#🧬│ai-chat
1 messages · Page 45 of 1
yes
I just purchased a physical cd made with an ai voice. Legally. With cash
if thepiratebay is still around, we can survive
ah I cleaned some of my servers and was confused on why AI Hub wasn't in my servers, thought I had left it on accident
bro how trying to make a ai voice illegal
It was someone who got banned for posting rat porn came back and spammed copyright content and reported it
what happens if that voice actor is dead
that someone infiltrating us to destroy the server
damn he's petty
The software is legal, open source and educational. Using it for impersonation is what they want to make illegal
oooh
aint no way what
good to know that the server is back
I don't get why people actively sabotage it, we're getting information and RVC help for free here
and more things
it was Drake's henchmen
some people are so ungrateful fr
It's stuff that goes beyond that
the sticker, there’s a version of it on this server but the original one is got it from is xunact
Egos getting in the way
It's complicated
what's up? did they have beef with the owners or something
cuz there dumb
Is this the new one or the previous server was deleted?
Notice how there's only an ai originals channel and not a covers channel?
new one
previous was deleted
its cause ppl are making $$$$$$ off selling voice models and they want that money
lmao
money, power what power anyway lol
Ohh I see, I thought I was banned lol I didn't notice it right away
haahaha
wow
clout chasrs
losers
prob dumb question, but what's the best thing to use to make ai covers right now? idk what to use since colab got banned :/
Local is the best
local
yea
if you have a decent gpu you can at least inference
for training you need more than a decent gpu if anything
It takes a ton of resources no one wants to give away for free
bro the latest mangio realtime is so good if you got a nice GPU
super low latency, soo close to be able to record live
still can't tho
Mine is like half a second with a rtx3070ti
run 20 games while having the realtime open
@amber light has reached level 20.
yea same im with 4070ti

Yeah running a game with it gets laggy
I think its gonna be delayed until they can get quality conversion with small chunks
@leaden bear has reached level 3.
imo i’d rather use okada
i prefer mangio
It will be about the same latency as an international discord voice chat anyways
can I use the RVC program with a 6700 XT ?
You need cuda
okay! do yall mind sharing a download link to what you use for local inference? 0
yup
-rvc
Hey, @tiny sentinel!
👇 Here are some useful resources to help you learn how to make ai covers
🇺🇸 English Documentation
Credits: AI HUB France
🇺🇸 English FAQ
Credits: AI HUB France
🍏 Applio-RVC-Fork: Guide
Link: Google Docs
Credits: IA Hispano
How to install RVC on windows
Link: YouTube
Credits: iroaK
RVC v2 Tutorial for Paperspace
Link: Google Docs
Credits: LollenApe
- 🇫🇷 French Documentation
- 🇮🇹 Italian Guides or
-it_g - 🇷🇺 Russian FAQ for RVC/Voice changer
- 🇧🇷 RVC Guides in PT-BR (
-rvc_br)
damn why
thanks for replying so fast btw/in general, i appreciate it a ton!
doesn’t support amd as of now as far as i know
We're here for educational research. The kids make the shitposts
Nvidia only for training models
interesting, I wonder what's so different about AMD GPUs that doesn't allow them to run the same tasks
i think Okada you can use AMD right?
AMD & intel works for inference
and what's different with nvidia
Missing software to actually make use of the GPUs
ah
WE LOVE AI VOICES & AI IN GENERAL
Intel has their own libraries, nvidia has their own which is the most used, and AMD also has something in the works
is applio better then the gardio one
this one
https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI
cuz this is what i used
i already installed this
i'll give rvc a try then! thank you!!! have a good day 🙂
ill install Applio one see which i like more
makes more sense now, thanks for explaining
the interface UI for Applio is better
okeee
Current situation in regards to model training :
- Nvidia just works
- Intel might work, not sure if it does on windows, but it requires a dedicated ARC GPU
- AMD doesn't work on windows for now, and support on other operating systems is unstable
at least more user friendly
yea cuz holy shit my brain is like "wut is this?"
true, it's a good interface
@tiny sentinel has reached level 1.
On the other hand, inference works on just about any GPU because devs bothered converting code to something that works on all modern cards
I had trouble tho, i couldn't convert voices without an index and was getting some errors
not happening in realitime..
but it works nice anyway, i just used some random index
hello ı am cem
ello
@glacial zodiac has reached level 2.
its cuz i have perms
😭
jk
u can send things in verifed when u get level 5
hey guys, quick question, from outside the gui, what's the folder to drop my .pth and .index files in applio rvc? It seems they're on the log folder. Is this permanent?
copyright

de echo de reverb model
thanks
dose it come with UVR or do i have to download it
download it in the download center
oke
under vr arch
thanks
found it your a god
lmao
yea exactly lol
banned
lol
i feel like a scrub losin access to verified
i just idled in last Ai Hub cuz i was too lazy to set up stable diffusion,TTS and RVC
where do i download mp3 file after using AI-Cover-Gen-No-UI google colab?
So any news on potential updates to RVC or newer program(s) like RVC that might be better than it lately? It seems like we stagnated on the AI voice cloning side of things now.
Rvc has the advantage of being relatively easy to build models
And being free and open source
Are any one of those possible replacements for RVC V2 in terms of better quality models, faster to train/inference, don't glitch out & get confused when there's background vocals in songs etc?
Making your own or just looking for professionally made ones?
SynthV renders instantly
But it's paid, and has 30+ pre made voices to choose
Diff singer is very good quality but not as easy to use. Free and open source
hi people
What about neutrino & cevio ai, never heard of them before. Are they free, open source & as easy to use like RVC but better overall?
Both of those are paid. Currently the best quality on the market is synthV
Ace studio has a 14 day free trial. Synthesizer V has an unlimited use basic limited feature version for free
SynthV can train models like RVC can? I thought it was more of a vocaloid like thing to make songs with.
@lucid lion has reached level 4.
Correct it's like vocaloid. In fact there's several vocaloids that are now available on synthV. You can train your own model for diffsinger
Hey, didn't there used to be an Adult Gohan voice model?
Hows the training quality comparing to RVC2?
More involved but more control
Thanks man
Well, I mean something more like RVC but like about as easy to use & free & open source but improved in ways RVC isn't yet. Anything out like that yet?
Rvc is going to be the best as far as open source, free and ease of use
Whats the best easy way to use a voice model?
Kits.ai used to be but they later introduced monthly subscription fees.
If you're going to pay you may as well get a professional legal one
DONT YOU DARE
Isn’t RVC2 free though? but kinda difficult to use?
Rvc is relatively easy to use once configured
Use applio if your pc is good
Do you have a video tutorial on how to use an uploaded voice model that has already been trained?
I don’t understand how to use a voice model from weights.gg and applying it into RVC2 locally?
Teto runs without a gpu
Yeah but you can’t create sick ai cover of Cupid
@storm meadow has reached level 29.
And it’s incompatible with any RVC programs
I use synthV for the acapellas then realtime change with rvc
I can’t seem to find anything about using voice models, I only see how to train them?
So the models work with RVC too?
I CONTROL the voices with synthV in real time
Dang, so we really have stagnated in updates/improvements to voice cloning. Hasn't been really any new developments since RVMPE a few months ago.
I can adjust lyrics and pitch instantly
Well the server being attacked didn't help.
How accurate do the voices sound
Anyone know about this when installing applio?
wait i cant send pictures
bruh
omg
nvm i think ik how to fix it
RVC v3 got cancelled a while ago too after it seemed like a pretty big upgrade & improvement, so that unfortunately went out the window.
Pretty good because it's clean acapella
A lot of those upgrades got applied to v2 and rmvpe
...
It be nice if we saw another improvement in RVC again or some other thing like it that could replace it possibly before the year ends.

I was 10 minutes late to my meeting cause of this AI shit
f
If you want better just go professional studio made ai voices
We got Disney princess voice actors, Broadway performers. Famous voice actors and singers in multiple bands
All legal
Something I'd like to see in RVC or otherwise is fixing the issue of it getting confused & glitching out when/if there's background vocals in songs. That'd be a big upgrade.
That's why you go with direct control diffusion,
Not just copy an existing vocal
Can it do text to voice ?
oh nvm i see.. yea i notice rmvpe is really nice quality
https://youtu.be/7NCD2CAXwmA?si=zllEaijWlVvxYZvw look closely, they also have a normal TTS too
Feel free to support me on Ko-fi: https://ko-fi.com/adamedmond
Uploading this first because DEATH cover got blocked many times, probably because of the music video thing so I'm doing the plan B video right now 👍
SVP distributed as well if anyone would like to try it out with other voicebank! Feel free to use it and don't forget to give credit w...
Mangio crepe still works good for some voices
Yeah but it will sound kinda robotic
There's elevenlabs for tts
i just wanna say, i'm still existing
oh maaan
sorry to tell you
gex with kalomaze
poop in a box n mail it to Kalo
haha woops wrong reply
💀
@ ADMIN trust me daddy
i'm getting the silent treatment
Wassup I'm Automaze!
- My prefix in this server is
-(customizable with-prefix) - Interested in how I'm built? I'm actually open source!
- Feeling a tad bit generous? Buy me a coffee!
-trust me daddy
Is elevenlabs fundamentally different than RVC or is it just a service running on top of RVC?
hmm
@plain wasp has reached level 2.

It works different
Interesting. Is it proprietary? I haven't found much on how it works. Maybe I need to read more academic papers
i read this as, let chatgpt read it and make me code for it
furry

elevenlabs Is not an open source service
Your a boykisser
Lol true, might as well
Hold on guys
Yea gotcha, I'm curious what they do behind the scenes. I guess that's their secret sauce
If you had a method better than everyone else would you sell it?
SynthV ai diffusion is top secret
guys can someone tell me which ghost face model is better? this https://www.weights.gg/models/clm736de51py7cctcyjfap7ds or https://www.weights.gg/models/clmfc2fnl000owsoud6onz5pp ?
id go with the most likes
Never in my life I ever seen someone make a ai cover of happy
Yea makes enough sense, i'm just a curious dude
i guess ill do it
From who?
Get Pharrell's album G I R L on iTunes: http://smarturl.it/GIRLitunes
Get Pharrell's album G I R L on Amazon: http://smarturl.it/GIRLamazonMP3
Get Pharrell's album G I R L on Google Play: http://goo.gl/fwLdhZ
From the Despicable Me 2 soundtrack. Check out http://24hoursofhappy.com to experience the world's first 24 hour music video! Directed b...
Sita
who tf is sita
could i send u the vocals i converted with both of them and you can tell me which one sounds better? cause i dont really ever seen the movie tbh lol
What's direct control diffusion? Never heard of it.
That's what synthV uses
Search it up its some singer
@wicked pollen has reached level 1.
i havent watched it either 😭
How to post photos here
try find some on youtube
Send me the link here if you find it
If I put up thumbs up it's cor-
trusted role
I have, it's my favorite horror movie series so I'm very familiar with his voice.
SEARCH UP happy by sita
w
i found out that if I sing something it makes the ai sound like way better. i made a joke cover of the song then ran an ai I couldnt get to do anything through it and it sounds way better
use the Ghostface Killa one, thats the best
the song or a cover of it?
The song
im talking about the fictional character not the artist
killah is an american rapper
yea just playing 🙂
It's bc the vocals are clean & not extracted though a song's instrumental.
Provided to YouTube by Jive
Happy · Sita
Happy
℗ 2001 Zomba Record Holdings B.V.
Released on: 2003-07-29
Producer: Douglas Carr
Producer: Ulric Johansson
Composer, Lyricist: Tobias Gad
Composer, Lyricist: Jacqueline Nemorin
Auto-generated by YouTube.
could you check dms?
👍
Well that’s the thing. I lost the original separate files. So I ran it through another extraction and it still sounded fine
@elder willow has reached level 2.
Oh cool
All users have left the voice channel. The player has been paused.
@wicked pollen has reached level 2.
Oh, I thought you sang the vocals yourself with your own mic. But vocals that are recorded without any processing on them come out sounding the best most of the time.
I did sing them but i just didnt have the orignal vocals isolated just the full mix
very uwu post
Anyone know any good deep feminine voice models
Dua Lipa if any of hers have been archived. She had a few on the first server before it got wiped & her voice sounds pretty deep to me at least.
Ok ty
You're welcome.
The scary thing is how I am not dating them yet
how to make model on Applio (Mangio-RVC-Fork)
can some on tell me
or find the video for me
i can find on youtube
Check on #1159513888199540817
its not there
Then ask on #📑│making-models
I'm back.
@last gyro has reached level 1.

:0
There's a female v model
But way too obvious
If ur tryna troll
й
Cyberpunk protagonist
aight I'll check that out ty
I'm sad all the models from the old server are gone for now
@plain wasp has reached level 3.
They should be on Weights.gg
everybody friend v3rku on discord and spam him with dead dog memes, hes a pedo so he deserves it
💀
🐕

Any way to get his personal information
To report him to the authorities
his name is daniel stevenson
You are actually a child
😭
The mature thing to do
This pizza slaps tho
Could you dm me the address, contacts or just an irl photo
😭
the guy in the r kelly fanclub didnt give me his address
i think he has loli stuff
I don't think saying 'he's a pedo' alone will make a good case
he wants to legalize lolis
@elder willow has reached level 1.
woah
can someone help me.
@stiff rune has reached level 1.
How can I post an AI Album without it getting took down?
Omg haven’t checked here for a while and old server is nuked
Post it on youtube then you can link it here.
hi
Old server got shot down because some loser malded too hard
Also I tried /resync but doesn’t seem to work
So it's an Album but It's songs all in one video but youtube keeps taking it down.
hmmmm maybe try to post one song at a time.
dang I wanted to do it all in one but I'll just do that
yeah that way you can find which song is getting it taken down. You can put them all in a playlist after you're done.
@placid garnet has reached level 4.
K thanks
is there gman model
does AICoverGen, by SociallyIneptWeeb , running it trough colab, only support nvidia card ? i got this error today saying it couldnt find my nvidia driver, but i never had nvidia , i ran it trough my amd and it always had worked until today
HOLI
Anime women arent real
true
guys can i get a tutorila for the voice changer ?
anime women are real
Scitzoprenia doesnt make them real
tortilla
tutorila
@fierce charm has reached level 1.
becuse if i believe that anime women truly exist even if hypothetically they didnt then for me they are real in my mind they exist and its real even if its a fake reality its still my reality taht anime women are real
anime women are real
hi
someone knows why i cant listen the audios from the voices models?
wonder the same
So the RVC no ui inference only requires NVIDEA GPU, is it possible to run it only on CPU? I am not worried about the time it might take
are you trying to make a colab use cpu?
No, I connected a local runtime
then why are you using no ui?
guys
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
are you using it on ur pc?
Yes
@surreal edge has reached level 3.
how tf are you using no ui locally
Well actually running a colab, but using docker to have a local runtime, so all the processing is done on the local computer
So you know how colab does processing in the cloud right?
yes
dms
ok
Hey, long time a chat.........!!! 😁
hi
Wow, nah that's 50 bruh. Come on 2 people rank are insane level..!!!
Are there new for models is coming up...
.
????????????
gamer
Blows out of mind........!!!!

:0
the rank is ranking
typing
Speed up typing from PC
i am always here
i am here so much when i cant use pc i am on mobile
Vote Booster: 10% (1 hour remaining)
✨ Tip: Use /card to modify your rank card
AHHHHH..!!!
lol
NNNAAAAHHH no way...!!
PikachuuWu just take rank 2 too far..!!
real
:0
theres no gman ai models yet 😢
make one
e
hii
Haiii >.< Hiiii ^-^
i dont wanna
@sullen sigil has reached level 1.
Yes
real
WE'RE BACK
WE ARE SO BACK
Hello everyone.
Apologies for something negative, but I'm having issues right now and I'm trying to resolve them.
Could anyone maybe try helping me a little.
@fallow pecan has reached level 1.
my sticker :0
i wrote that
Hello my friends
jello:3
Kalomazin', it's amazin', it's amazin', so amazin'. Kalomazin'...
repeat on loop
there was
Does anyone have any suggestions for software on which to use the voice models? (Windows)
I MADE MY FIRST AI COVER where should I post it
I need to lvl up
Don’t
Don't what
Don’t post it, unless your posting a link to STREAM it
no downloads to stuff with instrumentals
So what should i do
Is it on youtube?
❌ You have no rank. Keep chatting to earn a rank!
Not post a download
random vibes
Ok
For my gpu since I have a 7800xt which one do I select because I’m prompted cpu then gpu 1,2 and 3
Can RVC models be used for text to speech? Or are they just voice to voice?
Wth all you said made no sense
With easygui you can enter eleven labs key or use google voice and do tts.
I got a script that uses edge tts, but any combination I try sounds nothing like the character, am I doing it wrong?
Perhaps not a very great voice model.
Don't reveal when a fake name is fake
My name is Jeiafeia
Hi everyone!
Are there still any pure TTS models around? I can't find any others on weights.gg
@stuck crane has reached level 1.
It's my final piece for a Morgan freeman desktop agent
Can anyone like explain how to do all of this, sorry if its a bother
Or a video would be great too, just to learn the basics!
Cause I couldnt find any online
huggingface
Anybody here have an AMD Gpu using the DirectML Version of the AI voice changer
Like how do I even get my gpu to be used it's only using my cpu
Is there like a single introductory video
@chilly cliff has reached level 1.
SORRY FOR SOUNDING DUMB Omg
So does anyone know exactly how much feature index ratio & consonant protection effect the output of song conversions & what settings are a good baseline to have them on?
@lucid lion has reached level 5.
looks like google killed colab infering
and im out of kits.ai minutes.. goddamn it i need to make a sample
Using GUI or NoGUI
nogui
Uh oh
the one on -colab
@peak dome They killed the NoGUI now
training gotta still work right
@amber light has reached level 21.
If inferring is dead
damn it
Training probably is too
wait lemme try
please translate this
it blocked me on the model downloading part
i put in my hugging face link, then when it was done downloading, it disconnected me
can someone infer for me
also why does covergen not support zip files without folders in em
all my old models dont work anymore for covergen
and im NOT remaking 40 model zips
Wordband
local wont recognize my gpu
What gpu
God damn haven't seen a 20 series in ages
Hi chat
Anyone have a clue?
Just play around and figure out
i havent touched consonant protection but feature index ratio lets you make it sound more like the person with "accents". 80 is pretty good for a decent dataset
I was wondering. How does 50 epoch sound like?
it depends on the model
Like 50 is too low no?
some sound good at 50, some sound good at 1000, it all comes down to the lowest point on tensorboard
i think a safe amount would be 250 with a 25 save frequency
then on tensorboard, search "g/total" and set smoothing to 0, then look for the lowest point in the graph. go to that timestamp in the logs, and find the closest save to that time period
Yes,idk tho depends on steps
this prevents undertraining and overtraining
higher =/= better, and lower =/= worse
Why no embed in chat
Gyat damnnn fauna model sounds kinda good
Size means everything
@red portal has reached level 12.
you gotta be a trusted user or something like that
How do I continue training?
For a 50 epoch model
Not in #✦│chat
It got to 240 after 3 hours
And disconnected
Which files is it? Going to put them to an alt
What GPU do you have in your PC
Like which account you chose to store
It was colab
Is your google storage full?
Still tryna see if you can run local
Snapdragon 8 cores
Nope
Ok you should be fine
Just run the colab again,connect google drive and enter the same name you typed before
Lmaoo
And continue training
Gonna upgrade my laptop lesss gooo

@novel haven has reached level 12.

R7-6800H, 16GB DDR5, 1TB SSD, NVIDIA RTX 3070 Ti
Is training locally possible?
Not on a phone 
What if
What if it was running Linux
💀
Then it would be limited to CPU only and probably also burn down your house at the same time
Bro would be the new Oppenheimer
Rvcppenheimer
what are we talking about
"Cannot connect to GPU backend"
Nuclear bombs
Nah, that's the former owner of the server bc according to him, making this into a place about making & sharing models is the same as the invention of the atomic bomb & will ruin hundreds of thousands of lives.
@solar jacinth has reached level 2.
Menh said what?
Haven't touch training in a while
Oh u meen snoop?
Yeah, the snoop dude.
I dont agree with that
Snooppenheimer if you will. I don't agree with it either, it's basically what he implied before he left the server.
I mean like sharing models won't end the world
Obviously, but he didn't seem to see it that way almost in his last message before he left.
I mean people have opinions
hallo
introducing myself
.
@brisk escarp
tks
That Oppenheimer stuff was a joke 💀can't believe people took it seriously
He planned to leave for a long time as the server was going to shit
that's what snoop said
poopingheimer
im starving
Read my message after
I miss more BTS vocal models 
hola soy santiago juju
aooooooooooooooooooooo
Why can’t I find a Mark Hamill model? I know we had one before.
theres a joker version of him on weights.gg
How do I get there again?
@narrow terrace has reached level 1.
I have a quick question
You type it into google
@split lynx has reached level 6.
https://youtu.be/UAh44gOz1FA?si=u8FJb7ia-dWyWbQ5
Which neco arc voice model was behind this cover? There’s too many neco arc voice AIs and I can’t distinguish which one
The one with the cat
I mean they’re all cats lol
Just use one then
They all the same
Or u could use each model and find out which one u like most
yall i have a quick question
why does it say that i need a model and an input audio even though i put it already?
Rip to my pc
@hallow dew has reached level 1.
could that be possible?
I mean text to video is a thing
I just don't think the technology has a mature enough to be really that useful
huh, interesting
i just thought it would be cool for ai to make full on animation other than art
why does everytime i try to convert it says "Please select a model and input audio file"
@karmic briar has reached level 1.
hey..
since this server is a place for ai stuff, i wanna ask another question
could there be.. something better than character.ai?
idk
alr, fair enough
@neat slate has reached level 1.
lmao noobie, i love your bio
is there like a tutorial for doin rvc2 models i thought there used to be one in the server but i cant find it and none of the youtube ones are helpful
Yes
risuai.xyz + mancer.tech Mythomax API
sillytavern
Runs on your pc, supports nsfw, and a bunch of other cool stuff
why cant i convert a model and a audio file cause a stupid error called "Please select a model and input audio file" in rvc gui
You need something to run an LLM though
Google Colab banned LLMs
why did Google colab get pissy over ai stuff though
@storm meadow this is a way easier thing to setup in my opinion
Oh ok
Yep
It was using up a lot of their GPUs
pov: a multi trillionaire mega corporation cant handle a few thousands of ai makers
hi people
You really don't know how much it's using up
Google Colab is intended for academic research
lebron james
LLMs still work in Colab
@gleaming oak has reached level 1.
yall help me please!1!1!!1
@karmic briar has reached level 2.
i keep getting this error called "Please select a model and input audio file" even tho i got a model selected and i have an input file in there
oh yeah its in rvc gui
ss?
what
screenshot?
No I’m asking you to send a screenshot
r the voice models safe in #1159289700490694666
in dms? or anywhere else?
I was talking about the NSFW ones
Here if you can, if you can’t dms
i sent it dms cuz i couldnt send here
r the voice model safe
How and why would they have viruses
it has 1melon virus 😎
@undone hedge has reached level 1.
for applio when it sayus installing depencencies do i pick nvidia cuz i have nvidia gpu
cuz the video says pick runtime
Yep
thank you
no one has uploaded a singular TADC ai cover on youtube, that would be a pretty good investment at the moment
😳
investment? We invest in those now?
views
I need help with weights gg, I can't copy the model link
hi
for applio should the dataset be in zip or wav
i don't like dall-e 3 anymore
@quartz cloak has reached level 1.
cool
The old links don't work anymore on weights.gg
Which links?
The ai voices before the old ai hub got nuked
@elder willow has reached level 1.
They work?
The download links?
I don't think so for the old ones.
Works for me, might just be the ones your downloading
How would a virus even install via RVC?
Everything here is open source
If your using windows
You don’t need a antivirus
It’s just another thing that slows down your pc
Windows defender is all you need
Hey guys i was just thinking.. what if there was a model for a drum kit then you did beatboxing with your mouth???
ohh
make drum beats with just your mouth............ hmmmmmm
bro you are amazing thanks
@leaden bear has reached level 4.
i wonder if it would work
is it single drums? i imagine like a whole drum kit in one
It includes many drum models
Not sure, never tested it
lmao its easier to just make a drum beat , beatboxing is hard
i think its only single sounds
oh nvm actually
they extracted the all drums from songs then made a model from it so i think you have to feed actual drums into it and it'll change the kit i guess
Oh ok
btw incase anyone needs to extract vocals from songs, I got 'Ultimate Vocal Remover' aka UVR and its amazing, local and takes 10 seconds to extract... someone in here mentioned it earlier
and free
probably perfect for making models now that i think about it.. with the de-echo n de-verb
any other options?
mdx23c and foxjoy reverb hq are both the best models for vocal isolation and clean up cuz they have basically no frequency cut off
Does it remove the inst from the vocals too
mhm
Im not familliar with it so im not sure but the model is "UVR-MDX-NET Inst HQ 3".. apparently its new and it includes all the high frequencies
i tested it, its pretty dang nice imo
it has 'inst' in the name lol
its not that new
@sour vigil has reached level 23.
the newest uvr version has mdx23c which is really good but resource heavy
if you want a good model for both vocals and inst use voc ft
instrumental
ah word yea i saw you can choose the instrumental too to extract

pretty nifty tool, glad to find it
cool i found the X23C, ill try it out later
thank you 🙏 @sour vigil
is it on mvsep? iirc there's multiple versions mdx23c?
oh yea so you have to remove background vocals too to make a model right? can you do it with UVR?
@leaden bear has reached level 5.
I found the answer here in the guides:
https://docs.google.com/document/d/1HmkG9cmL8SLX7-vJcPT1-1KgUQtCrwXB8CicYmG4LW8/edit#heading=h.ub20t3fkdp39
I guess they use KIM-Vocal after isolating the vocals from the instrumental to then isolate the lead vocal from it
kim vocal 2 is worse than voc ft and inst hq 3
yeh kinda figured i'll probably have to mess around but atleast kno that UVR can do it
sry gues i should bring it to making-datasets channel
htdemucs_ft seemed to work best for me to isolate the vocals
I made an english voice for Yuya! https://www.weights.gg/models/clntz4f1y0003l608z03i8np9
mehhh thats good for stems in generap
but its not that good for vocals
anyone wants a model made? (paid)
i tried the recommended ones in the guides but unfortunately they didnt seem to be able to remove the sound of a saxophone while the person was speaking in the audio clip i was using
Hi
hiya
@quick sequoia has reached level 3.
I cant believe the originall channel got banned, I was there from back in march @timid musk
damn sad man
thanks! i'll try it out
@grizzled sapphire has reached level 2.
hello i cant find any kid laroi models anyone got any?
enable gpu conversion
@ivory lion has reached level 1.
hi bros
no idea then lmao
best way is to ensemble that with mdx23c and voc ft
For vocals mdx23c, voc ft and Kim vocal 2
Are there any good models for german rap? Just for fun.
oh interesting idea, thanks for the tip
gota make ur own shit now 😦
im lazy af
think its ez if you can get all the right tools
good luck friends, wish me luck too lol
Disclaimer: this is for personal research purposes
do it, you will feel smart if you figure it out haha
lmaoo true
i felt soo smart just getting the local version to install
for Mangio RVC
err Applio
its simple just
- get a bunch of songs
- process songs to isolate vocals, remove effects and silent parts and noise; this is a dataset
- Then put the vocal files into RVC tool and run for 6 hours and get your model
hey does anybody know how do i get more points in the bing image creator?
@elder willow has reached level 3.
One message removed from a suspended account.
No 2 is not simple lol
@twin fractal has reached level 5.
Its so hard to remove random noise like table smacking or somethin like thst
Its even harder if you are getting a dataset from a streamer that only does gaming streams, imagine isolating game sfx sounds, music, and sometimes other voices from the game
where do you typically install Applio? is Desktop fine?
hi
speaking from experience

Should i make chief keef
Diokat is a cock
A part of a tocic side of the SFM Gmod and TF2 communities
He never changes
a snitch niga thats that shit i dont like
hello!
hi
hi
whats a working google colab page?
does anyone have it? the one i used seems to be broken now
where is help tab
The server a bit inactive innit?
Bro fr said innit
I've learned British English so let me enjoy my crisps and beans on toast with fish n chips
And you go enjoy your "cheese burger"
Fine
Just say cheese burger 
👋
Nah that's way to American
Hi
hows everyone doing
So wait is wanting to cook furries bad?
hi chat, lost motivation to innovate. what do i do
i need a convincing girl ai voice
Renovate
and im not to able to find it so where can i find it
Just keep searching on #1159289700490694666
they are only 50mb each, you can get them all and test them one by one
maybe we need a bulk download button
@blazing lintel this one sounds good, i have no clue who this person is tho https://www.weights.gg/models/clmnyczp6019kwsz7943pwi2f
@keen fable has reached level 3.
Gg
@rare sorrel trust me daddy
Wassup I'm Automaze!
- My prefix in this server is
-(customizable with-prefix) - Interested in how I'm built? I'm actually open source!
- Feeling a tad bit generous? Buy me a coffee!
@tired flume has reached level 6.
Gg
I need a guide for the AI-Cover-Gen-No-UI Google Colab link
I'll send you a yt link (tutorial)


why did my msg got del?
@solar coral has reached level 1.
I saw this so much that when I look at it, I know what it is
HI
It doesn't work
Yo it's the Dio guy
oh damn i'm ai noob
Levels have reseted
Some bastards edited their messages to contain links with copyrighted material
We were basically sabotaged
Discord deleted the server because of that
eehh
@median cairn has reached level 1.
damn miss you guys
good ol days
just a few months
i had to wait november for me to get back at uploading
Make Dio sing Spectre (We Live, We Love, We Lie) by Alan Walker
aight
The public ones are broken but there are secret ones that work
provided by the info channel?
Imao I'm actually subscribed to this guy
i've heard it works only in brazilian or something

even the disconnected training colab?
No the training one works though
ah good then
business as usual for us lmao
i need rvc colab workin aaaaa
@median cairn has reached level 2.
@median cairn you can ask @abstract ridge for the working inference Colab link
yesss yesss yesssss
done
Now we can hear Dio sing We Live, We Love, We Lie
I made Emperor Palaptine sing Darkside a few weeks ago lmao
hi
@abstract ridge has reached level 17.
gg
idk i might be busy for a while and i have a date with my uhm at the weekend and
wait no gg
@abstract ridge you dont deserve levels
Lucky
you hate onion
the entire chorus got mistaken as backup vocals and got cut by MDX tho
Bruh
sadge
how long does it typically take to install Applio?
Depends on your internet speed
how long'd it take you?
For him it would be a few seconds because he has insane internet
Seethe filthy heathen
my internet is so good
Steam estimated it would take me 143 days to install Detroit Become Human at one point
I wish I was kidding
feels like i've been stuck on the "extracting runtime.zip file...." portion for a while now. is that normal?
💀
You should probably just wait
It's a huge zip file after all
good to know. thanks!
Why
They're so delish
And nutritious
Eat onions, your body will thank you
Whats RCV2 and is there an option for it in the software?
i lost my verified 
It's already in the software
So I dont really need to do anything when a model says RCV or RCV v2?



