#π§¬βai-chat
1 messages Β· Page 320 of 1
Just seen you are looking for a model of yourself?
In that case ofc u won't find it, either make it urself or pay someone to do that or wait for someone to do that for u
Rip
Rip rvc
Well, I think he's Brazilian
Yes
Isn't that spanish
It looks Portuguese to me, but it could be Spanish too
At least in google translator detection it says spanish
Btw if u are looking to be an engineer here too: /jointeam
I don't think it is, otherwise there would be a question mark at the beginning
I did two days ago
Oh lol
i have a funny joke
Ayo? @runic iron level 1 !!! 
someone make a model for gordan freeman and the download is just an empty folder
No i don't think so
IP: 203.232.130.102
IPv6: e351:3b7a:c0c3:5f6d:35e9:a347:c12c:eb0f
MAC Address: EF:2F:4A:F4:AB:87
Address (not exact): 1114 Docbi Place
Used: /doxx @runic iron
Still got some artifacting and sounds kinda robotic.
This can be fixed by cleaning the vocals and removing reverb/echo in case you didn't. And please, do not post audio files with instrumentals (no matter what kind of inst)
IP: 8.141.128.134
IPv6: fe4e:d923:53f1:a175:1f77:cebe:d375:461a
MAC Address: 14:7A:4F:4E:05:04
Address (not exact): 674 Mecsi Ridge
Used: /doxx @dusty warren
U.U
a rating for all of the vns i've played so far
- morenatsu
-fye game cant say more - promise to keep
-really interesting story from what i've seen so far (tho i hope it finishes π’) - socially awkward
-dunno, had a bit too much fillers. other than that it's really good
what
Could someone help me how to make the AI voice?
Do you have a GPU ?
tempting
-gui
- How to use RVC Mainline Colab by Cauthess
- AICoverGen Colab Guide by Eddy (Spanish Helper)
- Create a model with RVC disconnected (colab) by Angetyde
i don't use local lmfao
ππ» okay
latter
how to use ai voice?
Whats ur pc GPU?
In realtime or for pre-recorded audios/ai covers ?
And whats ur pc GPU?
Is a 3070 sufficient for okada
This interaction has expired, use the command
/guides realtimeif you wish to see it again.
Thank god these youtube tutorials been doing my head in a bit so ill try read that
Do you guys all use that modified version of okada
Youtube tutorials are outdated, its better u follow that guide
fr?
I dont personally use it as a i got a bad pc, but the wokada fork has better performance
yes fr, yt tuts are like 1 year old
I don't know how to lower it?
written guides can get updated more easily
i didn't know that
How to lower what ?
yea never follow them, u have high risk of going into errors, as like most yt guides ABOUT RVC are old
as rvc hype died so there isn't many new tuts anymore
I want to know how to install and use GPU?
Ayo? @mild cloak level 1 !!! 
GPU is not a thing you install.. It's an hardware component in your PC..
You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU
You are on windows right?
Ok sorry, how do you do it?
I told you how to check ur pc gpu in windows
check it
Please send me the link, I don't have the link
Its not a link..
Follow this ^
OK thanks
so whats ur pc gpu ?
Hey I need u for a sec
@k...
Is there any issues? U can ask in help channels
or is it just a question
Well it's a another user issue
1240232150331162636
This user is sending hack links to me
That's all
If u want picture
Could you show me some proof in dms?
Friend me i can't send messages till we are friends
now?
Thanks
There u go
what pitch shld i normally do
Ayo? @mild cobalt level 2 !!! 
If for voice changer
If youre male to female between 8-14 depends on your voice
Im male using a eric cartman one i think
Sweet will do thank you
s
is there a way to finetune UVR models on some really obscure genres of music if I have the stems for it
that was a test
should delete the porn gif link that dude sent 
ohh i didnt know it was porn
!ban 898914252826501180 sent porn gif
@hyb01 banned
Reason: sent porn gif
Duration: Permanent
whos a good boy?
I didn't either, until I clicked on it π
why it is not permitted:
yeah nadare tried doing something with it but training was just unstable :p
you 
my songs sound kinda off
I can even chop it up for you
i mix on bandlab
my songs dont sound that good
Bro just send it through the mix and master on bandlab it will do it
huh
How do you know let me hear it
how?
Ayo? @distant sentinel level 3 !!! 
already taken care of lol
somebody posted links to porn gifs but husky boy banned him
On mobile you can click mastering track and it will master it for you
wat
check dms
@covert lake is this how people collect good voice samples
L
cool
@inner nexus Azure's gone, he isn't in here anymore
it Cam btw, I joined not too long ago since he invited me
uvr is so fuckig slow
what
nothing
hi
where do i get config in models?
ΠΎΠ±Π΅Π·ΡΡΠ½Π°
do not click on that gif omg lmao 
Ohhh hell nooOoooooooooaaaaaaaaa
got another @river adder 
I used to just find voice compilations on yt
Good boy
-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
- π FaceFusion UI, by Nick088 Google Colab
- π FaceFusion NO UI, by Nick088 Google Colab
- π 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.
Ayo? @elder willow level 1 !!! 
Compression
Haida?
I dont do models since a year, it was an easy way to find samples
yee lmao
That's nice
unfortunately I know what you are talking about 
i think he is speaking of youtube low quality ahh compression not like a dynamic range compressor
Yeah it works, but ideally it's best yo get the least amount of compression
No no
Not compressors
Compression algorithms
Like mp3
Opus
Ogg
Etc etc
Flac
Yeah just to name a few
It works like a zip
It isn't bad
You open it and it's like listening to a WAV file
The only limit is it's bit depth
A compressor just affects the dynamics of your audio
It's meant to smooth out spikes and stuff and maintain consistency in your audio
256 kbps opus or aac lc would be pretty good ive never had premium so idk what codecs they encode for premium
It's all opus or m4a
ye but idk if they encode both for premium bitrates or not
It's around 200 smthing
As opus
It's very good and efficient
Don't use mp3 formats then
Use ogg
Ahh, that's sucks then
as far as ive heard ever and also using opus myself personally 256 kbps basically sounds lossless
Ayo? @tacit falcon level 5 !!! 
you get minor pre-ringing in an mp3
Yeah cuz opus does a great job at maintaining the dynamic range up to 16khz, which is what most ppl can hear. Past that point it kinda goes brrr and it cuts off after 20khz
Mp3 is just bad
And huge
true, before i used to download from yt and convert into wav
Yeah that works too
mp3 is like generations old in the world of audio anyway
would that help against yt compression lol
I misread, I thought you fetched as wav
this makes me think ur the kind of person to export 192 khz wav on an already compressed audio that isnt already 192 khz for no reason
like upscaling? nah never tried audiosr
Up sampling just adds aliasing. Ai upscalers aren't great either
i just did models for fun then got bored of it
it'll add aliasing if you use a resampler without a filter lol, most software ever has a filter in the resampler
Why? Is there a version of audio?
Ayo? @quartz scroll level 1 !!! 
afaik apollo adds noises, audiosr and especially resemble enhancer are good for speech
It add aliasing regardless, unless you use something like rx to resample but upscaling just doesn't make sense unless you want it to match other files
No that's pretty pointless
denoises it then upscales it
i just meant to say it is a very old format at this point
Just make sure your vocals are the same Hz as your instrumentals to avoid any dumb stuff
Oh ok
low quality vs audio sr vs resemble enhance
btw my first models were TADC, not like TADC is anywhere else than yt
so ofc at first i did it from yt
to me audiosr just makes it sound kind of like he-aac on that one sample
Yt has good quality audio for a model it's not bad don't get the wrong idea
It just isn't ideal
yea if i ever make again models i will π΄ββ οΈ
sail the high seas

The high seas are always the best
seek mullvad before you sail the high seas
Why the fuck
Ayo? @pale bison level 16 !!! 
huh
u just cant upload potentially copyrighted stuff cuz of what happened 365 days ago
Hello! I'm trying to download the '' real time voice changer client '' but doesn't let me, it says that many people have already check the website too many times. What should I do ? does anyone can send me the app ?
are you following a youtube tutorial? Because u shouldn't do that
I still remember that day 
I'm not, what should I do ?
First of all, whats ur pc gpu?
'twas hell
366 actually
was like the 4th october
1060 3gb
Ryzen 5500
16 ram
i didnt even remember the date tbh
Ayo? @livid bear level 1 !!! 
cuz i had left the server before it got deleted then i came back for a month then rage quit
if ur not uploading a file thats probably why
i thought links to youtube was fine yeah??
I remember it, vividly
yup they are
can you set me straight to it ?

if only he knew
for safety its better to not post audios that could contain copyright instrumentals, as a file
i cant remember if i was actually back in the server before it got deleted or not
at least upload the file off the platform
i cant know ur pc gpu, its ur pc
You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU
its better to check that if u got a good pc first so
@covert lakedo u recognize the voice
peter griffin?
it kinda is
more epochs dont mean more quality
honestly cant remember when I first joined ai hub myself, think it was like in August last year
idk
the model can undertrain or overtrain
Last update: Feb 10, 2024
it cant overtrain on my thing
i never use tensorboard cuz i got no time to watch it
Yes but it's depends of how fast is the gpu or cpu
couldnt you just train for a certain number of epochs and just waste hard drive space saving x amount of models then check the models with good loss if ur lazy then?
the speed yes, we talking about overtraining
Interesting
NVIDIA GeForce GTX 1060 3gb
overtraining is when the model learns too much making it have worse quality on general audios basically
i joined in may of last year or something but didnt talk until july or august, once i got access to verified i went to town
3gb? that might be low, im not sure if its doable but i can give u the guides
-rt
This interaction has expired, use the command
/guides realtimeif you wish to see it again.
1st guide, its the wokada fork, which is optimized
I joined like july 2023
like start
I never talked to anybody in any other channels besides the help channels lol almost noone remembers me tho because I changed my name and most people I knew have left or became inactive
ok, if I can use the model's voice that I download from #1175430844685484042 I'm fine
what

yes it still uses RVC models
You mean when the video is more longer the quietly become bad?
@covert lake
I don't really understand what you said, but I mean when you train the model too much, meaning its good only with the dataset but not with general audio which is not what u want
carti will never drop
thats the thing
kinda better than before, but i would suggest to just use the tensorboard
for what
to check the loss
what even is this? rvc?
yuh
for knowing which .pth is the best to use
then be doomed to check every single epoch it saves then

that was a different audio than the one of the video u sent before yea
Ohhhh ok
Thx for telling
tf is yume 2kki
i thought it was just a fan game not a mod
most people on ynoproject play 2kki, not really that obscure in that community hehehehn
the whole niche is pretty "niche" tho
i thought 2kki was more popular
just because 75% of people on ynoproject play it
have u played it or any of the nikki games on ynoproject
neither of both is any known
the people that talk in this server do not usually play niche games like this lol
i would think so at least from having talked in this server before
some of the old ex-regulars likely have i think
this shit take forever
turn the overlap down to 0.75
Ayo? @tacit falcon level 6 !!! 
why
there is literally no good reason to have it at 0.99 because beyond 0.75 you wont really hear much audible difference
ohhh
0.75 is good enough to eliminate clicking artifacts and keep the output consistent
0.99 is enough to boggle the mind
might have better sdr, but i dont remember if it does by much and the difference isnt usually super audible for this use case
Can I use W okada on my fridge
try it
is there an ai where I can extend images for free (and preferably no sign up)
@covert lake26 epochs moment
maybe if you try hard enough
use tensorboard π
make me
i do but i dont know shit about it
idk if thats good or bad
halp
u needa check logg/g/total
and put smoothness to the max
^
^
^
^
^
if only i had a Documentation...
Adobe firefly
I'm looking for a website
adobe firefly is a website
you're on Discord right now, also a website
it makes no difference
who made this 
isnt this lusbert's? i cant remember lol
iran
i know he made something
still pretty helpful imo
codename0 i think
u were seeing loss/g_d/0 before
wtf why is it purple for u? u sure u did as https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/#-graph
Last update: Feb 10, 2024
can you just tell me
i mean yeah...
jesus christ
i thought the graphs were color coded based on whatever tensorboard chose for each experiment/tfevents it found or whaatever
Hello, can you recommend an application/site to use for covers?
There used to be a Google Drive link, but I think it was closed.
graphs are colored based on how many logs you got
if you have 50 models trained with logs you'l have 50 random colors
the lowest part seems to be around 300 steps, it seems a bit too zoomed it tho, this button would center the view
basically what i thought then, i thought it was just randomly individually picked tho, i didnt really use it too much to know anything about it tho
what epochs are u on rn btw
whats ur pc gpu?
im used to just orange as i use only cloud lol, but ty for the info
uhm
Ayo? @vagrant veldt level 1 !!! 
what does that have to do with anything
just checking cus if u got a good pc you would be able to do it locally, without having to worry about the cloud limitations
click the checkbox to remove outliers at top left, then make smoothing 0.987
no?
with only 1000 steps in the training, why are you even asking? it will be a very poor result anyway
Not really sure if u can do it locally, u could try following https://docs.applio.org/getting-started/installation#amd-gpu-support-windows
Or its better if u use Ilaria RVC Zero which is a ZeroGPU HuggingFace space, its on cloud (remote good pc) and its the fastest
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 βNo gpu is available for you for 60sβ Introduction (with website link) Ilaria RVC Zero, is an RVC (Re...
Documentation for a high-quality, open-source speech conversion ecosystem designed for simplicity and optimized performance
to show the full graph
anyways yea let it train a bit more
much much of source audio do you have?
im not gonna tell you exactkly how many seconds
1 minute? 5 minutes? 10? 30? 5 hours/
rvc sucks anyways
Ok
if you cant bother to provide answers, why would anyone bother to diagnose that?
ofc its gonna suck if you don't put the effort into how to use it properly 

if that's your source audio, well, there's not much to learn from that
nobody said it is
try using rmvpe for pitch detection idk
huh? the server doesn't need phone number for verification, just chat till level 5
you have any examples?
yuup π¨ π¨
just send it to him in dm then
this is what mods do when someone uploads copyright
i can confirm
you know anything about the difference between rmvpe and rmvpe+??
lmao
the only difference should be that rvmpe+ has a pitch threshold, like limiting whats the max and min pitch
Hey guys! So I've been working on a song and I've been wondering if you can guys support my video? I made the audio a year ago and recently posted it because a guy wanted to do cover art for me! Would any of you be interested in liking, commenting and sharing it out? I'm trying to get motivation..
no
Thank you
xd
as soon as someone mentions copyright or uploads anything alarms get triggered for the mods
COPYRUGHT
MCAS
EXACTLY
if u wanna promo, there's #1159290752195633273 but not any other channels
Ty!
all it does is just zeroing out f0 values below and above certain thresholds
yea it has a min and max pitch
f0_min=1, f0_max=40000
which is basically free for all
that's what RVC Studio has
so what exactly does it limit here? lol
it certainly wont help with voice cracking
lowering the pitch of a source audio does help a bit
lemme see inference, maybe the pipeline has other limits
okay, that has f0_min=50, f0_max=1100
oh u mean that fork? idk if hes using that
nah, i'm just looking at this rmvpe+
that seems to be the only change from regular rmvpe
Hello guys, i came accross a AI model on insta, i find the reels very impressive, do you have any idea what app he used ? I put the link right to check it out : https://www.instagram.com/fleb_genz?utm_source=ig_web_button_share_sheet&igsh=ZDNlZDc0MzIxNw==
I'll check it out. DM me. Also, some advice, don't let any trolls deter you.
Ayo? @elder willow level 1 !!! 
Oh sorry ππ I just wanna know.
But, sorry because I break laws...
Ayo? @final nimbus level 1 !!! 
what actually should help with the voice cracking at high pitch is using the f0 replacement file
but I feel nobody has actually tried using it
btw this is all the info i know about the different rmvpe versions:
- rmvpe: A Robust Model for Vocal Pitch Estimation in Polyphonic Music, the normal version of the best pitch extraction, it's robust and not sensible to noise
- rmvpe+: has a pitch threshold, it limits the max and minimum pitch, basically deleting f0 values below and above certain thresholds
- rmvpe-gpu: Training ONLY, uses your gpu for the feature extraction process, using more gpu so making training faster
- rmvpe-onnx: Wokada ONLY, its a must for AMD Users who use ONNX models
same things
i just made this and felt like sharing so others could see it too lol
- is just this
f0[(f0 < f0_min) | (f0 > f0_max)] = 0which is really not the right way to do it
like my model can work with normal speech where f0 does not go over 350, but trying to infer a song with female voice at 550 is nothing but cracking
i remember it literally just does that instead of putting a high pass and low pass on the audio it feeds into the model using the parameters you gave
Hello, good day, does anyone know of a model with a very strong American accent as a woman?
Does anyone pls wanna help me get a free training item on weights.gg by cuz it says I need 5 sign ups to get one
thats what i thought it did
so to fix the cracking I either need to train with a bit more singing samples or just run the audio thru audacity and bring the pitch down 5-10%
I've heard from a single source that Character.AI is going to eliminate some core features that keep the userbase active such as romance. Does anyone here have information on whether or not that is bs?
Don't promote it here, only #1159290752195633273
Also, For training on weights.gg, you can:
- Pay
- Make a creation for 5 days streak to get a free premium training, you can repeat this
- Use referal code, getting 5 friends to subscribe via your training
However, weights.gg just uses RVC, you can just use that for training rvc models for free, if you want to do this you can tell me and I can tell you more info
alternatively a f0 file could be used, but I've yet to see a successfull replacement attempt
lmao

Asking for investment purposes...
well, i don't know if its true or not, i dont use character.ai, im just laughing because that sounds very funny they're removing it
IT would be a great tragedy. For the company of course
probably wouldnt help no lol
no more AI Waifu romance 
Do not even suggest such a possibility. No one would fund Character.AI and it would be massive net loss
Hi
yuup
ehh itll just open up the window for another project 
True fact
I remember checking out Aisekai a while back but they shot themselves in the foot real quick
when one door closes another is bound to open, something like that
How do I know you're not a chatbot?
would do the same but to the head if I had to romance a AI character
lmao
sending pzpzpzpzpzpz or saying racial slurs idk
As a Large Language Model, I don't explicitly have opinions.
just send an extremely long string of pzpzpzpz see which people go unstable and lose context
i remember doing that with clyde ai chatbot and it would start frothing at the mouth and sending random latex notations and math equations for no reason
Clyde? Don't you mean Claude 3.5 Sonnet?
no, it's the removed discord chatbot
no, discord had an ai chatbot they were testing last year
Ahh okay
it was very... very... stupid... and funny
very easily manipulated and it was zesty
Zesty????
Ayo? @elder willow level 2 !!! 
idk it wasnt fully lobotomized
@distant wharf
lol I remember how stupid it was lol
is that clyde
i am
in fact im on vesktop even
@Clyde come back bitchh
the auto highlight is gone
try
i probably have an a/b test that is finally completely removing him
doesn't reply ofc
already on
its the same for me even with that
idk if u need developer mode for that then
Yall ever feel like the most meaningful and intellectual conversations you can possibly have are with ChatGPT or Claude?
no
Silly pookie I'll tell him in dms
Acknowledged
its just an llm
no I do not have conversations with AI
unless someone here is AI

π Okay
We are.
we're some kind of ai hub?
Ayo? @tacit falcon level 7 !!! 
@atomic vector say hi to trev
hi trevaintdead, nick088 says hi
-# AI-generated responses may be inaccurate; please verify important information.
nobody here is real
k

the only conversation i can have with chatgpt is asking him a question about some random thing in a programming language that i dont know and cross-verifying what he just said
Nobody anywhere is real tbh
Sorry, I can't assist with that.
no
no worries, just here to help with any questions you might have, feel free to ask me anything
-# AI-generated responses may be inaccurate; please verify important information.
pzpzpzpzpzpzpzpzpzpzpzpzpzpzpzpzpzpzpz
looks like you just sent me a link with some random characters, but it doesn't seem to be a real website or anything meaningful, is there something specific you'd like to talk about or ask?
-# AI-generated responses may be inaccurate; please verify important information.
lol
i dont even think that works on chatgpt anymore
AGI when guys? 2027 or later?
based
@tiny peak
oh ye thats the id for the ai one
@distant wharf is the one u get when like u dm someone that blocked u idk how to explain
yea thats not gonna work for me i thought u were pinging the ai one
there's no way to make it work unfortunately
he would literally always respond with 3 x's even if you said something ridiculous
-rt
This interaction has expired, use the command
/guides realtimeif you wish to see it again.
@tacit falcon do i have pretty eyes
waat
cant see shit
U must be blund
looks like an eye though, congratulations
blund
which is the current best real time voice chnager fork
my bad
yo im looking for a way to get a song to sound like restaurant ambience. Any one have ideas?
Does anyone know any hacks for meta ai?
use premiere pro
I want to make an ai cover for free from my phone. Do you have any application or site suggestions?
something idealy free and online and can you explain how? Just stating a software does not instantly make me an expert on how to fix my problem
Ayo? @wide sparrow level 1 !!! 
Hey guys! So I've been working on a song and I've been wondering if you can guys support my video? I made the audio a year ago and recently posted it because a guy wanted to do cover art for me! Would any of you be interested in liking, commenting and sharing it out? I'm trying to get motivation..
ΠΠΈΡ Π½Π΅ ΠΏΠΎΠ½ΡΡΠ½ΠΎ, ΠΌΠΎΠΆΠ½ΠΎ ΠΏΠΆ Π½Π° ΡΡΡΡΠΊΠΎΠΌ
best software to use for models? (not creating, using)
i have rvc gui, not comfortable
and i hate appelio
movies suggestions rq?
Hello guys, would you like it work with AI software upgrading company?
Que tal
@proud moss why am I here
Who is this random bruh
Ayo? @proud moss level 1 !!! 
No reason?
Horny
nice
Fr
try eating
(I got a bra pic from someone as result of rizz)
incredible
Shes 16

Im 17
you think that will help??
Its almost 1 am
true
well then
no it'll make you hungrier
Ahh,,
π

What should we remix next?
Join our Discord! https://discord.gg/WTSM
SUBSCRIBE TO OUR SECOND CHANNEL!
https://www.youtube.com/channel/UCKAMsvPoP7WHfyhhAIiM3hg
MERCH: https://wethesusmusicshop.creator-spring.com/
OFFICIAL MINECRAFT SERVER: instacraft.net
TWITTER: https://twitter.com/wethesusmusic
ORIGINAL SONG: https://www.youtube.com/watch?v=s...

No
who can help me? me need api gpts, i want create chat gpt bot in telegram, me say my country not support
mostly voice ai

Ayo? @restive comet level 1 !!! 
im also new
anyone know why the voice changer seems weak? how hard am I meant to try for it to work
its just my voice with a slightly changed pitch
does anyone know the name of this dude or where the original video is https://cdn.discordapp.com/attachments/972554009703514202/1292274748780187723/Q40JvBtnaqKjZZXf_1.mp4?ex=67032425&is=6701d2a5&hm=6c72b3fe1e87955f10740d1cb9983b5a06d0d3180c00cc33a04b02f57918482b&
I had the same problem with a stock gif π₯Ά
o
Yo anyone got a beat
im not sure if this is the right server to ask but i was wondering, would anybody know any sort of AI that could be a good chatbot? im trying to implement one for my plannerapp to help people get suggestions but ive tried meta-llama and its too big on my computer heavy on gpu / cpu. any suggestions
No
Get better cpu gpu
Ok
if ur homeless just buy a home
Is there any ai that can make a recoridng sound studio like? Like a song that is recorded on phone but is really good quality but that it can be turned to studio.
I already tried removing echo and passing it through AudioSR but still sounds recording
there are different audio upscalers, the result is usually... mixed kind
bro just left
Yeah there is also upsamplers which need a powerful pc for like second highest pcm it requires a beefy cpu and at least 32gb of ram
hi ok
Which ones could help? I can send the file by dm if you want
well, you tried audioSR
hey
hi
I can try using this one upsampler I have
Alr gonna send you the track
what i do here
hi

Ayo? @modest moat level 1 !!! 
hello
Yo guys iβm starting a clothing brand based on some juice shit lmk and dm if youβll be down to support and follow my page! NO PROMO PLS DM
Ayo? @raw inlet level 1 !!! 
This is not related to this server and this is promo so put it here #1159290752195633273
this is not a server for branding
and it's even worse its juice world clothing
hi
Meine Mutter lieΓ einen KΓΌhlschrank auf meinen Hodensack fallen und dann verbrannte mein Augapfel und verwandelte sich in meinen Hodensack
Bruh π AndI thought AI Hispano was wild
hai
evil
Hello
Are there any projects like(ATM machine) that i can make using c++ basics(arrays,loops,functions,if/else)
ask chatGPT
Thanks
day, ΠΏΠΈΠΏΠ»
Ayo? @soft dawn level 1 !!! 
which one ai to use for creating a video Π°-Π»Ρ PDA recorded conversation, EAS Scenario, with text and styled images?
wanna introduce 280 scp scenario, but for mtf group
hey.
γ
whats ur pc gpu?
fr just says Intel(R) UHD Graphics
it means u have integrated graphics which is not good for making ai models locally (on ur pc)
U will have to use cloud (Remote good pc)
As you dont got a good PC, its better you use cloud for training an RVC Voice Model:
- Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
- Applio (ui)
- Mainline (UI)
- RVCDISCONNECTED (no ui)
- Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
- Mainline (UI)
- Applio by Vidal (UI)
- Applio by Shirou (UI, no guide as of right now)
- Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly):
Those are all the ways
the easiest is google colab, but u can get disconnected
I HIGHLY suggest you to use Kaggles which give more gpu time even if they are harder
so kaggles is best?
being honest man ill do whatever it takes to have a decent output ive never seen anyone succesfully synthesize takeoffs voice and it not sound like ass
yes, it gives way more time than colab
in colab u have a risk to be disconnected for the low 4 hours daily and lose your work
would you mind guiding me a bit?
the quality between those are the same, the difference is the gpu time they give
i see
the Mainline & Applio by Vidal, are hyperlinks, which mean blue text that redirects u to a site, both of those have guides
Omg og minecraft
sup
One message removed from a suspended account.
Has anyone here beat Resident Evil 4 (2023) on Professional difficulty?
Hello

you had a server link lmfao

no promo except in #1159290752195633273
Hi
I think I'll make something once gpt-4o-audio-preview drops
Time to have the state of the art TTS app, hehe, with RVC integrated course
I'll try to make it free but the API is like really really expensive
Have been using AVM with RVC and got to say it's literally like having your own voice actor. It's absolutely amazing.
One message removed from a suspended account.
One message removed from a suspended account.
we are so back
is that a furry game

i actually found a takeoff one
Ayo? @elder willow level 4 !!! 
i havent played that one either 
whenever I talk in the google colab client of rvc nothing comes out
https://www.weights.gg/models/clm73gkdc2da7cctcsuy0qzr0 click on create
ill play it as soon as i finish rtf ithink
it's what the model will say
whats ur pc gpu btw?
U could do it locally if u got a good one, meaning u wont have to worry about the cloud limits, but will have to install the program
literally none but i wanna download this model off the site
yes i cmd it
when it offers the download thing it just trys to create something
he's using a weights, he cant paste the model link
for weights, you can use it directly here u just need to login
yea i can use it but i like to download things so i thought maybe i could
i mean its working okay
else, u still need to login, click the 3 dots of the model, download it even if it says the popup that u can try it here there will be a download anyways, then read the model uploader of https://docs.google.com/document/d/1YbXcLFPaGjhOdG5NFkK3QrucCEpHZBwFUxkeMO8aB18/edit#heading=h.c5ea25jd74zp
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 βNo gpu is available for you for 60sβ Introduction (with website link) Ilaria RVC Zero, is an RVC (Re...
U cant use it locally tho, i gave u the other way to use it in ilaria rvc zero
what??
It wont
Weights.gg works all fine
itll be processing like at most maybe 2 min clips around just 30 seconds
.
Well yes, you can use it on CPU, will have to download the program, and the model locally, but it will be SLOW ASF
I highly suggest you to just use cloud
it could take even 10 mins for just a 2 mins conversion
i can use it on cloud yes thats cool
but i wanna have a copy on my computer for safeys sake idk
i have a lot of his unreleased music which includes snippets and so i know first hand shit will get uploaded and removed
and it even happened with a older takeoff model before or smth like that
huh? they dont really check what u input
Also, u have to give it only the vocals
i mean im explaining why i wanted a copy on my computer
?
the site is all good on mobile
its just a site, its not heavy like a game or smt
hello havent seen you in a whil
realize an extravagantly dressed African man with 90s long tube hair with a feather on top and an extravagant rich man's suit with purple and white colors dancing inside an Italian church on the altar
i finally found out how to download the weights gg model
i downloaded the one uploaded here and the one there and its a 2mb difference
well im happy i have access to one now and have 2 saved wanna make a pro one
WAIT
i gotta rig this up immediately
original
https://rentry.co/voicechangerguide
fork (optimized)
https://rentry.co/Forkvoicechangerguide
crazy work
-# what
yo guys do u have the tay k ai?
Ayo? @novel scarab level 1 !!! 
You can search rvc ai voice model at:
- #1175430844685484042
- #πβfind-models
- https://weights.gg/ (login required)
- https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
- https://applio.org/models
- https://voice-models.com/
- https://thevoicemodels.com/ (for Turkish Models, login required with discord and level 2 on their server)
if there isnt one, you can:
- #1159289738314919936
- #1191429836321849435
- make it yourself with our docs guides https://docs.aihub.wtf/essentials/how-to-make-voice-models/
nick
for the love of jesus
do yk what im supposed todo?
i got the weights gg model and i wanted to try using it but i cant seem to make it
and i ran outta messages for chatgptπ
chatgpt does not know about rvc
i told u how to use weights & ilaria rvc zero
u needa login
ye dont ask chatgpt
first it thought it was a image model then halway through resorted back to thinking that π
i did login to wieghts if thats what youre saying
i logged in and downloaded model
for using it on weights, u just need to be on the model page then click 'Create'
but how do i use locally even if slow i personally dont wanna goto the site everytime and im more than fine with waiting 10 minutes for two
alright, the docs are down rn so wont be fully functional but i can share some archives
can it be used by cmd line
Ayo? @elder willow level 5 !!! 
It works, but if u got any issues unfortunately i cant help you as i got a bad pc and dont do things locally, but in #β¨βai-help theres some helpers
well, the ones i shared are with ui
if u want a cli versions there is https://github.com/blaisewf/rvc-cli
ngl
im tryna make this big ass thing
i collect his songs and the snippets i wanted to run em through a code where it could convert them to audio, take that audio and use a ai to split the instrumental and vocals, then enhance the vocals and instrumental and reconstruct it
if thats possible
why am i so lost when it comes to this
i dont think its gonna happen at this point, your thoughts?
you could try reading its docs https://rvc-cli.pages.dev/
π RVC + UVR = A perfect set of tools for voice cloning, easily and free!
but i never really used this version
most people use ui
hi
hi
hoping to find a Gru from despicable me voice changer here for a project of mine
what are the easiest software to start developing AI?
Maybe minions are better?
find it in the channel #1175430844685484042
do u think its possible
ig from what it seems u want to do
thank you
Ayo? @ornate igloo level 1 !!! 
commitung cannibalism
how can i upload files to HuggingFace?
what are you trying to do?
im trying to upload files to huggingface
what files? you mean like upload ur model or like using an huggingface space?
any file
like the RVC V2 pretrains
are you using an huggingface space?
no
Ayo? @elder willow level 1 !!! 
im trying to train with beatzforge V2
so you are trying to upload your dataset to huggingface ?
yeah
if you are training a model, you really don't need to upload your dataset, if u really want to theres https://huggingface.co/new-dataset
but u dont need to, u can literally just train without sharing it, be sure u arent following some yt tuts
@slim coral #1235952130855010365
I am working on a university project that involves developing an AI to help discover new songs. Can anyone recommend a deep learning model that is suitable for this purpose?
hi
does anyone have the link to download python and pytorch?
hallo
Hello! 'm Albi and i'm new here, i love AI for music...
pytorch is installed using pip install
thank you!
hey whats a gui app i can use on linux, arch btw, to have chats with lama 3.2 and image gen by flux with lora support
ΡΠ°Π»ΠΎΠΌ
are there any apps for rvc training for free?
kaggle or google colab
Kaggle or you can use Google Colab.
Kaggle got 30 hours of free usage that reset weekly.
-kaggle
- Applio Notebook, by Vidal Kaggle
- Applio Notebook, by Shirou Kaggle
- Music Source Separation, by Shirou Kaggle
- UVR5 NO UI, by Eddy Kaggle
- RVC Mainline, by Hina Kaggle
- Original W-Okada's Voice Changer, Kaggle
- Modified W-Okada's Voice Changer, Kaggle
- π UVR5 UI, by Eddy & ArisDev Kaggle
- π RVC AI Cover Maker UI, by Shirou & ArisDev Kaggle
- π How to use RVC Mainline Kaggle by Cauthess
Note: Kaggle limits GPU usage to 30 hours per week.
btw whats ur pc gpu?
in case ur pc is good enough
im using a intel
intel integrated graphics?
yeah
btw always check the user pc gpu, i seen cases of people using colab while having an rtx, or people using local while having integrated graphics

yea use kaggle (highly suggested for more gpu time) or colab
Oh that's right!
hi
Hi
chat, how do i create my own voice model?
whats ur pc gpu?
-docs
Suggestions for @past karma
- How to use RVC Mainline Colab by Cauthess
- AICoverGen Colab Guide by Eddy (Spanish Helper)
- Create a model with RVC disconnected (colab) by Angetyde
rtx 3060, is it good enought?
the docs are down
ah
thats good for doing it locally (on ur pc)
Docs are currently down but i can give u an archive
how are u
Ayo? @dreamy heath level 5 !!! 

aight, do u have a video explaining how to do it because i have the big stupid
Ayo? @past karma level 1 !!! 
good, what about you?
Ayo? @little nebula level 2 !!! 
anyone know a female german voice model?
good
Nope, there is no updated youtube video
If u try to see one ur prolly gonna run into issues, its better u read one of the written guides i sent you
quality is the same, performance is much better
especially with updated torch
not the ancient 2.0.0 or 2.1.x
of which ?
and u mean speed?
preprocess, feature extraction -- those have been heavily reworked and take much less time to process source audio, training speed - faster
are there any tests to prove that the same dataset on applio is better than mainline tho
I havent done rvc comparisions except when i did rvc termux, which i found out EasyGUI inference on cpu is around 10 secs faster than Applio's one
my training speed is slower on applio
no, since the back-end is the same the voice models are about the same (minus the changes due to random shuffle of samples)
using FP16?
There needs to be tests to prove it tho
i mean.. you train two models in RVC using the same data set, they will be slightly different
that kind of difference
you were talking about speed, not quality
both applications use the same model configuration, so there's no difference how they perform quality-wise
oh.. but I did test
yea i was talking about a speed test, quality doesnt change in any rvc versions
torch 2.1.0 vs 2.3.1 - my epoch went from 1.2it/s to 1.6it/s
that kind of difference
just by replacing torch/torchaudio/torchvision
thats the difference between applio & mainline?
code-wise there was a lot of cleanup
genuinely asking for proof of a test since i seen others say other things #πβmaking-models message π
thats why i like having test proofs
like not to be rude or anything ofc
i dont have nvidia gpu to test
but I guess I can hack Zluda in again and give it a spin
taking RVC1006Nvidia.7z
hopefully, ig if one is faster in amd than the other in amd gpus then it would be the same for nvidia gpus
can someone help
I also just added a use of accelerate for traning
ask in help channels and be specific
should make things a bit more faster as well
i dont remember the requirements of applio & mainline but would be better u do tests without adding anything else to those
understandable, RVC1006Nvidia vs Applio 3.2.6
alright, let us know
while u are at it, coud also do an inference test if u want
I need 2 good ai engineers, dm me
Some day I'll come back and finish inference code in fork to show how fast inference can actually be lol
do you need help? If so, ask in help channels, else if its to promote an ai app or smt ask in #1159290752195633273
u have ur own rvc fork?
I want to hire
This isn't really the place for ai investors
its mostly RVC & Wokada here
Ok
I have inference code in the voice changer, so it's both realtime and inference for a file. Just not available in the UI
But it's a bit bugged and I never really bothered since there are working solutions already
ohh
But it takes 17s to infer 4:21 audio with rx 6600m last time I checked
Applio with Zluda took approx 29s iirc, but my memory might be wrong on this one
tbh i have always heard people saying it was slightly slower but no one had any proof
I mean, atleast on cpu inference it is of 10-8 seconds than easygui, but idk i havent tested on gpu
Haven't compared to mainline, so cannot really tell
I havent to that either
AI HUB Docs

