🧬│ai-chat

[dim minnow]

ik kan geen tut vinden op ytb

[storm meadow]

That’s not a NVidia GPU

[slate epoch]

Is there a way to get around that?

[storm meadow]

Buy a nvidia GPU

or

What is he trying to do?

[slate epoch]

Generate AI Cover using the Hina WEBUI

[storm meadow]

Isn’t the hina WebUI a colab?

[slate epoch]

yeah

[storm meadow]

Make sure they are connected to a runtime

[slate epoch]

Will that work with the GPU he has?

[storm meadow]

It’s not his GPU it’s colabs GPU

Google colab doesn’t use your GPU

Make sure he is connected to a GPU on colab (top right)

[slate epoch]

Right next to the three green dots?

Which option should I select?

[coarse vessel]

make sure it uses T4 gpu (cloud gpu not your own) as shown in the status in top right, and hasn't reached daily limit.

[slate epoch]

The very first option?

It says it wasn't possible to connect to the execution ambient

[coarse vessel]

then wait tomorrow for daily limit cooldown

[slate epoch]

Umm, ok

I'll tell my friend to try it

Thanks @coarse vessel and @storm meadow

[prime shale]

hi can somoene guide me how i can find a model that turns male voice to regular americna english voice? im so confused around here

[coarse vessel]

[prime shale]

wtf

[tight cypress]

@timid musk @forest quarry someone is spamming

[odd knot]

this is what the message supposedly translates to: "Oops, what's in it for me, I dealt a bunch of cocaine and then fucking stuffed it in my kid's mouth"

[timid musk]

wild

[tight cypress]

I think it is different from XTTS

probably just another zero-shot TTS

XTTS is based of Tortoise TTS
We don't know which architecture or based-of-TTS of XTyde2 yet but I can bet that it is self-trained

like Tortoise TTS

💀 OK my brain is melting like butter. So I gone checking it

The app.py and download link is directed to XTTS v2. Could it be the wording issue with new TTS?

even the HF documentation is sending people to XTTS v2's documentation

woah what happen

https://github.com/BoltzmannEntropy/xtts2-ui

it have a Colab and HF space compared to this XTTS v2 UI made by Boltzmann

so welp

[stone coyote]

Pal of mine somehow set up ai voice models w/ voicemeeter or something like that. Said I should ask here if I want to set it up too. Could anyone help me w/ that? Apologize if this isn't the right place to ask.

[hexed imp]

Hi, I'm sophie here! Hope to see a lot of amazing people in this community!

[elder willow]

looking to hire someone to make a dataset (must know how to seperate vocals with ai accurately)

lmk

[slate epoch]

Is the Hina WEBUI working well for everyone?

I suddenly started getting errors while trying to run it

[icy ingot]

hey

[wanton vapor]

👋

[modern wedge]

https://github.com/hydralauncher/hydra/releases/tag/v1.1.0

[modern widget]

can you please suggest best software in which i can use these models

[covert lake]

Check our docs https://docs.aihub.wtf/

[covert lake]

???

#1159290752195633273

[covert lake]

You can js check our docs https://docs.aihub.wtf/ , else u can pay someone to make you the model #1159289738314919936 or #1191429836321849435

[covert lake]

XTyde2 is xtts2

Its not a new trained model

[tight cypress]

Thx for the response. Although your wording can misleading some.

[covert lake]

Wdym misleading? Im just saying that its not a new tts model or anything, its just xtts2

[tight cypress]

It doesnt have a clear image that it is XTTS until you hook up the Colab and downloading stuff or navigating your README.md and clicking documentation.

I think it should be New TTS GUI rather than TTS as the whole because it can mean other thing like it is a completely new thing

[feral halo]

Ayo? @tight cypress level 16 !!!

[covert lake]

@dapper trail what do you think about this?

[twin fractal]

What modifications did he do?

[covert lake]

I mean he can say them

I didn't honestly look at the difference between it and for example https://huggingface.co/spaces/coqui/xtts

[tight cypress]

I already put a counterpart there

[covert lake]

From what i remember he said he made it inference faster

[covert lake]

Oh lol

[covert lake]

Btw it can't be used to stretch datasets for rvc

[twin fractal]

I still don't understand what that means lol

Did he mean like using the 0shot to create new audios which can be used for rvc training?

[covert lake]

Yes

Which is not possible

Bc it has lower quality

And quality is better than quantity

[twin fractal]

possible but yeah low quality

Also no pitch variance

[covert lake]

low quality would make the model worse

its better to use 1 min good quality than using 5 min extended low quakity dataset

[elder willow]

Hi everyone! I'm Lisa.

[red copper]

1

[plain hemlock]

Got a question

[dapper trail]

Ability to upload files from links, available languages (last space not mine had only English), supports mp4, mp3,flac,wav, now works with gpu and other various changes

[dapper trail]

Puoi spiegarmelo in italiano che non ho capito

[covert lake]

Essendo che xtyde2 è solamente un ui modificato e non un nuovo modello tts, quindi pensano che sia meglio che sia definito come nuovo gui tts, perchè alcuni pensavano che era un nuovo modello come quella persone affinche lui ha visto il codice e ha capito che è solo xtts2

[covert lake]

But isnt that already supported?

look at https://huggingface.co/spaces/coqui/xtts

[dapper trail]

Non é questo

C’è non fa la stessa cosa

[covert lake]

si è xtts2

[dapper trail]

Eh ma dove clona gli audio

Ah sai ho visto

Scusa

E vabbè

Hahaha

[dapper trail]

É andata così

[covert lake]

sarebbe meglio che togli la parte del nuovo tts e che dici che è solamente un nuovo gui

e che non puo essere utilizzato per allungare dataset

perchè alcuni si stanno confondendo

[civic bison]

xtts is interesting

[feral halo]

Ayo? @civic bison level 1 !!!

[civic bison]

im not a big fan, i think hierspeech has better results but that could just be the neural upscaling they do

[strange oyster]

Кхуй

[covert lake]

whats that tts?

[plain hemlock]

i broke the app somehow

[feral halo]

Ayo? @plain hemlock level 4 !!!

[plain hemlock]

it won't launch properly

[covert lake]

!howtoask

[pine acorn]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[covert lake]

also, use help channels

[plain hemlock]

when i go into the http bat it starts up the CMD but like the actual voice app won't launch

it was working fine last week

It doesn't give me an error it just says intialising

[civic bison]

https://github.com/sh-lee-prml/HierSpeechpp

i think its just the upscaling thats tricking me

[leaden compass]

btw i tried some scout 1750 epoch voice model but idk why i still sound awful when it comes for me with the mangio rvc or aka rvc v2 voice conversion

[covert lake]

Ask in #1192011222023950368 or #🔍│help-w-okada

[leaden compass]

i mean i see some good scout voice models but in reality i sound awful and not rlly accurate when it comes for me for the scout voice

with mangio rvc or rvc v2 voice conversion/voice changer

[outer cairn]

when i choose the model there is always an error any fix?

[covert lake]

!howtoask

[pine acorn]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[covert lake]

ask in help channels

[covert lake]

mmmm tbh looks kinda abbandoned, but will see it

dont think it should be better than gpt so vits tho

[civic bison]

probably not

its faster though from what i can find

i might try to set something up with openvoice, if i can find a way to stream it it would be kinda cool

[covert lake]

bc its 0 shot

0 shot is faster but less quality than normal training

[civic bison]

yeah but like its 100+ times faster

im gonna have to do more testing, i don't think using the hf space works well

[cedar jay]

god i need a triple h ai so bad

[velvet cape]

@dapper trail
you should probably add accents like american and british

[feral halo]

Ayo? @velvet cape level 6 !!!

[fading kestrel]

I want skibidi toilet

[feral halo]

Ayo? @fading kestrel level 1 !!!

[fading kestrel]

i want a diarrhea voice model

[flat rover]

[covert lake]

[civic bison]

im gonna see if i can rig up the upscaling part in xtts

[slim summit]

:3

[loud oak]

https://flashcardfy.lol great ai study tool

[deft thistle]

hi

[brisk magnet]

[verbal stirrup]

hi

[tight cypress]

hi

[rose chasm]

hi, i have a question, which better big dataset and low epoch or small dataset and high epoch ?

[novel igloo]

Hi all

[forest quarry]

Bigger dataset, lower epoch.

But don't let that take over you tbh. Always train for the maximum amount of time you can... training until it overtrains

[rare monolith]

@formal hound

[wispy pier]

Hi!

[stuck oar]

hi

[rose chasm]

thx🌹

[sullen mortar]

Level

[still quest]

what is this server for

[forest quarry]

AI stuff

[fallen ruin]

Average Inuyasha shipper

[forest quarry]

that's @river adder

not me

i do point out helpful members tho

[astral herald]

selamlaa

sasasdasa

[dire acorn]

hi

[terse urchin]

what ı should do?

[forest quarry]

it's still doing something else, wait for mount drive to finish

[zenith hamlet]

Fazla türk yok aga

[thorny token]

hi

[eternal goblet]

hi

[chilly tiger]

h

[craggy cove]

manually turn it into mono

put it into rx, select one channel, paste it over to another file ( same samplerate )

cause, iirc, rvc's set processing code is doing kinda joined mono / averaging ( can be wrong tho! )

and that might actually hurt the set as isolated stuff, samples, aren't 1:1 on both channels

[bitter yacht]

helllo everynyan

[rare dirge]

hau are yu? i am fine fank you

[signal cairn]

no wonder this generation cant make folders

[glossy charm]

Hello can anyone help me with an ai cover I'm working on?

[elder willow]

https://www.twitch.tv/inprocess поддержити беднягу

[pine mango]

парни есть кто русский?

можете помочь настроить программу пожалуйста

[forest quarry]

Привет. Я не русский, но думаю, что это руководство может помочь с настройкой устройства смены голоса: https://rentry.org/VoiceChangerGuide

UVR BVE or Bandit Plus....

maybe Demucs DNR

[craggy cove]

lmao fr.
What's so hard about right click on the bat, edit, copy paste into chat gpt if you're too noob to understand stuff and prompt it with " is the code safe? "
lol

[signal cairn]

-rvc

[rare sorrel]

Suggestion for @glossy charm

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[indigo anchor]

or you can look at the bat file yourself and determine if it's safe

[glossy charm]

I'm trying to make the wiggles (the og ones) sing Beatles songs

[feral halo]

Ayo? @glossy charm level 1 !!!

[craggy cove]

bruh, that's literally what I said " if you're too noob to understand stuff "

🗿

[tiny coyote]

🫤

[elder willow]

yo

[native escarp]

Can someone help me making duet ai cover

I have no idea how to do it

[indigo anchor]

render each singer out individually and process both as separate files

if they're singing at the same time, then you wouldn't be able to use the original stems

[native escarp]

i tried

It just don't sync correctly.

[tepid basin]

rvc doesnt change the tempo/speed of the input vocals, probably an issue elsewhere

[native escarp]

I use the same singers and do it in audacity but it doesn't sync right.

[covert lake]

fr

[coarse vessel]

what do you need for help? the separation? Let It Be is prob the easiest you can do on

[signal cairn]

dont crop out any silences or manually sync

[craggy cove]

but you know, it won't ever be 1:1 synced
( as long you ain't talking about tempo, bpm and such )

[native escarp]

Can you help me making it?, i will credit you, I will send you the model and other information for it, it's not a long song.

[craggy cove]

each model has different characteristics and ways of how they articulate and so on

[signal cairn]

just use audacity to manually sync it, way easier

[native escarp]

I tried and it's really doesn't sync, sometimes it's just cut out

[glossy charm]

Need help with basically everything

[native escarp]

Or turns into cursed.

[signal cairn]

follow the guides

[glossy charm]

???

[signal cairn]

-rvc

[rare sorrel]

Suggestion for @glossy charm

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[signal cairn]

@glossy charm https://docs.aihub.wtf/essentials/how-to-make-ai-cover/

[native escarp]

Can i send you the audio? You don't have to use the ai i got everything ready.

I just can't get it

The sync is too hard

[craggy cove]

Listen @native escarp you're indirectly told to: read, put some effort in and do it yourself

Admin would be the last person here to do that kind of a job

aside:

[native escarp]

I will try again on it

[feral halo]

Ayo? @native escarp level 4 !!!

[craggy cove]

Open up fl, set the right bpm for whatever you work with
sync the vocals1

sync vocals2 to match vocals1
take waveform for referencing if that makes it any easier for ya

[iron onyx]

Does anyone know the fix to this error

File exists, proceed with creation of creds and client

[coarse vessel]

go learn how to mix and anything, esp this tutorial:
https://www.youtube.com/watch?v=r7SDWpSk1h4

[native escarp]

Alright thank you

[fallen ruin]

Truly the most beautiful costume

[knotty spindle]

Hello all

[forest quarry]

what

[glossy charm]

Imma do get back

[dry dove]

I got strike by UMG even though I used no copyright instrumental, they really suck

[forest quarry]

good luck

now explain why you hate mj

/j.

[glossy charm]

• coughs *
  Leaving neverland, I got scared for life after watching it

[tacit marlin]

the rain

[feral halo]

Ayo? @tacit marlin level 1 !!!

[forest quarry]

Well, read this:

https://medium.com/@justhoughts/leaving-neverland-debunked-in-10-minutes-or-less-35d2017469ba

It was straight up defamation

[dapper trail]

Hi

[forest quarry]

idk whats that

sorry for being uncultured

[tardy vector]

heya

[forest quarry]

pikachu!

[latent rampart]

Hi all

[zenith hamlet]

Are you guys talking about mixing the vocals

[rapid oxide]

im lazy to search the server, where's the tutorial for the ai voices

[forest quarry]

https://docs.aihub.wtf

[craggy cove]

Nah

The person's struggling with matching 2 vocals ( duo ) with a song or something of that sort, idk

[zenith hamlet]

Then what

[zenith hamlet]

Ohhh

Just put the og vocals and inst and match the vocals with the og and then delete the og

[craggy cove]

smh, that's what I more or less said

[tall loom]

hello, how to made my own voice model??

[covert lake]

check our docs https://docs.aihub.wtf

[verbal yacht]

hi where can i get help with the voice changer

[torn halo]

I hope everyone has a great day :3

[fickle haven]

hi

[covert lake]

Hi, you can ask in #1192011222023950368 or #🔍│help-w-okada

[elder willow]

Hello everyone👋

I'm looking for some artificial intelligence tools, but I can't seem to remember their names. Maybe you can help me out:

1. Artificial intelligence which is actually a plug-in inside Premier Pro, an open source (completely free) that makes cuts for me in quiet sections. similar to (ai ai captain)

2. An AI tool that analyzes TikTok pages, evaluating the videos and their impact. I recall seeing this used by a popular content creator on TikTok, but the name escapes me.

Many thanks to all the helpers (:

[verbal yacht]

thank you i asked in forum

[craggy cove]

if there's literally too much across the dataset, ye @elder willow

found it the hard way trying to clone Hatsune Miku from her raw vb ( fused tons of small 'phoneme' samples

in other words, if there's too many of em, compared to the rest of the set

yea

if you need some reference; Kurisu uses laughting, softer speaking, screaming, loud talking, crying ish speech, sighs, breathing

anything that has a meaning to you and your model - will work

oh?

oof

well, if you're fine with model struggling to reconstruct breathing / and fine with kinda artificial breathing's dynamic

if not, I'd not touch those at all

in fact, you should not play with dynamic range at slightest all ( any form of X-control; in this case, breath control (( and others too, such as loudness control etc

You need to expose the model to how your speaker ( your set's voice ) handles the dynamic range for all sort of sounds, pitches and so on

making it " uniform " or controlled will only make things worse, in fact, model will 'tear' quite a darn lot ( unless you're lucky, sometimes happens

That's all still up to you, remember

but according to my experience, it is a big nono

[coarse vessel]

as long it has enough clarity, keep it. basically keep it natural as possible.

[craggy cove]

Yea, reason why we don't tinker with dynamic range and response is because then, rvc won't know how your model reacts to different unseen scenarios

and that will result in poor or lack of f0 control + tearing and other artifacts
( tl;dr, all that kalo used to teach people is wrong; esp the compression and transient taming )

yea?

[restive pine]

is there a way to make a.i voices sing?

[craggy cove]

🗿

Nah, we do not

the only scenario you can, is when the transients really really really do stick out
and I mean it, not by 4 to 8 maybe 9 db but some extreme differences such as 12, 15 or 20 on avg

Imma post you some helper pic

This is how you wanna handle your set ^ - yw

Can't help you any further with more complex stuff. Am busy trying to implement onnx infer to my fork ( am busy a lil

[keen salmon]

Hello!

I need a bit of help with something, is it possible to get the help here?

[craggy cove]

@keen salmon Reading doesn't hurt my guy

[keen salmon]

or is there a specific channel

[craggy cove]

well, internet is big 👀

[keen salmon]

Its more so about this client Im using

[craggy cove]

oh, understandable then ye
( I know it can be a pain in the butt, some aliucord release were screwed in that matter, for instance.

[keen salmon]

I cant seem to get it to work, with the models i ADDED

*added

[craggy cove]

well, either way, go to #✨│ai-help
pretty sure someone will help you out in there

[keen salmon]

Okk thanks

[craggy cove]

I mean, how else am I meant to help you? everyone starts off with basic terminology and basically lack of knowhows

back when I started up with vocaloid and music production in 2015/16, I had nothing at all
It takes time and patience, lots of research, guides, articles, yt videos

esp helpful is watching some streams / vids of famous producers doing the mixes
mastering insights, audio restoration videos ( about rx and audition
some advanced spectral works ( spectra layers )

etc, just take it easy
My transition from utau/vocaloid/music prod to AI / rvc took around 1 to 2 months ( to get a mid level of understanding and mechanics

Hmmm, well, RX is the main deal when it comes to such things, adobe Audition is quite universal and so is audacity ( but that's more for general use ), then there's daws for mastering and mixing, production and for plugins? yeah, I mostly know 'waves' stuff ( the de-reverber I use for example )
other than that, I probs can't help

yea, that's true ( also, I'd be a lil more careful about piracy and such topics in here ) some people are iffy lol. Don't ask ~

[coarse vessel]

depend on how audio issues are, but several individual RX 10 modules should generally work, the rest may need some manual labor

[minor blade]

Yep, ur right.

[craggy cove]

^ there

whenever possible, resign from automation for the sake of manual and controlled job

mhhh

the workflow should be;

1. fuse all samples into 1
2. dealing with spacing inbetween full_words ( full phrases, not inbetween phonemes or natural pauses )
3. tame the dc offset if one occurs
4. de-noising, de-clicking, de-humming ( if legit one occurs ) and other stuff such as: plosives taming and so on (( order for this one depends on the case and how each of the issues affect the rest ))
5. slight compression according to images, if some transients in some places really outstand from the rest ( by a ton )
6. normalizing to -3 or -2 db
   *** and ofc, always keep in mind the sample rate and bit depth. ( if you do a lot of copy pasting or exporting ) ***

Oh yea, and for the 'final product' just export it as: your og samplerate, 32 bit float for bit depth
( increasing bit depth to something higher than og won't hurt at all (( rvc does that too, all audio -> 32 bit float )) but doing it in reverse will definitely cause issues [ or can in cases of gt high bit depth / dynamic range audio ]

////////////////////////////
That's all I can say

Good luck ✨

: ) yw and remember, key is patience and you'll get rewarded ✨

[lusty violet]

how do i make ai models?

i have many audio files

[stoic ibex]

hi guys hows going? i really need your help easily creating a model, someone who could help me? ill be grateful

[forest quarry]

?

[forest quarry]

https://docs.aihub.wtf/
has everything you need to know

[stoic ibex]

can i write u in dm?

[forest quarry]

Like, do you want a model to be made by someone else or

Do you want to make a model yourself

[stoic ibex]

i rly need a voice from a song

if u could help me ill seriously be grateful

[forest quarry]

you can always #1159289738314919936

if you want to extract vocals... https://mvsep.com

just create an account, use bsroformer as ur model

[stoic ibex]

wrote u in dm, its very easy

[craggy cove]

lol

[gentle canopy]

hi

[craggy cove]

Nah nah

just use an eq

and repeat until dc is gone

( alt+d ) shows you stats of your audio

alternatively:

but I'd go with the first one as it's definitely less aggressive

better keep that " nice low " slight rumble in sub-bassy frequencies ( can be pleasant for soft / whispery voices or generally softer tonations

[coarse vessel]

yeah, though the analog option sometimes works better

[craggy cove]

I don't like how it plays with phase tbf

I'd personally stick with digital rather than emulate the analog eq

rvc doesn't need " warmth "

just type em in

and u can uncheck these too, just for the sake of being ( circles ):

[craggy cove]

generally, analog eq introduces the phase shift and yea, you don't wanna play with phase in case of rvc ( remember? keeping it natural is the key ) and yes, that phase shift is what causes the characteristic effect of the analog eq but still

[gilded edge]

hey guys

I need some help

[craggy cove]

tf?

what kind of audio you work with lol ( show me the stats

[gilded edge]

Im getting an error when going on the google colab JSONDecodeError

[craggy cove]

tell me the offset value

ps, when working with " stereo " or " isolated stereo " from mono, pick the one that has more consistent peak volume and such stats
and focus on just one. Don't work on 2 at once as it'll cause some issues in evaluation

Nah, I meant the stats ( alt + d )

pick the L channel

as the avg ( rms ) is closer to peak

ye, the L has no offset

in a short, revert the eq you were doing

copy the L channel to new file ( keep the new one mono )

and work with the dc offset in there.
Your step 0 is picking the channel, moving it to other file and continuing with that one. ( past that, you forget about the other channel

Because like I said, the channels aren't 1:1 especially in case of isolation

[feral halo]

Ayo? @craggy cove level 29 !!!

[craggy cove]

after isolating the thing, you pick the better channel and work with that one only

well not really rely but if it's offset, that's your way of telling what the value is

( also you'll see that weird / funky stuff at low low low hz area, [ 0 to 50 hz ] are disappearing (( can be lines or other smeary stuff ))

might be, ye

Overal, you can select just one and ctrl+a and then copy

paste it to new file ( same samplerate as original but keep it mono )

and work from there, ye

[ornate yew]

do you really use multicolour as your normal spectrogram colour?

like does it not hurt your eyes

[craggy cove]

no? ( I spent way too many months with it ) 💀

Mutlicolor is the actual way of working with sets

Because you didn't use the original sample rate

you have to keep it the same

if your " stereo " is 44100
you keep the "pasted mono" 44100

there's absolutely no other reason it happens

no

you have to resample the 48s to 44100

the isolators of any kind are 44.1k or less

nah, you just resample the 48s to 44.1 ( so it's just as ur isolated ones )

then all is uniform and not pitched

smh, hopefully they'll make 48 isolation models soon

[restive pine]

has anyone installed applio?

im stuck...

[craggy cove]

Prepare your nerves for that one 🤣

✌️ I believe in you

[elder willow]

Hi

[half horizon]

hi after cleaning dataset and normalize and all is necesary to do simple rate ?

[craggy cove]

no

You keep it 1:1 with what it was / is originally

[mint lava]

Can someone help me with this ai voice?

[half horizon]

i diont understand 1:1 to make u close i have 3 audios 9 min pero one i cleaned em all like doc say mvsep +audacity so now what i need to do ?

[half horizon]

what is ur quesition ?i know the first step how to clean datasset do u need that ?

[mint lava]

i think im stuck with the python and pytorch part

[half horizon]

idk sorry but u can try no cpu in googlecollab https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[feral halo]

Ayo? @half horizon level 5 !!!

[craggy cove]

well, aside some older models which had different sample rates sadly ( as in, were downsampling the audio ) now newer models output 44.1 khz regardless

so even if you put in 48khz audio and use, say, bsroformer, you get 44.1

so that's your sample rate = 44.1khz

[half horizon]

wil send u image that in simple rate

[craggy cove]

when you work on that set, in audacity or whatever, you export it just as that, as 44.1

just, after you're done with doing whatever to your set, export it as 44.1khz 32 bit float file

[half horizon]

yeah i did 44 in audacity

[craggy cove]

and 32 float, correct?

[half horizon]

24 bit

idont have 32

[craggy cove]

you gotta export it as 32 as that's what the mvsep outputs, at least in bsroformer and mdx23c's case

whether it's 44 or 48 and 24 or 16 bit, mvsep ( again, at least in case of mdx23c or bsroformer, so, best models ) it will output you 44.1khz 32 bit float

and that's what you gotta stay true to

[slim sun]

Ayo

[half horizon]

i passed my data set by mdx then mdx b and fowjoy and denose

[craggy cove]

I know the thing wow. the spek

[slim sun]

Im looking for someone or something that would make me a melody:( where should i look?

[half horizon]

after that i did audacity so last export was by audacity

for truncate and normalize and silence

[craggy cove]

Look, I don't have to see the spectrogram response or whatever, Imma just tell you that you gotta export your set ( if it's mvsep-origin ) as 44.1khz 32 bit float

that's all

There's no need to introduce complexity for non audio-savvy people

I could go on and brag about frequency range response and it's effect on the models but I don't have time or patience for that

[craggy cove]

• I am busy so ^

[half horizon]

just tell me if i can go next step or somthing wrong on what i ded to let me search more to understand whow i can do 32

im newbie

[craggy cove]

there

mono, 44.1, 32 bit float

that's your target export settings for audio you work with ( if their origin is mvsep and models; mdx23c and bsroformer

[half horizon]

idont have this in my audacity

[craggy cove]

and how am I supposed to help that?

I am just telling you what's the right and optimal way of exporting it

[half horizon]

no just tell me if 24 is good too or not i need to search how to make 32

[craggy cove]

I mean, you can export it in 32 bit float in any audacity

any tool or program that is meant to work with audio has that

[craggy cove]

I'll just say that you potentially degraded your audio with 24 bit

huh

you use some kind of outdated audacity?

aside, it can't be stereo, it has to be mono
( pick one channel and remove the other one

but I don't work on daily with audacity so can't help

[half horizon]

ok so one more question

[craggy cove]

yea?

[half horizon]

i started after downlidng mp3 from youtube by streo and i worked with flac stereo next on mcsep

so i need know strat from scratch with mono

or just the last flac file from mvsep when i put it in audacity must export it mono ?

iwil try to download new audacity version

so i need start from scratch or get last clean file that idid from mvsep

[craggy cove]

there is absolutely no reason at all to do mp3 - > flac
it doesn't work that way

mp3 is garbage and you shouldn't be using that to begin with

use a tool called 'yt-dlp' ( on github )

[half horizon]

i tryed but it didint work

[craggy cove]

you open up a commandline in wherever you put that tool and use this:

yt-dlp.exe -x url

didn't work? how so

perhaps it's bugged on win7?

[half horizon]

yeah

if i do somthing it closed

[craggy cove]

well rip, in that case you'll uhhh

[half horizon]

and say somtimes eror when i opned it

my laptop is garbage

[craggy cove]

oh, cause you do not open it
you use commandline for it

It is a 'cli' ( commandline ) app

[half horizon]

i opned it from .exe

[craggy cove]

yea, you do not do that.
it is a ' cli ' application
cli apps are used via commandline, like I said

[half horizon]

yt-dlp.exe

[craggy cove]

just because it is exe doesn't mean it is openable

imagine this folder contained the yt-dlp.exe
you type in cmd in the address bar

[half horizon]

can u give me the link

wil try again from scratch

[craggy cove]

and you do it that way ^:

yt-dlp.exe -x URL

then you gotta convert the thing you'll get, it is in .opus or .webm format ( you convert it to .wav with ffmpeg:

ffmpeg -i your_audio.opus your_audio.wav

[half horizon]

is there a video ion youtube for that bro ?

[craggy cove]

Not sure

@elder willow can you help this dude? I am kinda busy and can't continue it really

Thank you 🥺

[half horizon]

thank you guys

ty taewoo i remmber u too u helped me before

iwl explain to u to understand my case

now i understand how to clrean coretly data set

[tired spruce]

Sup

Who wanna see my feet pic

[half horizon]

but the problem is i cleand a mp3 from youtube

[tired spruce]

My brothers feet

[half horizon]

i download mp3 from youtube then i converted to flac and i started cleaned it

[torn arrow]

super confused on how this works, especially on mac since always an error and can never get past the google colab part

[tired spruce]

It's so clean

[latent summit]

https://tenor.com/view/ren-and-stimpy-duck-sweating-tongue-out-lick-teeth-gif-14153037

its over

[tired spruce]

I know

[half horizon]

so code man told me that i need downlaod videos using yt dlp as beging to no lose quality

so now i need do all from scratch

and alsso i need download new version of audacity to export 32b

[latent summit]

use stacher

[torn arrow]

Macbook Pro

[feral halo]

Ayo? @torn arrow level 1 !!!

[half horizon]

link plz

ty will check

[torn arrow]

M series, to be specific an M1

[latent summit]

just make sure you go through the settings when you do to set things to download in max quality youll figure it out

I fucken hated trying to use dlp by itself

shit just didnt work for me at all

[feral halo]

Ayo? @latent summit level 1 !!!

[torn arrow]

Even when trying and can't use it at all, maybe I'm doing sum wrong but it just does not work

[latent summit]

stacher just werks and is also dlp

exactly which is why I installed something that works better

[half horizon]

stacher worked for me

[latent summit]

give me gui or give me death ill fuck with a command line if I have to but I really dont wanna

[half horizon]

so how to corectly download and extract audio

is there a video for that ? what setings must fo

[latent summit]

two command lines too many

[half horizon]

i i know the part of mvsep i lerned it wich one to extract and clean just i need start with good file audio

[half horizon]

can u tell me wich seting must fo

yt dlp dont work my laptop i tied many every time not working i see that stacher is good for my case

[latent summit]

click stacher in the topleft and go to post processing set your audio quality to best for starters

if you wanna change your download format and file location its top right

[half horizon]

i did best

what second step

[latent summit]

do you just want the audio?

[half horizon]

yeah

[feral halo]

Ayo? @half horizon level 6 !!!

[latent summit]

if that's the case click on BEST and set it to FLAC

then just paste the link and it should just work

[half horizon]

i didint found where is flac

where i can found format types ?

[latent summit]

you see that pink BEST in the top right?

click that

[half horizon]

i see only only one best in the audioquality and i put it on 0 to be on best

[elder willow]

Sup y'all

what is this Discord Server about?

[half horizon]

yeah i see it

got the idea

ty so much is there more setings ?

[latent summit]

there is but nothing you have to mess around with

[half horizon]

yeah i changed it to flac

[latent summit]

cool you got the idea then

[spark raptor]

hey, I'm looking for somebody with a Google Colab account, as Google have suspended my free account for tonight because of too much RVC training and they accept none of my credit cards

I have all my source audio as WAV in a ZIP directory, so I'm not sure if #1159289738314919936 is the right channel for me, where voice creators are usually required to find and edit the source audio themselves

I'm willing to pay you via PayPal the amount of a one-month Colab Pro fee

[half horizon]

now ican get best possible audio so i wil do all from scratch

ty so much all of u guys

[latent summit]

for the record if you wanna download a normal video make sure its set to bestvideo ext=mp4

[latent summit]

not a problem meboy

[half horizon]

yeah but idont need video i only focused to learn training and make ai cover

or make somone say what i want

famous one on youtube

[spark raptor]

you mean I should get a different Googlemail account?

[half horizon]

iwil told u wich steps i do on mvsep to rate me if i need change somthing

[spring condor]

Hi

[covert lake]

you can:

• use #1159289738314919936
• pay #1191429836321849435
• just make another google account, or wait for tmr

[quartz roost]

ai covers are dead. vocaloid originals are goin viral

[covert lake]

rip ai covers

[half horizon]

will search the term ty

[quartz roost]

https://youtu.be/19y8YTbvri8?si=u565PP5jpjP6w0bE

fastest new record to 10 million views

[half horizon]

and after ? wich ones for example i was using mdx 8k 10 17 then mdxb then foxjoy then denoise

[quartz roost]

https://youtu.be/ijOVfRGOAno?si=hzkvV3TEZLsCkuo3

[fallen ruin]

We mammals we dont [REDACTED] eggs

[spark raptor]

okay, thanks guys, I'm trying it with a new account

also @elder willow

[feral halo]

Ayo? @spark raptor level 1 !!!

[half horizon]

this very nice igot the idea this is very nice

[quartz roost]

https://youtu.be/v1TaIfKctPo?si=LMRsL9TRqI0Y7LXj

teto is the best selling ai singing voice

[elder willow]

https://www.youtube.com/watch?v=8FYPGjSesZM

[half horizon]

i like the idea u just put bak of vocal then put lyrics the he wil sing em ?

[half horizon]

nice one

[elfin horizon]

Hi

[spark raptor]

hm, I'm afraid it may not work, as when making the new Googlemail account, it immediately asked for my phone number, so Google will prolly be able to tell it's me

@elder willow

[half horizon]

do u have tuto how i can do it as a newbie ?

[quartz roost]

this keeps dropping phonemes

[quartz roost]

https://kasaneteto.jp/sv/

buy her here

[swift abyss]

Hello everyone, I'm making models for R$5, whoever wants can call me in PV

[finite thicket]

bababooy

[half horizon]

ty but im moneyless i just have awindows7 and i want learn and make things in googlecollab

[lusty violet]

omg

/lvl

!lvl

[feral halo]

Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
✨ Tip: Use /card to modify your rank card

[quartz roost]

there is a free trial

[half horizon]

what i need for example to make somone that i have his audio to sing ?

i need only his audio ?

and put it then he wil sing ?

[forest quarry]

yeah

[half horizon]

i need no card debit ?

[frigid condor]

dunno why but this song gives me CH4NGE vibes by Giga

[forest quarry]

....no?

[quartz roost]

https://youtu.be/L94J1eKbdvE?si=SvLmJfztkPuGZHR-

[feral halo]

Ayo? @quartz roost level 68 !!!

[half horizon]

i mean for trying the site i need register they sk for my debit card or no ?

[quartz roost]

https://resource.dreamtonics.com/download/English/ download the "basic" version and a "lite" voice database. eleanor forte does english

[frigid condor]

+skull+

💀

the choreography is insane tho

[forest quarry]

what site?

[half horizon]

kasaneto that testki send for vocaloid

[forest quarry]

mmmm i don't think so

[forest quarry]

they've explained right there ^

[half horizon]

yea aleardy saved this link ty so much

[quartz roost]

you will need to buy pro for full features like all languages

[half horizon]

now i need start from scratch to clean dataset

[long sundial]

Cineva român?

[half horizon]

i saved the web site wil try this things after having clean data set

[long sundial]

Sunt nou pe grup

[half horizon]

ty so much tetski

[quartz roost]

the voices are already made, it will not work with your dataset

[half horizon]

i want make my friends sings not famous people

for example

or somone in my country that i have his audio on youtube

[quartz roost]

you're free to do this with permission from the voice provider with rvc

[half horizon]

yeah i have perpission of course where i can have their audios

making pzople sing is good things

[quartz roost]

synthV is made with paid professional singers with a contract

[lime mortar]

hi

[half horizon]

ah i got it me im just amateur i want make parody or meme

[sleek spruce]

so I’m kind of new to all of this, but is there any other program/website that’s better than Suno?

[lime mortar]

im tryna figure out where im supposed to use the models and stuff

[half horizon]

i download audio using stacher i use format flac but the file downloaded was on webm format is that normal ?

[forest quarry]

there's Udio... dunno any other good alternatives

[forest quarry]

You can use them in Colab - check https://docs.aihub.wtf

RVC > Cloud section

[lime mortar]

ok

[feral halo]

Ayo? @lime mortar level 1 !!!

[static sorrel]

korno

[elder willow]

whats the best girly voice model

[short sandal]

Please help me, what programs are there to raise your voice separately from voice mod?

[half horizon]

mvesep maybe ?

[short sandal]

thanks

[crude eagle]

Anyone know any good voice models for narration? Like really basic voices?

[covert lake]

what? btw we don't give support for voice mod, we use wokada here

[tardy vector]

3500 euros for shit

[short sandal]

Do you use the wokada channel? Or did I misunderstand

[feral halo]

Ayo? @short sandal level 1 !!!

[elder willow]

anyone here can recommend a shop for egirl models?

[feral halo]

Ayo? @elder willow level 1 !!!

[forest quarry]

ilaria's shop in #1191429836321849435

[random robin]

🤰🏻🧍‍♂️
🤰🏻 🧍‍♂️
🤰🏻🧍‍♂️
🧍‍♀️👶🏻🧍‍♂️
🧍‍♀️👶🏻🥛🧍‍♂️
🧍‍♀️👶🏻 🥛🧍‍♂️
🧍‍♀️👶🏻 🥛🧍‍♂️
🧍‍♀️👶🏻 🥛🧍‍♂️

[forest quarry]

😭

[elder willow]

there is no swae lee models

???

[slate epoch]

So, which WEBUI have you guys been using to make AI covers?

Seems Hina mod is unusable now

[forest quarry]

https://docs.aihub.wtf/rvc/cloud/applio-colab

[slate epoch]

Thanks, I'll try it out

[half horizon]

i use stacher to download from youtube but i dont understand i choose flac but audios downlded are format webv idont understand why

someknow why ?

[loud oak]

hi

[feral halo]

Ayo? @loud oak level 4 !!!

[livid mural]

hello general chat

[rotund sierra]

I love brazilian music
https://www.youtube.com/watch?v=AtWW_b2H1n4

[velvet monolith]

y'all is it the vb-cable thing safe?

[proven ember]

Excuse me, but do y'all guys have a realistic Sans AI Voice?

I just need to make an AI cover of him singing Life is a Highway

[velvet monolith]

y'all is it the vb-cable thing safe?

[forest quarry]

you can always search up whether its safe or not

no problems on my end

[velvet monolith]

do you have it?

[feral halo]

Ayo? @velvet monolith level 1 !!!

[proven ember]

I see someone else joined beside me

[north spoke]

[half horizon]

the files downladed as a webm format . and this dont work on audacity i choose flac but files still downlded webm

[half horizon]

any help guys if i need to do somthing to make feiles flac

[north spoke]

Ola

[half horizon]

hola

[dusky wagon]

ahola

[silk robin]

Intriguing

[loud oak]

https://flashcardfy.lol great study tool

[sand forge]

oo

silltness

[jovial badger]

guys i need help #🔍│help-w-okada

[sand forge]

what website should i do with the ai model?

[summer grove]

enviar un mensaje

[storm meadow]

I would download RVC instead of using a website instead of google colab or HuggingFace when possible

For inference/covers you don’t need a GPU

[sand forge]

like downloading a rvc seems hard for me

[feral halo]

Ayo? @sand forge level 1 !!!

[storm meadow]

Applio has a installer

[sand forge]

OOoo

[storm meadow]

-local

[rare sorrel]

🖥️ Local stuff

• 🍏 Applio, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub
• GPT-SoVITS, by RVC-Boss GitHub

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

More commands: -audio, -colabs, -overtrain, -realtime, -rvc, -help

[sand forge]

tysm!

[elder willow]

hello

[ivory drum]

hello

[half horizon]

guys im still stucked on how to download flac audio from youtube ytbdlm dont work for me and for stacher.io i dont know why the format is always webm

ichoose flac but still output is webm

[storm meadow]

Try https://cobalt.tools

[half horizon]

ty so much

[feral halo]

Ayo? @half horizon level 7 !!!

[storm meadow]

@half horizon settings button

[half horizon]

yeah i did there are no flac and for best chooice the format is opus dont work on audacity wich one i need to have best quality for dataset

[dull void]

anyone got the link for sketches voice for rvc

[storm meadow]

Use .wav

[storm meadow]

#1175430844685484042

[half horizon]

is that lossless and good for dataset ?

[storm meadow]

.wav is true loseless

[half horizon]

ty somuch

[storm meadow]

But YouTube compresses audio anyway so if you can get it elsewhere I would probably

[dull void]

didnt work it just gave me the voice for M Gaby Sketch

[storm meadow]

Check weights.gg

If it’s not there train it yourself on RVC or #1159289738314919936

[half horizon]

i have only one source is yt

[storm meadow]

Yeah then do it then

[half horizon]

ty i wil is my first prject ty for the help

[west marsh]

the gojo voice won't work for me how do i use it

actually none of them will work for me

[storm meadow]

What are you using?

[west marsh]

the site

[storm meadow]

Which site?????

!howtoask

[pine acorn]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[west marsh]

https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=OVQoLQJXS7WX&uniqifier=1 sorry im new

[storm meadow]

Outdated

-colab

[rare sorrel]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

More commands: -audio, -local, -overtrain, -realtime, -rvc, -help

[storm meadow]

Use applio I guess

[west marsh]

thank you

[feral halo]

Ayo? @west marsh level 1 !!!

[fallen ruin]

RIGHT!?

https://cdn.discordapp.com/attachments/1115866252729139221/1238975784505446471/92c.jpg?ex=66413d9e&is=663fec1e&hm=5c675bc96cd9c27257bc6ee9881b2859365b2e92ef72e4ba04b1f1beb28cae02&

[rotund venture]

@blissful citrus it does sound really good

@blissful citrus

can u send me the female voice changer file

ok

[feral halo]

Ayo? @rotund venture level 1 !!!

[rotund venture]

im gonna get a mic soon

[west marsh]

i have one more question how do you download the file from the site

im using AICoverGen-NoWebUI [English]

[spark raptor]

@elder willow @covert lake okay, so far I've found that one account equals one voice before Google suspends it again

[rapid oxide]

can i make ai covers on mac

[storm meadow]

[covert lake]

It doesn't really work like that, it depends user to user how much time of free daily gou you got, it could be 1 hour, 4 hours, 50 mins, it all depends on how much u use it

[modest lance]

Any guide for RVC 😄 ? The app uses like 80% of my CPU on default settings (GPU setting is set on my GPU tho, RTX 3080)

[storm meadow]

What fork are you using

It should not be using mainly your cpu

[spark raptor]

one training usually takes 90 minutes for me

after that, Google immediately suspends me when I try again

[modest lance]

MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a

[covert lake]

Training time can also depend too by the dataset lenght and the epochs

[modest lance]

Is there any guide for the settings in RVC ?

[covert lake]

Btw its not really "suspending", its just that the free daily gpu time

[strong haven]

no doofenshirmitz ai voice changer, guess i gotta hold hostage Dan Povenmire

[storm meadow]

Did you select your GPU

[covert lake]

Btw

[storm meadow]

And that’s w-Okada not RVC, (can use RVC models tho)

[covert lake]

-rt

[rare sorrel]

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

[modest lance]

yes, the GPU: came default on "cpu", but i select it on my gpu

[modest lance]

where to download the RVC ?

[storm meadow]

Yes select your GPU

[feral halo]

Ayo? @modest lance level 1 !!!

[storm meadow]

RVC is for covers

[modest lance]

oh i see

[storm meadow]

There is realtime but Okada is pefered

[modest lance]

i wanna do realtime trolling

[storm meadow]

Send a ss

[spark raptor]

4.5 minutes, 500 epochs, that equals 90 minutes of training on the Colab

sigh I only wish I had more HDD space, then I could do all this by means of my own GPU

[storm meadow]

I’ll tell ya what to change

[spark raptor]

but thank you for letting me know it's a daily limit, so I don't have to wait until next month or so

[covert lake]

Yea dww

[vital herald]

does anybody know how to upload any kind of hugging face or drive model to weights.gg

[half horizon]

iwill wait whole life ?

its 20 min and still on epoch30

??

[devout wagon]

Weights.gg only download the pth file and index when you include the huggingface link in #1175430844685484042

[devout wagon]

Not really, you can look at something else while you train but you must keep track of time. It'll tell you how long you have left when you click on the "session resource". I mean, it will kick you out so

[vivid crest]

sa

[half horizon]

i can wait 2 hours more is that enogh ? dataset 28min

[devout wagon]

https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/#section-13 you must save your model using this step before Google kicks you out

when you end the training, you find the steps and epoch in your weights folder on the left side of colab. Mangio fork > Weights >

[half horizon]

i feel dumb ididnt get the idea its my first experience is there avideo explain it im on no cpu mode google collab

[devout wagon]

nope, you ran out of gpu? You have to do it on another account

and it didn't save because you didn't save your model steps

[half horizon]

so i need to do forexaple 100 epoch ? for my case ?

[devout wagon]

No, you follow the tensorboard here to detect overtraining https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/#--monitoring

[half horizon]

i do i folow now

[devout wagon]

oh my bad this is #🧬│ai-chat

[sour kindle]

hi guys, what collab do you use? i tried to make an ai cover again after a couple months, but the one I was using didn't work

[drowsy bison]

guys for realtime voice changer what better? 4090 gpu or tesla t4

[tight cypress]

4090

Tesla T4 have less VRAM

and it is older too

[primal vapor]

hi

[foggy narwhal]

does anyone know a hakari voice?

[elder willow]

when the nettspend ai model dropping

[amber light]

when the beat fire till nettspend starts rapping

🗣️🔥🔥

[lost wraith]

Tfw rap

[plain warren]

Any Ai Artist / YouTubers ?

[storm meadow]

#1175430844685484042 has voice models of artists and yotuubers

[tight cypress]

oh yeah, something is definitely cooking

[zenith hamlet]

Ilaria rvc

[sturdy ocean]

uhh

[plucky lark]

1200152556266070180

[feral halo]

Ayo? @plucky lark level 4 !!!

[plucky lark]

MINOR VAPING IN VC

get him out @signal cairn

[acoustic gust]

made a ai character https://character.ai/chat/WOUkCNPEKM3kouUmupO68quChxrWxO6da0qQwqYCdGE

[elder willow]

What vocal extractor was a game changer for you guys? What came out the cleanest for you?

[frozen egret]

hot debate, suno is better than udio
suno:
-more length in the created music

• gives more credits for the free trial
  -much better versatility
• there's even an explore section to find new genres

and weird genres to say the least

[tight cypress]

Suno Home library have improved in like what, under 2 months?

with them adding new design, composition, categories, stuff

[coarse vessel]

vocaloids https://www.youtube.com/watch?v=vG_iT-f-Q_I

[errant portal]

I wonder why most of the models on kits.ai are locked now

I don’t know how it works

[storm meadow]

You select a model, use this voice

Is it really that hard?

[errant portal]

Well I’m just stupid then

[ember current]

this si like the 80th server why does no one have Herbert's voice from family guy

how do i do that

[storm meadow]

Check #1175430844685484042 or weights.gg

You can request the model in #1159289738314919936

[errant portal]

Is it better to use the app or website for weights.gg because the apps us just a black screen

[lament gulch]

Hello

[storm meadow]

I use website

[errant portal]

Ok

[storm meadow]

But I really only use weights for models, I use RVC for ai covers

[violet igloo]

ᶻ 𝗓 𐰁ᶻ 𝗓 𐰁

[lean blaze]

.wk

[indigo pike]

XHRIS2EAZY in AI HUB

👑  ⁦mox⁩ - 1559 plays
** 2.**  ⁦gilbert⁩ - 903 plays
 3.  ⁦Kainé⁩ - 313 plays
 4.  ⁦exan⁩ - 231 plays
 5.  ⁦yeresin⁩ - 187 plays
 6.  ⁦p⁩ - 136 plays
 7.  ⁦epilogue⁩ - 99 plays
 8.  ⁦fantuh⁩ - 97 plays
 9.  ⁦sqwab⁩ - 94 plays
 10. ⁦kittyevelynyan⁩ - 82 plays
 11.  ⁦1234567891231⁩ - 53 plays
 12.  ⁦00.q⁩ - 47 plays
 13.  ⁦luminal⁩ - 45 plays
 14.  ⁦beef⁩ - 42 plays

sigilkore Artist - 77 listeners - 4377 plays - 56 avg

[coarse vessel]

its point is only to tame harsh peaks with subtle compressor settings, while keeping the dynamic range natural as possible. watch the peak and max RMS stats before doing it.

[feral halo]

Ayo? @elder willow level 18 !!!

[coarse vessel]

model maker role? then do /submit your polished model in #1159514067187277865

[alpine granite]

pretty sure thats a level role

[storm meadow]

#1159290752195633273

#1159290752195633273

[red torrent]

hallo

[flint jacinth]

and if there is a mellstroi model?

there's just nothing on search

[broken raft]

You can ask in #1159289738314919936 Someone might be able to make a voice model of Mellstroi for you.

[coarse vessel]

embed fail

[lime zodiac]

Who can help with writing an indicator in pine script, I'm a beginner and I have an idea for it

[thin raven]

Hello, Im XtronXI

[elder willow]

[cunning swan]

Hey guys! I’m Ictguy. Glad to be part of this.

I prep’d a Medium article on how to get Gemini-Advanced thru cookies

Wanna read?

[feral halo]

Ayo? @cunning swan level 1 !!!

[covert lake]

promo only in #1159290752195633273

its not even for free lol

[cunning swan]

You gotta pay a Lil penny

[covert lake]

theres no reason to put 'for free' then lmfao

[abstract ridge]

And home security cameras should be illegal because you're taking away the main source of income for burglars

[loud sedge]

who gives me duckus e girl model?

link

[atomic vector]

An error occurred

Sorry, I could not find models that match your search "recep tayyi erdogan"

[winged breach]

is there anyone who knows how I can find Turkish voice models?

[winged breach]

AI HOME a nasıl katılabileceğimi bileb varmı discord a çok yabancıyım, biri yardımcı olursa çok sevinirim

[zealous field]

Oi

[remote owl]

HI

[vague elbow]

hi guys do you know how to edit the cover so that you can’t hear the ai?

[wispy widget]

ai hub

[worldly oasis]

how to ai

[wispy widget]

you do the ai

[plain hemlock]

got a question

would you recommend using like the Server mode over Client mode when using Okada because i find that like even if i get rid of nearly every issue in the client i will still get some small robotic inflexations in my voice

[forest quarry]

if server makes it better, sure, go for it

[plain hemlock]

there any dangers to using the server mode?

[forest quarry]

...don't think so?

[lime locust]

hi, i have a problem and i need help

i get an error while executing the step 2 one the google colab

[forest quarry]

that's an outdated colab, use a newer one, like Applio:

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb

if you want a guide for it: https://docs.aihub.wtf/rvc/cloud/applio-colab/

[lime locust]

thanks

[forest quarry]

yw :)

[plain hemlock]

Ok the cmd when connecting to the server is giving me an error saying the sample rate isn't supported

[elder willow]

which software to make these ai cover things

No cached install found..

Download completed successfully!

[plain hemlock]

WOAH wtf

the server system is so clear!!!!! it's like a river stream

[forest quarry]

https://docs.aihub.wtf/ has a few local softwares

also guides on everything

[forest quarry]

hmmm

that's interesting

[ionic pumice]

holy shit timbo back

[elder willow]

sup chat

[elder willow]

thx

how to do ai tts?

[forest quarry]

applio has the TTS tab, just use that

[elder willow]

alr

[plain hemlock]

I wonder

Does gain in and out do anything at all or is it just a volume setting

[feral halo]

Ayo? @inner mist level 1 !!!

[snow sedge]

Wdym by "do"? They amplify volume, that's what they do

[elder willow]

im so lazy to search for it what is the best thing to get vocals cleared bla bla

yk like getting the vocals

[covert lake]

https://docs.aihub.wtf/rvc/resources/vocal-isolation/

[night fox]

a few year ago there were Viktor Reznov Voice model

They removed it from everywhere

if anyone have it can send me?

quickly

[bold saffron]

https://weights.gg/models/clm73bgt8222ucctc0me78xvt

[night fox]

thanks

[dense ether]

where are the tutorials

[feral halo]

Ayo? @dense ether level 1 !!!

[ionic pumice]

oo aa oo aa

[glacial lintel]

hi

[heavy cipher]

Give me ai voice model of rosè

[storm meadow]

#1175430844685484042

[covert lake]

https://docs.aihub.wtf/

[storm meadow]

-docs #1159513888199540817

[rare sorrel]

Suggestion for @dense ether

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[stark tapir]

всем привет!

[ionic pumice]

hi timb o

[storm meadow]

#1159290752195633273

[heavy cipher]

This is so confusing

So there's a random chat here

And I dk anyone

[steel spear]

heyy

[storm meadow]

All servers when you first join

[ionic pumice]

yeah i mean

[torn halo]

Good morning everyone hope yall have a good day/night

[heavy cipher]

I want to make ai covers

[ionic pumice]

-rvc

[heavy cipher]

And I m so fucking dumb I can't even do that with tutorials

[rare sorrel]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[ionic pumice]

if you just want to use models it's p easy :)

[heavy cipher]

I made it!!!!

[feral halo]

Ayo? @heavy cipher level 1 !!!

[heavy cipher]

Yeahahahahh

[stable cloak]

Hi, are the "bot assist" voice models trustable?

[ionic pumice]

elaborate?

[abstract trellis]

have somebody a good girl voice?

[elder willow]

@limber elbow whyd you block me lucas?

[ionic pumice]

ilaria has free & paid e gorl voices

[hushed raft]

@limber elbow

why did you block us

[elder willow]

yeah lucas

[coarse rampart]

hi quick question can u use ia voice for commercial use ?

[elder willow]

who is stopping you?

[ionic pumice]

i don't think this is um

the right place for this

[hushed raft]

@limber elbow text us back

[feral halo]

Ayo? @hushed raft level 1 !!!

[elder willow]

he wont dm me

[elder willow]

so yes it it

[ionic pumice]

if it's a model from an artist: yea no.
if it's a model made from users using their own voice(acted or not): ask them

[ionic pumice]

what he do

[coarse vessel]

ethically, ask the voice owner for permission, otherwise dont monetize it

[elder willow]

stuff

[feral halo]

Ayo? @elder willow level 1 !!!

[timid musk]

so close to being correct

[hushed raft]

he no lie it true

[timid musk]

Ethically, using someone’s voice for anything without permission is seen as bad

Legally, if you claim it’s the actual person and try and monetize that you can get in trouble legally and get sued

If you want to make your favorite streamers voice into a voice model to use just for fun there is no repercussions 👍👍

For example someone made an ad with a Nicki Minaj ai model and she sued them apparently, idk if she ever followed thru with that

[fervent sapphire]

How to use the ai voices

[ionic pumice]

-rvc

[rare sorrel]

Suggestion for @fervent sapphire

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[quartz roost]

Yes there are legal commercial ai voices for sale or get written permission.

[tulip gate]

very very new to AI voices

I just trained a model with RVC now how do i use it?

[amber light]

https://docs.aihub.wtf/

[tulip gate]

many thanks

[unreal basin]

@limber elbow

[unreal condor]

how can i use the voice changer in game

[stuck frigate]

Virtual cable

[unreal condor]

can you possibly dm me to show me how

[feral halo]

Ayo? @unreal condor level 1 !!!

[unreal condor]

if not np

[crude drum]

link?

[stuck frigate]

Sent

[stuck frigate]

Google

[unique moth]

Hi does anyone know any voice clown ai for free or how to use those ai paid shit for free ???

[stuck frigate]

Vb cable if I am not wrong

[unique moth]

I want to start making content ai youtube content creation and everything is paid licences

[outer spire]

[ionic forge]

The best ai for singing doesn't exist

I thought Appelio is gonna be the best

but I found a HUGE problem that cannot be solved

I am pissed. a lot.

[wicked badger]

yo

[tulip gate]

sweet this told me how to save my models once i've trained them (which i didn't know)

how do i use it
I used VRC to train but i want Bark for the voices

[quartz roost]

This is the best selling singing ai voice

[ionic forge]

AI Voice Changer duh

[ionic pumice]

real

[ionic forge]

that's Utau

[ionic pumice]

nuh uh

[ionic forge]

yuh uh

[ionic pumice]

nuh uh

[unreal condor]

didnt work.. guys i need method to use voice changer in call of duty any help?

[ionic forge]

Anyway

[quartz roost]

You said singing not realtime

[ionic forge]

I hope somebody can help with this problem #✨│ai-help

[ionic forge]

I thought Teto is not from Vocaloid

[zinc rune]

where can I convert