🧬│ai-chat

[left bronze]

wtf is teto

[left bronze]

wtf 😭

that mf ugly as hell no disrespect bruh

1000 yard stare 😭

[wanton bloom]

anime is NOT ❌ real

[left bronze]

no offense guys

[covert lake]

bro what yall saying

[left bronze]

js anime is not real

[wanton bloom]

debate

[covert lake]

you can in #🔍│find-models or #1175430844685484042 or go to https://weights.gg, can also search it on https://huggingface.co/models but watch out as not all models arent for rvc

[cobalt dust]

Tanks man

*thanks

[left bronze]

yo what ur yt

[covert lake]

check my connections on my profile

[covert lake]

yw

[wanton bloom]

Nick088

[left bronze]

huh

[quartz roost]

same girl

[elder willow]

I postd in help but is anyone able to help me finish installing RVC? I'm extremely confused

[feral halo]

Ayo? @elder willow level 1 !!!

[left bronze]

do you guys have the jjk cast audio logs i need it for elevanlabs 😭

[left bronze]

same opinion

[quartz roost]

https://tenor.com/view/kasane-teto-teto-meme-fast-dont-care-gif-25221703

[left bronze]

i feel like im in a bottom set in class

[covert lake]

you can in #🔍│find-models or #1175430844685484042 or go to https://weights.gg, can also search it on https://huggingface.co/models but watch out as not all models arent for rvc

if the model isnt there you can request it in #1159289738314919936

[left bronze]

@covert lake

do you have voice logs that can be used for elevenlabs

of anime chcartors from jujutsu kaisen

[covert lake]

yww

[covert lake]

uhh you mean a dataset for him?

nope

[wintry sinew]

its a evil spirit

if you fall into the trap

then your done

[covert lake]

you become tetski

[left bronze]

ill never fall for that ugly ass mf

no offence

[wintry sinew]

[elder willow]

https://tenor.com/view/madison-r-cute-kitty-cute-cat-gif-25737639

https://tenor.com/view/cat-good-morning-cute-so-cute-run-gif-5560961944863499511

[tardy vector]

deepfake

[quartz roost]

https://youtu.be/6k0Yc6CtV8U?si=yKH5uGij6xXFUwgJ

[idle light]

D0tcommie released a new song

[lethal onyx]

Any MJ model Recormendations?

[idle light]

@forest quarry @elder willow what do you think?

[lethal onyx]

yeah pls

i need a good mj voice

[left bronze]

Stop sending me this mf

[feral halo]

Ayo? @left bronze level 4 !!!

[forest quarry]

my models

[forest quarry]

what era

[lethal onyx]

anything but ofw

im making a fake album to cure my MJ disease

[forest quarry]

eh ig the zafue models r pretty decent

thriller era n shit

oh wait those r rvc v1 models lmfao

ehhh honestly idrk new models 😭

[celest cargo]

we’re kinda out of the golden era for ai albums now unfortunately

[fallow coyote]

any stable diffusion programs i can download instead of using a website?

like for ai genorated images

[wintry sinew]

try asking in #🔍│help-ai-art

[fallow coyote]

cheers mate im blind

[elder willow]

Hey, i need a little help. Is there any site that you can make your AI voice model for covers and that like stuff, just asking

[signal cairn]

a storm fucked up my house

[tardy vector]

House insurance?

Is that a thing?

Face tracking

[signal cairn]

no

[tardy vector]

ouch

then what

[signal cairn]

why tho?

[tardy vector]

I wanna do fun 3d things to know how to do fun 3d things

[signal cairn]

isn't there software that already do that?

[tardy vector]

most people would do it like that]

and probably the industry

[celest cargo]

i’m sorry to hear that

[covert lake]

Rip u😭

[light bone]

shading

[signal cairn]

stuff is going back smoothly

[tardy vector]

shading view?

[covert lake]

There are free ways to do that on our docs https://aihubdocs.github.io/en

[signal cairn]

Just hope anything electrical doesn't get fucked

[idle light]

is dat you?

[light bone]

im not talking about shading view

[tardy vector]

yeah

[tardy vector]

Are you going to have to pay for the damages

or

[signal cairn]

Welcome to Brazil

[signal cairn]

Now use an anime girl

[tardy vector]

wat

[tardy vector]

what

[signal cairn]

dont you mean shadows?

[quartz roost]

You're in teto territory

[tardy vector]

shadows exist

[quartz roost]

What do you mean I just got several new albums and it's popping off more than ever

[signal cairn]

i dont see any

[wintry sinew]

exposed

[tardy vector]

thats shading view

[covert lake]

Pikachu face reveal?!

[signal cairn]

also try using a filter so the glasses don't look FHD and the camera quality is 720p

[light bone]

nah just glasses shader could be better

[signal cairn]

So that they blend with the camera footage

[tardy vector]

i went for crappy because the quality is crappy

[light bone]

what

[signal cairn]

Replied to the wrong one

[tardy vector]

its just reflective and seethrough

[light bone]

ye i can see it

[tardy vector]

is it good vfx? no
is it meh? i guess

idk its my first time doing this kind of thing

and i didnt wanna render for 8 years

[light bone]

i mean tracking aint that bad

its better than buildin blender tracking

[tardy vector]

built in blender tracking probably would have done better if I couldve attached it correctly

[light bone]

I tested with this plugin and with builtin and the result was practically the same

[raw perch]

Not that anyone does or will care, but I finally got Tensorboard to work on RunPod.

If interested, I can make a short guide + python script to make it easy.

[ionic rover]

thats cool

how much are you paying monthly?

[signal cairn]

What GPU does it have

[mighty timber]

Ill buy nitro to whoever helps me

[signal cairn]

And how much

[mighty timber]

Please

[signal cairn]

What do you need

[mighty timber]

I need to know how to run rvc / voice cloning on macbook

[signal cairn]

you can't, you will have to use colab

[mighty timber]

Can i use colab on mac

[signal cairn]

Yea you access it using your browser

[mighty timber]

How do i do that?

[signal cairn]

https://aihubdocs.github.io/en/

read the guides on RVC Disconnected

actually just follow the guides

if you want to train or make ai covers

[mighty timber]

Can i take someone elses voice and clone it

Using that tool?

Im trying to have a male voice for gaming purposes lol

[signal cairn]

Yea

you'll have to use w-okada on colab tho since Mac can't run it properly

[tardy vector]

meow

[mighty timber]

Ojj

[feral halo]

Ayo? @mighty timber level 5 !!!

[tardy vector]

ojj

[idle light]

hehe it makes me happy when people share the docs

[mighty timber]

Oh okay in checking it out rn

[idle light]

thanks m0dz

[mighty timber]

Sorry my keyboard spazmed

[raw perch]

they have 3090s 4090s and even A100's (80gb VRAM ones). It's a cloud-based service, so not free. Usually just run a 3090 (can do multiple GPUs too)

[mighty timber]

One q @signal cairn

Im on aihub tingy and it says i have to download rvc

But u said rvc doesnt work on mac?

[signal cairn]

use the colab one

[signal cairn]

these ones

[mighty timber]

Ohh okay

[signal cairn]

Also check DMs

sent u something

[mighty timber]

Kk

[ionic pumice]

[modest mountain]

How do i use the models

[raw perch]

@modest mountain Either https://aihubdocs.github.io/en/rvc/cloud/inference/ilaria-rvc/ or this https://aihubdocs.github.io/en/rvc/cloud/inference/aicovergen-no-ui/

[sly valve]

the google collab one is kinda buggy for me for a few days now

[left elm]

Isn’t that a kep1er album?

[unborn ermine]

.FM

[indigo pike]

Now playing - patchy319.eth

Oh Yeah

Foxygen • We Are the 21st Century Ambassadors of Peace & Magic

2845 total scrobbles

[tardy vector]

yes

[abstract ridge]

Burnt thick bread and unmelted hard bricks of cheese

Bon appetit

[tardy vector]

gordon ramsay knows how to do it

[celest cargo]

apple

[placid flame]

is colab safe to use again?

[feral halo]

Ayo? @placid flame level 1 !!!

[signal cairn]

only use UIs if you have the paid plan

otherwise use the noUI

[placid flame]

ok

[odd knot]

or if you want to use the inference

just use ilaria rvc

[elder willow]

dude if its 2 pth files what do i do for index

[odd knot]

but if you want to train, use either easy colab by rejekts or rvc v2 disconnected

[odd knot]

train an index

also the pth file should weight 50-70 mb

[elder willow]

hhow

[odd knot]

are you training on local?

or on colab?

[elder willow]

whats that

i js found the model

on weights

[odd knot]

oh

i thought you were training a model mb

[elder willow]

it gave me

2 pth files

[celest cargo]

@tardy vector

[tardy vector]

?

okay

[celest cargo]

delete :3

[olive stratus]

automod nonexistant

[celest cargo]

pretty much ig

[tardy vector]

cant filter every n word

people get creative

[idle light]

What about AICoverGen?

[raw perch]

@idle light I am going to make a .py sript / a written guide for RunPod. Tensorboard works. Would you be able to add it into the guide? https://aihubdocs.github.io/en/

[idle light]

Depends on various things. First, how many people on this server do you think are going to use it?

[raw perch]

I am not sure. You have Paperspace on there and mostly everyone (If not everyone) uses Colab.

I have it here. If you would like to add it.

[signal cairn]

kinda volatile tbh

[idle light]

Volatile?

[idle light]

Nah tbh

I feel like I'll waste time

Might be helpful in the future

[raw perch]

How would you waste time making an alternative solution? Just curious. Edit: Alright, cool could've just said no.

[gray mirage]

HAS ANYONE MADE THE "missile knows where it is" VOICE?

I just thought of it and it would be funny af

dude

everytime i try to select a voice, it says interaction failed

[signal cairn]

might get disconnected

[trail coral]

does kaggle have free gpu?

[olive stratus]

yes

[unborn ermine]

Pretty sure yea

[timber shard]

@dense current 📉

[elder willow]

https://www.youtube.com/watch?v=NwHIa6IpN-k

[river verge]

♫ Your defense is in trouble... ♫

[acoustic scaffold]

can anybody make a model of the mayor from bugbo and the red brother

[tight totem]

yall what happened I was here like a few months ago and there were alot more models?

[signal cairn]

server got nuked, but most can be found with search commands on #🔍│find-models btw

[tight totem]

alright thank you : )

[whole axle]

would W okada ever make its own site, and sell the voice changer?

if it had a affiliate program itd be such easy money to promote lol

[signal cairn]

no

w-okada is a open source software

they can't sell the voice changer

people like me sell the voice models tho

[umbral hazel]

i want all the mickey related voices dont ask

[elder willow]

erm their all in #1175430844685484042 🤓

[quick garden]

#✨│ai-help message can anyone help me with this?

[dusk leaf]

where did the guide channel go?

[signal cairn]

#1159949278270193734

also check

https://aihubdocs.github.io/en/

[mighty steppe]

Hi, i have a questions, are there other repositories like kaggle to search data, because i wanna search img refer to pneumonia in dogs, but in the most popular repositories i cant find.

[feral halo]

Ayo? @signal cairn level 55 !!!

[signal cairn]

no idea

[signal cairn]

@elder willow Ilaria RVC only convert audios

for good covers you need high quality models and clean vocals

[tacit shoal]

i havent used an ai thing in a while

whats the best method/program

rvc?

[tepid basalt]

Ya

Hello

[ripe hill]

Hi

[trail coral]

hi

[trail coral]

yeah

[peak dome]

do you have a link to it? we can see where else to add haha

[timid musk]

https://tenor.com/view/cat-cats-explosion-explodes-cat-explodes-gif-10131775219258044191

[raw perch]

I have to send files. I didn't make a website or anything.

[topaz grotto]

hey

[peak dome]

it is
difficult to keep tabs on literally dozens of diff progs 💀

[peak dome]

hmmmmmmn maybe github?

ig you could send, but we're generally wary of the whole "download random files off a public discord server"

[sharp agate]

yo guys does the ai thing work the same as it did before

with the collab thing

[raw perch]

it's cool. I literally do models for free. I can dm if you don't mind.

[peak dome]

Okie!

thank you!

[devout wagon]

Has anyone weights.gg model been blurred recently? I can't even change the image

nvm, it's happening to everyone

[lethal swan]

@fleet ingot

[fleet ingot]

@lethal swan

[wintry sinew]

[timid musk]

i love these guys

[wintry sinew]

Fr

I see em daily

[glad garnet]

unlimited vc goin crazy

hol up wrong chat

[daring turret]

hello

[abstract ridge]

Same

Surprisingly wholesome for this place

[plain plover]

im trying to install models but not working

Oh yeah sorry

[feral halo]

Ayo? @umbral tulip level 1 !!!

[stuck lava]

i just treid to train my first model. it sounds absolutly nothing like the person i treid to use

does anyone know why that happens and a fix or work around?

[feral halo]

Ayo? @stuck lava level 1 !!!

[dapper lintel]

can someone give me a good vocal and instrument remover?

[light bone]

Uvr

[dapper lintel]

a website, not a app

[light bone]

mvsep or xminus

[terse lotus]

beni bnla 31 kr

[plucky thicket]

bro ghost pinged 2 helpers

[covert lake]

see this https://aihubdocs.github.io/en/vocal-isolation--datasets/vocal-isolation/

theres the colab of uvr

[lean blaze]

wrong link

[light bone]

What

[lean blaze]

kurwa

[idle light]

@wintry sinew THEY KEEP DOING THIS

[wintry sinew]

Yea

Daily

[idle light]

THEY PING EACH OTHER JUST TO SAY THAT BURGER EMOJI

What's going on

[covert lake]

its not the wrong link wdym

[umbral tulip]

I have 26 minutes of clean, voice isolated audio. how do i know how many epochs to use? i have tensorboard but im just not sure what to try

[covert lake]

check this https://aihubdocs.github.io/en/rvc-resources/epochs-overtraining--tensorboard/

[stiff delta]

ok so here the point just after 1.5k is the best one
right?

can you help me nick

[umbral tulip]

i just looked over this but im still not sure what to do. like should i just not worry about it over training and just use the save points?

[covert lake]

yea worry about undertraining or overtraining

[covert lake]

yea the one near 1.5k steps

[stiff delta]

so i keep the model running

[covert lake]

no dont need to

it over trained after 1.5k steps

[stiff delta]

and now does the point before 5k is the best?

man what the hell is rvc 😭

[ancient swan]

a graph

[topaz grotto]

anyone got musical ear if yes could yall getting which inference quality is better

[stiff delta]

can you tell what am doing wrong : (

[ancient swan]

show it

[ancient swan]

what's wrong?

[stiff delta]

like which point is the best (at which i should stop training)

i have targeted 500 epochs

shall i wait for completion and then look for the lowest point?

or what 😭

[ancient swan]

iunno, just listen to different epochs, use whichever sounds the best

[stiff delta]

what do you prefer like a min target epoch count?

or what do you set it to when you are making a model

[ancient swan]

i stopped using those funny graphs 2 months ago, cus they literally do not help to find the best epoch

[topaz grotto]

save lowest epoch and wait for completion

[stiff delta]

the target epochs count

[topaz grotto]

and test both epoch

[stiff delta]

whar

then what do you use

[topaz grotto]

the last one and the lowest loss one

[ancient swan]

ears

[stiff delta]

teach me master

[ancient swan]

listen to different epochs and choose whichever one sounds the best

i usually save each 50 epoch and listen to 50, 100, 300, 400, 600 and 1000 epochs

[stiff delta]

oh

[ancient swan]

1 sounds slightly more stable, tho they both sound pretty noisy, i'd denoise the dataset and retrain it

[stiff delta]

also is there any way where i can find what epoch count was there by looking at this

[ancient swan]

where

[stiff delta]

by looking at this?

[unborn ermine]

,fm

[ancient swan]

ah, usually mainline/mangio tells you the step count in the model file name like e100s500

applio tells just the epochs count, so ig there's no way to tell which epoch at which stepcount on it

not that it matters tbh

[orchid wing]

Guys

[ancient swan]

yo

[orchid wing]

Yo

[stiff delta]

i do it like at

[feral halo]

Ayo? @stiff delta level 6 !!!

[stiff delta]

the graph is at

[orchid wing]

Can I ask you something?

[ancient swan]

sure

[stiff delta]

then do the maths and find the epochs 🤓

[ancient swan]

ye, i suppose

[orchid wing]

Okada voice changer not working for the android?

[stiff delta]

is okada available for android too?

[ancient swan]

no idea, never used it on android

[stiff delta]

wharrrr

[covert lake]

They are pretty much helfup tho

[ancient swan]

should work with colab theoretically

[ancient swan]

nah

[twin fractal]

Android arent even strong enough to run conversion let alone realtime conversion

[covert lake]

What if it overtrains or undertrains?

[ancient swan]

ears work better, at least in my case

[orchid wing]

IAM using okada voice changer in android but with some lag

[stiff delta]

wtf

[covert lake]

What if he doesn't have good ears?

[orchid wing]

??

[twin fractal]

What do you do with okada and no vac

[stiff delta]

[ancient swan]

depends on the dataset, if the dataset is dogshit, you'll just be able to hear the undertraining and overtraining, if the dataset is godlike, overtraining won't affect it too much

[ancient swan]

rip

[orchid wing]

So you don't know any voice changer for android?

[covert lake]

The w okada colab should work

[ancient swan]

eh, iunno, but i think some ru guy was able to launch it on android somehow

[covert lake]

There are different occasions tho

[ancient swan]

nah

[orchid wing]

Lagging and I can't use okada voice changer in discord or any other apps

[covert lake]

Ask in #1192011222023950368 or #🔍│help-w-okada ,i never used it on mobile sorry

[unborn ermine]

Android doesnt really let you do voice cable or anything

Maybe if you root but

[topaz grotto]

i had did level 2 filtration i see i gotta go level 3

[ancient swan]

dunno what that means, but sure, do it

[topaz grotto]

i mean there r some parts which i compromised so i have to remove them fully

except that everything is done

[open zealot]

-rvc

[rare sorrel]

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)

[ancient swan]

ah, reasonable

[topaz grotto]

i just need a good voice trainer i cant rely on someone cuz i have been replying on one guy a lot lol

[ancient swan]

a good voice trainer, u mean model maker?

[unborn ermine]

.fm

[indigo pike]

Now playing - patchy319.eth

Hope

John Frusciante • Curtains

2954 total scrobbles

[elder willow]

How would I use the ai to change my voice is it possible in BandLab?

[unborn ermine]

If your on PC and using the program based one, you can run it through virtual audio cable

Or if ur using collab, dowload the file and upload it to bandlab

[elder willow]

I got a phone is all lol

[signal cairn]

okada isn't build for okada anyway bud

[signal cairn]

dm me

[covert lake]

You wanna do ai covers?

[covert lake]

Btw there are no vac for mobile?

[opaque spire]

Is there a channel for stable diffusion

[covert lake]

#🔍│help-ai-art #🏙│ai-images , we also offer free dall e 3 if you wanna try that or you can just ask help or questions in the first channel, the 2nd one is just for showing the art

[signal cairn]

dont think so

you can't really choose a different microphone on android

[drifting totem]

https://tenor.com/view/walterwhite-walter-ballin-walterisballin-heisenberg-gif-26404773

[feral halo]

Ayo? @drifting totem level 1 !!!

[covert lake]

We should write on the colab wokada guide that it isnt for mobile then😭

[abstract ridge]

Local inference and training is best done on raspberry pi

[inner atlas]

[mighty meteor]

there any lil tecca voices?

[inner atlas]

Get out Japanese

[light bone]

what

[inner atlas]

Your distortion on YouTube is even worse than English’s.

[ancient swan]

[inner atlas]

Ay show speed Hello

[light bone]

what

[inner atlas]

Do you love cristiano ronaldo

SIU

[light bone]

bro is delusional

[inner atlas]

Do you love messi

[light bone]

https://tenor.com/view/sigma-gif-24293069

[inner atlas]

You're just a stupid Messi lover, now shut up and get out of here (I'm just a fan of ishowspeed I just want to become like him)

[wintry sinew]

Ligma

[abstract ridge]

My good sir

With all due respect

[light bone]

or ill just mute you if you continue to be toxic

[ancient swan]

виталик, время остановиться

[inner atlas]

._. окей

X_X

[abstract ridge]

What the fuck are you talking about

[inner atlas]

I went to die

for a whole month

[shrewd dawn]

Would it be possible with a external Mixer and USB capabilities?

[feral halo]

Ayo? @shrewd dawn level 1 !!!

[topaz grotto]

just a trainer

i dont need dataset maker

lol

[ancient swan]

ah

just use colab

[topaz grotto]

collab recommend?

there r whole lots of collabs

[ancient swan]

easygui by rejekts

[topaz grotto]

idk

[ancient swan]

you just upload the dataset and click funny grey arrows and it does magic

[young glacier]

https://tenor.com/view/doggy-golde-retriever-gif-6626272048167568250

[idle light]

Can you train a model in easygui and lmk if TB works?

It's not showing any graph

It's been like that since yesterday

Can you check pls?

[ancient swan]

I'm a bit busy rn

[frosty condor]

DAMNIT! THEY RUINED VOICEMYAI NOW!

The subscription virus got them!

Now it's terrible

[signal cairn]

no

[signal cairn]

welcome

they figured out gpus arent free

[modest pasture]

yo, anyone could help me config my ai voice changer ?

[unborn ermine]

^ pretty valid

Im not a fan of subscriptions, but ill pay for one that gives me value.

(idk what voicemaya is)

[covert lake]

voicemyai sucks anyways

[covert lake]

It sucks, just use rvc its all free

Check out more abt rvc here https://aihubdocs.github.io/en

[signal cairn]

i mean you either buy a 500$ GPU or pay 8$ to use their service

[unborn ermine]

fr

I mean, I pay $23/mo for chatgpt

I'd pay for amazon prime if my mom stopped paying for it

[covert lake]

I use bing

[frosty condor]

Let's just hope they won't get FakeYou and Weights next or I'll go insane...

[low sundial]

hello

[frosty condor]

So FakeYou and Weights are safe for now

[low sundial]

does anyone know where i can find niffty from hazbin hotel?

[covert lake]

You can try using the /search command from @atomic vector or @heady grove or the /find command from @fresh narwhal or @hidden grotto to check if someone already made this model. Alternatively, you can check the #1175430844685484042 channel or use https://weights.gg/, or you can also use https://huggingface.co/models but watch out bc there arent only ai voice models here

[low sundial]

thank you ❤️

[covert lake]

Btw i already found it https://discord.com/channels/1159260121998827560/1203544962906595328

Yw

[low sundial]

thank you so muchhhhh!

[covert lake]

You're welcome

[unique path]

me pica la polla

[covert lake]

#🌍│español and dont talk rude

[unique path]

my cock itches

[covert lake]

Dont talk like this.

[wintry sinew]

avarage general chat

bruh

no more

[slender elm]

bro wtf

[abstract ridge]

Never been happier with splitting the channels back

[wintry sinew]

just ping me if you see any trolls in general

[abstract ridge]

So once every 30 seconds

[unborn ermine]

mf couldnt even stick with what he said, had to delete it

[topaz grotto]

yo any dataset master here

i want assistance with checking if my dataset clean when i complete

it

[wintry sinew]

bet

[abstract ridge]

Nuh uh

[wintry sinew]

[abstract ridge]

Troll'd

[spark flax]

hey guys, does somebody know where i can generate ai voice - text to speach with voices like king leonidas, I tested open voice and the ai voice aint accurate

[quartz roost]

https://x.com/vocaloid_yamaha/status/1759533296533340605?s=20

ボーカロイド - VOCALOID6 (@vocaloid_yamaha) on X

[Update Info]
Today, VOCALOID6 Updater ver6.3.1a is Released. Applying this updater carries out the following updates:

• Added 2 new voice banks (MICHELLE, LUCAS)

▽More Detail https://t.co/bqnhWia626

Please enjoy your music creation with new voice.

#VOCALOID #ボカロ

Twitter · Feb 19, 2024, 11:00 AM

[quartz roost]

It was always terrible

[quartz roost]

君は実にバカだな

[signal cairn]

Baka mitai

[quartz roost]

https://youtu.be/D808VoCbDVo?si=wMjdhZS2nwwkZaB0

[wintry sinew]

[tardy vector]

ola

[wintry sinew]

hai

[spark flax]

hey guys, does somebody know where i can generate ai voice - text to speach with voices like king leonidas, I tested open voice and the ai voice aint accurate

[covert lake]

you want a tts? or tts with a voice model you want? or make an ai cover?

also dont spam it

[spark flax]

sorry

[covert lake]

its fine

[spark flax]

tts with a voice model my frient.

[covert lake]

online or locally?

[spark flax]

like to give the ai my files and from that files the ai is able to generate with my files voice speech in a tts variante

[feral halo]

Ayo? @spark flax level 1 !!!

[spark flax]

dont know what the difference is my friend

[lime solstice]

gugu gaga

[covert lake]

locally you do it on your pc, if its good, online you dont need a good pc and can do it on mobile

[covert lake]

so you mean you wanna make an ai voice right?

and then use it as a tts?

[spark flax]

than locally

[spark flax]

yes, i downloaded the voice of leonidas (300 movie ) and i want him to speak my text

[covert lake]

you can use any tts you want, personally i use ilaria tts which is an online fork of edge tts

https://huggingface.co/spaces/TheStinger/Ilaria_TTS

sorry but i dont know about a local tts, maybe you could try edge or bark tts

and then after you generate a tts of the same gender of the voice u wanna get, you can just convert it with rvc to the voice model you want

you can check out more about rvc here https://aihubdocs.github.io/en

[spark flax]

thank u friend, but how can i put my file into that

[covert lake]

you just generate the voice with ilaria tts

you cant just put your file

generate the voice with a premade voice

then use rvc to convert that voice to the voice model you want

but yea you cant directly generate tts with the just a file

[covert lake]

do you got a voice model of leonidas?

[spark flax]

yes

[covert lake]

i dont mean just an mp3

[spark flax]

no its an mp3

[covert lake]

do you got the .pth and added.index?

[covert lake]

that isnt a voice model, its just the voice

[spark flax]

how can i change my file to that

[covert lake]

you would need to train a voice model

but before

You can try using the /search command from @atomic vector or @heady grove or the /find command from @fresh narwhal or @hidden grotto to check if someone already made this model in #🔍│find-models . Alternatively, you can check the #1175430844685484042 channel or use https://weights.gg/, or you can also use https://huggingface.co/models but watch out bc there arent only ai voice models here

see if the model has been already made

if it hasnt

then you will have to train it

[spark flax]

when i finished the training, what is next step

[covert lake]

you can also check on how to train it here

[covert lake]

if you found the voice model, or trained it if there isnt one, you have to use ilaria tts select a premade model that is of the same language and gender of the voice, then download the tts done by it, and convert it using rvc to the voice

[spark flax]

what is the meaning of rvc

[covert lake]

Retrevial-Based Voice Cloning

its an ai tool, the same one used for making ai cover

[covert lake]

also, whats your pc gpu?

[spark flax]

2070

[covert lake]

2070 what? rtx?

[spark flax]

rtx

[covert lake]

then you should be able to do it locally yea

find the model first tho

and check the docs i sent you

if you need more help theres also #1192011222023950368 and #✨│ai-help where you can ask help

[spark flax]

you the best

[feral halo]

Ayo? @spark flax level 2 !!!

[spark flax]

thanks for helping

preciate

[covert lake]

You're welcome

[abstract ridge]

朽木不可雕也

[unborn ermine]

.fm

[indigo pike]

Now playing - patchy319.eth

Crawl Away

Tool • Undertow

2974 total scrobbles

[latent hamlet]

He

[left bronze]

put me in there n she wont be able to walk for 2 weeks straight

[unborn ermine]

[left bronze]

lightwork

[wintry sinew]

huh

[idle light]

?

[left bronze]

?

[rare sorrel]

🎙️ Realtime Voice Changer (Online)

(Online) W-Okada - Colab Version

A free, online version of W-Okada that allows you to run code on Google's powerful GPUs.

Link: How to use W-Okada's Voice Changer

Credits: Antasma245

[unborn ermine]

[elder willow]

I AM MAKING THE WORST BRAINROT

[gentle canopy]

hi

[covert lake]

Also nvm the ilaria TTS space is having problems, you could use the tts inside the ilaria rvc space, and you can also directly doing it online using the models if you want https://huggingface.co/spaces/TheStinger/Ilaria_RVC (duplicate it first)

[elder willow]

Does anyone have a donk voice from cs

[steep yarrow]

whats the voice changer link again

i lost it

[half pond]

Someone knows site of selling followers for instagram/ TikTok and all that shit?

[dapper trail]

?

I don't think the server is for this, but no I don't know any

[half pond]

Thanks anyway

[fading terrace]

voice tayler derden plz on russian

[idle light]

#1175430844685484042

[fading terrace]

voice tayler derden plz on russian i dont find him

[idle light]

#🔍│find-models

[signal cairn]

good luck getting shadowbanned

[wheat hazel]

selam türkçe kişilere nasıl ulaşabılırım ?

[signal cairn]

https://discord.com/channels/1159260121998827560/1159595680659157132

[wind cipher]

So question. Is there a surefire way to take one voice out of a chorus with a male and female singer?

[river portal]

wowie

[random ibex]

Voice models jungkook

[feral halo]

Ayo? @random ibex level 1 !!!

[quaint wharf]

does anyone have a good deep voice model

[unborn ermine]

The burger king foot lettuce one is pretty neat

i cant remember the guys name

[quaint wharf]

lmao

[unborn ermine]

But uh

[quaint wharf]

where that be

[unborn ermine]

With ur pfp, you should check out ProjectM music visualizer tbh

[quaint wharf]

wdym with my pfp?

[unborn ermine]

its swirly

[quaint wharf]

yeah what's the problem?

[unborn ermine]

idk, I have a copy of the model ill dm u a copy

[quaint wharf]

uh sure

[unborn ermine]

Nothing, ProjectM is cool

It reminds me of ProjectM music visualizer

[ivory warren]

guys does anybody know how to do the Gegagedigedagedago animation

[feral halo]

Ayo? @ivory warren level 1 !!!

[atomic vector]

@wheat hazel, I have found 1 results that match your search!

iboskom12

Fırat Albayram [Noluyo Ya?]

• Uploaded: <t:1692316800:d>
• Likes: 0

Epochs

400

Technology

RVC

Algorithm

Rmvpe

Requested by huper2296 · Feb 19, 2024, 6:37 PM

[analog copper]

Does anyone know how to replace a voice on a song on mobile

[unborn ermine]

-settings

[jade parrot]

hi, I need help

alr bot, got it

does rvc work fine with uhd graphics?

[elder willow]

how do i do realtime i wanna sound liek a man

[exotic torrent]

wheres the rvc model list

[unborn ermine]

Did a new rvc video release

so many new ppls

[jade parrot]

is it safe to use a rvc from any notebook?

[atomic vector]

@wheat hazel, I have found 2 results that match your search!

mustagfa

Ceyda Kasabalı [Noluyo Ya?]

• Uploaded: <t:1692144000:d>
• Likes: 0

Epochs

400

Technology

RVC

Algorithm

Rmvpe

Requested by huper2296 · Feb 19, 2024, 7:07 PM

[wintry sinew]

#🔍│find-models

[unborn ermine]

.fm

[indigo pike]

Now playing - patchy319.eth

Marmalade March

Psychedelic Porn Crumpets • High Visceral, Pt. 1

3003 total scrobbles

[unborn ermine]

@inland dagger my last fm is 90% drug music

[inland dagger]

[dapper drum]

dose anyone know how to make harmonies in fl studio?

[unborn ermine]

[elder willow]

Hmm

hello. when training a model with ov2super. do the other people using it have to also use ov2super

[idle light]

OV2Super is a pretrain. Only works for training

So yeah anyone can do inference with it

Unless they use a really really outdated version of RVC hehe

[elder willow]

i see

[harsh forum]

I hate perter

[idle light]

https://aihubdocs.github.io/en/ here's everything you need to know =3

@shy juniper please don't spam

[elder willow]

Can i get a working svc link

Ik svc is outdated but I wanna use the juice model

[idle light]

@wintry sinew @forest quarry I love this so much

😭

[wintry sinew]

-svc

[rare sorrel]

https://cdn.discordapp.com/attachments/1159822429137412177/1194008385113301003/svc.gif?ex=65c13ef6&is=65aec9f6&hm=ce1f57bc316793f641c80f9b359ba9db7d1312af6918b1aed9c8468c74035bac&

[unborn ermine]

Is this svc

[light bone]

yes

[unborn ermine]

Well it works rlly well

[left bronze]

guys

does anyone have audio logs of jjk chacrators like gojo and sukuna

that can be used for elevanlabs

like i aint looking for the voice ai

just the audio logs

[quartz roost]

We are using diffsinger svc now

[quartz roost]

Finally something good

[elder willow]

I need a colab link for svc that actually works

[forest quarry]

I've sent one for you, last time I checked it worked so idk... Maybe something broke down after 5 months

[elder willow]

I tried it. It wouldn't let me add a audio file

[forest quarry]

oh well

[elder willow]

The svc juice wrld model did a good job for high vocals

[vital venture]

Ayo i heard you can somehow "outsource" your gpu for companies to use for money, what do yall think of that?

[quartz roost]

It runs local

[quartz roost]

Terrible idea

[vital venture]

Oh shi, why

[unborn ermine]

I mean, thats prob more profitable than mining crypto on gpu now

[vital venture]

Im talkin undervolted and stuff

[elder willow]

I need a colab link because I'm using a Chromebook

[vital venture]

From what ive heard at least

[quartz roost]

If you're playing with ai you need the right tools

[unborn ermine]

But if ur doing that for profit, you need to be careful

[elder willow]

Yeah chrome browser and colab

[unborn ermine]

Plenty of people buy ASIC miners that are like 10,000$ and then they end up being unprofitable or the break even ends up being way longer. Specifically KASPA

[dapper trail]

Tell me about the pictures to do with the colab

The one adapted by me and hina

[vital venture]

I was talking abt just like a 4090 tho

Like 1

I heard it doesnt take too long to break even, but you can tell me in more Detail maybe if it does take quite a bit?

[arctic gulch]

hey guys, i'm looking for a billie eilish model that works, for some reason I keep getting this error: ModuleNotFoundError Traceback (most recent call last)
<ipython-input-5-b114748cdea6> in <cell line: 4>()
2 #@markdown Link the URL path to the model (Mega, Drive, etc.) and start the code
3
----> 4 from mega import Mega
5 import os
6 import shutil

ModuleNotFoundError: No module named 'mega'

[idle light]

Please ask in #✨│ai-help

This is gen chat

[arctic gulch]

got it, thank you

[elder willow]

Welp this sucks

[unborn ermine]

fr

[dapper trail]

We adapted a spectacular colab

[elder willow]

I can't find a working svc colab. The thing is ik it's outdated but I wanna use the svc juice model

For my high vocals

Rvc doesn't work well with my high vocals

[quartz roost]

https://youtu.be/NYdwXaPhx18?si=Bbxc0Sk6vJKMJtcC

That what tet is for

[elder willow]

google colab is bugging me rn

tryna download a train index file but it doesn’t show up in my downloads folder

so over for me

I just need a svc colab link

[vital venture]

Why u got obama pfp

weirdo

[quartz roost]

https://github.com/MoonInTheRiver/DiffSinger

[elder willow]

what about obama

[vital venture]

Smh he lookin kinda hairy in that pic

[elder willow]

Is it colab

[quartz roost]

This is a more advanced svc editor local only

[elder willow]

Then I can't use it

I'm using a Chromebook

[quartz roost]

Maghni ai aims to be more compatible and may be able to run on a chromebook

[elder willow]

I just need svc.

Sovits svc

Are you a bot?

[forest quarry]

they're gone

[elder willow]

Im getting really frustrated

[forest quarry]

yeah uh... idk what you can find tbh

sovits is gone for good sadly

[elder willow]

Rvc doesn't do good with my high vocals

[forest quarry]

could be the model, most likely

but eh

¯_(ツ)_/¯

[wintry lark]

just made a model of the new baldi's basics plus character (dr reflex, his voice clips just got released on the wiki)

his voice is goofy af

[idle light]

Did you try using mangio-crepe?

[elder willow]

I want svc. The svc juice model sounded realistic. Anyways all the good rvc juice models are private. So???

It's really frustrating

[elder willow]

Also I can't even train my own model cause my Chromebook

[idle light]

You can train models in any device with RVC Disconnected

https://aihubdocs.github.io/en/ learn more here

[elder willow]

I've tried it just disconnects me

Idk why

-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

🤗 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[elder willow]

can someone give me a disconnect prevention method for colab

[unborn ermine]

.fm

[indigo pike]

Now playing - patchy319.eth

The Last Hymn

John Frusciante • Going Inside

3040 total scrobbles

[unborn ermine]

rare frusciante

[elder willow]

im training rn

i tell yall how it goes

[signal cairn]

just leave the tab open lol

[elder willow]

it disconnects if i dont do anything

[signal cairn]

you will be disconnected after some hours

its not infinite

i would say 4~6 hours max

[elder willow]

i get disconnected after 30 mins

[signal cairn]

you have to watch out for the captcha

but wait

which colab are you using?

[elder willow]

rvc v2 disconnected

ima train 200 epochs at a time

atleast i will try

[valid sage]

can GTX 980M and 32 GB ram can hold voice changer in real time?

I can easily convert my recorded voice into the voice from the model thanks to RVC webUI, but what I care most about is being able to modify my voice in real time

[feral halo]

Ayo? @valid sage level 1 !!!

[compact crater]

how do you convert 40k model to 48k (and vice versa)

[elder willow]

can i train the same model after training it once

[alpine granite]

i can do that locally. Im pretty sure you can do that on collab too if it imports the log files

[elder willow]

im training 200 epochs at a time

[feral halo]

Ayo? @elder willow level 10 !!!

[elder willow]

oh shit my model is almost at 200 epochs

someone help me real quick

which one do i download

nvm

how do i train a model i just trained

@alpine granite

[umbral hazel]

guys where can i search models

[feral halo]

Ayo? @umbral hazel level 1 !!!

[umbral hazel]

guys in what foroum i can search models

[elder willow]

^

[wintry sinew]

#1175430844685484042 #🔍│find-models or https://weights.gg/

[elder willow]

do you know how i can train a model i just trained?

[umbral hazel]

thanks but first one doesnt allows me to search

[wintry sinew]

local or online?

[wintry sinew]

it should it might be a discord issue

[elder willow]

online

[umbral hazel]

mmhh

[paper estuary]

I tried to generate a 10 word sentence using one of those .pths in AI-voice-cloning and it's lagging my computer even after 5 minutes of generating while the tutorial video did it in 3.5 seconds,

specs:
5950x
64gb ddr4
rtx 4090

What could be the issue?

[wintry sinew]

go to the bottom of this https://aihubdocs.github.io/en/rvc/cloud/training/rvc-disconnected/

[elder willow]

aight

[quartz roost]

On the shelf between the keyboards and microphones

[umbral hazel]

🎵

[wintry sinew]

the teto propaganda is real

[quartz roost]

Why aren't you using teto

[dapper drum]

I hope the doc have the Advanced TB use

[elder willow]

What mics were there tho?

[wintry sinew]

[forest crypt]

blehhhjolly

[quartz roost]

Pretty nice selection ngl.

[elder willow]

I think you sent the wrong image lol

but nice guitars fr

[quartz roost]

I didn't take a picture of the microphones

[elder willow]

Well what was the name of the shop?

I could just look at their stock

???

can someone explain

[quartz roost]

Shimamuragakki

[wintry sinew]

you need to put the step of your last save

[rigid oyster]

why theres no horny voice models dude

[wintry sinew]

if you ticked the only save last cpkt think it should be 2333333

[abstract ridge]

https://tenor.com/view/pee-panicking-what-terrified-gif-15348578

[wintry sinew]

🤨

[abstract ridge]

Mods looking at this be like

[rigid oyster]

how tf am i gonna turn this zip file to link as model

[elder willow]

Juice WRLD_e200_s6000.pth so it would be 6000?

[abstract ridge]

Wha-

[wintry sinew]

yea if you didnt do this #🧬│ai-chat message

[quartz roost]

There is but I'm not telling you who or how

[tardy vector]

I love onion and garlic

[rigid oyster]

why

[feral halo]

Ayo? @rigid oyster level 1 !!!

[tardy vector]

@wintry sinew loves beans on toast

[wintry sinew]

use the 2333 number

[quartz roost]

Because I can already tell you don't respect voice actors.

[wintry sinew]

its good tho

dont hate

[tardy vector]

im not

[rigid oyster]

voice actors?

[tardy vector]

tea loving. fish n chips addicted

[quartz roost]

These are real peoples voices. They must be treated as such.

[rigid oyster]

what are you talking about dude

[quartz roost]

Every time you use a voice clone you must treat them as if they are an actor on a stage. With dignity

[elder willow]

do i tick this too

[rigid oyster]

man i just cummed

[wintry sinew]

tf

[rigid oyster]

gotta go

[wintry sinew]

nah you already did it before

[tardy vector]

this dude what

[elder willow]

is this good or bad

[odd knot]

you should check

loss/g/total

and/or loss/d/total

[abstract ridge]

• Joins server

• asks for horny voice model

• says they just cummed

[abstract ridge]

• leaves

[odd knot]

naaaaaaaaaah bro

what even is that message

you shouldn't have said that in a server

[abstract ridge]

Welcome to #🧬│ai-chat

[elder willow]

i have never trained a model before

[odd knot]

screenshot "loss/g/total" and "loss/g/total" graphjs

graphs*

[abstract ridge]

https://tenor.com/view/nsfw-discord-gif-21330281

[elder willow]

where is that?

[rigid oyster]

https://tenor.com/view/free-the-shawshank-redemption-freedom-raining-gif-13933748

[odd knot]

in tensorboard

[abstract ridge]

https://tenor.com/view/embed-fail-embed-perms-level-5-discord-embed-perms-at-level-5-gif-14188677204049426720

[odd knot]

did you close your colab already?

[odd knot]

go to scalers

then click on a search bar and type "loss/g/total"

[river verge]

https://tenor.com/view/nichijou-daiku-igo-igo-soccer-my-ordinary-life-gif-23602888

[odd knot]

uncheck "ignore outliers in chart scaling" and set smoothing to "0.9"

[elder willow]

aw shit my runtime ran out

what is d and g.pth

[odd knot]

i don't think those things are important

but from what i remember

d = discriminator

g = generator

[elder willow]

my runtime ran out

[odd knot]

but i don't think you need these to put these in a zip file

[odd knot]

at least you have saved the model, so it's ok

[elder willow]

idk how it sounds

[odd knot]

let me tell you how to hear how your model sounds

first, download the "juice wrld.pth" file and the "added_IVF740.index" file

[elder willow]

ok

done

[odd knot]

then zip them

create the huggingface account

[elder willow]

i dont want my model to be public

since idk how it sounds

https://tenor.com/view/prata-gamer-prata-games-sonysta-drake-sincero-gamer-sem-regras-gif-10515327006541893322

can i make the link unlisted?

@odd knot

[odd knot]

do you have the huggingface account

[elder willow]

yes

[odd knot]

click on your profile then "new model"

pick a random license (or pick "openrail"), make sure private is checked before creating a "model", since you don't want your model to be public

[candid violet]

Anyone know when google collab no ui will be back ip

[elder willow]

done

now what do i do?

[odd knot]

click on "files and versions"

then click on "add file" - "upload files"

[elder willow]

done all that

already

[odd knot]

check the zip to make sure you put the pth and index file in here

[dull venture]

[elder willow]

its 105MB

[odd knot]

ok

drag and drop or click on the box

find the zip file

[elder willow]

-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

🤗 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[odd knot]

then click on "commit changes to main"

[idle light]

😭😭😭😭?

[paper estuary]

Anyone know what causes such long generation times?

[elder willow]

index file is not popping up

in applio @odd knot

[odd knot]

click on the little triangle arrow

[elder willow]

i did

nothing

[versed loom]

everything is a horny voice model if you put your mind into it

[odd knot]

try refreshing the thing

[rigid oyster]

i got what you mean

and doing it

[elder willow]

oh my model sucks

OMG

LMFAOOOO

[signal cairn]

i have some

[elder willow]

@odd knot

[signal cairn]

[signal cairn]

first time?

[versed loom]

Sable's Juice Wrld model is pretty clean though im pretty sure its private

[elder willow]

yes

[elder willow]

yeah

it is

[signal cairn]

most singer models are private

at least for those you see on tiktok and stuff

[versed loom]

i dont watch tiktok so i wouldnt know

[elder willow]

yall wanna hear what my model sounds like

its bad

[versed loom]

but what does the input sound like

[rigid oyster]

what is it based on

[elder willow]

studio sessions

[signal cairn]

sure

[versed loom]

how many minutes

he got a ton of those

[odd knot]

yep

[odd knot]

as expected, so much noise

[signal cairn]

mfw dude didn't clean the vocals

[elder willow]

how do i fix it

[signal cairn]

clean the acapella

[odd knot]

denoise your dataset

[signal cairn]

that shit is nasty

[signal cairn]

not even the dataset

[versed loom]

how'd it get so much noise on studio sessions though

[elder willow]

idk

[signal cairn]

it shouldn't generate that noise when not talking

clean the vocals lol

[devout wagon]

studio sessions are still noisy

[elder willow]

how do i clean the dataset

[signal cairn]

-audio

[rare sorrel]

Suggestion for @elder willow

📚 Audio Guides

• Perfecting Audio Isolation on Low-End Rigs: A Practical Guide, by Litsa The Dancer and Faze Masta Google Docs
• Gathering and Isolating Audio, by SCRFilms Google Docs
• Creating Datasets for RVC using iZotope RX, by LUSBERT Rentry
• Vocal Mixing Tutorial, by Roomie YouTube

🛠️ Tools

Audio Separation/Isolation

• https://mvsep.com/
• https://multimedia.easeus.com/vocal-remover/
• MDX-NET Colab

More commands: -colab, -uvr, -karafan, -overtrain, -help

[elder willow]

so i gotta train all over again????

[signal cairn]

yea

welcome to rvc hell

[rigid oyster]

is there any channel i can send pics

[elder willow]

bruh

[versed loom]

It is what it is. You can either commission someone (inb4 shameless promotion) to do it or pay for sable's bot if your not satisfied but that includes money which is ass

[feral halo]

Ayo? @versed loom level 3 !!!

[signal cairn]

i think you should be able on #📑│making-models

[elder willow]

im broke

so thats not gonna happen

[versed loom]

thats why its ass

[signal cairn]

beware that only model masters take paid commissions

do not take or pay model makers as we cannot guarantee quality nor trust

[versed loom]

i dont like doing commissions

[signal cairn]

yea ik

just a quick notice if anyone wants to ask for some

[elder willow]

i only used 68 audio files to train it

[signal cairn]

amount is not quality

[rigid oyster]

can you look at dm

[feral halo]

Ayo? @rigid oyster level 2 !!!

[versed loom]

i heard sable's model was below 20 minutes dataset

so its possible

[elder willow]

how long does it take to denoise everything

[versed loom]

pretty quick in uvr

unless your gpu is ass

[elder willow]

i gotta do that to 68 files total...

[elder willow]

im using a chromebook

a samsung chromebook

how do i use this colab

[versed loom]

just merge all 68 files together and run it through mvsep or some colab that uses uvr models assuming it doesnt have a limit on length

[elder willow]

im using mdx net colab

how do i use it

[elder willow]

did you have the noise problem when you created your first model

[plain cove]

??

[elder willow]

???

[plain cove]

I Used my voice I had 5-7 Files that have 1 minute

And it works well for me

250 and 500 eph

[elder willow]

im training 200 epochs at a time