and instead of getting asked to take down something, i've got praised for what i've been doing lmao.
so i should be fine with the poko model, dunno about the rest of idol talents i've made voice models on.
You don't have nothing to worry about.
You can keep making Idol talent models
what if i'll get caught?
They will ask you to take it down
What are you scared off like genuinely lol not like they're gonna knock on your door, at most it would be a personal gmail to take down the model
Yep, that's it.
But that's pretty unlikely
And if you don’t comply maybe you will get into legal issue but not likely
I have lot of indie vtuber sources sitting on my SSD
Right, it's not that serious
Because unlike VA models, i have seen little to no Idol AI cover content on YT. And that's how the VA can find you and tell you to take down something.
RIAA get my ass off of this shithole country plsss
i got a fckton of protool sessions and stems on my external hard drive.
earlier this year i used to ask some vtubers if ai voice models of them are allowed to be made or not
but now i just don't care
Me neither.
A PERSONALIZED EMAIL FROM MY IDOL, IDOL CORP OMW
Again people might find it poor taste because you can argue it's unauthorized
Imma just say "I don't care anyway".
I think as long as you don't use it for malicious shit, just go on ngl
That's what i've been trying to say.

I mean what are they going to do to you? You’re in like a different continent from where they are

They must understand the difference between using their AI voice for commercial use and simply wanting to hear your favourite character sing your favourite songs
completely different spectrum
Yep, there's a lot of people that refuse to understand.
Look, even some latin artists got impressed by the work me and ProodbyNG did.
Instead on issuing takedowns on us
Like god and comet skill! Veteran and Top skill.
I have 7+ years of experience in Web development services with a success in leading projects from concept to development.
Collaborative and proactive, Thrive on new challenges and continuously upgrade my skills. With extensive knowledge of Front-End development , Back-End design, Mobile, Testing / Deployment.
My sill set are follow as: T3, Next, React,Redux, React Native, iOS, Android Github, Reanimated Figma, Tamagui, Svelte.js, Typescript, Flask, Python, OpenAI, Stripe, pinecone, Java, Spring Boot, Spring MVC, Apache Kafka, Zookeeper, AWS, microservice, RESTful API, MySQL
I promise best quality & clean code and in-time delivery.
Let's contact and make something remarkable together.
I have 0 year experience in web development. 
Is this a fucking resume
Ayo? @serene shadow level 1 !!! 
prob another AI bot
Yea u dont say
AI Hub isn't where you find your job, baby. 
Ai in general isn’t where u find your job
And if you’re trying to do that you will just be replaced too just like the people you replaced as a job tbh
I did it for like 6 years or so
I just read the name of that youre ill
Ok. 
No
@sour night hi, meet @opal marsh
hi ! i'm new here ! can I use this program to change voice while streaming games with a vtuber ? i doesnt work for me the program dies in games not working properly works only standalone
i already know him very well
yea ai replaced u
might be that ur pc gpu is not good enough to do ai + heavy gaming
its i5 13600k with rtx 3060 12gb
That's a crazy and decent specs your PC has. 
yes and the voice changer dies strugle with games on and vtuber
only works for voice recording and discord - light apps that not share gpu
what do u think about it, I am frontend developer with bunch of experience. you have any project?
Lmao im aight
Ayo? @serene shadow level 3 !!! 
I don’t share my stuff im weird
bro think this is a job seeking hub
Your word is very shark!
Ayo? @vernal flare level 1 !!! 
your word is very 🦈
ngl, no idea what's up with such people recently
The frequency of those " offering pro services " seems to increase gradually
server's called ai hub and alot of people are trying to get ai related jobs sooo lol
fr
oof
actually i dont think its increasing i think its the same and that we are just noticing it more bec less people are chatting now
oh, hmm.. good point
how to create my own model?
what's ur pc gpu?
rx 7600
guys what is the best ai for humanizing my text
you're on windows right?
yes
read https://docs.applio.org/applio/getting-started/installation#amd-gpu-support-windows to install applio
applio is a fork (modified version) of RVC (retrieval-based-voice-conversion) used for inferencing (using) Speech To Speech models with pre-recorded audios and train (make models)
me totally not leaking windows stuff that im not supposed to
Are you a windows insider-insider
lol
we got the same monitor
no i am eggman, the person chasing sonic
Wake up, everyone! A new RegalHyperus drum model just released!
Kung Fu Fighting (Drum model no. 544)
what?...
Ayo? @elder willow level 1 !!! 
hi
Ayo? @patent dome level 1 !!! 
do you use windows
I don't daily drive Windows
you might soon
I don't think I will
:<
i only use it for niche things like gaming and music production
i could use it if i wanted to
thats great
Hi, I need help. I used to make AI covers with a Google Collab, but it just stopped working. Can someone help me? I was using Jammable, but it's really bad and only converts a random part of the song, not the whole thing.
youll get it soon
perchance
how come you didn't do that on windows 11
that would make more sense to me
i dont use windows 11
and fit better
ok
ee
also this was vfx
i had the feeling it was
cuz the stuff added on looked too bright to me compared to the monitor and it looked slightly too sharp
but thats pretty good
its been aa long time since ive done any vfx but im too lazy to get into that again now
it felt uncanny to me
if u recompressed that it would be hard to tell
a lil outdated but alr
Btw Helen, you shouldn't handle it like that as you inevitably ping others too
Instead just use " copy message link "
It'll redirect to a given msgs without pinging
is it legal to post a song with ai on spotify or sum
like with a diff artist ai
and just not credit them
-applio
Not available yet
Ayo? @elder willow level 1 !!! 
hello
I was wondering what are some of the best guy voices I tried looking but couldnt find any that I liked
you could check https://rentry.co/ForkVoiceChangerGuide#voice-models-to-try-out
Guide style is in the same as Blanc_dot's. Thanks Blanc_dot for corrections. Most technical information comes from deiteris.
Last update October 6th, 2024: Multi PC setup explanation added
Translations added for:
German: https://rentry.co/ForkVoiceChangerGuide_de
Turkish: https://rentry.co/ForkVo...
also be sure to use the wokada fork and not some yt tuts
other than those suggested models, you just have to try and test if you don't like even those,
there's no way to know it's quality before testing it (like more epochs won't mean more quality or smt if u get what i mean)
ty
yw
Pikachu is still around? I don't check this server anymore
Tetski is still around? I check this server every 5 minutes
Tetski has proven their relevance
Hey I have quick question is it possible to create text model out of person messages so it will write like the person? If yes to not waste more time where to find guide more info about something like that and maybe also how much resource intensive that would be
am pixelpioneer
Ayo? @oblique badger level 2 !!! 
use the astra labs bot
t
-rvc
- How to use RVC Mainline Colab by Cauthess
- AICoverGen Colab Guide by Eddy (Spanish Helper)
- Create a model with RVC disconnected (colab) by Angetyde
hi
Hi, everyone
did you guys know that this channel was the off topic channel
it's been so long
good old days
hey there does anyone have a pth file of peter coyote's voice
good clear trained file
You can search rvc ai voice models at:
- #1175430844685484042
- In #🔍│find-models , Do /find with @hidden grotto
- 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.ai-hub.wtf/essentials/how-to-make-voice-models/
:wave: @covert lake, How can I help?
Available Commands:
• @weights find <query> or /find <query> - Search for RVC Voice Models
• /create - Create an AI Cover
• /image - Generate an Image
damn u alive
add an llm to this lol
Scroll a little bit in channel section, and you'll find #1175430844685484042. If a voice model you wanted doesn't exist, you can find it at Weights.gg. Otherwise, you might wanna request anyone to make a voice model in #1159289738314919936 instead.
do you know what the best image to video / text to video model free or even paid is
Ayo? @velvet drift level 3 !!! 
what's ur pc gpu
text/image to video AIs:
- Locally (runs on ur pc):
- pyramid flow (Image/Text to Video
- cogvideox 1.5 5b: Image to Video, Text to Video
- Cloud (remote good pc, running on an online website for example, easier to setup):
- Weights.gg (paid only)
- lumalabs
- Hailoua AI
would you say lumalabs is better than hailoua ai or are they same quality because hailoua is very good from the text to video it generated its great
I don't really know tbh
alg thjanks anyway
i see no ones saying happy bday to @covert lake
thanks 
I need help, I am trying to find the artist named “DECO27”song being Rabbit hole. I am trying to make the cover for “Rabbit Hole but it’s 2009 Roblox song” and blending with orginal artist “DECO27” it’s not showing up for me to do a ai song for it

https://www.youtube.com/watch?v=eSW2LVbPThw u mean this one
「死ぬまでピュアピュアやってんのん?」
Listen & Download: https://karent.jp/album/3592
2024年11月27日発売
9th Album 「TRANSFORM」 特設サイト
https://nbcuni-music.com/deco27
Lyrics & Music: DECO*27
Arrangement: tepe
Bass: 堀江晶太(kemu)
Movie by OTOIRO
Director: yuka fujii
Illustrator: おむたつ
Designer: yuka fujii, lowpolydog
CG Artist: あわしま
Composite / Motion Graphics: Yuma Saito...
Yes and the “Rabbit Hole but it’s Roblox 2009 Roblox song”
Ayo? @fiery ridge level 1 !!! 
It’s all instrumental to that

AI HUB Docs
