🧬│ai-chat

[polar flax]

yea finish it

[ionic flower]

https://tenor.com/view/filthy-frank-joji-japanese-singer-facepalm-gif-17490402

i was trying to say if its ok for THIS server's rules not yt rules

[earnest dragon]

no music is allowed here only #1159290752195633273 and it must be uploaded to YouTube

[polar flax]

you just better avoid attaching audio/videos that may be copyrighted, as it seems discord's policy regarding the mass bombing incident on our old server by RIAA exactly last year

[west finch]

Hey! I'm looking for some directions. I'm trying to get my text to speech with donald trumps voice. I'm trying to find a free version that doesn't have any limitations on characters. is there anything out there atm? thanks in advance

[earnest dragon]

I know one that is free that but only let's you do 250 at a time but as many times

[west finch]

which one?

[earnest dragon]

I can send the link to the voice

[earnest dragon]

@west finch https://talkingavatar.ai/#/voice?id=12526723644780037

[ruby vessel]

Is there any difference in quality between Ilaria and Applio?

I see people saying Ilaria is better but From what i see Applio has more options

[earnest dragon]

applio is better but they are pretty close applio has fcpe and ilaria has rmvpe+ and applio has split audio and ilaria does not

[solar torrent]

I have been run RVC CPU locally on some of PCs in my college. The fastest one I could find is Dell Vostro with Intel Core i3 eighth generation DDR4 8GB RAM.

Since I don't know how much time it used to inference a single audio, unlike the Automatic1111 which it tells on both terminal and GUI, it should probably take up around 4-10 minutes on that PC.

[feral halo]

Ayo? @solar torrent level 2 !!!

[earnest dragon]

what gpu does it have and i3 8th gen is not good

[solar torrent]

There's no GPU, only CPU.

[polar flax]

8 GB ram is bare minimum for modern requirements with win 10/11, and not recommended for AI workloads without debloating ur windows or even running under linux distro

I'd strongly recommend running online with ilaria RVC: https://docs.google.com/document/d/1YbXcLFPaGjhOdG5NFkK3QrucCEpHZBwFUxkeMO8aB18/edit

as comparison, doing cpu inference on a laptop with intel i7 1165G7 and 16 GB ram took a bit more than the input audio duration

and running SD 1.5 model through some comfyui workflow is already demanding, not counting on the hi-res fix and upscaling process

[lofty citrus]

alguien tiene la voz ia de alguna argentina???

[tired jasper]

https://huggingface.co/QuickWick/Music-AI-Voices/tree/main
good luck

[tired jasper]

also you can try to find some in #1175430844685484042

[lofty citrus]

thank you papu

sos good

[tired jasper]

np hijo

[lofty citrus]

no estan los que necesito

[feral halo]

Ayo? @lofty citrus level 1 !!!

[lofty citrus]

😦

[solar torrent]

I'd rather buy newer PC for this. I don't really care how fast they are, I just wonder how it's finished. Playing Minecraft with my friends while running RVC on another PC, and the AI has just finished like that. That's fast for me.

[dapper drum]

Dose anyone even use hp6 anymore?

[solar torrent]

Because some people kept telling me to run it locally last year, it'll take over hours to finish for a single audio on my laptop, but I can wait.

[polar flax]

ryzen 5700X3D or 7800X3D, and with 3060 or 4060, and 32 GB ram, could be great for you to run vanilla mc/with fabric+iris, and wokada voice changer

though too much mods may torture ur cpu

[dapper drum]

Isn’t it the rtx 5090 and new CPU coming out layer this year

[devout bluff]

although w okada is pretty insane in terms of voice changer evolution is it still kind of eh? I tried it out today with some voice models and it just doesn’t sound that good

[polar flax]

50-series should come in the next year, and better be delayed to implement GDDR7 3 GB chips rather than sticking on 5070 12 GB and 5080 16 GB

Nvidia being self-advertising as "AI-powering" should give more vram lol

[chilly lake]

initial release gonna be 2GB chips and 192b bus

then maybe in a year or two there's gonna be TI Super with 3Gb chips

enjoy 5070 with 12GB lol

[earnest dragon]

these are probably are not that good rvc1

[earnest dragon]

yeah a ces launch https://wccftech.com/nvidia-geforce-rtx-5090-32-gb-rtx-5080-16-gb-rtx-5070-12-gb-ces-2025-launch/

[timid mango]

-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

[spare linden]

Hi i have a question how do i download the song i generated the ai cover ?

[earnest dragon]

where

[tight cypress]

I am back from grave

[earnest dragon]

who are you

[tight cypress]

someone

[spare linden]

i figured it out

[warm patio]

How to terminate the .bat file

[earnest dragon]

taskmanager

[warm patio]

thank you

[feral halo]

Ayo? @warm patio level 1 !!!

[maiden copper]

hi

[polar ore]

hello

[feral halo]

AI HUB's leaderboard

Want to view more than the top 10 users? Click here

[tepid basin]

@eager viper https://applio.org/

[eager viper]

tnx

[thorn fern]

@indigo wedge phốt nhớ bổ sung phần dox nhé, phốt mà toàn than khóc buồn v

[tropic carbon]

is it possible to turn off the hear myself?

[earnest dragon]

yeah set monitor on w-okada to empty

[tropic carbon]

its alr on none

[earnest dragon]

then what about output what's that set to

[elder willow]

hey

[earnest dragon]

what

[elder willow]

did you have ai voice setup ?

[tropic carbon]

fixed thanks, and the voice auto applies to my default mic?

or i have to change sum in settings

[earnest dragon]

you have to set it to the virtual audio cable as output and then in voice chats set mic as the virtual audio cable

[tropic carbon]

what if i dont have that..?

[earnest dragon]

then follow the w-okada guide to get it

[earnest dragon]

-rt

[rare sorrel]

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

[elder willow]

guys help me

[earnest dragon]

with what

[elder willow]

i need ai voice how can i do it ?

[earnest dragon]

voice changer you mean

[elder willow]

male to ai voice changer

[tropic carbon]

so which one do i follow?

[feral halo]

Ayo? @tropic carbon level 2 !!!

[earnest dragon]

-rt first link

[rare sorrel]

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

[tropic carbon]

the third one?

[earnest dragon]

no first link then in there they have a part about the virtual audio cable

[tropic carbon]

@earnest dragon ty for the output i can select my original headset and for the input the cable?

and now my mic does not seem to work

with the wb

and yes it is set to wb in settings

for example in discordf

[earnest dragon]

output on w-okada is the virtual audio cable input is mic and then for games and voicechat input is virtual audio cable

[tropic carbon]

yeah i did all that

nothing, can i send you a pic in pm

i cant send here

[earnest dragon]

then you have to start the voice changer voice to be able to speak

[tropic carbon]

yeah no wonder, it is started

i can hear myself as before BUT NOT ON DISCORD

can i send a picture?

[harsh mantle]

Greetings, I need some help.
How can I make a voice model using a recording I already have

[covert lake]

what's ur pc gpu?

[harsh mantle]

nvidia gtx 1650

[earnest dragon]

not good enough

[tropic carbon]

@earnest dragon

[covert lake]

you can't train a model on that locally (on ur pc)

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)

[thorn fern]

@indigo wedge nãy có con chó nó cắn này

[feral halo]

Ayo? @thorn fern level 1 !!!

[thorn fern]

[tropic carbon]

does anyone have a good old mans voice? Or some old womens voice?

[covert lake]

#1159290752195633273

[tropic carbon]

anyone? or is there any other site i can look for voices

[feral halo]

Ayo? @tropic carbon level 3 !!!

[dense nacelle]

whats rvc v2

[tacit falcon]

why is she saying that

[grand swan]

a

[winter storm]

goida

[polar flax]

meanwhile 9800X3D may come in november this year, earlier than 9950/9900X3D
https://www.techspot.com/news/105107-ryzen-7-9800x3d-could-first-3d-v-cache.html

[mint quest]

Okay, I understand your rules, but I am looking for a place where I can talk about these things.

[covert lake]

Not the right place.

You won't find a NSFW place here

It's a SFW community

[mint quest]

Yes. Can anyone suggest somewhere else I can go to find what I am looking for so I can stop bothering you nice people.

[feral halo]

Ayo? @mint quest level 1 !!!

[mint quest]

Is Ayo a server here, or a website, or what?

[mint quest]

So apparently, "ayo" is a term used to get someone's attention. It is not clear to me what Arclame was trying to get my attention to.

[lethal flax]

chat is there an AI for remixing songs?

also why did replicate become pay to use

what?

that's like, vocaloid right

I love that there's a UI that is exactly for what I'm tryna find https://replicate.com/sakemin/musicgen-remixer but it's on replicate so I'd need to pay them

I have legitimately 0 money whatsoever

I'm trying to find an easy way to change the instrumentals to a song, and I've tried just regularly remixing songs before and they all sounded absolutely dogwater awful

I'm not your "Bub"

this server sux ngl

[feral halo]

Ayo? @lethal flax level 9 !!!

[lethal flax]

people always seem to beat around the bush instead of actually answering questions

[covert lake]

cant really help u for that, idk

[covert lake]

its a bot saying u levelled up in the server

[mint quest]

Okay, thanks

But if someone has the answer to my main question, feel free to DM me.

[lethal flax]

dawg you're in a server for ai

and you're shocked that people want to use ai programs

🤦

[covert lake]

seems like that project is open source, if u got a good pc u could do it locally https://github.com/sakemin/cog-musicgen-remixer or try to make/see if theres another cloud computing service for that

[lethal flax]

I'm not very good with running code that doesn't have a UI implemented with it

[elder willow]

@covert lake Is there a colab for this

[covert lake]

i don't think so afaik

[elder willow]

is anyone willing to make a colab for this if so maybe @tender tundra Can ....

[humble solar]

can someone reccomend me female eng voice model for talking

[vast maple]

Hello who know were I can found the frensh serv

[covert lake]

from what i remember its down since months for copyright, @left cloak u got any news?

[cunning otter]

Fortniteburger

[covert lake]

tried but seems using cog and docker, and seems not updated in 7 months, didnt work

[heavy inlet]

hi

@covert lake

[covert lake]

?

[heavy inlet]

Are you good at making Ai cover songs?

[covert lake]

are you looking for someone to make it for you, or to make ai covers urself?

bc its free to make ai covers

[heavy inlet]

yes

[feral halo]

Ayo? @heavy inlet level 1 !!!

[covert lake]

yes to what

i asked u 2 things

[heavy inlet]

How do I learn my voice?

[covert lake]

Do you want to train (make a voice model) of your voice, to then later on use for an ai cover ?

be more specific on the request pls

[heavy inlet]

Please make it! I really need to learn my voice

[covert lake]

so...

Are you looking to make an ai cover with your voice?

[heavy inlet]

yse

[covert lake]

then, you need to train a voice model of your voice, then use it to inference (use the model) to make an ai cover

[covert lake]

what's ur pc gpu?

[heavy inlet]

in my voice

[covert lake]

i can tell u how to do it but won't do it for u

[deft sleet]

he's a troll 100%

[heavy inlet]

Yes

[covert lake]

i assure u there are people who thought GPU was a thing u install

idk why there's SO MUCH tech illiteracy

[deft sleet]

lol

[covert lake]

!howtoask

[pine acorn]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

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

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

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

__**BE PRODUCTIVE.**__ 🤝

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

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

[covert lake]

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

[covert lake]

i swear people think ai is just a thing u do at the snap of ur fingers

ai does NOT work like that

[deft sleet]

tons of boomers trying to use ai rn so yeah

[covert lake]

i see even young people like that

cus the average person buys shit only to watch tiktok all day

[deft sleet]

true lol

[covert lake]

like when someone buys a 4k gaming pc just for roblox

[deft sleet]

I got mine to play csgo like 4 years ago for 5k

[feral halo]

Ayo? @deft sleet level 1 !!!

[deft sleet]

3090 rtx i9 10900

The price was supposed to be like 3k but they kinda scammed me

[covert lake]

dang

i see 😭

[deft sleet]

That's what you get when you buy retail lol

[lethal flax]

is it just me or is everything on weights just the logo rn

oh

I guess it was

the pictures are loading but now the song I'm tryna convert is not downloading

[polar flax]

zoomers* ?

[hushed birch]

Boop

@dawn temple

[tacit falcon]

Around the world, members of Generation Z are spending more time on electronic devices and less time reading books than before, with implications for their attention spans, vocabulary, academic performance, and future economic contributions.

[covert lake]

e books

[tacit falcon]

their books are drm protected and needs saliva spit sample and email address to be opened

[tepid basin]

I so dearly wish we could go back to a time where people can socialize in the living room without short form video

[covert lake]

who ever said about buying them

[tepid basin]

Same with car rides. Demimishes to tiktok

[covert lake]

sail the high seals

[tacit falcon]

i just said it for no reason

just because you mentioned e books

[covert lake]

build a time machine to 2016

[tepid basin]

No we need pre-vine

Vine is classic but it was the start of the downfall

[covert lake]

no we need when undertale was famous

[tepid basin]

Fair

[tacit falcon]

vine's kinda old though didnt it only start gaining traction in like 2014 or something

[covert lake]

with meme animations

[tacit falcon]

old as fuck even

[tepid basin]

Yeah old af

[tacit falcon]

that should have been it really

[tepid basin]

Damnit I'm realizing that shits like a decade old and I've been on the internet that long

[tacit falcon]

no more of those apps once it pissed off

except apparently vine tried to come back from the dead anyway and they apparently did so, but not for very long

[random flare]

HI

[tacit falcon]

Goodbye

[proper seal]

why yall ban takeoff

[polar flax]

them when given some written articles/ebooks: "I ain't reading allat"

[green wolf]

[covert lake]

their skill issue tbh

[ionic pumice]

https://tenor.com/view/kitty-eepy-gif-10295980600460052163

blep

[tacit falcon]

I ain't reading allat

[graceful talon]

Hola

[ionic pumice]

boku no saiba sanda saida

[main pebble]

Is there a Dork Alina here?

[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.aihub.wtf/essentials/how-to-make-voice-models/

[tacit falcon]

https://tenor.com/view/kasane-teto-bread-gif-957594897712705805

[elder willow]

yo guys

is anyone down to vocalize an AI travis mixtape?

i'll mix the vocals n shit and I have some beats ready

[raven hornet]

yall know any really good. voice changers that are free besides voicemod and dubbing ai

[covert lake]

wokada which uses rvc voice models, whats ur pc gpu?

[raven hornet]

tbh idk

im using a macbook air

so i dont think its much

[covert lake]

u should be able to do it locally (on ur pc), especially with the fork (modified version) which is optimized, check https://rentry.co/ForkVoiceChangerGuide

[raven hornet]

can we go on a call so i can screenshare it

[feral halo]

Ayo? @raven hornet level 1 !!!

[raven hornet]

bc this is confusing

go in #unlimited vc

[covert lake]

sorry but i can't call, but in the written guide there's everything u need

following it u should be able to use it, if there's any issues u can ask in #🔍│help-w-okada or #1192011222023950368

[lavish marten]

Damn, I'm just wondering what could be the best trained female voice I could use. There are a lot here...

[oak forge]

hey

how is called the AI program for voice changing?

i forgot

nvm i find it

[polar narwhal]

hi

[solar torrent]

Whoever telling to run RVC locally should've run it on their PC too.

[chilly lake]

what's the problem with that?

[elder willow]

hi is there a mod around

[solar torrent]

Well, that was when last year.

[river adder]

sure how can i help

[elder willow]

@yrntakeoff is asking why they got banned

[feral halo]

Ayo? @elder willow level 1 !!!

[river adder]

the should have a message from @pine acorn

[solar torrent]

Both Automatic1111 and RVC would deflate their memory usage when leaving idle for so long on my laptop, and then inflate back again when there's inference going on.

[elder willow]

thank you I'll let them know!

[river adder]

found their mod case

[elder willow]

I showed them, is there a way they can appeal or?

[placid orchid]

I need an ai voice for an exe im gonna make

[lethal flax]

chat how do I get a streak in weights

[candid fox]

hi

[stable storm]

yes

[carmine hatch]

I need this model of star butterfly
针对T-ponynai3 v6的 星蝶公主lora

it shows up on civit ai but not found

there's another website that has it but disabled the downloads :/

[plush hearth]

is there currently anywhere to get beatrice v2 models?

[heavy zephyr]

not work link https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[hard oar]

hi chat

[covert lake]

yea the docs are temp down #📰│dev-updates message, use https://docs.ai-hub.wtf/essentials/how-to-make-voice-models/

[covert lake]

Not really, there's lot of rvc v2 models tho

use those

[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/

[hoary prairie]

who plays fn?

😭\

[heavy zephyr]

Is there a video explaining the tool?

[covert lake]

No, there are no youtube videos that aren't outdated

what's ur pc gpu?

[heavy zephyr]

rtx 3060ti 8gb

[covert lake]

you could do it locally

[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

[covert lake]

wait, laptop or desktop?

[heavy zephyr]

desktop

[signal lotus]

Anybody use Intel arc Gpus? Wondering if an RTX 3060/3070 would be better than that because of CUDA

[covert lake]

Alr, u can read the docs then

U should be able to do it

[covert lake]

Afaik Intel arc GPU sucks, especially for ai it would be better to use nvidia

Nvidia is the most compatible for almost every ai project

[signal lotus]

Then, should I get 40 series or 30 series?

[covert lake]

An rtx 3060 12gb would be good, remember also that memory matters too

For RVC training it's suggested to have 8 or more vram

Idk if u are gonna try other ai projectd

[pine cargo]

hi

[lethal flax]

Chat is there an AI for making a UI for code
or porting replica UIs to Colab

[cunning otter]

Dantes ai voice funny moments featuring the mod team

[ionic flower]

Making a submissions sheet for some ai cover album.

I will put "Various AI" instead of "Various Artists".

[elder willow]

lmao

-svc

[rare sorrel]

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

[elder willow]

the 1 percent of bacteria be like 💀

[ionic flower]

Does it make sense or?

[elder willow]

look at the gif

[ionic flower]

Comedy

[analog cosmos]

[elder willow]

hey!

[ionic flower]

Oh and another thing, I'll add some quality control to it because I am not gonna see low quality ones

[gloomy bay]

hi guys

[zenith stump]

where's the link to create models?

[covert lake]

what's ur pc gpu?

[zenith stump]

GTX 1660

[feral halo]

Ayo? @zenith stump level 1 !!!

[covert lake]

your pc is too weak to train (make models) locally (on ur pc)

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)

[zenith stump]

yeah ty

[covert lake]

Google Colab = Ez but low gpu time so risk of losing ur work
Kaggle = hard but wayyyyyyyyy better gpu time

[zenith stump]

I remember making some models like 2 years ago on a colab

[covert lake]

its way more suggested to use kaggle now

gives 30 hours weekly

[zenith stump]

isn't 4 hours enough for a 11 minute recording?

[covert lake]

gpu time on colab is not granted, its around 4 hours, could be even lower

• u have to use the tensorboard, which means train to 500 epochs and see which .pth is the best

so u COULD be able to but also have a risk of losing ur stuff

use Kaggle atp

[covert lake]

you used the tensorboard before right?

check https://docs.ai-hub.wtf/rvc/resources/epochs-tensorboard/ & https://docs.ai-hub.wtf/essentials/how-to-make-voice-models/ (the docs don't contain about kaggle as its not updated, but it shows about dataset cleaning)

[placid parrot]

Hey folks, I was curious about something and I felt that I've scoured the entire internet trying to find the answer. I picked up DigitalMate for Steam which is an AI desktop avatar - I've got the whole thing working aside from trying to get customized TTS. It seems like it only supports a handful of options for that, Windows TTS, VITS-API, Edge TTS, and Fast VITS.

The VITS API seems like it'll be the easiest way to go but all the variations of VITS seem overly complicated compared to RVC - which I can't think of a way to force an RVC mode into the software. Creating custom models for VITS (so I can use the VITS-API) seems like just a much of a chore, aside from paying someone to create the model. Which I would, but it seems like most people create models in RVC.. so I'm in a bit of a boot loop in trying to get a customized voice for the software. Anyone have any suggestions by chance?

[chilly lake]

depends on a language

[placid parrot]

Yeah, I've found that most of the VITS / SoVITS / BERT-VITS2 are generally all in Chinese

[marble current]

Hello, I am a beginner artist and I am looking for a person who will make a short animation for my first song. Would anyone be interested? 😀

[queen onyx]

hola ¿como están? ando buscando una voz de un personaje, pero no la encuentro, la voz que busco es la voz japonesa de black rock shooter del videojuego, la voz la hace Maaya Sakamoto, si me pueden ayudar se los agradecería, gracias por su comprensión.

[tiny bridge]

which voice changer is the best btw?

[chilly lake]

for english, speech-wise Edge TTS seems superior, but it goes thru MS servers with all privacy implications and it is rather bland

[placid parrot]

@chilly lake Yeah, and although it's not bad, you can't upload custom voices to it

[feral halo]

Ayo? @placid parrot level 1 !!!

[chilly lake]

can run RVC model on top of it

[placid parrot]

You can?

[chilly lake]

models that use eSpeak-ng are quick, but somewhat robotic and make more mistakes with prononcuation

xtts - zero shot, great quality, but sometimes hallucinates and lacks reproduceability

[placid parrot]

Yeah, and unfortunately, I don't see this software working well with it.

[chilly lake]

with a python script you can run TTS, get an audio out, call RVC inference

[placid parrot]

Noobies, would you mind if I shoot you a DM, you have given me the first glimpse of hope I've had trying to get this thing to work and I don't want to spam the chat if it's just me chatting it up

[chilly lake]

k

[icy pendant]

https://rentry.co/forkvoicechangerguide

[tiny bridge]

thanks

is it compatible with rvc voices?

[chilly lake]

unless you dig out some old wokada, it should be able to handle both v1 and v2 RVC models

[tiny bridge]

Sweet I can now dump Voice.AI

[feral halo]

Ayo? @tiny bridge level 1 !!!

[covert lake]

https://cdn.discordapp.com/attachments/1159290193619189821/1283652791931506759/sponge-bob-night_.gif

[covert lake]

#🌍│español

[covert lake]

this is an ai server, but ig u could see in #1159290752195633273

[solar agate]

chat!

why isnt my voice changer thing working

i need some help

[covert lake]

#🔍│help-w-okada or #1192011222023950368
and be specific

[solar agate]

y- yes alpha

idk how to work this jawn

[covert lake]

wtf.

[solar agate]

my fault gangalo

[solar agate]

like how do you get the voice changer to voice change. im a noob at this ai stuff

[ruby vessel]

Can i use Applio on a drive thats not the C drive?

[chilly lake]

yes

just dont use spaces and weird characters in the path, those may result in weird stuff

[ornate pilot]

is https://docs.aihub.wtf/ down

it says "Your domain is expired. Renew the domain to activate the website" for me

[charred trench]

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

[ornate pilot]

thank you!

[feral halo]

Ayo? @viral tundra level 1 !!!

[ionic flower]

why is the username that?

[naive coral]

Hi

[placid parrot]

Is there anything against the rules (I can't seem to find any at the moment) that says you can't request a voice model in VITS format? And even if it isn't, is it worth trying to put it out there since I don't see many people using it?

(Obviously, I wouldn't mind paying for it)

[earnest dragon]

you mean sovits models or do you mean tts

[placid parrot]

VITS, BERT-VITS2, SoVits - really something that has the .pth and config.json

Sorry, still learning the whole lingo, lol

[earnest dragon]

so bert-vits2 is tts sovits is sts

[placid parrot]

Gotcha. I had no idea, what I had figured was that SoVITS was more singing / attention to pitch but still had TTS with fine-tuning, BERT-VITS2 was more pronouncation.. but I guess that makes sense.

I'm more specifically looking for VITS or BERT-VITS2 then.. something with a character that has TTS

[lethal flax]

hey so I have a video of an unaired TV pilot that isn't going to be released, however there's this loud, obnoxious laughter occasionally throughout the audio, any tips for removing the loud laughs?

[ionic flower]

good thing i know what sts is now

[covert lake]

#📰│dev-updates message

[steady oasis]

Hi I'm new hopefully we all get along ☺️

[west finch]

Hi chat. Any1 can help me convert a text to trump or biden speech? or point me to where I can do it for free without download

[feral halo]

Ayo? @west finch level 1 !!!

[charred trench]

Will having scoffs or breaths in my dataset ruin the dataset?

[ionic flower]

Uhhh you mean sts?

[hallow haven]

which preset is to sound like a man

[west finch]

Basically just need a trump voice generator from text or speech that is free

[ionic flower]

You could find one but.... your luck

[chilly lake]

it aint hard

[jagged prism]

guys, i might need some help right now

i can’t find a good ai cover website

[tired token]

@hollow tangle sorry for the tag, thank you for the sale of the goth mommy pack. its soo good!

[wise quiver]

this MIGHT be Ilaria on an alt account

[charred trench]

is it better to use a pretrain or not use a pretrain

[ionic flower]

....you ever used haysay?

[feral halo]

Ayo? @ionic flower level 6 !!!

[earnest dragon]

I already gave you the link to free Donald Trump tts that is high quality if you want Biden also I can give that also

[west finch]

not keen to download anything I have a high risk computer

[real thunder]

May I have the link for that/

[placid parrot]

So, before I post a consignment request up, are there still people who create VITS models now'a'days? 🙂

[fierce hinge]

hi

[earnest dragon]

no it's a website no downloading needed

[earnest dragon]

which one the trump or the Biden

[astral shard]

yo

[polar flax]

is BERT-VITS2 a new thing?

[placid parrot]

No, I actually believe it's a bit older, seems that it was pre-RVC from what I can gather and is primarly chinese

[feral halo]

Ayo? @placid parrot level 2 !!!

[fervent cliff]

hi

[polar flax]

sovits-svc sibling but for TTS?

[placid parrot]

Sounds right

[fervent cliff]

Essay about keyboard spam made my ai that i gave it

[polar flax]

the recent ones are RVC (STS) and GPT-sovits (TTS)

[haughty holly]

Could someone make a Fred Astaire model?

[gray rover]

For memo~~
*.pth - pytorch models

it isn't -vits or such exclusive

[timid wave]

if i want to make ai covers online what website is best to use?

[fervent cliff]

ai

[placid parrot]

@gray rover Sorry, a bit confused on that - still learning some of the extensions. So a .pth file (pytorch model) and a config.json arn't exclusive to VITS?

[gray rover]

config is just a configuration and .json happens to be what most AI related stuff uses for storing configuration, hyperparameters and so on

but yes, in case of AI, .pth means it's coming from pytorch based environment

but in case of rvc, those .pth models are weight models

[placid parrot]

Oh. That would explain why some of the things I find on HuggingFace don't really seem to follow the .pth and .json pairing even if it's listed as VITS.

[fervent cliff]

Gemini funny funny funny funny

[feral halo]

Ayo? @fervent cliff level 1 !!!

[gray rover]

well yes, however I can't speak for vits and so-vits and so on

[fervent cliff]

i just make this account four minutes ago

[gray rover]

There's also some other " secure " format people turn the .pth into as well

or something like that

[chilly lake]

there's safetensors, but rvc pth is just weights, no code

[placid parrot]

Gotcha. Thank you. I'm just trying to figure out a way to have a custom voice model that can be loaded into a program that I have which says it supports VIT and requires a .pth and a .json file

[gray rover]

^ there

[fervent cliff]

i have a ai essay

[gray rover]

isn't it gpt-sovits that you mean by any chance?

[fervent cliff]

nooooooooooooooooooooooooooooooooooooooooooooooooooooooooo i lost it

[placid parrot]

I thought it was possible, but my understanding is that SoVITS is Speech to Speech?

[fervent cliff]

it's gone

[gray rover]

hold on,
https://github.com/jaywalnut310/vits
https://github.com/auspicious3000/contentvec/
these may interest you

[fervent cliff]

why Gemini why do you delete my essay

[gray rover]

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/tree/main
Generally, at the bottom of the repo are all components, systems etc

Worth researching if you're into grasping more on those

[placid parrot]

I appreciate it, and yeah, I'll take a look. I'm a bit sure that it's beyond me, but I'm trying to figure it out as I go. I'd mainly dabbling in it with the effort to get this program straightened out, lol

[gray rover]

That's fine, but at least you'll get the base idea on what's what

[fervent cliff]

made new ai essay

yay

IT's called my ai essay gallary one because i lost first one

[gray rover]

V3 most likely won't ever happen

and the " v2.2 " isn't actually in terms of changes but more bout code structuring and all of that

Let's say main dev decided to revert some or most of " troublesome " changes others done to their code

[fervent cliff]

🪂

🤸‍♂️

[earnest dragon]

2.2 probably will never happen

[gray rover]

yeye, I meant the v3 never happening

the guy mentioned an " update " of rvc being v2.2 so I just corrected em

[fervent cliff]

nooooooooooooo

fallll noooooooooooooooooooooooooooooooooooooooooooo

[gray rover]

but ye, been into rvc for almost 2 years now and Imma tell you that v3 just isn't happening and people should stop hoping for it.
Heck, it isn't even a proper " v2 " in fact

the v1, v2 and potentially v3 meant just pretrains' version

not the rvc ( platform )

[fervent cliff]

🪂

[feral halo]

Ayo? @fervent cliff level 2 !!!

[gray rover]

Calm down that's for one, I am trying to explain the misconception here

[fervent cliff]

🤸‍♂️

so far

noo

[earnest dragon]

but isn't rvc and sovits and all of that all using vits

[gray rover]

But back to the topic, the versions people hope for aren't what most of y'all think.
All it is, is just pretrains' versions.
Legit rvc " update " or step-up will be once hifigan is replaced with newer tech such as bigvgan or potentially-in-future implemented EVAgan
That's all

[fervent cliff]

🪂

[earnest dragon]

you basically could make a vits based sts and then call it rvc 3 as long as it better and as easy to use then it would technically be rvc3 just not made by the creator of rvc

[fervent cliff]

one

two

four

[gray rover]

well, you " could " as long it does what rvc's meant to do

[fervent cliff]

seven

[gray rover]

voice cloning based on retrieval

[untold furnace]

all my homies hate the internet archive data breachers

[gray rover]

but still, hifigan has too many issues. Phase being one, fidelity in upper harmonics being the 2nd one ( spectral smearing )

[fervent cliff]

🤸

no fell so far

[earnest dragon]

yeah because all they are is just vits based sts

[gray rover]

yup

[fervent cliff]

AI Essays

[gray rover]

Now, excuse me but I gotta go back to training my model

[earnest dragon]

so technically a modified version of sovits could be considered a rvc3 if it was better and did what rvc does

[gray rover]

yea, again, could but rvc is just one. I wouldn't carry the same name really.

Tho, if I was the one to make it and was more of an actual dev, I'd implement an additional transformer-based helper and verifier - just that alone would make it something else than rvc is

[earnest dragon]

yeah but I feel like true rvc 3 or equivalent would be a zero shot sts

[gray rover]

Maybe, but my stance on zero-shooting is that atm, it can't be as accurate as carefully and lengthy finetuned stuff

at least in audio

[earnest dragon]

it could if the model was trained with enough audio and variance and stuff

[chilly lake]

some tts do a decent 0-shot, others are pretty bad (tortoise)

[gray rover]

Still haven't found a good one ^

that I'd deem HQ enough and do the style cloning job just right

[chilly lake]

tested F5-TTS today, it has some limits

[earnest dragon]

the only sts that is zero shot is seedvc and its not that bad for a zero-shot

[chilly lake]

if you dont care about 100% match it is decent

[gray rover]

ye, well said, those who don't care.. and I'm one of those details-freak lol

if it ain't perfect or close to, it's a nogo

and it'll stay like that for a while, sadly

[earnest dragon]

yeah but you see how faceswap is going the zero shot route I feel like sts will go the same route

[chilly lake]

quality though, very good

[gray rover]

Well, pictures, images and generally pixels are way way easier than audio

Audio is the most complex one

[chilly lake]

comparing to XTTS that output is smooth af

[gray rover]

Phase, bands, all the harmonics

spatial response

[earnest dragon]

still the best zero shot tts I have tried is talkingavatar.ai but it's paid to create the voice but free to use the voices people make public and stuff.

[gray rover]

too many variables compared to just " imitating pixels well "

I personally hope for elevenlabs-level opensource system eventually

I just love how they nearly perfectly handle style copying

[earnest dragon]

just for zero shot tts most of the models you have to do a separate voice per emotion

[gray rover]

Had at least 2 or 3 times on tiktok where I couldn't tell if it was AI or not and I am ultra sensitive to this, given it's been 2+ years for me of experience and work with AI gen. audio

so I am def impressed, still

[gray rover]

yea

[real thunder]

Trump TTS

[earnest dragon]

okay you have to make a account but it's free to use the ones that are public https://talkingavatar.ai/#/voice?id=12526723644780037

[real thunder]

@earnest dragon thank you

[earnest dragon]

each generation will sound different and if it has background noise because of the sample just use mvsep bs reformer

[weak scarab]

hi I wanna train my own rvc model

is 25 min of clear, raw voice data enough?

or what should I aim for

[charred python]

Hello everyone,

I am excited to introduce myself and express my interest in potential collaboration opportunities with your esteemed companies. My expertise in Artificial Intelligence (AI), Machine Learning (ML), and Full Stack Development, combined with my experience in supporting various startup ventures, positions me as a strong candidate to contribute to organizational goals.

My professional background includes assisting startup companies in creating their Minimum Viable Products (MVPs) and fostering functional work environments. I believe my skills and experience align well with the dynamic and innovative cultures represented in this Discord channel.

For your consideration, you can review my professional profiles at the following links:

LinkedIn: https://linkedin.com/in/joiner-cantillo
Portfolio: https://joiner-for.vercel.app/
AI Researcher Experience at BlackCube Lab: https://blackcube.digital/team

I am eager to explore potential synergies and contribute to the success of your companies. I look forward to the opportunity for further discussion regarding how my expertise can add value to your teams.

Thank you for your time and consideration. I am keen to connect and explore potential collaborations in the near future.

Best regards,
Joiner Cantillo

[solar torrent]

Onikata Kayoko is my favorite woman from Blue Archive.

[unreal girder]

Hey guys,sup?

[valid ridge]

hi I wanna to find some voice models

[jagged prism]

no

[solar torrent]

Hi.

[cursive osprey]

use audiomodify.com

[earnest dragon]

Thays trash

[valid ridge]

thanks

[solar torrent]

voice-models .com

[earnest dragon]

trash also

[solar torrent]

All of them are trash? What the fuck.

[earnest dragon]

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/

[earnest dragon]

and then use

[feral halo]

Ayo? @earnest dragon level 34 !!!

[unreal girder]

I hate to say this but,I'm done supporting certain things

[earnest dragon]

-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

[earnest dragon]

All websites like audiomodify and stuff are just paid rvc and stuff and is trash when you can do the way I just said for free and way better

[solar torrent]

Model file name inside a ZIP that I downloaded from Weights gg is just model.pth. I have to manually rename them.

[feral halo]

Ayo? @solar torrent level 3 !!!

[unreal girder]

Ai stuff that is

[earnest dragon]

yeah you should if it's that generic but ilaria zero you don't need to

[polar flax]

weights.gg stores models from #1175430844685484042 and also user-created ones that are downloadable

[solar torrent]

Voice models I've downloaded from voice-models indeed retains the original file name in its ZIP instead of model.pth.

I'm just stuck in 2023.

[unreal girder]

Looks like no one's paying attention

[polar flax]

use a time machine to go back

[cursive osprey]

I’m creating a presidential parody video. Audiomodify had a couple of good voices for Obama, but anyone know where to find AI models for Trump as well?

[solar torrent]

I use Weights mostly to make people sing BBL Drizzy.

[cursive osprey]

[feral halo]

Ayo? @cursive osprey level 1 !!!

[earnest dragon]

it does but weights.gg has more then this server because it has this server and more

[earnest dragon]

nah weights.gg has the best voices

[earnest dragon]

tts or what

[plush hearth]

anyone able to help?

[feral halo]

Ayo? @plush hearth level 3 !!!

[cursive osprey]

oh, ig i should check it out, thanks for the info dude

brb

[cursive osprey]

yep

[plush hearth]

im getting some errors

[hard oar]

what kinda?

[feral halo]

Ayo? @hard oar level 1 !!!

[plush hearth]

[Input]: ServerAudioDevice(kind='audioinput', index=18, name='Microphone (HyperX Cloud III)', hostAPI='Windows WASAPI', maxInputChannels=1, maxOutputChannels=0, default_samplerate=48000.0, available_samplerates=[]) None
[Output]: ServerAudioDevice(kind='audiooutput', index=13, name='CABLE Input (VB-Audio Virtual Cable)', hostAPI='Windows WASAPI', maxInputChannels=0, maxOutputChannels=2, default_samplerate=192000.0, available_samplerates=[]) None
[Monitor]: None None
[checkSamplingRate] Error opening InputStream: Invalid sample rate [PaErrorCode -9997]
Sample Rate:
[Model]: 32000
[Input]: 192000 -> False
[Output]: 192000 -> True
Sample Rate is not supported by device:
Checking Available Sample Rate:
[checkSamplingRate] Error opening InputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening OutputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening InputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening OutputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening InputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening OutputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening OutputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening InputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening OutputStream: Invalid sample rate [PaErrorCode -9997]
[checkSamplingRate] Error opening InputStream: Invalid sample rate [PaErrorCode -9997]
Available Sample Rate:
[Input]: [48000]
[Output]: [192000]
continue...

[solar torrent]

Did you actually make it to output 192khz?

[plush hearth]

i just set it as that yeah, I was told to keep it on 48000 but it was giving the same issue

what should my input output be set as

[earnest dragon]

the rvc is not for you unless you want to say the lines your self if you want tts then I can send you a high quality free tts of trump

[hard oar]

hmm... like a big problem

[earnest dragon]

you can't make it do that

[plush hearth]

yeah lol, whats the next best input output setting if I cant use windows wasapi

[solar torrent]

192khz for something like this is crazy. Maybe try set the output the same as the model?

[plush hearth]

my SR is set at 48000 now

idk how to set it as the model, should I use mme?

[thin gazelle]

my computer is 32GB+1TB

😎

[hard oar]

I’m looking for an AI model that can do a Yoda impression for a little project, anyone got an option?

like a web..

[solar torrent]

The team behind the Red Archive game canceled their game the same day Yuno Miles announced he quit doing music. Even though Yuno has recently back to business for whatever reason, the Red Archive has just instead become a manga for a con in Japan.

Been interested in this game for nothing but because of a joke I made long ago.

[polar flax]

why tf did u set 192k output? 💀
go back to client audio device setting in the voice changer if not sure

[plush hearth]

I didnt mean to set it to 48000

and clients sound more shit than mme server

im trying to improve the quality and limmit the choppynes

idk how to get it to work with windosmsiapi

[polar flax]

it's basically MME with Sup1/2 option and also a bit more latency

[plush hearth]

yeah exactly

I want it sounding better than mme

if possible

[polar flax]

depend on if you connect your headset through it's USB cable or directly to the pc's audio jack
also VB cable is shit, try VAC lite instead

[polar flax]

nope, unless you try some suggestions

[plush hearth]

my headset is connected to usb, is vaclite diffacult to use? or just install and set to it

[weak scarab]

im training a model rn

500 epoch is enough?

is there a limit to it?

[polar flax]

https://software.muzychenko.net/freeware/vac470lite.zip

one thing is, to run the setup64.exe, not 64a which is for some arm snapdragon laptop

[weak scarab]

I have 22 minutes of raw voice data so idk how many epochs I should do with that

[plush hearth]

yep, got it downloaded. do I just do the same thing by setting it in the output?

[plush hearth]

ah

seems to be working

[placid parrot]

Hmm, quick question for anyone familiar with it - I was just able to create a voice module for GPT-SoVITS, I've got the .pth file, but it didn't create a config.json - anyone know how you go about doing that by chance?

[valid ridge]

thanks

[feral halo]

Ayo? @valid ridge level 1 !!!

[hollow tangle]

thank you!

[tired token]

i did have a question about it, but i sent it in the website 🙂

[hollow tangle]

dm me

[polar flax]

the ones in #1175430844685484042 are subset of the weights.gg listing but have gone through QC staff review

[elder willow]

Anyone who have chatgpt prompt shorten answers of questions? Academic purpose

[feral halo]

Ayo? @elder willow level 1 !!!

[solar torrent]

I have Kanye bear plush.

[earnest dragon]

weights.gg has all the models this server used to have and currently has and also ones posted only on weights.gg

[torpid anchor]

hi

[tepid basin]

Graduation?

[signal lotus]

I've never used any AI models before, I'm getting a new GPU for AI voice modelling, specifically real time voice changing. Now, I can get a 3060 12GB or 6700XT (12gb) for same price. I'm confused which one to go with. Or if there are any other suggestions in the same price range please do suggest.

[uncut aurora]

Good evening everyone!

[brisk finch]

Hi

[covert lake]

You should use the tensorboard

Where are u training

[covert lake]

There are different Text To Speech (TTS) AIs:

GPT So Vits: RVC isn't as good as GPT So Vits for tts, but gpt so vits (few shot tts, which means needs just a lil training for models) can't use rvc models (and viceversa), and its only limited to: english, chinese & japanese, if you wanna check gpt so vits instead, read https://docs.aihub.wtf/tts/gpt-sovits/

Freemium 11labs: A easy way to do TTS is https://elevenlabs.io/, you can't use RVC model on this but its a mostly premium easy way for good quality TTS

FishSpeech: FishSpeech is a 0 shot (no explicit training needed) TTS, if you got a good pc you can use it locally else use their site

With RVC Models:

RVC is natively for Speech To Speech, but forks such as ilaria rvc mainline & applio have built in tts (using Microsoft Edge TTS to make a generated tts audio, which i suggest you to choose a tts model that is the same gender and language of the rvc model you wanna use, and then convert it with rvc)

If you wanna do tts locally with RVC Voice Models (if you got a good pc):

• You can get Applio in our docs
• While Ilaria RVC Mainline here (no guide as of right now)

If you don't got a good pc you can do tts with RVC Voice Models on cloud:

• Ilaria RVC Zero (Running on A100 GPU, free fasted rvc on cloud) and the guide

• Use Applio UI Colab (with google colab T4 free daily limit gpu)

• if you don't wanna use edge tts, you could try another tts ai from our tts index and use the output as an input in rvc

[covert lake]

Technically those are noises

[covert lake]

Use Ilaria RVC zero #🧬│ai-chat message

But if you got a good PC would be better to do it locally as you won't have any limits

[covert lake]

To use a pretrain, in general there's the OG pretrain as default, bc else UD need TONS of hours

[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.aihub.wtf/essentials/how-to-make-voice-models/

[covert lake]

If you really want to do it in a website it's better to use Ilaria RVC zero

[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]

But, if you got a good PC, I'd suggest you to do it locally (on ur PC) so u don't have any GPU limit

[covert lake]

RVC V3 IS RELEASING:

NEVER

[polar flax]

the latter is fine, not a good idea to remove it unless not enough clarity, even there are decent ASMR models already

[covert lake]

Fish speech 1.4 seems better than it, I tested it too

[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/

[charred trench]

Ahh gochu

Thanks!

[floral remnant]

Hello

could someone help me?

[covert lake]

ask in an help channel and be specific

[floral remnant]

where is that

ive never used it and i want to how to download it and stuff

[covert lake]

specify what you want to do

and whats ur pc gpu

There isnt a single program for everything

#1159290752195633273

if you share it in any other channel it will get deleted

@gray swan

Why do people even join just to promo

[tacit falcon]

why do people join and ask ridiculous questions and then leave after they get ignored

[covert lake]

people think everything is a 1 click shit

thats not how internet works

[novel storm]

whats the best english realistic voice changer

[covert lake]

For realtime voice changer in calls, wokada that uses the rvc voice models, specifically the fork (modified version) that has better optimization

whats ur pc gpu?

[covert lake]

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

[novel storm]

AMD ryzen 7, and '

[feral halo]

Ayo? @novel storm level 1 !!!

[covert lake]

thats a cpu

check the gpu tab

[novel storm]

rtx 4060

[covert lake]

alright thats good enough

-rt

[rare sorrel]

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

[covert lake]

1st link

[novel storm]

what would you suggest for best realistic voice changer for games

[covert lake]

the wokada fork

follow the 1st link

its a guide

[novel storm]

is there a link for that to download

[solar torrent]

A jeep that will go to Gensokyo.

[covert lake]

yes, inside of the guide

click the first link

its a guide, u need to READ it

[agile axle]

waddup

I've been trying different sites to play around with AI vocals using my isolated vocals, but they all sound odd, like weights gg. Maybe I'm using the wrong settings I don't know. Anyway, I heard about hugging face, is that the best option?

[covert lake]

all of those ai sites just use RVC

we have partnershio with weights.gg, it uses RVC too

[agile axle]

I'm a noob twin, idk what any of that means Rawr XD

[covert lake]

the quality depends on the model and the cleaned vocals

[covert lake]

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

huggingface is the biggest ai platform in general, it has spaces to try out some ai projects like RVC, the quality depends on the model you are using and the cleaned vocals you are giving it

you need to use a good model and clean better the vocals to have higher quality

you can also use RVC locslly (on ur pc) if u got a good pc, which is better than cloud (remote good pc, like those ai sites) as u wouldn’t have any limits in terms of gpu time

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

[covert lake]

@agile axle

[agile axle]

I'm here

[feral halo]

Ayo? @agile axle level 1 !!!

[agile axle]

I'm on mac bro

[covert lake]

alr then its kinda bad, mac is bad for ai

mac can only inference (use models) locally (on ur pc), not train (make voice models)

do you wanna do it locally or use cloud?

i could give u the easy ilaria rvc zero way which is a zerogpu huggingface space, the fastest using A100

[agile axle]

Whatever is the simplest, best quality way, sure

[covert lake]

Ilaria rvc zero then

[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]

the quality matters by the model and vocals u give it tho, doesnt matter if u use those sites or other things as its the same program, u need to be sure to use a good quality model

[agile axle]

I have sound-treated/De-noised, stemmed-out vocals from my own tracks. I know they gotta be fully dry as-well.

Do you have any tips on finding the best-quality models? I'm looking for Future's In particular

the rapper Future

[covert lake]

that’s good, u can also see vocal isolation here https://docs.ai-hub.wtf/rvc/resources/vocal-isolation/

[covert lake]

try and testing

thats the only way

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.aihub.wtf/essentials/how-to-make-voice-models/

[ocean folio]

Y'all is there a cheap mic RVC?

[covert lake]

ofc u can also get commisioned models by model masters

[agile axle]

Bet, will watch the guide asap!!! ty

[covert lake]

yw

[agile axle]

What do use Ai for bro if you don't mind me asking?

specifically Audio

[covert lake]

for converting a voice to another? RVC which is the best out there

[agile axle]

noooooo, I'm saying what do you personally use it for?

[feral halo]

Ayo? @agile axle level 2 !!!

[covert lake]

use for what ? Could u elaborate more pls

oh i got it

well, personally i used it for ai covers

i used to upload them into my yt channel lol

[agile axle]

I wanna hear em

It better not be no carti stuff lmfaoooo

[covert lake]

I don’t really do that anymore tbh

[covert lake]

i made a playlist https://www.youtube.com/watch?v=VA50k6XCFqo&list=PLkJf7YTZRCGRe9mZ-b0q939eBNCLiMYdN

[covert lake]

nah mostly cartoon or video games lol

[agile axle]

Lmafo, I see now. Minecraft villager cover is crazy work bro

[covert lake]

lol

[solar torrent]

I might make the BBL Drizzy pt.2 if my video get more attention, but then I've also made more people singing Drake/Diddy/EDP/Kris Go Away by Rizz Records.

[agile axle]

Where's the Diddy voice model?

[solar torrent]

The Diddy model doesn't seem to be here.

[agile axle]

You should get a custom model

Yo, quick question, do dynamic vocals make the Ai models tweak out?

[chilly lake]

does it do voice cloning?

[solar torrent]

Is dynamic vocal even a thing?

[covert lake]

Yes, its 0shot

look from this convo #🔊│ai-development message

F5 is better than fish speech 1.2, but 1.4 (the latest open source) is better than it

idk if the site uses 1.5 or smt, surely better than it

[chilly lake]

so it needs a reference audio and then also a text from the reference? and not just audios, but np arrays

you're killing me, bill

[gray rover]

Naturally

[covert lake]

Iirc it gets automatically the text reference

Like in F5

[chilly lake]

it does not in F5

you have to provide it in the script

[tacit falcon]

it was very clear it wasnt coming long long long ago

[chilly lake]

@covert lake this Fish is taking literally forever

cant say the quality warrants such time

[covert lake]

In the HF Space at least yea

[chilly lake]

on local I have to run 3 scripts - one that converts reference to npy, another for llama with reference .npy, reference text, text to generate - that takes very unreasonable time, then another script to infer .npy from llama

[covert lake]

Huh? I run it on hf space and my own colab notebook was pretty fast

I just used the webui

It's easier

[chilly lake]

i'm trying local

[covert lake]

Ye use their webui

[tight cypress]

gonna do comparison between Fish and E5 F5

[chilly lake]

this morning f5 did something stupid

[wide jolt]

hii

[ionic pumice]

https://youtu.be/-wt2q_jmIz4?si=AIZmTvw5bAgfihdL

skibidi

[chilly lake]

yeah.. this fish is taking way too long to make it usable

[covert lake]

yea fish is better

[covert lake]

what's ur pc gpu?

u using their webui right?

[chilly lake]

using a script, 6700xt

patched zluda in as usual

[covert lake]

amd sucks for ai

on T4 worked pretty fast

[chilly lake]

it is what it is, xtts does the same thing in 1/6 time

[covert lake]

fish speech has better quality tho

1.4 prolly the best 0shot

[chilly lake]

perhaps... but not at 6x speed

what does it spend time on anyway? result of llama is lik 30kb list of tokens

[covert lake]

that's an amd skill issue ngl

want me to share u the colab notebook?

[chilly lake]

audio generation is freaking fast, llama though... just kill me

fish

the usual suspect

[tight cypress]

well, E2 F5 is leaking to the RAM in my setup

[chilly lake]

anyway, sound quality in fish may be good, but quality if the read is 7/10

pauses after commas are not quite right

[covert lake]

better than F5 tho

f5 sounds more unnatural

[chilly lake]

I had much better results from f5 yesterday, today it makes some weird pauses

[covert lake]

it has period sometimes

[covert lake]

jokes aside ig it hallucinates

fish speech never hallucinated for me, wbu?

[tight cypress]

I mean the Gradio F5 automatically enable silent cutting

[chilly lake]

different generation method

fish has inference in llama for tokens, aftet that it just voices the tokens

others convert text to tokens and do inference on tokens

which usually leads to hallucinations

or at least to some variance in the output

fish is rather bland read

xtts sound can be fixed with rvc on top of it

at least improved, the default is 6/10 in my opinion

maybe using a better reference quality can help, I was just screwing with differently pitched samples

[covert lake]

and determinat-

xtts one has some variations

u using fish speech 1.4 right?

[hard oar]

any real time voice changer for dc vc?

[covert lake]

what's ur pc gpu?

the best one is wokada, that uses RVC models, especially the fork (modified version which is optimized

[hard oar]

is there any websites?

[covert lake]

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

its better to check ur pc gpu first

if u do it locally (on ur pc), u won't have to worry about any GPU time limitations as the websites can't be free unlimited time forever

[real iron]

Guys any indian girl voice model

[hard oar]

bhai tharki nikla

[hard oar]

yeah, thanks for the info... just gimmie few mins ill brb

[covert lake]

alr, tell me the gpu name when u do it

[real iron]

Dude i want one voice model yaar

[hard oar]

yep

[feral halo]

Ayo? @hard oar level 2 !!!

[hard oar]

jk, if i got any, ill let yk

[real iron]

Dude this e girl vc models are soo bad yaar

[feral halo]

Ayo? @real iron level 4 !!!

[covert lake]

are u on mac perchance before u do that?

[real iron]

Do you have any girl voice model recommendations/

[hard oar]

yeah im on a mac.... and im a bit afk rn..... not doing anything rn

[covert lake]

oh u dont have to check it then

[hard oar]

fr they are not suited

[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

READ the guide, the mac part

[hard oar]

try smtng in #🔍│find-models

[covert lake]

(there arent any yt tuts)

[hard oar]

oh

thanks for the help

[real iron]

Did yaat but not that good

[hard oar]

y dont u try hindi singers voice?

[covert lake]

yw

[real iron]

Eh where can i get

[hard oar]

@covert lake you know any hindi girl voice models?

[covert lake]

no

[hard oar]

oh ok

[toxic nova]

Hello, I would like to share a music video to know your opinions. Which channel should I put the video in?

[covert lake]

#1159290752195633273

i suggest to upload it on yt and share the link here

dont directly upload the file for copyright issues

[toxic nova]

Can I just put the TikTok or YouTube link?

[covert lake]

yes, upload it on tiktok/yt and share the link here

[toxic nova]

in the canal with the rocket

[feral halo]

Ayo? @toxic nova level 1 !!!

[covert lake]

yea #1159290752195633273

[toxic nova]

thank you it's done

[covert lake]

yw

[toxic nova]

il ya des fr ici ?

[covert lake]

https://discord.com/channels/1159260121998827560/1159520446799155350

[toxic nova]

t'es fr ??

[covert lake]

no, i used the translator

[toxic nova]

have you seen and listened to my video?

[solar torrent]

John Cena and Kai Cenat got taxed by Fanum.

[elder willow]

I know this might sound crazy but if you can make AI play animal crossing, you can basically make AI replace humans

[heavy zephyr]

I have a problem with gui easy, if I create an index it is fast, but the model takes an hour

[covert lake]

not really

[covert lake]

you mean easygui google colab?

That's not an issue... it's normal that it takes hours for training a model

not everything is super speed fast and easy

[heavy zephyr]

yes

[covert lake]

i suggest u to use kaggle btw

its harder than colab but gives 30 hours weekly, so u dont risk losing stuff

[covert lake]

yea and its completely normal

its not a problem

[heavy zephyr]

I saw on YouTube fast why is it slow for me and I want the same but fast

[covert lake]

Youtue Tutorials are mostly outdated for rvc

What tutorial did u even watch?

[frozen egret]

i just wanna say, use weights.ai ai models and their voice converter

i've been using them and so far it's way easier and much managable to use than rvc

[covert lake]

he's making a model

not using one

[frozen egret]

oh, well nvm then,

[covert lake]

For training on weights.gg, you can:

• Pay
• Make a creation for 5 days streak to get a free premium training, you can repeat this
• Use referal code, getting 5 friends to subscribe via your training

but weights.gg just uses RVC so

[heavy zephyr]

I want to apply the same easy gui interface and fast or website

[feral halo]

Ayo? @heavy zephyr level 2 !!!

[willow cape]

hi

[frozen egret]

yea but rvc gardios takes longer to set up and most of the time give differing results

i made needle (bfdi) sang havana and in weights it was good but then in the gardio rvc it was BAD

[heavy zephyr]

I left the computer for two hours and it told me error

[frozen egret]

i didn't really care with training

[covert lake]

explain the error in #✨│ai-help (its in the colab output cell, not the ui)

and also send the google colab link in it

and also the yt video u watched, bc its prolyl edited if it 'trains super fast' lol

[covert lake]

gardio? you mean gradio ? gradio is a framework for building uis, its not an rvc version, its what rvc uses for uis

and the difference doesn't matter in quality as its the same program, unless u didnt clean the vocals, cus from what i remember weights.gg auto cleans them (which u can do that urself)

[wild vine]

I am looking to generate some ai song cover, I found some audiomodify site but maybe do you know some better alternatives ?

[covert lake]

audiomodify is shit, all of those sites just use RVC

which is open source and free

[lean viper]

hi im new

[covert lake]

what's ur pc gpu?

[covert lake]

hi new

[lean viper]

guys i need a problem

[covert lake]

u mean need help?

[lean viper]

where can i send image?

[covert lake]

you can't
You need to level up to get perms, by chatting

1: get perms to send images in help channels
5: get perms in also other channels

in like 5-10 messages u will be able to send them in help channels

[feral halo]

Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote

[lean viper]

but, i need a problem whit the voice changer,

and i dont know...

[covert lake]

yea explain ur issue in #🔍│help-w-okada

very soon u will reach level 1 and be able to send messages in that channel

[feral halo]

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

[covert lake]

yea like in 1 message

[solar torrent]

The big leak from Game Freak, I think the name of this game company has another whole new meaning now.

[rare sorrel]

🖥️ Local stuff

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

Credits to Faze Masta and Antasma for compiling these links.

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

[solar torrent]

[young otter]

[timid wave]

what rvc is best for non english language

[barren star]

Guys is dubbin ai good or rat?

Or what realtime voice changer you recommend?

[plain cove]

supervisory board 2

[plain cove]

-help

[rare sorrel]

Command has expired

[plain cove]

-realtime @barren star

[rare sorrel]

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

[covert lake]

what's ur pc gpu?

[covert lake]

rvc v2 is the best version of rvc

which is what everyone uses

depends on the mode quality

[feral halo]

Ayo? @timid wave level 1 !!!

[timid wave]

im really carefull when i comes to downloading stuff on my pc how can i assure that im downloading the right one and not a virus

[covert lake]

what's ur pc gpu?

and also don't follow RVC Yt tuts as most of them are outdated

[timid wave]

3050 i think

[covert lake]

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

and also the memory gpu, and if u could tell me if its a laptop or dekstop gpu

[timid wave]

umm its a RTX 3050 OEM 8gb gpu memory and desktop

[covert lake]

yea good enough

-rt

[rare sorrel]

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

[timid wave]

hwat should i select?

[covert lake]

1st link, its wokada, which uses RVC models, its the fork (modified version) which is optimized in performance

And its safe, its even Open Source meaning anyone can see the program code

[timid wave]

its this it?

i can send ss mb

[brave anchor]

what is the voice program to download

I'm new

[covert lake]

what's ur pc gpu?

[brave anchor]

AMD Ryzen 7 3700X 8-Core

[feral halo]

Ayo? @brave anchor level 1 !!!

[brave anchor]

?

[covert lake]

that's a cpu

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

[brave anchor]

gtx1650

[covert lake]

alr that's good enough

-rt

[rare sorrel]

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

[covert lake]

1st link

[obtuse wigeon]

@twin otter I just realized that the channel is for 2 people.

I'm sorry

[still inlet]

which collabs let me upload multiple audio files for inference, like a folder with 5-20 second speech clips?

[verbal moat]

is anyone good at ai prompting? I am looking to pay someone to generate some ai art for me, i'll give you the details & pictures of what i want

[covert lake]

https://discord.com/channels/1159260121998827560/1218902018673086524 can help

Also maybe u should ask in #🔍│help-ai-art or #1159290752195633273

don't share instrumentals here for copyright issues

Upload the file to yt, and share the link in #1159290752195633273

[covert lake]

you can try applio colab https://docs.ai-hub.wtf/rvc/cloud/applio-colab/ , rvc has batch inference but there's no guide about it tho

[polar lava]

runway ml down for you too?

[covert lake]

its basically inference for multiple times as the same time, but never used as there could be times u need to play with the pitch

[covert lake]

i don't really use that but the site works

[glass junco]

i made the beat tf

[tall root]

Hey y'all, where can I share my first ever video AI creation?

#1159290730200707153 message

Have shared here; appropriate?

[polar lava]

thanks, seems to be only when you log in

[devout bluff]

is there any fix to the slight glitchiness when talking in w okada? or is this just to be expected at this level

[polar lava]

works again 🙂

[blissful pier]

Hi guys,

Can I ask how to setup tect to speech, how to make a charecter panic?

[covert lake]

There are different Text To Speech (TTS) AIs:

GPT So Vits: RVC isn't as good as GPT So Vits for tts, but gpt so vits (few shot tts, which means needs just a lil training for models) can't use rvc models (and viceversa), and its only limited to: english, chinese & japanese, if you wanna check gpt so vits instead, read https://docs.aihub.wtf/tts/gpt-sovits/

Freemium 11labs: A easy way to do TTS is https://elevenlabs.io/, you can't use RVC model on this but its a mostly premium easy way for good quality TTS

FishSpeech: FishSpeech is a 0 shot (no explicit training needed) TTS, if you got a good pc you can use it locally else use their site

With RVC Models:

RVC is natively for Speech To Speech, but forks such as ilaria rvc mainline & applio have built in tts (using Microsoft Edge TTS to make a generated tts audio, which i suggest you to choose a tts model that is the same gender and language of the rvc model you wanna use, and then convert it with rvc)

If you wanna do tts locally with RVC Voice Models (if you got a good pc):

• You can get Applio in our docs
• While Ilaria RVC Mainline here (no guide as of right now)

If you don't got a good pc you can do tts with RVC Voice Models on cloud:

• Ilaria RVC Zero (Running on A100 GPU, free fasted rvc on cloud) and the guide

• Use Applio UI Colab (with google colab T4 free daily limit gpu)

• if you don't wanna use edge tts, you could try another tts ai from our tts index and use the output as an input in rvc

[runic niche]

I want to find mortal combat raiden voice model, can anybody advice where?

[clear phoenix]

just joined and i got a new pc with 4070ti super. whats a good speech to speech voice changer for real time?

[rigid granite]

👋🏻

[runic niche]

I am looking to generate some AI song covers. I found some audiomodify (https://audiomodify.com/) sites, but maybe you know some better alternatives.

[covert lake]

audiomodify is just one of those scam sites that just use RVC (free and open soruce) and make you pay for it lol

[undone hill]

hiyaaa

[covert lake]

What's your PC GPU and what are u lookin for

#1159290752195633273

[elder willow]

hi guys

[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.aihub.wtf/essentials/how-to-make-voice-models/

[covert lake]

wokada fork

-rt

[rare sorrel]

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

[covert lake]

1st link

[rugged garden]

its been a while since i've made an ai model, how do i make one? there seems to have been large modifications on how you create your own ai voice model

[covert lake]

i mean rvc didn't change much

What's ur pc gpu?

[rigid granite]

Okay which one is the best english girl voice for real time