🧬│ai-chat

[rigid granite]

I've been looking thru a lot of them but dont know how to judge which ones are good and which ones are not

[rugged garden]

cant be bothered checking

[feral halo]

Ayo? @rugged garden level 1 !!!

[covert lake]

You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU

How can i help you if u dont give me the info?

[rugged garden]

the name of it, right?

[covert lake]

yes, example: RTX 3060

[rugged garden]

GTX 1660

i think? idk

[covert lake]

Alr, you can't train models locally (on ur pc) with that

[rugged garden]

ive done it before

[covert lake]

As you dont got a good PC, its better you use cloud (remote good pc) 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):
  • Mainline (UI, No guide as of right now)
  • Applio (UI, No guide as of right now)

[covert lake]

one sec i needed to send the message lol

[rigid granite]

Hey nick 🙂

[feral halo]

Ayo? @rigid granite level 1 !!!

[covert lake]

I sent you all ways to do it on cloud:

Google Colab = Ez but low gpu time

Kaggle = SUGGESTED, it's harder but way more gpu time

[covert lake]

hi

[rigid granite]

You probably get this question a lot but what's the best one to sound like a english female?

[covert lake]

i dunno, most e girls models are paid made by model masters #1191429836321849435

[rigid granite]

Got any free ones to test out? Cus I've never used a voice changer

[covert lake]

Nope, i don't really use e girl models

[rigid granite]

👌🏻 I'll stick with markiplier than cus it's funny

[devout bluff]

lmk if you find any been looking

[devout bluff]

is there a way to make one yourself that is much better??? I dont understand the process well for making rvc models but does bigger dataset = better model?

[rigid granite]

#1213260351702044732

This one I guess, let me know if you find a better one

[covert lake]

well, it really depends, for example for a dataset u need only 10 mins

anything over an hour has no difference in result

it also depends by using the tensorboard

what's ur pc gpu? to see if u needa do it locally (on ur pc) or cloud (remote good pc)

[devout bluff]

3080

[feral halo]

Ayo? @devout bluff level 2 !!!

[tame reef]

hello, what is the best setting for Chunk and Extra (GPU:4050)

[devout bluff]

is that one high pitched/unnatural sounding though? I feel like the best model would be something that is just sounds normal

[spiral timber]

How come this server gets a new member every minute?

[rigid granite]

It's high pitched but works, but if you find a more natural one please @ me I'm in desperate need

[covert lake]

As you got a good PC, you can use RVC locally, you can choose between:

• Applio: A fork of RVC with some extra features like Applio TTS, same quality tho
• Mainline: The original RVC

Check also https://docs.ai-hub.wtf/essentials/how-to-make-voice-models/

[devout bluff]

I see thank you, I also have another question, is there a large benefit in making your own model or is it more just finding the voice that sounds right

since there’s already so many models available ^

[devout bluff]

I have found ones that sound more natural but ngl they all have a slight unnatural undertone that you can kinda tell it’s ai

[solar glen]

hello

[covert lake]

well, it depends on how good u are at making models ofc

[devout bluff]

what is the difference between a generic model you’d find here and a paid one exactly? wondering how they can make it that much more realistic

[covert lake]

Well, paid ones are made by model masters, which got more experience and approved by QCs for making high quality models

[devout bluff]

yeah I understand they have more experience but what exactly do they do to make a model a lot better than an average guy??

[chilly lake]

gee... how much are they paid

[covert lake]

well they got more experience, and care more on cleaning the dataset using iZotoper RX for example

[covert lake]

check #1191429836321849435

depends really on their commission

[devout bluff]

so their better at it because they clean it better? I’m mainly asking all of this bc I want to know how I can make a good model myself

[covert lake]

depends by model master but usually around those prices ig

[devout bluff]

since if I do take the time to make one I want to do it well lol

[chilly lake]

i'm sure after trying a dozen times you can make something good

[covert lake]

the thing is, they got experience which u can achieve by using it much

[devout bluff]

hmm ok

I’ll start creating some models then and figuring it out Ty for the help

[covert lake]

like its not really easy someone would be model master at first time, i think u get what i mean

[covert lake]

Yw

[devout bluff]

Yeah I understand that, I was just mainly trying to ask for specifics so I can keep that in mind while I try to make one myself

But Ty so much now I have the resources to do it

[novel nymph]

Hi Nick

I have a question about uvr5

Which is the cleanest vocal mode?

I use kim vocal 2

[tepid basin]

Nah not these days

Do you have bs-roformer UVR beta? Its the best rn

[novel nymph]

Oh
I don't

What is that?

[tepid basin]

https://github.com/TRvlvr/model_repo/releases/download/uvr_update_patches/UVR_Patch_4_14_24_18_7_BETA_full_Roformer.exe

Use BS-Roformer under the MDX-Net category. This is a early beta of UVR from the devs

You can just update your existing UVR with this exe. Or uninstall it/reinstall if you want

The same model is also available at https://mvsep.com/en , if you don't mind waiting for the queue

Higher SDR = better

[novel nymph]

Okeyyy

Thanks

But
I don't know which mode is better

[tepid basin]

Like which model in UVR is best?

[novel nymph]

Kim vocal 2

I guess

[minor blade]

The best one isn't Kim vocal 2.

It's pretty outdated compared to the new ones

[novel nymph]

🤦

[tepid basin]

Kim 2 is pretty outdated by todays standards

[minor blade]

Facts

[novel nymph]

Then what is the best🥹

[minor blade]

BS Roformer (11.31 SDR version) on MVSEP

And Mel Roformer (Bas Curtiz version with 11.18 SDR) also on MVSEP

[tepid basin]

BS-roformer, available on MVSep or that github link I sent (for your local UVR)

Mel is also valid

Not smart enough to figure out how to get that Melband karaoke on UVR tho. I neeeeed to do that

[feral halo]

Ayo? @novel nymph level 19 !!!

[minor blade]

Yeah, that one

[novel nymph]

Okey

[novel nymph]

The sdr is highest
Right?

[dreamy kiln]

why dont vac's work for me?

[ivory kayak]

rn im using so-vits whats the one that takes INDEX files

[minor blade]

RVC

[ivory kayak]

tanks

[dreamy kiln]

how do i get my vac to work, ive watched tutorials but they wont work

[ivory kayak]

i downloaded and extracted the mainline rvc what now

[feral halo]

Ayo? @ivory kayak level 1 !!!

[rotund niche]

how do i make a model of my freinds coice

im severly stupid

[rotund niche]

plz @ me when you respond

[minor blade]

Read the docs buddy.

https://docs.ai-hub.wtf/

[ivory kayak]

ok but pls respond

[haughty gyro]

hello guys
I’ve been exploring AI-generated music and came across audiomodify.com. Does anyone know other good sites? i need to generate britney spears voice

[earnest dragon]

why not just use pre compiled applio or if you are just doing inference them just use ilaria zero hf

[ivory kayak]

isnt appolio hacked or sumn

oh nvm lol

[earnest dragon]

Audiomodify don't be using use the stuff reccomend here it's all using rvc.

[ivory kayak]

thanks cxsmo i suck at this ai stuff

and you saved me...🥺

[earnest dragon]

no it isn't and was never it was only the discord account but that got fixed

[haughty gyro]

thank u bro

[chilly veldt]

Type shi

[proper egret]

Hello!

[surreal hemlock]

Don't click: ||if you are reading this, it is already too late. You have been infected by the curse of pee pee poo poo man. If you don't copy and paste this on 5 different servers, you will face the consequences. I was a victim like you, trying to be free||

[ivory kayak]

FUCK

[surreal hemlock]

haha

[proper egret]

Fine I wont

[surreal hemlock]

https://cdn.discordapp.com/attachments/1276346952262619301/1295280819610452008/attachment.gif

[feral halo]

Ayo? @surreal hemlock level 1 !!!

[surreal hemlock]

Oh no.

[proper egret]

Can we talk about regular AI things here?

[feral halo]

Ayo? @proper egret level 1 !!!

[karmic sigil]

Just saying hello here, creating wonderfull music with AI myself

[earnest dragon]

like what is regular ai things voice and face are regular ai things

[jovial veldt]

Hi, does anyone know how to make an autotune voice model of rock music from 10 years ago (maybe with a bad quality), and doesn't get distorted and sounds like the original voice?

[earnest dragon]

what do you mean auto tune voice model it doesn't work like that because it doesn't act like autotune

[jovial veldt]

What happens is that I created a voice model with weights, of a rock band that uses autotune most of the time (Blue Stahli), but the result does not resemble the original voice, I tried not to use the parts with autotune, but I was not satisfied with the result

[feral halo]

Ayo? @jovial veldt level 2 !!!

[karmic sigil]

Couldnt you just record the entire track, and add some kind of referb / autotune after the fact?

[proper egret]

Like talk about new AI models and stuff...

[jovial veldt]

Do you mean to add reverb to the audio tracks after having created the model? or before

[vocal flare]

whats the best app for duke dennis AI voice

[jovial veldt]

Maybe that can be done with a synthesizer

[jovial veldt]

There two here https://voice-models.com/

[vocal flare]

thank you

[ivory kayak]

yo is there like a tutorial for applio

under voice model do i put the file path to the pth file

holy shit this is wilddddd

[static hill]

قال قحاب

[feral halo]

Ayo? @static hill level 1 !!!

[earnest dragon]

no stop promoting these trash websites that just use rvc when we already have everything here hf and weights.gg

[thorn cove]

how do i train an ai model on collab

[broken raft]

#🤖│bots message

[polar flax]

which AI? SD1.5? XL? Flux? LLM?

[covert lake]

Colab or local?

[covert lake]

RVC, what's ur PC GPU?

[covert lake]

Yw

[covert lake]

I was sleeping lol

[novel nymph]

good morning lol

[covert lake]

goodmorning

[solar torrent]

I went there to download the voice model that actually has a link to Hugging Face and not some inference thingy. But since you keep calling these trash, sick, ass and garbage, I had to bye.

[polar flax]

iirc we here have been recommending that particular site alongside weights.gg and #1175430844685484042

anyway more choices (that are downloadable) are always better

[orchid slate]

there is a better channel for this #1159290752195633273

[solar torrent]

Hey man, I'm sorry. But I'm only here to post my own creations, not looking for a job.

[thorn cove]

thanks?

[covert lake]

#1159290752195633273

[covert lake]

its better to use kaggle btw

As you dont got a good PC, its better you use cloud (remote good pc) 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):
  • Mainline (UI, No guide as of right now)
  • Applio (UI, No guide as of right now)

kaggle gives more hours

[tender flare]

Hi, can anyone help me? It asks me for a configuration file type in json format, and since it is in Japanese I have no idea what to do, the voice rar files do not have any json so I am a little lost.

[covert lake]

are you using RVC?

[tender flare]

what is RVC?

[feral halo]

Ayo? @tender flare level 1 !!!

[tender flare]

sorry im really new

[covert lake]

what program are you using for making ai covers?

RVC = Retrieval-based-Voice-Conversion, its the best program for Speech To Speech models

Could you tell me what are you trying to do and following what guide?

[tender flare]

I try to be able to speak in real time using AI, and I was barely guided by a video

[covert lake]

don’t follow yt tuts

most of them about this topic are outfated

from what i seen ur tryna do realtime voice changing for calls

what’s ur pc gpu?

You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU

@tender flare are u on mac perchance btw?

[tender flare]

I noticed that they are outdated, the interface that was in the videos is very different from when I downloaded the program etc.

[covert lake]

yea, never follow them about this thing

please give me answer to my questions so i can help you better

[tender flare]

in windows, i have a rx580, intel i7, 16 ram an 1 tb rom

[covert lake]

alr ty, your pc is good enough to do it locally (on ur device)

-rt

[rare sorrel]

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

[covert lake]

1st link, its wokada, the program specilized to use RVC models in realtime, specifically the fork that has better performance

unfortunately there is no yt tut, so u have to read the guide

if u run into any issues, u can ask in #🔍│help-w-okada

[tender flare]

Well then the first one is the one that has the answers?

[covert lake]

yea its the guide, follow the 1st guide

[tender flare]

tysm

[covert lake]

yw

[rugged veldt]

Hello, there. I'm mainly here so I can change my voice to a space marine while playing. 😄

[covert lake]

what’s ur pc gpu?

[rugged veldt]

AMD Ryzen 7 5700U with Radeon Graphics

if that is what you want to know 😄

but i play on PS5 and call with friend on PC

[hazy pine]

Yoo hi

[covert lake]

seems like you got integrated graphics

which is kinda bad

if u really want to do it locally (on ur pc), u could try the cpu mode in https://rentry.co/ForkVoiceChangerGuide

or else use cloud (remote good pc)

For Realtime Voice Changing for Calls on Cloud (remote good pc for those who don't have a good one, YOU CANT DO THIS ON MOBILE):

• Google Colabs (4 hours daily of free T4 gpu, easy to use, require only a google account) :
  • How to use Original W-Okada's Voice Changer Google Colab
  • Modified W-Okada's Voice Changer Google Colab
• Kaggles (30 hours weekly of better GPUs, T4x2 & P100, harder to use, requires an account and a phone number)
  • Original W-Okada's Voice Changer Kaggle
  • Modified W-Okada's Voice Changer Kaggle

Google Colab = ex but low gpu time
Kaggle = harder but more gpu time

[rugged veldt]

And what role does graphics play in voice transformation ? isn't it about the microphone ?

[feral halo]

Ayo? @rugged veldt level 1 !!!

[covert lake]

in literally everything

ai takes ALOT of computing

The program ur running, runs the AI on ur pc using its power, and ai requires much power, think of it like running games

Ud need atleast a decent gtx 1650 or rx580 for running it on GPU, but as u got integrated graphics, ur only options are either cpu (but not sure if that would ne really good to do) and cloud (remote good pc)

[rugged veldt]

i see video where they use this app: https://voice.ai/ i dont know anithing about that. i just want to install some app ane start it.

[covert lake]

voice.ai is a scam

don’t use it

its basically paywalled wokada (the program ur using that is open source)

[rugged veldt]

oh ok. thaks a lot

[rugged veldt]

you save me

[covert lake]

yw

[rugged veldt]

I just want to soud little bit robotic it doesn't have to be ai voice just regular voice changer. i dont want to pay for something that trivial. bcs on discord there is something like that but it is not free.

[worn moth]

to use a voice model u need to pay?

[covert lake]

might not sure about a non ai voice changer

[covert lake]

No, unless u commission a model by a model master

what are u looking for? And whats ur pc gpu

[worn moth]

4060

[covert lake]

u want to inference (use models) on a pre-recorded audio or realtime voice changer for calls?

[worn moth]

ye

[covert lake]

yes to what?

i asked u 2 different things

[worn moth]

2

its simple i download a model from #1175430844685484042 and i put in the app and i see the train bt

[feral halo]

Ayo? @worn moth level 1 !!!

[covert lake]

so, just use them in realtime for calls?

-rt

[rare sorrel]

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

[worn moth]

ofc

[covert lake]

then follow the qst link, its wokada that uses rvc models, its the work which is better performance

just read it, and no there are no yt tuts bc all of them are old

[elder willow]

For the applio, can someone tell me what folder to put models in to register in the dropdown box selection

[orchid slate]

you dont.

go to download tab

[elder willow]

For downloaded models, where do I put them

[orchid slate]

use the interface

[elder willow]

Oh i see. I can drag and drop in there. I'll try it out. thanks

[orchid slate]

👍

yes

[elder willow]

Ah it works. Thank you again.

[orchid slate]

make sure after download you hit the refresh button top right

it will refresh the lists

[elder willow]

Do all voice models work with TTS or are those separate ones

Oh nvm it works yea

[orchid slate]

no not all. Read carefully what you download. Sometimes it explicitly mentions like - ONLY TTS , not rvc

[elder willow]

Ahh okay. Thanks. I'll be on the lookout in the future.

[orchid slate]

if not mentioned - you basically need to test it out and see if it works.

[rustic marsh]

https://tenor.com/view/wink-anime-anime-wink-peace-peace-sign-gif-3929988093931252469

[hidden ice]

hey all ❤️

[faint breach]

Why I don't speak

Why mute. Me

[tacit falcon]

what bro

[faint breach]

I got muted

Un mute me

[feral halo]

Ayo? @faint breach level 1 !!!

[faint breach]

Un mute me 😡

[solar torrent]

It seems like some mods were trolling around.

[feral halo]

Ayo? @solar torrent level 5 !!!

[solar torrent]

Or if you did actually trolling around their space, then I think you deserved.

[faint breach]

Un mute me

@tepid basin unmute me

Ohhhh

What the hell un mute me

[orchid slate]

small recap why you got muted and lateron unmuted:

You joined and started to blow clean your microphone with mic on.
Next time clean your microphone before you join a voicechannel. Blowing out your microphone hurts our ears.
After that you were ASMR-ing ... please get the microphone out your mouth while talking - position it properly.

[worn quartz]

Robots

[delicate stone]

@arcenovum you can dm for that..

[orchid slate]

basically hearing someone inhale to fully blow in the mic as hard you can. Fine i can understand 1 time ... but he repeated this like over 5x. Most of us in chat did lower the volume already. Since it continued i gave the temporary mute. He stayed and i unmuted him 😉 Hope this all clears up the confusion.

[jovial veldt]

The voice models be created with weights have more or less quality than rvc models created with google collab?

[orchid slate]

the quality depends on the time spend on providing the audio file to train on.

[stoic brook]

hi

[ionic pumice]

https://youtu.be/2Ysfrrtc-5k?si=OI19odFtEU0OwUm-

[stoic brook]

does the link in the guide work

it says it sent an invalid response

[worn quartz]

Free text to image ai?

[stoic brook]

https://docs.aihub.wtf/

this link doesn't work

[ionic pumice]

#📰│dev-updates message

[stoic brook]

thanks mate

[feral halo]

Ayo? @stoic brook level 1 !!!

[stoic brook]

how do I download an RVC from github

is there a guide for that

in this server

[orchid slate]

go for rvc help to #✨│ai-help , #🔍│help-w-okada

maybe you already find the answers to all your questions there if you scroll up

[little pelican]

what's your take on audiomodify guys?

[dapper drum]

Dose anyone even use hp6 anymore?

[tacit falcon]

Your device software cannot be checked for corruption. Lock the bootloader

[ripe raven]

how i can create my own voice model?

[limber mortar]

Hello all. New member here. Nice to meet you all. 👋

[crisp dragon]

ello Kira

nice to meet you as well

https://tenor.com/view/anime-girl-waves-anime-girl-anime-wave-gif-13842840944751638187

forgot no embeds XwX

[crisp dragon]

#🧬│ai-chat message you can check this one

[polar flax]

https://tenor.com/view/anya-embed-fail-gif-25814121

[crisp dragon]

AAAAAAAAAAAAAAAAAA

[tacit falcon]

https://tenor.com/view/fail-gif-10054675685696503316

[tired holly]

sup chat

[solar torrent]

Hi chat.

[honest drum]

hello

[pale patrol]

u can relate to that and its me :3

[fickle flame]

guys
I downloaded w okadada voice changer
it just shows
a
screen
thats letting me choose mmmvc13 rvc and thing
did I download the wrong one

[harsh heron]

hello

[reef lion]

.

[snow gulch]

is there a guide for mac user?

[covert lake]

for doing what?

[covert lake]

be sure u aren't following a yt tut btw

[snow gulch]

for rvc okada

cant seem to run it and almost no guide on youtube sorry man

my pc got bsod today 😭

[covert lake]

rvc is the program used for training (making ,in this case Speech to Speech, models) and for inference (use models) on pre-recorded audios

Wokada is the program that uses rvc models in realtime for voice changing for calls

99% of the guides for rvc or wokada are like a year old lol

I guess you are looking for wokada right?

[snow gulch]

wokada ?

suddenly i thought of wakanda 🤣

[covert lake]

yea i explained u the 2 different programs above that message

[snow gulch]

oh thanks

i only know how to run the settings on pc nvida but no idea at all on macbook m3

[covert lake]

yw, so you want to use wokada right?

[snow gulch]

yep

[covert lake]

-rt

[rare sorrel]

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

[covert lake]

1st link

[snow gulch]

gonna test how bad it is on mac

[covert lake]

its a guide to the wokada fork (modified version), has better performance

[snow gulch]

thanks nick

[covert lake]

and yes its on mac too

Yw

[snow gulch]

lets goo time become john cena again

[covert lake]

should be all good as people run it even on a GTX 1650, for any issues its better u ask in #🔍│help-w-okada

[snow gulch]

thank you nick 🫡

[feral halo]

Ayo? @snow gulch level 2 !!!

[snow gulch]

nick thank you

[covert lake]

yw

[twilit quail]

anyone on here running llama with vision locally?

[covert lake]

dunno, i mostly do cloud ai things in general as i got a bad pc, and here its mostly about RVC & Wokada

[twilit quail]

Ah, like wokada the voice changer?

[covert lake]

ye

[hard oar]

🤣 where is that guy who was talking abt audiomodify web for trump voice 🤣

hi chat

[covert lake]

btw, i could suggest u to use ollama

[twilit quail]

Ah yes, I was planning on using that as a client. I took one look at CLI interface and said no that's icky

[feral halo]

Ayo? @twilit quail level 1 !!!

[covert lake]

iirc ollama is the best for llms locally tho

https://www.reddit.com/r/Oobabooga/comments/1dmw6cu/why_is_ollama_so_much_faster/

[rich sable]

hey guys im DP

[twilit quail]

I just wanted to see if anyone had been using and what hardware they have been running it on if so. I think my system could handle, (3070 RTX, 64 GB, AMD Ryzen 5900X. I know it's VRAM that it really boils down to though. I was granted access to all their models but wanted to see which versions people had been utilizing commonly

[twilit quail]

you're double pen?

[polar flax]

hi DisplayPort 2.1 53 Gbps

[covert lake]

u should be totally able to run it localy, sorry but idk what else to suggest u other than ollama, i mean there's also the oobabooga text generation web ui but its slower from what i heard

[solar torrent]

Ok, dual port.

[twilit quail]

LOL DP on top

[solar torrent]

Don't worry.

[glad tree]

[blazing root]

FELLAS CALL ME THE DRINK

[elder willow]

@covert lake Is there a ai that can change the style of a song? from rick astley to the beatles?

[twilit quail]

Like this? LOL: https://www.youtube.com/shorts/-1aU72sImUk

[covert lake]

not really sure about that sorry

[ionic flower]

https://tenor.com/view/your-mom-spy-tf2-team-fortress2-sheesh-gif-21838783

[amber light]

@fiery crag #1159290752195633273

[wide lava]

ppl

[tulip sapphire]

#✨│ai-help

[lost arrow]

hello is flux best ai image gen now

[covert lake]

https://medium.com/@lompojeanolivier/flux-models-the-best-comparison-flux-schnell-vs-ideogram-vs-sd3-medium-68d1a5f19ad1, yes

[lost arrow]

i see but how can i use flux 1.1. I want to be able to use reference images

[covert lake]

locally or cloud (depends on ur pc gpu)? also ig ur looking for img2img

[lost arrow]

either one works i think

is img2img the one

actually im most likely gonna use cloud

[covert lake]

img2img : u give it a reference image with a prompt
text2img : just a prompt

[covert lake]

yea that depends if u got a bad pc

[lost arrow]

ye i got a pretty old pc

[lost arrow]

so where can i get img2img

[feral halo]

Ayo? @lost arrow level 1 !!!

[covert lake]

Well flux is a very resource demanding ai, the only ways i could find on cloud is https://replicate.com/bxclib2/flux_img2img

the thing is that flux doesn't have a diffusers img2img pipeline so the only way is with a comfyui workflow

[lost arrow]

alr

what would u need for flux to run locally

[covert lake]

for the dev version, more than 12gb vram

[lost arrow]

damn alr

[covert lake]

what's ur pc gpu, ram and vram?

[lost arrow]

ill check

I have 8 gb of ram , a 1060 and i think the 1060 has 6gb of vram

I have very old specs

i think it was 8 or 7 years ago i got this

[earnest dragon]

yeah 1060 only has 3gb or 6gb models

[earnest dragon]

which one the 3 or the 6

[lost arrow]

the 6 one

[earnest dragon]

did you check

[lost arrow]

Ye

virtual memory 6gb

[earnest dragon]

either way it's not enough reccomended minimum is 8gb vram and 16gb of ram

[lost arrow]

alr

[feral halo]

Ayo? @lost arrow level 2 !!!

[lost arrow]

This PC is getting old now

so ill get a better one in a few years

I did find that rubbrband does have free flux 1.1 pro

[covert lake]

yea nvm rip

[covert lake]

actually, there's one https://huggingface.co/docs/diffusers/main/en/api/pipelines/flux#diffusers.FluxImg2ImgPipeline

But idk why no one made a cloud version about this

[earnest dragon]

a few when it's already 8 years old by the time you upgrade it will be over a decade old and games nowadays are starting to reccomend 16gb ram minimum and 32gb ram reccomended

[covert lake]

not sure how much more resource consoming it would be

[lost arrow]

yes but it will be a big upgrade

[covert lake]

idk that site, but btw its flux.1, which there are 3 versions:

• pro, closed source, available only for paid
• dev, best open source model but also much resource demanding
• schnell, less resource consuming and open source, but less quality than dev

[earnest dragon]

there is glif which if you create a lora which is easy you can then get a artist account which gives you a good amount of daily credits to use

[lost arrow]

wait one min

its says flux dev, flux 1.1 pro and flux schnell

[covert lake]

i sometime use civitai, btw also weights.gg uses flux and lets u make loras

[covert lake]

1.1? iirc it should be flux.1, unless im wrong lol

nvm im wrong

[lost arrow]

I can show ss if you want

[lost arrow]

alr

[covert lake]

yea i was wrong, idk why only the pro version was named 1.1 instead of .1

[lost arrow]

anyways this model is incredible

I remember when Dalle 3 came out on bing

[covert lake]

yea its the best on for images atm

[covert lake]

its closed source only btw, flux is better

[lost arrow]

ye

but on release it was uncensored except for nsfw so u could legit generate celebs

however they soon put restrictions and made the model worse

[earnest dragon]

yeah but glif is the best for flux loras and stuff only thing that really is better is pulid for flux with ic-light

[covert lake]

btw for using flux for free u can use also weights.gg or civit.ai, just that i dont think those have img2img

[covert lake]

ic-light?

[lost arrow]

alr

[earnest dragon]

a example of ic light on a normal picture is this ic light basically add the lighting from the image you give it to the image and replaces the background with it

[covert lake]

mm interesting

[blissful bluff]

hi

[slow bridge]

hi! I’m a fan of img2img and sketch2img in realtime!

[light bone]

https://fal.ai

[covert lake]

yea that's an example of things id never use

[light bone]

lol what

[covert lake]

can't catch me ever spending money

[light bone]

its free

[covert lake]

i use my own kaggle notebook for flux, but in text2img

[covert lake]

oh it is?

i directly went to pricing and seen no free tier

[light bone]

yeah i think u have 1$ for start

so its free

[covert lake]

what u gonna do with 1 dollar

[light bone]

0.035$ for one megapixel

with lora

0.025$ for mega pixel without lora

[covert lake]

okay i actually didn't expect that but i still use all of those for free 1bluecookie

[covert lake]

ye u can do that for free too https://huggingface.co/spaces/multimodalart/flux-lora-the-explorer

[light bone]

but u have to wait 10 minutes

[covert lake]

u can still make ur own kaggle notebook and have 30 hours weekly for free

It takes just 1 or 2 mins per generation on flux.1-dev fp16

[light bone]

now what about training

[covert lake]

well..
Maybe can do it with lightning.ai (they give 15 credits for free monthly and have no restrictions)

Also maybe with kaggle T4x2 (32gb of vram in total)

And for surely can do it for free on civit.ai via credits, iirc on weights.gg too

my best friends are:

• 0$
• piracy

[light bone]

u cant train with 2 gpus

maybe in kohya idk

[sturdy crown]

do someone have og enzo vocal

[elder willow]

What is yall opinion about jews

[covert lake]

[elder willow]

@covert lake How do you use lightling.ai for rvc?

can you make a guide?

[covert lake]

like voice models?

[covert lake]

there are 2 notebooks, but no guide about them

but, genuinely its just better to use kaggle

gives more gpu time than lightning.ai

lightning.ai gives around 22 hours monthly while kaggle gives 30 hours weekly

the reason why i like lightning.ai more sometimes its bc its less restrictive, like no need for code obfuscation, bc rvc, web uis, sd and other things aren't banned on it, unlike they are in kaggle/google colab

for example deepfaking is extremely banned in google colab, while lightning.ai allows u to do it

[elder willow]

where are the lightling.ai notebooks?

[covert lake]

-lightning

[rare sorrel]

Suggestions for @elder willow

⚡ Lightning AI

• Applio, by Shirou Lightning AI
• RVC-WebUI, by Shirou Lightning AI
• FaceFusion UI, by Nick088 Lightning AI
• 🆕 UVR5 UI, by Eddy Lightning AI
• 🆕 UVR5 NO UI, by Eddy Lightning AI

More commands: -colabs, -kaggle, -hf, -realtime, -rvc, -help

[elder willow]

can anyone make a guide for this?

[light bone]

tf is lighitng

[covert lake]

lightning.ai bro

https://lightning.ai/

not much gpu time, but genuinely u can run basically anything u want without restrictions, comfyui, webuis, etc

[covert lake]

id use it only in case a project is highly prohibited on kaggle/google colab, like deepfakes and sd web ui stuff

[light bone]

damn prices are crazy af for premium shit

[covert lake]

premium tier or gpu time hourly?

i use it just for free

also they have an interruptable shit which makes all price cost bby half

[light bone]

premium

[covert lake]

oh ye forgot it was 50 dollars

[tulip cedar]

Hi. Any idea where I could make a chatbot that is basically me that either learns while I talk with if or that I can feed my chats for free?

[earnest dragon]

you would need to find a lightweight enough one that can run locally and also then you would have to make a lora of yourself for it

[tulip cedar]

Yeah I guess infortunately I dont even know where to start lol

[slate snow]

realperson

[amber light]

realperson

[feral halo]

Ayo? @amber light level 57 !!!

[kind drift]

gotta question

[feral halo]

Ayo? @kind drift level 1 !!!

[amber light]

ask away

[plush hearth]

wtf why is my thing saying "trial" randomly

[amber light]

what thing

[plush hearth]

okanda

[feral halo]

Ayo? @plush hearth level 4 !!!

[plush hearth]

first time this has happened

when I speak randomly in the sentence you hear "trial" in the background

[amber light]

#🔍│help-w-okada ask in here, a helper will probably know whats causing it

i dont use okada

[plush hearth]

noone is online

[dull saffron]

How do I install a voice model

[haughty gyro]

AudioModify seems good for AI song covers

[tepid basin]

Install it where

Have you downloaded RVC yet?

[dull saffron]

No

I just joined

Last month

[tepid basin]

Are you trying to do realtime (convert your microphone for Discord/games), or do you want to convert audio files (for memes and edits)?

Also do you have a GPU? You can use online GPUs if you don't but its suckier

[dull saffron]

No like download it

[feral halo]

Ayo? @dull saffron level 1 !!!

[earnest dragon]

audiomodify is trash they charge and are just using rvc which you can use for free and better I feel like mentions of paid rvc websites should be not allowed

[dull saffron]

Why is this saying Ayo😭

[dull saffron]

Can someone help ? Whenever I push a link it says it could steal my personal data how to fix this?

@tepid basin

[tepid zinc]

hi

anyone know what voice changer i can use

[wind stratus]

its normal, thats discords safeguard, but links on this server are clean

[dull saffron]

Oh

[bright bolt]

hi yall im new so i gotta write sum so just ignore me 🙂

[earnest dragon]

-rt first link

[rare sorrel]

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

[lofty hornet]

🤟

[rich sable]

yeah there is an ai called suno and if you input the lyrics of a song and type in a style you can achieve a similar outcome

Guys I was just curious if anyone knows a good live celebrity voice changer thats free? Also if possible I'm on a mac so maybe also a live deepfake FOR mac

[feral halo]

Ayo? @rich sable level 1 !!!

[rich sable]

???

[earnest dragon]

live deepfake really isn't good on mac

[dapper drum]

is the default settings for BVE 2 okay?

[earnest dragon]

what's bve 2

[dapper drum]

x-minus.

[earnest dragon]

mvsep bs reformer is better

[dapper drum]

k

[mystic seal]

yo

[arctic eagle]

Who knows good website to create the Main image (character) but with difference scenes (background, poses etc) It seems no user friendly AI can do it? To create/keep constant character from one Main image?

[elder willow]

ПЕРЕЕЗД НАХУЙ

[broken raft]

Weights.gg has something like that. You can generate an image with the subject (or "character") with different scenes depending on which prompt you enter.

[shell heath]

This voice changer?

[broken raft]

No. This is not a voice changer. This is a software used to isolate vocals.

If you're looking for a voice changer, then you should use W-Okada.

[earnest dragon]

-rt first link

[rare sorrel]

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

[elder willow]

yo guys how do you get like those ai voicovers on meme videos and stuff

like i used to have an old program for it

got from somewhere i forgot but it was so fun to use but when i reset my pc i lowkey lost the name

[arctic eagle]

it can generate constant character from one reference image?

[feral halo]

Ayo? @arctic eagle level 1 !!!

[tired jasper]

This upscaler model was able to do it even when it was not trained for restoring texts at all. Imagine doing a finetune on it for texts. Even a model finetuned for this task got worse results

I'm quite impressed

[broken raft]

Weights.gg can generate a constant character from many reference images. Not one but many. Just make sure to enter the character's subject name in the prompt.

[elder willow]

guys someone give me a program i beg of you :sad

[feral halo]

Ayo? @elder willow level 1 !!!

[broken raft]

What program are you looking for?

[elder willow]

anything that can voice over videos or mp3s somethin for meme videos

[broken raft]

Wait. I think I know what you're talking about. You're looking to do an inference.

-hf

[rare sorrel]

<:huggingface:1179800228946268270> Hugginface Spaces

• UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by thestingerx Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• 🆕 FaceFusion UI, by Nick088 Huggingface Spaces

More commands: -audio, - colabs, -kaggle, -local, -overtraining, -realtime, -rvc, -help

[broken raft]

I'd suggest getting the audio file of the video and then isolating the vocals using MVSEP.com Make sure there are no reverbs and/or echoes from the vocals. Then go to the Ilaria RVC Zero Huggingface Space, select a voice model, upload the vocals and then generate. Next, you'd need to mix the A.I. generated vocals with the audio from the original video and then you're done.

[elder willow]

thank you!

[vivid veldt]

hi

[arctic eagle]

hm, so if I got one good reference image, pretty much it will not generate constant character?

[vivid veldt]

@arctic eagle hi samosa

😳

[broken raft]

Oh. Wait. I think I misunderstood what you were saying. Sounds like you wanted to train an image model. I think you need to have a minimum of 5 images to a maximum of 20. You gotta make sure the images for the constant character you're wanting to train to generate are high quality.

[polar flax]

tbh you can train a LoRA from even a single reference image, meanwhile tons of images might be preferred

[signal cairn]

welcome to spooky month everyone

[elder willow]

hi

[spark perch]

I was looking for the server, I kept hunting until I found it. Yes, the logo is in a different color. I thought I was banned 😂

[solar torrent]

Even if they changed the server profile pictire, it's still AI Hub.

[spark perch]

It's just that our mind already sees the server as the color blue

[north grail]

Hey all

Anyone participating in DevFest?

[feral halo]

Ayo? @north grail level 1 !!!

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

[frozen egret]

hey

[north grail]

Wassup

[frozen egret]

man just woke up at 1pm today lmao

[neat kayak]

anyone wanna play minecraft?

[chrome hedge]

Question: Does anyone know any sort of Video AI that you download into your PC to host like Tavern AI? Also, AMD compatible would be highly appreciated.

[frozen egret]

idk

[feral halo]

Ayo? @frozen egret level 3 !!!

[covert lake]

Like text to video? Maybe cogvideox 5b https://huggingface.co/THUDM/CogVideoX-5b#quick-start-🤗 idk if u can run it in comfyui or smt tho

[chrome hedge]

Was hoping something like Image to Video.

But this will work too, thanks

[covert lake]

I don't think there's an open source maybe, or maybe u could try see if there's a cogvideo variant

[slow nexus]

Hii everyone ,

I am working on the AI solution for the Indian legal domain for question answering and knowledgegraph and using llama3.1 8b models .

but due to less knowledge about indian legal it not provide the accurate answer , so thats why i want to finetune models ,
Could you please tell me guide me how to fine tune the llm models what methods could be use what types of dataset .
And which llm models is better to finetune for this usecase ????

[covert lake]

DevFest?

[north grail]

Yes

[solar torrent]

I'm not a developer. I can only go to UserFest.

[covert lake]

oh the google one, yea most people here ain’t devs lol

seems cool tho

[north grail]

I see

[north grail]

Yep

Contributing to open source is sure cool

[solar torrent]

OpenAI, but some of their products aren't open source.

[polar flax]

what kind of projects, first?

[covert lake]

”some”

[covert lake]

yea for sure

[solar torrent]

What I can do, for example RVC and A1111, is to edit some of their codebase to work.

[covert lake]

RVC is abbandoned tho

there has been like no improvement in quality

and i feel like for other 5 years we aint gonna see anything better, just like how it happened with DFL in terms of quality

[solar torrent]

Oh, so that's why the last update for RVC was October last year. Man, I feel like I'm stuck in 2023.

[dapper drum]

rvc 3 when

[covert lake]

never, rvc boss is working only on gpt so vits

[solar torrent]

That won't happen.

[clear phoenix]

Do the mics have a huge affect for speech to speech?

[polar flax]

ideally if there's someone successfully implements something better than hifigan, i.e. evagan or bigvgan-v2
currently there are some ppl trying different optimizers and some config params #🔊│ai-development message

[covert lake]

i mean, forks like applio get updated, but it doesnt change quality

its just new features like TTS, but nothing that makes models quality better

[solar torrent]

If your microphone is noisy, it sure affects the audio.

[covert lake]

and the rvc hype died

[covert lake]

hopefully

[clear phoenix]

got any recommendation on a good mic? I kinda just use a razor wireless headset so far

[feral halo]

Ayo? @clear phoenix level 1 !!!

[solar torrent]

I have some models saved on my hard drive, alongside with Stable Diffusion models.

[solar torrent]

Not really pro at selecting a microphone.

[covert lake]

we don’t really do LLMs here, but maybe you could check the huggingface diffusers docs https://huggingface.co/docs/transformers/index

[clear phoenix]

dang, cause i bought a new PC with a 4070 TI Super and stuff but my audio sstill cuts off or sounds liek crap. im using okada

[sturdy crown]

ÿes if u have send pv

[covert lake]

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-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.ai-hub.wtf/essentials/how-to-make-voice-models/

[sturdy crown]

ok

[solar torrent]

LLAMA, but can only output 4 letters in a minute.

[covert lake]

llama 3.1?

[solar torrent]

I don't use LLM, but I assume if my laptop would run that.

[covert lake]

whats ur laptop specs?

maybe u could run the new llama 3.2 1b and 3b models locally, i run the 3b locally on my phone and its 5 tokens per second

[covert lake]

how tf dont u got even integrated graphics 😭

the cpu should be the same as my laptop lol

i mean u can juse cloud, or if u want run small models

[solar torrent]

This integrated Intel GPU is too old to show up on Task Manager. This laptop even has a space for Nvidia GPU but it is missing.

[feral halo]

Ayo? @solar torrent level 7 !!!

[covert lake]

😭

[solar torrent]

Too sad.

[graceful dagger]

where do i find epochs in the client??

[radiant nest]

You know? I wish TTS further more characters for songs

[covert lake]

are you talking about wokada realtime voice changer for calls?

you don’t need that, epoch is a unit of measurement of the training ai cycle, its just an info about the model training, and no, more of them don’t mean quality, just look at https://docs.ai-hub.wtf/rvc/resources/epochs-tensorboard/

[covert lake]

Use #1159290752195633273 , Or else i will delete everywhere else u send that promo

[rough arrow]

?

[covert lake]

promos can be exclusively only in #1159290752195633273

[rough arrow]

what

ohh my bad

[covert lake]

its fine, maybe needed to word it better lol

[tight cypress]

Damn

i3 275M is like 11 years old

[sharp pike]

hello

[tight cypress]

Unless you are down poor, i3 12100 is only 122$ MSRP

And it is two years old

[sweet python]

bro

Can the Lenovo Thinkpad T400 use this voice changer?

[covert lake]

this?

hell no how tf are u even using chrome on a 2gb ram laptop 😭

[polar flax]

win xp era laptop 💀

[tight cypress]

You probably wont able to play 95% of modern games

Like really it is just for office uses these days

[covert lake]

you are even lucky if it boots up

[sweet python]

4gb bro

[tight cypress]

That a little better

[sweet python]

yo bro im got trojan

[covert lake]

Anything under 8gb of ram is shit

[covert lake]

what??

[tight cypress]

I wont suggest 12GB either

[covert lake]

dont even try to run it on the pc

its too weak

[tacit falcon]

128 gb minimum

[covert lake]

You can’t do it locally (on ur pc)

u can try cloud (remote good pc) tho

[tight cypress]

Workstation investment 👍

[covert lake]

For Realtime Voice Changing for Calls on Cloud (remote good pc for those who don't have a good one, YOU CANT DO THIS ON MOBILE):

• Google Colabs (4 hours daily of free T4 gpu, easy to use, require only a google account) :
  • How to use Original W-Okada's Voice Changer Google Colab
  • Modified W-Okada's Voice Changer Google Colab
• Kaggles (30 hours weekly of better GPUs, T4x2 & P100, harder to use, requires an account and a phone number)
  • Original W-Okada's Voice Changer Kaggle
  • Modified W-Okada's Voice Changer Kaggle

[tacit falcon]

its gonna start being the norm in the future

[covert lake]

this is ur only way

[tacit falcon]

probably

[sweet python]

321$ get what laptop?

[tight cypress]

300 dollars... Oh yeah i am outside North America so sorry I cannot help

[polar flax]

most likely 96 GB DDR5 and that's plenty already

[covert lake]

desktop pc are better and more powerful, and thats kinda low

ud need atleast a GTX 1650 gpu

[tight cypress]

I mean...

[sweet python]

Even though my laptop isn't good

[tight cypress]

Not everyone want to sit at their spot and run AI all days. They got lives

There's a reason for Lunar Lake to appear you know

[sweet python]

but when playing GTA multiplayer it's good to get 100fps

[polar flax]

wokada run under lunar lake NPU when

[covert lake]

u need a powerful pc for local ai

u can’t do much about that other than using the limited cloud or buying a new pc

[covert lake]

gta 5 is a decade old

[sweet python]

I need advice bro

[covert lake]

use #1159290752195633273

[sweet python]

I'm not promoting, I just need creativity and ideas

[tacit falcon]

what is happening bro

[feral halo]

Ayo? @tacit falcon level 12 !!!

[polar flax]

you may consider getting an ROG ally or steam deck

[tight cypress]

As far as I knew there already efforts to support NPU by llama.cpp devs

[feral halo]

Ayo? @tight cypress level 17 !!!

[tight cypress]

And as well run Image Generation models

[covert lake]

it would be better u use that channel

[tight cypress]

Btw it is primarily centralised at Intel NPU rather than AMD's

[covert lake]

are there github projects that support the NPU

[crisp dragon]

budget mics?

[tight cypress]

https://github.com/openvinotoolkit/stable-diffusion-webui

[polar flax]

any decent $100/less headset mic

[tight cypress]

As far as I can tell OpenVino already have toolkit for NPU

[covert lake]

oh thats good

what about comfyui?

[tight cypress]

As long as ComfyUI have OpenVino backend

Which it dont have

[covert lake]

🔥

[tacit falcon]

https://seelectronics.com/products/v7/
https://www.audio-technica.com/en-us/at2020
He he he he
if nothing else try a cheap ass fifine mic
https://fifinemicrophone.com/

[elder willow]

What is "decent" and what's your usecase? For me for example Jabra Evolve2 (wired version) works great as a budget headset mic for meetings and calls, but I couldn't play in it for hours, it's uncomfortable.

[polar flax]

as long it's usable, imo hyperx cloud II is good enough

[elder willow]

Looks much better indeed.

[deft kiln]

Wheres the rvc collab thing that I use to make the ai songs

[covert lake]

its better to use Ilaria RVC Zero

its a ZeroGPU (A100 faster than google colab T4), HuggingFace (biggest ai platform) Space

[stray merlin]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

Don't forget to thank Ilaria if you find it useful! 💖

[covert lake]

you can read the guide above

if you really dont want to use it and want colab instead even if its slower, there’s applio colab https://docs.ai-hub.wtf/rvc/cloud/applio-colab/

[deft kiln]

I forgot how you make them can someone help

[covert lake]

read the guide, its explained there

[deft kiln]

https://huggingface.co/spaces/TheStinger/Ilaria_RVC)

[feral halo]

Ayo? @deft kiln level 5 !!!

[covert lake]

yes, thats the link of the space

read the guide above on how to use it

be sure you give it clean vocals of the song

[deft kiln]

Link doesn't work

[covert lake]

it does

Click -> https://huggingface.co/spaces/TheStinger/Ilaria_RVC

[deft kiln]

"Sorry we can't find the page your looking for:

[covert lake]

Where are you from?

might be its geo restricted

Even if it would be super weird

[deft kiln]

OK works now maybe

how do I find voices?

[covert lake]

yea ig it was a wifi issues

[deft kiln]

W

[covert lake]

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-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.ai-hub.wtf/essentials/how-to-make-voice-models/

If you don’t remember how to clean the vocals, check https://docs.ai-hub.wtf/rvc/resources/vocal-isolation/

@deft kiln may i ask what’s ur pc gpu btw? If u want (and got a good one’ u could do it locally (on ur pc) without any limits

[deft kiln]

I'm on mobile 😭 lol

[covert lake]

it’s actually doable locally on ur phone cpu but would be slow asf lol

Just use Ilaria RVC Zero then, remember to separate the song vocals and instrumentals, clean the vocals, and use a good quality model

[deft kiln]

W

[covert lake]

yw

if u run into any issues, u can check the troubleshooting part in the guide

or ask in #✨│ai-help

[deft oyster]

@sudden coral u can update your ag eternal sunshine model with new acapellas from live versions?

[deft kiln]

@covert lake

I am still confused

How do I change the voice to someone else's ai voice

on that link u sent me

[covert lake]

did you read the guide i sent?

[deft kiln]

Yeah but still confused 😕

[covert lake]

there’s the link to the guide here

[covert lake]

on which part?

[deft kiln]

turning the voice into the ai

[covert lake]

so, everything?

Okay, first of all, did u separate the vocals and instrumentals of the song? Or are u not doing an ai cover?

[deft kiln]

I separated it

using the vocal remover website which separates the sound and vocal

[covert lake]

ig ur talking about mvsep

anyways, after doing that

go to the ilaria rvc, and then to the model loader tab, and download whatever model u want

[tropic jay]

how to access voice models?

[deft kiln]

#1175430844685484042

[deft kiln]

How? (sorry for the overload of questions lol)

[covert lake]

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-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.ai-hub.wtf/essentials/how-to-make-voice-models/

[deft kiln]

Wait a sec I might have found

[tropic jay]

thanks!

[covert lake]

in the 2nd tab

u need to use either model downloader or uploader based on where you found the model

[deft kiln]

wait

[covert lake]

Model downloader for basically all the models in the server, huggingface links, etc

Upload a model for models found in https://weights.gg

[deft kiln]

whenevwr I pressed download after putting in the link and name of the model, and now it gave me 3 links;
Model pth file
Index pth file

whenevwr I pressed download after putting in the link and name of the model, and now it gave me 2 links;
Model pth file
Index pth file

2 links, not 3.

[covert lake]

what’s ur pc gpu?

[covert lake]

it shouldnt give u link, u should just put the download link, it tells u it downloaded the pth and added index, and then ur done. Can u show the screenshot in #✨│ai-help ?

You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU

[mossy shale]

anybody able to create X voice memo AI?

we got a lot of em already but not the one im lookin for 😞

[covert lake]

are u lookin for a voice model

[mossy shale]

i mean i already looked into the models available on the server but it doesn't have the specific one im looking for

[covert lake]

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-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.ai-hub.wtf/essentials/how-to-make-voice-models/

[mossy shale]

oh alr

i'll look into em thanks

[covert lake]

yw

[mossy shale]

You actually helped. No wonder why they wanted you for the mod. Thank you once again 🥳

[feral halo]

Ayo? @mossy shale level 1 !!!

[covert lake]

Ay, your welcome!

[craggy vine]

So sigma

L skibidi

[gilded siren]

What free tool to create a voice model (epoch, index...)

[feral halo]

Ayo? @gilded siren level 1 !!!

[covert lake]

Epoch is a unit of measurement of the ai training cycle

Pth is the file containing the voice

Index is the file containing the accent

RVC is the best program for Speech To Speech AI models

What's your PC GPU?

[heavy zephyr]

Hello, what is the best application to make a voice model?

[rare sorrel]

A model is the result of training on a dataset. You can learn more about it in the Applio Docs

[covert lake]

RVC

What's your PC GPU?

[heavy zephyr]

rtx3060ti 8gb

[covert lake]

@turbid mulch it auto detects some words?

Btw why is it only the applio docs instead of the ai hub one

[covert lake]

Alright should be able to, one sec

[covert lake]

As you got a good PC, you can use RVC locally, you can choose between:

• Applio: A fork of RVC with some extra features like Applio TTS, same quality tho
• Mainline: The original RVC

@golden brook I deleted your "suggestion" in #1159516963014451302 , for requesting models use #1159289738314919936

[turbid mulch]

I'm waiting the docs link gets fixed or it will be the one you deployed to vercel?

[covert lake]

Yui got a temp domain #📰│dev-updates message , IG it will be broken for a while but idk u should ask Yui when it's gonna get back

It's been more than a week so idk

[balmy mountain]

hi, this voice haven't support AMD yet huh?
mine is AMD RX6700XT 12GB and the voice come out 10-15 seconds after I say something

[north sphinx]

can someone drop good voice model training colab

[polar flax]

you should use fork wokada https://rentry.co/forkvoicechangerguide

[balmy mountain]

and I can use any voice from this server voice mod on this voice changer?

[polar flax]

read the guide carefully before asking

[oblique siren]

does kaggle kick you out if you're afk ?

or is it more leniant than colab

[covert lake]

As you dont got a good PC, its better you use cloud (remote good pc) 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):
  • Mainline (UI, No guide as of right now)
  • Applio (UI, No guide as of right now)

Colab = ez but low GPU time
Kaggle = harder but higher GPU time

WAY MORE SUGGESTED TO USE KAGGLE

[covert lake]

It shouldn't, but watch out cus there could be a warning after like an hour, sometimes happened to me but sometimes not so

[covert lake]

I swear the average user here hates to read

[tacit falcon]

no wrinkles

[oblique siren]

is it normal that it takes 5 minutes to train the index on kaggle

i don't get it

[timid musk]

how do you make an ai cover

@elder willow

[modern reef]

помогите пж

[sudden coral]

was thinking of making a new model w her new live performances 🙂

[lost arrow]

is chatgpt still the best chatbot

[arctic eagle]

great, can share link or full name of the site?

[cosmic coral]

https://dreamforgestudio.itch.io/the-haunting-house

do check it out

new horror game

[blissful bluff]

hi
guys
I need help
stable-audio-tools-main
I have downloaded both files.
Stable Audio Tools
Stable Audio Open 1.0 Model
So what should I do next?
I clicked setup.
Nothing happened
For the record, I have no technical experience in this field.
please help me

[feral halo]

Ayo? @blissful bluff level 1 !!!

[wispy widget]

ai hub on fire

[shy kraken]

hello yall

[deft oyster]

thx 🙏

[unkempt bluff]

https://www.youtube.com/watch?v=CitBvShJZG0

[last current]

hey , where / how can u find English Voice Models?

[covert lake]

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-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.ai-hub.wtf/essentials/how-to-make-voice-models/

[winter furnace]

hey can anyone help me with this one?

i got[Voice Changer] wait for loading model... -1 -1
[Voice Changer] [Exception] 'No model for model_common loaded. Please confirm the model uploaded.'
[Voice Changer] wait for loading model... -1 -1

[floral abyss]

hi guys i need a female ai voice can someone givve ?

[ionic pumice]

[wintry delta]

hi

[modern latch]

hi

[stone forge]

hi come to voice

[earnest dragon]

no not helping a guy with wanting to use a girl voice.

[floral abyss]

dude and my voice is lagging ?

[floral abyss]

@polar flax

[earnest dragon]

okay did you use the reccomended settings and also what gpu

[floral abyss]

rx580 (8gb)

[feral halo]

Ayo? @floral abyss level 1 !!!

[polar flax]

I'd recommend fork wokada https://rentry.co/forkvoicechangerguide

[floral abyss]

can you guide how do i install ?

[floral abyss]

i have installed it how do i open it now ?

its giving failed smth in cmd

you installed which one

[floral abyss]

this ?

don't know

staff mem could help

[earnest dragon]

android and iPhone will never happen because virtual audio cables are not a thing that works on them

[covert lake]

never gonna happen for sure

no one is gonna compile it into an .ipa or .apk

i installed rvc locally on my phone via termux (linux terminal emulator), no one would ever compile it and rvc hype is ded

[feral halo]

Ayo? @elder willow level 1 !!!

[covert lake]

surely never

its been like a year people waiting for rvc v3

there has been no improvement in rvc and the creator rather work on gpt so vits instead

i feel like rvc will be like DeepFace Lab, no improvement and none beat its quality in 5 years

voice mod doesn't have a realtime voice changer for phone

only on pre-recorded audios

yea and ur asking for a mobile version, while there is no realtime voice changer on phone

if u could get even just a laptop u could do it in cloud even if its bad

u could do it on cloud

cloud = remote good pc

locally = runs on ur pc

tho locally is better cus cloud has gpu limits as gpu is MUCH expensive, but i think 30 hours weekly are enough for free lol

For Realtime Voice Changing for Calls on Cloud (remote good pc for those who don't have a good one, YOU CANT DO THIS ON MOBILE):

• Google Colabs (4 hours daily of free T4 gpu, easy to use, require only a google account) :
  • How to use Original W-Okada's Voice Changer Google Colab
  • Modified W-Okada's Voice Changer Google Colab
• Kaggles (30 hours weekly of better GPUs, T4x2 & P100, harder to use, requires an account and a phone number)
  • Original W-Okada's Voice Changer Kaggle
  • Modified W-Okada's Voice Changer Kaggle

whats ur pc gpu?

You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU

i mean, ud be good enough even with a GTX 1650

if u really care about quality, get a desktop

laptop gpus are weaker than desktop ones

rip then

I mean for now u could either try to do it locally even on cpu mode even if idk if that's good

Or use cloud

ohh understandable

[covert lake]

i mean, those are ur only 2 options as of rn

-rt

[rare sorrel]

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

[covert lake]

1st link

its the fork (modified version) which has better performance

its the first link, which is https://rentry.co/ForkVoiceChangerGuide

[elder willow]

Hey I am looking for an Ai that makes songs where I can cover We found love by Rhinna but I can type in my own vocals?

[earnest dragon]

Paid Suno ai

[meager glen]

bye chat

😔

[elder willow]

Hello!
Web3 platform is expanding our team and looking for: developers, moderators, beta testers, analysts, and designers. The salary depends on the role.
You can write without experience. I'm waiting for you in DM.

Hey I am looking for an Ai that makes songs where I can cover We found love by Rhinna but I can type in my own vocals?

[feral halo]

Ayo? @elder willow level 1 !!!

[pseudo eagle]

i need help im using audacity to try to remove certain parts of the vocals so i can do an ai cover the song im trying to do is I got 5 on it by Luniz im trying to remove the like the chorus and the verses im trying to have 2 voices sing

the voices i want to use are cranky kong and donkey kong

i know someone did a cover of hank hill and bobby hill singing that song so i want to something like that

i have UVR and RVC as well as audacity

[wheat burrow]

orange ai hub 😰

[plush hearth]

does anyone have any ideas of when some quality beatrice v2 models will come out

[vestal topaz]

Anything better than wave 2 lip out there currently? Also any working google colabs or hugging face? all the ones Ive tried seem broken

[dapper solar]

hawk tuah

[earnest dragon]

https://tenor.com/view/itsascam-scam-meme-itsafraud-itsahoax-gif-23353011

[earnest dragon]

I already told you

[earnest dragon]

Yes there is if you want a video showing off one of the best zero shot lipsync have used I can send the video here

[nocturne latch]

https://tenor.com/view/discord-discord-perms-discord-gif-discord-permissions-no-perms-gif-19575106

[tacit falcon]

https://tenor.com/view/discord-image-perms-flex-image-flexx-gif-20201451

[hybrid goblet]

Mt=y name is jahid

[feral halo]

Ayo? @elder willow level 1 !!!

[pure sentinel]

Where do i get a Dr. House AI voice? cant seem to find it anywhere

[gray rover]

That's the sad reality.
hifigan v3, bigvgan, evagan to be implemented, faiss2 but nahh, majority of people got " bored ", are too lazy to do anything or simply are overlooking technical aspects over the QoL or ui bullshit lol

[agile axle]

Yo chat, should I be flat in performance for AI model accents to perform the best, or lean towards the tone of said AI model (Like a deeper voice registry for a person with a deeper voice like Future's )

& by flat, I mean not dynamic in tone.

[copper apex]

yeah man get a desktop, a rtx 4090 laptop would run like a 3070 desktop

[earnest dragon]

no it runs like a 16gb 3090

[copper apex]

ima say 3080

[elder willow]

which one does duckus use?

[cunning otter]

when is the huak tuah podcasts new episode dropping

[feral halo]

Ayo? @cunning otter level 1 !!!

[ionic flower]

no

[earnest dragon]

nah more 3090 from what I have seen if you get one of the full power limit laptops not the lowered ones

[earnest dragon]

overrated

[copper apex]

Either way laptops perform less than desktops

[ionic flower]

so Cxsmo, hows life?

[earnest dragon]

yeah itx or mini itx with a portable monitor and keyboard and mouse is the way to go as long as you have a outlet to plug into and then Vesa mount monitor to side of case or with magnets

[earnest dragon]

good

[naive knoll]

hi

[crude agate]

guys wheres the tutorial?

[ionic pumice]

skibidi

[earnest dragon]

we don't need a ai history lesson here

[vestal gale]

Jesus Is King

[earnest dragon]

stop with the history lesson

[ionic pumice]

https://cdn.discordapp.com/attachments/1055090217687724053/1287381133520142426/heart-locket.gif

[ionic flower]

this dude...

[feral halo]

Ayo? @wintry glen level 2 !!!

[elder willow]

theres this bug i cant fix can someone help

[earnest dragon]

https://tenor.com/view/gordon-ramsey-get-out-gtfoh-leave-gif-16002844

[earnest dragon]

what bug

[elder willow]

basically i have no friends

[modern socket]

ima post this and then leave the server

By combining this data, de Vries calculates that by 2027 the AI sector could consume between 85 to 134 terawatt hours each year. That's about the same as the annual energy demand of de Vries' home country, the Netherlands.

[earnest dragon]

that is not a bug that just how you are

[amber light]

nah we good

don’t flood

[elder willow]

guess ill kill myself or something

[ionic flower]

is that yappathon user a ai?

[amber light]

nah

js tryna be annoying ig

[tiny bridge]

is there a rvc archive with user ratings

need to be able to sort the best ones

[earnest dragon]

both he is using ai stuff in his messages

[amber light]

bro said ima post this then leave the server, might as well take the leave right now

[earnest dragon]

it will always be the racist ones or the egirl ones so not acurate.

[amber light]

wait

fuck

sorry

sorry @tiny bridge

[tiny bridge]

np

[amber light]

thought the dude came back and bypassed the timeout lmao

got confused

[tiny bridge]

just here to try to achieve my dream of becoming an egirl

[earnest dragon]

https://tenor.com/view/double-dutch-sal-salvatore-salvatore-vulcano-get-stick-bugged-lol-gif-21338641

[thorny wren]

crazy

[fair knot]

https://media.discordapp.net/attachments/907925466462183434/1081052668866330694/attachment.gif

[ionic flower]

what a fail

[fervent sigil]

Yo can someone invite me to a guild dms

[ionic flower]

uh no

[river sigil]

alguien tienen la voz de max vertappen

[earnest dragon]

what do you mean by guild dms

[ionic flower]

you mean that red bull driver who is rigging f1?

[lyric narwhal]

how i use this?