🧬│ai-chat

[terse pawn]

???

f

whats the best sets for a girl voice changer dosent have to sound like a egirl

or model

[jovial depot]

Ew

@terse pawn you don't understand the concept do you

[terse pawn]

???

[jovial depot]

It's not always about which model is the best

[terse pawn]

i js said

whats a good set or model.

[jovial depot]

It's about how well your GPU runs it

[terse pawn]

which sounds like a female

i have a 4060

[jovial depot]

Any model is fine

[terse pawn]

it dont gotta sound like a

egirl

[terse pawn]

do u have any recommendations

or

?

[jovial depot]

You meet the minimum requirements but this isn't going to get you the highest performance

I think a 3060 12gb is minimum right

[terse pawn]

i believe

so

[jovial depot]

It's only one gen more and runs worst

[jovial depot]

So I would suggest trying out a bunch

Markiplier

[terse pawn]

i tried alot

[jovial depot]

V2 is my favorite

[terse pawn]

but

idk how to

the sets

dont sound rig ht

[jovial depot]

What criteria were you looking for

[terse pawn]

js female only

[jovial depot]

Turn the delay higher

Wait hop into a call

[terse pawn]

yes

1 sec

let me get my mic

[gentle trench]

What voice changer do u even have currently?

[terse pawn]

js discord

only

[terse pawn]

voice changer demo

rn

cilent

yes

js discord

[gentle trench]

Yikes

That's old

Like ancient

[jovial depot]

Get your mic and hop in

[jovial depot]

Yuh

[gentle trench]

U can help em out of u like but I don't help out the girl voice model trolling people

[jovial depot]

Ok

I think he might

But I'll try to wsk

Ask

[gentle trench]

Give em Vonovox or smth

[jovial depot]

Yeah it's decent

[gentle trench]

Do the -rt command in AI help chat btw since this is just a normal chat but for ai

[jovial depot]

I like a silly fork that works and is easier to install on arch or cachy os

Deiteris I think it's called

I feel it's simpler

Than the main branch

[gentle trench]

I use tg fork honestly it's a bit better than deiteris

[jovial depot]

Anyways I'll move this to the other chat

[jovial depot]

Alr send it over

I don't use it often

[gentle trench]

U have Nvidia or AMD?

[jovial depot]

And it's only to tell friends stupid stuff as markiplier

[jovial depot]

Dude

All amd

I use the changer

[gentle trench]

Alr

[jovial depot]

Just give me the main page

🙁

[terse pawn]

hi

i got muted

yes

can i send u my screenshots?

[jovial depot]

Go to ai help

[sweet folio]

dumb question, how do I actually use these voice changer files lol, any recommended programs

[gentle trench]

depends on ur gpu

[sweet folio]

i have a 4070ti

is that ok

[gentle trench]

yea

[jovial depot]

honestly not a bad choice

[languid basin]

hello

[broken pike]

Yoo

[solar torrent]

NVIDIA GeForce RTX 4070 Ti has 12 GB VRAM, a step up from RTX 4060, so not always a bad one.

For help about W-Okada realtime voice changer or Applio RVC, use #✨│ai-help.

[spring glen]

Yooo

[crimson furnace]

guys how i can train my own ai voice?

[solar torrent]

For help about training an RVC voice model, the #✨│ai-help is a proper chat to discuss about it.

[shy mica]

Hello everyone!

[gentle trench]

Is this Murphy's law ↑

[wise trench]

i cant believe its already been like 2 years since i discovered this whole rvc voice cover thing

[gentle trench]

ew

[gentle trench]

give 3 good reasons

[frail tusk]

Do you mind imparting some wisdom on me?

Also has anyone tried ace studio?
I’m trying to figure out what the best vocal cloning tool to use is

[gentle trench]

applio is the current best voice cloning tool

[polar flax]

mods were sleeping

[hidden surge]

https://tenor.com/view/bobby-lee-science-alert-mad-tv-eyeglasses-gif-17646670

[forest crypt]

https://tenor.com/view/bobby-lee-science-alert-mad-tv-eyeglasses-gif-17646670

[gentle trench]

u are not from half-life

[hidden surge]

I got out

[gentle trench]

G-Man is free

[balmy drift]

I'm trying it out, and I like it so far! It's even accessible to screenreaders! However, I wonder why it opens a page in my Internet browser instead of opening locally as a desktop app on my machine.

[polar flax]

because it's not a desktop app

but it can actually run without internet connection

[balmy drift]

So what you're saying is, I can run it in something like Firefox or Microsoft Edge while offline?

[polar flax]

have you heard of localhost or like 127.0.0.1 ?

Applio actually runs on that kind of address

[balmy drift]

No I don't think I have. Those names sound new to me.

[polar flax]

https://tenor.com/view/google-exists-google-search-it-geekboy-geekboy-cz-gif-25933207

[balmy drift]

I'm not sure what's in that image. As I use a screenreader (Yes, I am sight challenged), it doesn't recognize images, only text-based parameters.

[polar flax]

forget that, I don't actually get used to dealing with ppl requiring such accessibility tools

if Applio running in the browser actually supports screenreader accessibility, that wouldn't be an issue to talk on

but if you need help for something, please go to https://discord.com/channels/1159260121998827560/1159290139609137264

[still sigil]

@bleak fractal no advertising

[bleak fractal]

Where do I post videos then?

And it's an AI cover I made of an Indian scammer singing Mundian To Bach Ke

[still sigil]

as a file, in #🌇│ai-videos

[bleak fractal]

ok

[queen ferry]

yo what happened with that older rvc voice changer UI the blue one

[bold lily]

what blue one blawg

[queen ferry]

i think this one

it was so good

[minor blade]

Pretty much outdated and abandoned since long time ago

[solar torrent]

I disagree. This realtime mode in RVC GUI is long older than other "W-Okada" versions and even Vonovox.

[ebon nacelle]

What's the best zero shot singing voice conversion tool now? I tried seed-vc but it's got some weird crappy quality to it. I also tried a few new projects that are based on seed-vc but they're all broken

[quasi shuttle]

What model can run image to image on 3080?

[solar torrent]

NVIDIA GeForce RTX 3080 has 10 GB VRAM, though some rare variants have 12GB VRAM, so almost all Stable Diffusion (ranging from 1.5, XL to Flux) might work but with some optimization of XL and Flux.

[quasi shuttle]

I have 10 gig ones sad sad

[tender monolith]

If you're using a male voice what should you set the tune too?

[gentle trench]

Like if you're also a guy or if you're a girl

[tender monolith]

also a guy

[gentle trench]

Bc it's dependent on voice

[gentle trench]

Just keep at 0 for most voices then

[tender monolith]

well whats the usual ranges i guess

[gentle trench]

I meant this as in like, your voice not the ai model

Like if your voice is deep or higher pitched

But pretty much all guy voices are good at pitch 0

[tender monolith]

i cant ever find a good tune for most voices and just sounds like doodoo

[gentle trench]

Tune?

U might be using a old voice changer bc it's called pitch in the ones that are used now

[tender monolith]

wheres the link for that one sorry

[gentle trench]

It's dependent on what gpu u have, like AMD or Nvidia

But for help with that we gotta move to https://discord.com/channels/1159260121998827560/1159290139609137264

[tender monolith]

i messaged there :))

[oak oasis]

i cant hear myself on discord

[summer star]

h

hello

[uncut dagger]

Guys i have a 19 second video , i want to upscale it...

How can i do it

[covert delta]

im not getting any sound from virtual aduio cable

[night geode]

what if the real rvc3 is the friends we made along the way

[forest crypt]

https://tenor.com/view/cat-gif-24525442

[gentle trench]

Car

[humble lava]

Is this server about AI or just random things and random ppl?

Can someone please help me with n8n. Google sheet doesn’t work and idk why

[solar torrent]

"AI Hub by Weights" Discord server is about AI and things, but better check your internet connection before you start talking.

[solar torrent]

RVC³

[solar torrent]

For help about W-Okada realtime voice changer and Virtual Audio Cable lite, go to #✨│ai-help.

[solar torrent]

See #🧬│ai-chat message.

[fallow rock]

guys which is the best egirl model, i wanna troll my friends

i dont find the E-woman model whcih is recomended by many yt vids

[polar flax]

we don't support for such purpose

https://tenor.com/view/boykisser-you-like-kissing-boys-don't-you-cat-lowpoly-3d-gif-9960745507421464646

[fallow rock]

alr

[solar torrent]

That doesn't always mean it's a bad one. RTX 3080 is a tier above and faster than RTX 3060, although there's a variant of RTX 3060 that does offer 12 GB VRAM.

[solar torrent]

Bless you.

[dawn vine]

heyy can anyone tell me why when i try to start MMCserversio its like not working

[timid forge]

someone guideee me how to do the voice changerr

[solar torrent]

For help about W-Okada voice changer, because I've been saying this over 100 times, go to #✨│ai-help.

[bitter wagon]

If i train a voice model using weights.com (partner of ai hub), after training does it gives the model .pth and index file once its trained? so i can run on w-okada?

[unborn vigil]

👋

[gentle trench]

No please god no, don't train a model on weights! Use applio

But yes it will give both of those files I believe

[bitter wagon]

why? just bought credits lol

[minor blade]

Yeah, use applio instead.

Either on colab, local or kaggle notebook

[gentle trench]

It's just kinda bad

Like compared to applio it's worse quality, it's payed service which applio isn't, and just in general has limitations

[bitter wagon]

umm ok ill use applio, ill get refund from weights

[oak oasis]

i still cant hearmyself on discord

[gentle trench]

Leo, we saved a life today

[fluid mist]

@cinder fox yooo if i send you this ken carson model tomorrow can you upload it on here this #outdated-model-maker-role take to long

[gentle trench]

There's only a singular QC I think which means there's pretty much nobody ever to review models

Hopefully they roll out an update to getting the role via votes instead of QC (the extinct species in this server)

[fluid mist]

@polar flax hey seiso sorry for bothering you this fine afternoon but i just asking if you can check my #outdated-model-maker-role it been 6 days and Ihave been waiting patiently to be accepted

[grave remnant]

👋

[gentle trench]

Change ur password

[polar flax]

mods were sleeping smh

[polar flax]

I don't have permission to accept/reject the applicants

[polar flax]

and pls stop pinging random person that may not be responsible of such the thing

[solar torrent]

[last onyx]

Yio

[humble lava]

That didn’t help at all

[polar flax]

there could be helpful things here https://discord.com/channels/1159260121998827560/1159290139609137264

[humble lava]

So there are different ppl then here? 😂

[solar torrent]

Not a sarcasm.

[weary geyser]

guys do u know any ai tools make ur cover for any song but be free

[lyric sonnet]

-colab

[pine acorn]

📒 Google Colab Notebooks

Google Colab is a Cloud (Remote Good PC) Service. While the Free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

• **Applio**

by IA Hispano
Google Colab

• **RVC Mainline**

by Hina
Google Colab

• **UVR5 NO UI**

by Eddy
Google Colab

• **UVR5 UI**

by Eddy
Google Colab

• **Wokada Deiteris Fork**

by Deiteris & Hina
Google Colab

• **Hina's Modified Original Wokada**

Google Colab

• **RVC-AI-Cover-Maker-WebUI**

by Shiro & Eddy
Google Colab

• **FaceFusion UI**

by Nick088
Google Colab

• **FaceFusion NO UI**

by Nick088
Google Colab

• **Music Source Separation Training (Inference)**

by Jarredou & Makidanye
Google Colab

[haughty pine]

anyone know why when I use RVC it keeps like glitching?

I just wanna sound like a dude

[teal dock]

Yeah RVC does that sometimes It’s usually the model or sample rate try 44.1kHz and a cleaner input @haughty pine

[haughty pine]

okay I'll try it ty it just sounds like it skips then don't idk how to explain

[teal dock]

Yeah that skipping is pretty common It’s usually buffer or sample rate mismatch make sure your input and output are both 44.1kHz, and try increasing the chunk/buffer size
Accept my friend request @haughty pine

[polar flax]

you still haven't listened what I said #✦│chat message

[polar flax]

also

[haughty pine]

No I did lol it did not fix it

[polar flax]

https://tenor.com/view/spongebob-creed-in-your-dreams-dreams-gif-25984561

[polar flax]

then what is this channel?

is it as what I said?

[haughty pine]

maybe < 3

[nocturne temple]

I didn't knew my local AI can make 3D rendered images

[young canopy]

I’m working on an ai tv series for fun

[whole flower]

artificial intelligence

[proven belfry]

I have build few projects using Claude and Gemini. How can I showcase them?

[unborn vigil]

👋

[sweet quest]

i can't use w okada, anyone can help?

[abstract bane]

👋

[vast flicker]

👋

[fossil bay]

Would okada work with love dolls?

O.o curious

[gentle trench]

with what...

[fossil bay]

stuff

<.<

just asking since those do have a.i. apps and whatnot and having anime voices would be cool

[gentle trench]

what is love dolls

[gentle trench]

but uh ye wokada does have voice models of anime characters if u import them from here into the software idk about that tho

[mellow coral]

can speaker split io can also be use for songs and not broadcasting?

[umbral marlin]

Could anyone pls tell me which AI platform was used to create very high quality realistic AI videos like this?

https://vt.tiktok.com/ZSPvg7QQp/

Farzan Films

I want to find the right AI platform to generate AI videos exactly like those

[gentle trench]

U could just use sora since it's pretty good

[umbral marlin]

Not available in my region app store

[gentle trench]

I'm sorry for your loss

[polar flax]

there are some other alternatives like Grok, Veo, etc.

[dusky silo]

I found the song for English Male Singing 2

[gentle trench]

What is it?

[dusky silo]

You by Zekk, if it doesn't show up on spotify the album is AD:HOUSE 9

I have a playlist of the songs as well

[gentle trench]

Oh nice

What sample is it btw on weights? The male singing 2 one

I'm on phone rn so I can't check

[dusky silo]

yeah

[gentle trench]

what

found the song on youtube btw

not bad tbh

[dense ginkgo]

Helllo

[glacial swift]

Is the rvc thing realtime voice chanfer?

[dense ginkgo]

Any good site to say if a youtube video is legit or IA made ?

[gentle trench]

what gpu do u have

[glacial swift]

Rtx 3070

[dense ginkgo]

Rvc doesn't real time voice changer,almost everyone thinks it does at first lol

[gentle trench]

alr, to get more help btw we gotta move herehttps://discord.com/channels/1159260121998827560/1159290139609137264

[gentle trench]

yea that's wokada but all the new peopl who don't know the difference just call it all rvc

[gentle trench]

just use sora, it's free

[solar torrent]

I only heard of ZenMarket, the website that imports products from Japan to international customers from around the world outside Japan.

[icy badger]

Where is server tag ?

[light bone]

is it worth to train model on spin v2 and refinegan?

[sage carbon]

where's server tag

[eternal prawn]

Merry Christmas
Frohe Weinachten
Веселого Різдва
Feliz navidad
Mutlu Noeller
Buon Natale
Sretan Bozic

[gentle trench]

this is true

[eternal prawn]

The server is at lvl 2
must be at lvl 3

Server Boosten Please

[sturdy summit]

Yo I've a question.. I'm using Vonovox-1.6.9 but I wanted to ask if there is any new or better program than it?

[abstract bane]

👋

[gentle trench]

https://tenor.com/view/nuh-uh-beocord-no-lol-gif-24435520

[sturdy summit]

Alr

[steel canyon]

Hey guys, am trying create an offline AI assistant. I already got the model but trying to figure out how to integrate it into the laptop's system. Any ideas 💡

[night geode]

idk

try asking chatgpt

[prisma owl]

Hey guys Merry Christmas Eve!

[plucky trout]

hi everyone

[jagged cobalt]

What ai is free generates pictures?

[spark dock]

who have ai model ivanzolo?

[ruby wharf]

@glad nebula Hey bro I was wondering how I can get model maker role? I’ve been a model maker in the past just don’t know how I can become one again sorry for pinging you please let me know 🙏

[glad nebula]

#outdated-model-maker-role message

[ruby wharf]

It says no access

[glad nebula]

uhh maybe they removed the model maker applications? lol

no idea

[ruby wharf]

Is there someone I can ask to get the role? Or did they change something

I wonder why

[glad nebula]

@dawn temple

[jagged cobalt]

Paying people to make ai for me dm me

[dawn temple]

Done

[ruby wharf]

Appreciate you Brodie

Thank you

[rancid bane]

Anyone knows why google's flash-lite model wont work anymore?

[shy mica]

MERRY CHRISTMAS GUYS WISH YOU ALL A GOOD DAY AHEAD AND ENJOY UR TIME WITH FAMILY AND FRIENDS!!🤶🎅🎄🌲🎉🎉☃️⛄❄️

[chilly bolt]

am i allowed to use this ai thing for real time voice changers?

or is that not allowed

cuz tbh i just wanna get a nice male voice

[vocal hornet]

fam

does ur voice lags

while speaking using the voice changer

[hollow kernel]

Kaggle says that "the Preprocess completed in 0.01 seconds on 00:00:01 seconds of audio"

I have a problem

[buoyant karma]

Hey! I am working on a project which will provide the most robust and authentic humanization of AI generated text using any way possible. I am looking for collaborators - who have knowledge of AI, statistics, linguistics, mathematics or anything which can be remotely used in such a system. This is dual-use project only for demonstration purposes.

[crude root]

What are like good GPUs for this? 3080?

[night geode]

better the gpu less delay you have, as long as your gpu isnt ancient any gpu is good

[crude root]

Well I currently have a rx6700xt but voice changer + game is basically not really possible

I used to put in my GTX 1660 as my 2. GPU and that worked quite well but I don't rly wanna continue doing that

[lethal hawk]

👋

[wild tide]

👋

[real edge]

👋

[sonic sky]

hey, can you help me?

how to train voice ai model? You know google colab is already gone, and i searching other method

[stoic urchin]

I'd also like to know; my project is also stuck at this stage.

[stoic urchin]

If anyone enjoys making short videos using free AI, feel free to discuss and exchange ideas.

[reef bay]

can i talk to staff

or moderators

[still sigil]

dm @serene sluice

[vocal hornet]

what output do yall use with w-okada voice changer

because the voice changer voice is laggy and i figured that changing the output might fix it no?

i currently use vb cable

is there a better alternative that is for free too?

[still sigil]

@wispy granite no advertising

[wispy granite]

Ohh sorry😅

[solar torrent]

What's up, Adi. For help about Applio RVC which is a better RVC GUI fork, ask in #✨│ai-help, or type "-colab" or "-kaggle" in #🤖│bots.

[solar torrent]

See #🧬│ai-chat message.

[solar torrent]

For help about W-Okada voice changer, better ask in #✨│ai-help. Because when you ask anyone in chat they won't always you a proper or direct solution.

[solar torrent]

More like an LLM model, by the way.

[solar torrent]

For a serious inquiry about this server that you need to contact a moderator/admin, like a spammer from here spams in your DM, use the modmail. For help about an AI program otherwise, reply in #✨│ai-help.

[polar flax]

"humanization" sounds more like a joke, it can never be a human, just being crafted sentences or any kind of generated content data.

[surreal tundra]

@slim zealot

[solar torrent]

Not where you promote your thing, lil bro.

[buoyant karma]

Cyber-security and AI engineer, here for VC/DMs.

[rustic pumice]

How do i add voices to okada

im trying to add a villager voice for a video

[solar torrent]

For help about W-Okada voice changer, go to #✨│ai-help. You have been told in #✦│chat message earlier.

[clever meadow]

👋

[novel path]

who can make an ai model

for me

ill payu

pay

[clever tree]

Which ai model you want?

[leaden iris]

Is there any reliable ai 3d designer who can make sketch drawing or art into 3d mesh? With optimistic topology as possible.

Its for game dev

[solar torrent]

Paid request and commission aren't allowed here, though you can still request for a voice model in #1159289738314919936 for free.

[rustic pumice]

oh sorry i mixed it up

[gentle trench]

You should learn how to make one yourself, it's actually quite easy and completely free

[fossil pumice]

I completed a new PC build a few months ago. I had a decision to make between a 5090 and a 3090, I had both on hand. I don’t game much so i decided to keep the 3090. Fast forward a couple months and I’m getting into AI..I decided to run a local ai chat assistant with STT/TSS, persistent memory, web access, some other features..I’m now contemplating using both GPU’s. Just curious about the thoughts of more experienced people here.

I think the idea might be to do inference on the 5090 and TTS on the 3090

It’s a 1300w be quiet! PSU, x870E godlike mobo, 9950x3d, 96gb of DDR5

[full ore]

Just found out they brought back instrumental/separate vocals back. I doubt it's a sign that Weights is healing.

[gentle trench]

u can download them but can't download the seperated ai vocals still

yknow, the important part

[full ore]

I swear I saw Download Instrumental/Backing Vocals

[gentle trench]

u did

but it's only the actual song vocals and ect

[full ore]

It won't appear on mine

I think I need to somehow update

[gentle trench]

relaod the page ig

[full ore]

Nothing...

[gentle trench]

ur weights is fucked then lmao

[full ore]

And that means...

[gentle trench]

no clue

[full ore]

I think they're gonna shut it down for not giving them my credit card number

[gentle trench]

I didn't pay them either

[full ore]

Maybe it's the new update (Old Weights kinda comes back) and they're still rolling it out, and I haven't gotten the update... I need an explanation

[gentle trench]

I have 0 clue I woke up and it was different

[icy badger]

Who's that in your pfp

[gentle trench]

I think that's a woman

[fresh scroll]

can someone help me with the github download? from the duckus video i dont see what he sees

[gentle trench]

You're not trying to download a voice changer from a yt video are u?

If u are it's old and outdated

[fresh scroll]

ah

can you give me a up to date version

[gentle trench]

what gpu do u have? nvidia, amd, and what os do u have (windows ect)

[sturdy summit]

[real edge]

👋

[silk juniper]

👋

[polar flax]

did you buy the ram before the price started inflating crazily?

like around less than $400

[fossil pumice]

I did, thank god. It’s corsair dominator titanium..last I checked it was $1k for the kit on eBay

[real edge]

👋

[silk juniper]

👋

[polar flax]

I hope I don't see any of consumer folks buy ram kit in such insane price and let those scalpers bite the dust till the AI bubble bursts

[polar flax]

back to the topic, the 3090 seems good enough and it doesn't seem worth to invest for running larger LLMs locally, so renting some cloud gpus would be the option

[fossil pumice]

I already have the GPU’s

[left sphinx]

yea the prices will come back down eventually next year

Google is rapidly bringing about the pop of the bubble

[night geode]

you just jinxed it

16gb ddr4 for $1000 in 2026 here we come

[solar torrent]

What GPU? GeForce RTX 5090 or 3090?

[solar torrent]

For help about W-Okada realtime voice changer, tell your PC GPU in #✨│ai-help. Asking anyone for help in chat channels won't gonna be easy.

[solar torrent]

For help or information about Weights website, there's official Weights Discord server. When you ask for the same in here, you can expect another person to put the "blame" on the website for whatever it is. Instead, there's UVR5 for higher quality audio stem extracting; for basic AI cover there's Applio RVC, though this workflow can be more challenging than the easier Weights.com.

[fossil pumice]

Both

[ivory wren]

unless your utilizing the full memory of the 5090 there usually isnt an issue with loading both models at once, might be a tad slow if both are inferencing in parallel, but 🤷 really cant notice.

own a 5090 and did the same tts alongside a 30ba3b model.

and yes, if your concerned about speed, always inference on the 5090.

[solar torrent]

While it's not possible to make RTX 5090 and 3090 to share data each other, like those enterprise-level data center GPUs, you can still assign each GPU for separate tasks respectively; one for hosting an AI instance or gaming while the other for hosting another AI instance.

[fossil pumice]

My plan was to use the 5090 for inference, the 3090 for TTS

[polar flax]

inference with RVC and most TTS models should be viable on entry level gpus like RTX 2060 or 3050

and even performance difference of them to 5090 is still not too significant for most regular cases

[ivory wren]

I did not say you couldnt inference on a top tier graphics card?

[polar flax]

it's just like performance difference of budget SATA SSD compared to pcie 4 and 5 SSDs when coming to game loading speed

[ivory wren]

well for LLM's that is a bit of a misnomer

but among the smaller model size of things thats pretty fair

[polar flax]

I said that the performance difference is not so relevant for TTS

only except some LLM use cases like coding, etc. where the tok/s is relevant

but I think the vram requirement is still more important to be able to run certain LLM size

you wouldn't want to use 5060 Ti 8 GB over 3060 12 GB if the former can't run but the latter can

[ivory wren]

we already know what hardware he has, so I’m a bit confused what the disagreement is

percy has both a 5090 and a 3090, and he already said he might run TTS on the 3090

my earlier comment wasn’t meant to be only about the TTS model, sorry if it came off that way.

what I meant is on a 5090, depending on what LLM you’re pairing with TTS, you can usually load and run both on the 5090 at once without much trouble unless you’re close to maxing out VRAM or trying to run both workloads hard in parallel. As well as if your ever concerned about speed, to run it on the 5090.

because for inference, even if it just a tts model, if its a 7b model like the one originally from vibevoice for long speech lengths, you will want a faster card, because a 2060 is not going to compare whatsoever to a 5090 in hardware+software optimizations nor in memory bandwidth.

🤷

[polar flax]

TTS and LLM are different beasts and have different spec requirements & relevance of the performance difference across gpus

but for considering the GPU alone for various purposes, you may find upgrading from 3090 to 5070 Ti or 5090 might worth it

[ivory wren]

a dense tts model will still have the same memory bandwidth constraints as a dense llm afaik right?

compute can be different, but memory usually isnt differentiated between types.

[zinc oasis]

hey

something ingenius came to life yesterday

[wispy widget]

https://imgur.com/QfhjzpS

[full ore]

Turns out the old download options only appear on the newest creation

[quartz oasis]

guys uh idk if its the wrong channel

but

i got the ai model setup and stuff but when the voice plays

it kinda cut in the middle of my sentences

what do i need to tweak to make it fully track my words

[silk juniper]

👋

[gentle trench]

Press to cat

[fossil pumice]

Is there a rough consensus on the best local TTS model?

I’ve been using piper for development of a local assistant and it seems really bad. Perhaps I just need to work with it directly more but the default voice is very robotic.

[covert lake]

@gentle trench I saw you 🥀 reacting some scam/hacked users messages, if you want you could also ping mods to help ban them btw cat_roomba_exceptionally_fast

[gentle trench]

I try not to bother mods too much sometime:3

99% of those scam accounts can't even post the actual images anyways

[limber nexus]

you got no more nitro

[sullen dagger]

https://cdn.discordapp.com/attachments/1384992007901937710/1413249810093445271/togif.gif

[gentle trench]

[limber nexus]

<@&1159293140440723499>

[gentle trench]

Mods kill that guy

[limber nexus]

sighs

Guess all this comes with a server like thi

[gentle trench]

Sadly

They don't read

[limber nexus]

oh well

[sterile zephyr]

Hey

[left sphinx]

hello

[gusty depot]

If you touch your fat balls you will completely lose your grandpa and this happened because the fat jumped into the firey lava over and over again

[gentle trench]

what...

[solar torrent]

Fanum calculates Fanum tax (foods) of Akashi Junko.

By the way, foods that were owned by Tea Party and related Trinity students also got Fanum tax'd by the Gourmet Research Society (food eaters) team in Kivotos.

[solar torrent]

The #✨│ai-help is the correct channel for help about RVC (retrieval-based voice conversion) and W-Okada voice changer.

[gentle trench]

???

There's no advertising lol did u read the rules?
<@&1159293140440723499> kill him or something idk I'm sleepy

[restive crow]

Hi what is the best ai voice model for me?

Laptop rtx 3050 95W
Im a Male i want to have a Female Voice

[steep mantle]

im ai 67

[sudden hazel]

How can I make the ai voice changer sound more clear and less ai if that makes sense, at random points it blurts out random stuff and rubbish

[solar torrent]

https://tenor.com/view/sixseven-six-seven-six-seve-67-gif-14143337669032958349

[polar flax]

https://tenor.com/view/rule-six-seven-rule-66-67-ao-oni-gif-3085875715560017117

[sullen dagger]

speak better raise the pitch *for more femaleness( raise the chunk size raise the chunk and get a good mic

[solar torrent]

[sullen dagger]

u can also lower input sensitivy

[solar torrent]

The what?

[sullen dagger]

input sens

[solar torrent]

For help about W-Okada realtime voice changer, there's #✨│ai-help. The person has already ask for the same thing in that channel, by the way.

[sullen dagger]

alrr

balright

[night owl]

I’m a senior developer experienced in delivering production-grade intelligent systems, decentralized applications, and full-stack platforms.
Results come from structured execution, not unfinished ideas.
If you need end-to-end development from architecture through production deployment I’m open to collaboration.
I value disciplined engineering, transparent communication, and scalable solutions.
Previous work can be shared upon request. DM me anytime

[dry sigil]

hey there! I have recently finished a whole Music Video, that was created with the help of AI. It's already on YouTube... is here a place, where I can show it and discuss about it?

[pearl glen]

Hi everyone!

I'm looking for someone who can help me train a custom voice model using my own voice recordings.

• I have around 60 short WAV clips (10–15 seconds each) ready to go.

• I'd like the model to be compatible with RVC or So-VITS-SVC (.pth, .index, and config.json files).

• I can send the dataset via Google Drive or WeTransfer.

My main goal is to use the trained model to convert other voices (like Suno outputs) into my own.

I’m open to:

• Free help, if you're available 🙏
• Or a small fee via PayPal (or a prepaid Visa card if needed)

Feel free to DM me if you're interested or need more details.
I can also send a short sample of my voice if you'd like to check it first.

I'm new to this, so any guidance or support would mean a lot.
Thanks in advance!

---

[sullen dagger]

i made hella bands acting like a girl

this so easy

Your_Local_Worm if u see this thanks for setting me up with a working voice changer

alr made 60$

uh

[solar aspen]

Guys, im using gemini 3 flash on silly tavern but every few messages it blocks tge chat with filter, how can i jailbreak it

[sullen dagger]

kinda

[solar aspen]

Sorry im new to this

[sullen dagger]

you cant except u have a rediculous good prompt or just like steal the code and reprogramm it to not have a filter

i mean kinda but have luck finding one

last time i found one is when chatgpt was still dumb af

[solar aspen]

Thats what tgey told me in sillytavern server, i need a prompt yo put in the system 0rompt

[sullen dagger]

aight teach me 2

i need it bad

aight then slide in dms

[solar aspen]

Bless us

I need one for gemini 3 flash

[sullen dagger]

probally but u need it so you can downlaod it and if it has password a d license key and allat then u can prob use a programm to like decode it

alos works on websites (prob)

[solar aspen]

What? Do you mean the gemini jailbreak ?

[sullen dagger]

baright

lemme try

so what now

pasted it in

so just do my thing?

[still nest]

Can someone explain how do I make a own voice
Or like a yt video

[sullen dagger]

g oto youtube "how to make a own voice model on weights"

[still nest]

Thank you

[solar aspen]

Did it work ? @sullen dagger

[sullen dagger]

yes

cool

shouldnt say that in here tho

i make minecraft plugins and mods

alrightttt

qwha

what kind of backdoors

of apps?

trick them into downloading?

or like send someone a plugin to try it

or smt

wym click.?

aight

alright

deklete that

if anyone sees this of the mods ur cooked

baright

aight

aight i dont condome it but you can do it ig

aight

aight

i am NOT clicking that link

MHMMMM

.cf.

[compact copper]

👋

[sullen dagger]

MHMMM

then leak my ip

🙂

ip cant do shit

and IF u cant exactly locate where im at

what tf is neymar jnr1010

thanks

no bro

"m.yt.com"

fym megayt.com?

if you could check my dc dms gng my ip is alr leaked on so many servers

sybau kaffah

fuck is that im german gng

im not türkish or wtf thats supposted to be

türkish?

?

its not my name vro

wym tag discord

no i dont know them

ong

bdw i dont have a facebook🥹

and whats that?

whats VK?

me ban?

alr gng ur so far out ure not even NEAR my name or ip

i never scammed anyone on mc vro

sure send the proof in dms

send in dms if u actually have the proof

alr gtg bath

gng this is not whery i prey on kid

and i dont even do that

ON WHAT SITE

what stream gng+

what streams gng

WHAT STREAMS

[gentle trench]

<@&1159293140440723499> idk what's happening but uhhh

yea

[sullen dagger]

i made him a bet to make him find my e mail

[gentle trench]

?

[sullen dagger]

i did not participate in doxing or leaking private information

[pliant elm]

well they did

now they are gone

[sullen dagger]

yea kinda

[pure mural]

Do you know any collective works related to 3D generator models?
urgently needed

[sullen dagger]

delete my messages too please

i dont wanna look like im speaking with the air

[river gust]

https://tenor.com/view/dono-wall-gif-24048347

[gentle trench]

crazy people smh

[pure mural]

Do you know any collective works related to 3D generator models?
urgently needed

<@&1159293204038955078>

[cerulean orbit]

👋

[gentle trench]

[cinder granite]

ai hub

[viscid pasture]

Yo, im new to AI video and voice models, I'm curious if anyone here has experiance with instagram ai models, that knows how to keep the consistance in the look of the model throughout different reels. Please msg me in dm's if you can help me with the basic knowledge, thanks.

[worthy wren]

h'

[dark pulsar]

Sjjejeksjejzl

[brazen juniper]

😂👌💯🔥

[compact copper]

👋

[night moon]

someone know how to make easly and good a voice model from my voice?

any new method

[gentle trench]

applio

[night moon]

okay

any tips ?

or its so obvious

it will be doing fine with rvc gui for 2025 ?

[night moon]

@gentle trench can i make own voice model in applio? Do u know any good tutorial ?

My applio doesent starting

@gentle trench

[gentle trench]

-rvc

[pine acorn]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[night moon]

i have ERROR: Failed in default startup: Error: Traceback (most recent call last):

pls help

[gentle trench]

I can't help with errors I use it on kaggle

[night moon]

u know any easier method ? i only want to make a ai song with my voice

[solar torrent]

After the ban of another user, you talked like you schizo at my first glance even if you weren't.

[solar torrent]

For help about Applio RVC fork, go to #✨│ai-help. When you ask in chat no one here always gives you a proper answer, to much your frustration.

[polar flax]

https://tenor.com/view/bait-cat-is-this-bait-cat-is-this-ragebait-gif-10902073105296387701

[elder willow]

парни есть новая нейронка дани докенра 666

[gentle trench]

cryptic ah russian

[wary vault]

is there a video on how to install

[solar torrent]

That's a bad response, anyway.

[elder willow]

бля ебало закрой

[polar flax]

https://tenor.com/view/speak-english-pulp-fiction-do-you-speak-it-gif-16440534

[elder willow]

пошли в войс

[elder willow]

call me

[solar torrent]

If you can't speak English, use a translator.

[elder willow]

https://discord.com/channels/1159260121998827560/1159273424053608498

[solar torrent]

And by a translator, the text is the easiest to access around, while the realtime voice language translator (not realtime voice changer which uses RVC/Retrieval-based Voice Conversion voice model) has not yet available or limited to more powerful and paid services.

Sorry, but if it's nothing personal I don't accept your private direct message either way. Not everything I say would mean "I am interested in buying your accounts" type shit like that.

[solar torrent]

For help about voice models like RVC, that's one of the things I know, you better ask in #✨│ai-help or #1192011222023950368. When you ask in chat there's a chance of you either getting scammed or trolled.

[solar torrent]

As of now, I don't make 3D model because I don't have a powerful PC.

[solar torrent]

Google Gemini is not your typical roleplay LLM. For text conversation, I can only focus on certain basic interactions (like asking about meme crossovers). Gemini Pro is another model but this one is slower since it has to think more before respond anything.

[bronze ibex]

hello guys

[solar torrent]

Hi.

[fathom meteor]

need turkish egirl voice....

[solar torrent]

[smoky silo]

Hi

[left sphinx]

@bronze ibex@solar torrent@smoky silo hello 👋

[dry sigil]

hi... where can I check, which level I have?

[brazen juniper]

piper.pepper's active items

No active items right now

[dry sigil]

ahh... got it

[shy mica]

hi

[shy mica]

Is anyone good at coding here?

I am a good coder but I need some1 else

For this project

nvm

I will do it myself

[feral mica]

https://nextgenaiinsight.blogspot.com/2025/12/ai-innovation-redefining-customer.html

[solar torrent]

https://tenor.com/view/meme-horse-you-have-alerted-the-horse-pretty-derby-uma-musume-gif-15063465391514740847

[sullen dagger]

yap yapyapyapyaqpyapypapypapypapypayap

[undone stream]

wher do i load the gui

[leaden iris]

Bro what u want to make?

[hard crater]

is it me or dose meta ai makes hell of a good videos for being free?

[night owl]

Hi. I’m a senior developer working across AI systems, decentralized systems engineering, and end-to-end product development.
if want my help, please feel free to ping me anytime.

[primal ibex]

is there like a very realistic man voice?

[midnight elbow]

guys, what's the difference between rvc and w-okada?

[gentle trench]

wokada is the realtime voice changer

it uses rvc models

[ruby shard]

how can i make my voice changer sound clear

my mic is crazy good

[night geode]

are you using nvidia or amd

[ruby shard]

amd

sorry for answering late

[night geode]

have you tried turning on force fp32 mode and increasing crossfade to 0.15

[ruby shard]

No, i didnt.. i just downloaded the voice changer

And i also just went off my pc

But if you know any other things that can make a quick fix other then that, tell me

[night geode]

put extra to 2.7

[winged geyser]

is any1 active rn

how do i use weights website to download a voice model?

[winged geyser]

ty yea i didnt see the “download anyway” option bc it was so dim😭

is there any accurate eren yeager ai tts model? @gentle trench

i make among us tiktoks and i’d need it for that basically 😭

[gentle trench]

no clue I don't make anime models lol

[formal hornet]

does anyone know why my w-okada would literally ignore my input device setting and just only work through my headset and not even default to another input when i unplug it

[solar torrent]

You ask me like if it's a real issue.

[blissful mulch]

has anyone tried https://github.com/Plachtaa/seed-vc ? how is it?

[leaden iris]

nah blender is pretty fine on normal pc but if you want to make something with millions of polygons then it's an issue

[solar torrent]

[opal spoke]

what app do you need to make voice models

[solar torrent]

Applio RVC.

[opal spoke]

its free right

[solar torrent]

Applio RVC is the free program where you train (make) a voice model, not just for simple AI cover.

[opal spoke]

alr thanks

i ma make Salman Khan

i jus thought of something better.....🇩🇪 🖌️ 🎨

[solar torrent]

You sure those NVIDIA GeForce GTX GPUs can train model locally? It sounds possible to work but it would be real slower than more-recommended RTX 20 series.

[opal spoke]

well gtx 1080 is equivalent to rtx 2060 8gb but just a little less performance

[zinc oasis]

just finished creating unlmited free api for llm and unlimited free api for image generation ai it will revolutyionize industries

ping me for more

[solar torrent]

Do you mean RTX 2060 Super (8GB)? The regular original RTX 2060 has just 6GB, which is a bit below of GTX 1080, although there's another variant of RTX 2060 that has 12GB. For a comparable GPU, the RTX 2060 Super one would fit better, while any GPU above RTX 2070 would potentially outperform GTX 1080.

[ruby shard]

how do i even get to those settings

anyone know?

[solar torrent]

For help about W-Okada voice changer, better continue the discussion in #✨│ai-help. Because while some people here know a bit about the voice changer, they can only provide you some simple short-term solutions rather than much deeper issue.

[ruby shard]

i asked there already

[solar torrent]

Well, some other dummies made me forget about you, so you can try ask the same question there again.

[solar torrent]

Not everything on Earth is always that free, so if you think if I wanted your free APIs, repeat the first part of my sentence.

[zinc oasis]

but from today it is

u need interent

ofc

[leaden iris]

I'm sorry. really I'm very very sorry..

[lapis bluff]

👋

[plucky fulcrum]

lol, noone gives a shit?

<@&1159293140440723499>

<@&1159293140440723499>

[pliant elm]

gone

[lavish goblet]

I don’t think both needed a mod ping as they were above each other but thank you

[pliant elm]

dunno who or what did it

[plucky fulcrum]

sowwy :3

[lavish goblet]

[lavish goblet]

Just more hacked accounts

[plucky fulcrum]

[pliant elm]

I know

banned one of the two

you did the other

[lavish goblet]

I know

[pliant elm]

https://tenor.com/view/chomp-cute-crab-nom-anime-gif-12018776

[ocean vapor]

yo anyone in here who can help me install properly rvc for amd ?

(ima use it to make songs with an anime girl voice 😭)

did yall get hacked or what

tf

[gentle trench]

There's like at least 1 person who falls for pictures??? Idk how pictures get you hacked but at least 1 per day I see them send images here

It's a stupid scam

Free block for me

[ocean vapor]

yeah whatever

its weird

[gentle trench]

Ikr

[ocean vapor]

yo

[gentle trench]

Are you wanting to train or use a voice changer

[ocean vapor]

could u heklp me installing the updated vers of wokada

[ocean vapor]

use

[gentle trench]

O alr

U said u have amd right

[ocean vapor]

i need rvc for em songs and wokada for vc calls withj othe artists and stuff

[ocean vapor]

ya

[gentle trench]

Btw we should move to here for this
https://discord.com/channels/1159260121998827560/1159290139609137264

[ocean vapor]

oh yeah sure

[gentle trench]

Bc this is just a channel for chatting about ai

[reef minnow]

i want to dm you for AI voice maker

[gentle trench]

I don't really do requests

[reef minnow]

I wannt recomendation for ai that can change my voice

Ideal is women

[gentle trench]

eh

what is ur pc gpu

[ebon spruce]

hello

and how come i cant post my videos on Ai-video

[rugged island]

I think permission/rank? Not sure. I can't either

[sonic salmon]

how can I create voice model with just .mp3 file of voice?

[rugged island]

Elevenlabs

[sonic salmon]

for free?

[rugged island]

I think you get free creds so maybe with a certain amount.

[rugged island]

I think there are free ones now:

https://www.google.com/search?q=bye+bye+elevenlabs&oq=bye+bye+elevenlabs&gs_lcrp=EgZjaHJvbWUyBggAEEUYOdIBCDY2ODdqMGo5qAIAsAIB&client=ms-android-xiaomi-rvo3&sourceid=chrome-mobile&ie=UTF-8

[autumn wasp]

How to get access to showcase anyone 🫠

[delicate raptor]

Bon appetit

[sullen dagger]

https://imgur.com/QfhjzpS

[karmic copper]

Does anyone have a link google colab for train my model??

[simple basin]

Hi guys

I successfully predicted exam questions by accident while using a neuroscentific research workflow with AI in order to study for an exam lol

and I take great interest in AI's analytical abilities

Particularly predictive analysis.

"predictive" being used loosely of course

[leaden iris]

Yes it is. Many characters are free just don't create many accounts.

[leaden iris]

Is it like giving all the previous year question to the model and the model learns it?

Then it gives the prediction of the latest year's question (out feature). Did you made that way?

[solar torrent]

"Free ones"

Does it use Linux? If not, I'm not going in.

[gentle trench]

You use Linux?

[solar torrent]

I don't mainly use Linux.

[plain cedar]

does anybody know how to get rid of lag/glitching of the voice?

like settings that will help please reply to this message

[elder willow]

what AI do yall use for realistic pics?

[elder willow]

Poland

[proven plume]

anyone knwo a rellay good ai for images

[sullen dagger]

Happy new year!

[sullen dagger]

sora2

[jade knoll]

nano banana pro

[coarse trench]

Which real time changer is the best?

@polar flax

https://cdn.discordapp.com/attachments/434113367938957315/1286443496118554686/caption.gif

No image perms 😭

[amber pecan]

ur an helper for a reason do ur job and answer our questions in ai help

[coarse trench]

You're pfp is so cute

[amber pecan]

ty

[coarse trench]

No problem :3

[fluid mist]

i wishhhh i could a free role of model maker

[surreal tundra]

happy new year to everyone 🎊 🥳 🙌 🤝

[bleak shuttle]

I hope you guys know you are contributing to the downfall of all that is good and the sanctity of humanity 😊

[solar torrent]

https://tenor.com/view/drake-kendric-skibidi-rixx-rizz-gif-13501886725531173327

[solar torrent]

Aside from complaining about other staff members, I doubt if you still expect those who know better to provide an answer, as much as you can't think of a simple solution yourself or being any helpful.

[frigid oyster]

a

[tender siren]

which is better? spin-v2 or contentvec?

[solar torrent]

Not quite conclusive. For more help about Applio RVC or W-Okada voice changer, there's #✨│ai-help.

[atomic topaz]

m

@serene sluice sup man

[left sphinx]

that feel when people thought using AI was meant to make things mentally easier but it pushes you to try to do things you never thought you'd be able to do and becomes mentally exhausting

[lime raven]

is there a way i can show off my voice model here somehow?

[zinc oasis]

yo who programs ai apps and needs llm in their apps ? just ping me here

[thick forum]

any1 that use voice changer client know why voice repeats

[gentle trench]

Vc client is old, did u get it off a yr tutorial

[acoustic skiff]

d

[fluid zealot]

what voice changer should i use?

[gentle trench]

depends on what gpu u have

[fluid zealot]

i got a 5060 ti 16 gig

[fluid zealot]

which one is good for a 5060 ti ?

[gentle trench]

vonovox, go to https://discord.com/channels/1159260121998827560/1159290139609137264 for more help tho

[solar torrent]

For help about W-Okada voice changer, better go to #✨│ai-help.

[tropic hill]

hi

[willow belfry]

anyone has turkish voice models?

[solar torrent]

For voice models, in #1175430844685484042.

[rancid marten]

guys what is a good ai model that can watch a youtube video and accually analyze the visuals like to see if its like 3d animation or a video game or irl

[unreal quest]

God has anyone heard of Neuroplay by hexagon

Apparently it allows you to play games with a non invasive bci headset.

https://neuroplay-by-hexagon-ddf550e0.vercel.app/

You can check it out

Technology advancement is so impressive

[safe umbra]

hey guys

help me

how to make models

[safe umbra]

help me twin

[solar torrent]

For help about how to train a voice model in Applio RVC, better ask in #✨│ai-help or #1192011222023950368.

[stiff talon]

Hey I’m starting a new project for Ai agent to book and lead calls if anyone knows how to do that please dm me

[worthy rampart]

Is a 5070 good for this voice chnager

[lilac acorn]

can someone please help me whit Realtime voice changer client

[waxen cliff]

what is the best girl vc

[gentle trench]

ew

[bold lily]

yes

we hate women here

😻

[gentle trench]

men kissing

[true osprey]

https://www.youtube.com/shorts/Xh7PYgvj-QQ

[north pier]

what happened to the tag?

[gentle trench]

what tag

[night geode]

the tag

[gentle trench]

?

[glad nebula]

@rare totem but yea the new rvc stuff i saw is promising, provides better unseen high pitch generalization, so, better singing

@rare totem but yea the new rvc stuff i saw is promising, provides better unseen high pitch generalization, so, better singing

[rare totem]

How come this field is lagging behind image, video, etc, do you think?

[glad nebula]

a lot, rvc is from 2023 and uses 2021 tech

the alternatives are far worse, they're diffusion based, and diffusion kinda sucks for audio

i have tried them

[rare totem]

So all the commercial services like Udimee, etc are still based on RVC?

[glad nebula]

udimee? whats that?

[rare totem]

It's basically voice cloning specialising on singing

[rare totem]

Sorry, it's audimee

[rare totem]

I heard that internally they use RVC, but I don't know if it's true. There are several such paid services

[rare totem]

kits ai is another one

[glad nebula]

hmm it does sound like rvc, the quality is not that good

[glad nebula]

back then they were using rvc but now they're using an edited version created by them

[rare totem]

I think with RVC. it gives you what you put into it in the first place. I've noticed that if you feed pro quality acapellas into a decent RVC model, it will give you great results. But if your input vocals are not up to sctatch you end up with what any other of those commercial services gives you. That's why maybe audimee doesn't sound good

[glad nebula]

true, ive heard very realistic and good quality rvc models before, is not impossible to have one like that, but as you say, it needs good data

diffusion based ai audio models sound very noisy to me

[rare totem]

https://tenor.com/view/gitapro3-gitagita-gitapro1-gitajah13-gif-3421945902795778238

I don't want to take more of your time. I'm gonna go and finish curating my dataset part 2. And will give your LC pretrains a spin. Thanks so much for your help and a chat

[glad nebula]

sure no problem, safest bet is 32k

40k and 48k works and do provide better quality (if your dataset have such data) but they do have problems due to nsf hifigan being too old for high sample rate

[rare totem]

I meant to send that gif differently. I'll figure it out one of these days...😀

[rare totem]

So you think even though my dataset is at 32k, I still should try with one of your higher pretrains (40l,48k)?

[glad nebula]

no, use 32k pretrain if your data is 32k, training 40k and 48k wont improve the quality of a 32k dataset

they only improve the quality if you have true 40k/48k data

[rare totem]

Yeah, I thought so. My set is 32k and 40k which I'm planning to convert to 32k so it doesn't confuse the training code. But you mentioned about trying with one of your higher models just to test?

[rare totem]

Or I misunderstood?

[rare totem]

Like you were saying you never know how RVC will react?

[glad nebula]

yup you can try using the 48k legacy core 1.5 pretrain and train your dataset

if you notice too much unstability (voice glitching, robotic breaths, etc) use 32k

just for experimentation

[rare totem]

I think I'll try your 32k first cos its the sensible choice, then see if your higher ones give any improvement

[glad nebula]

yea do that, the reason why i told you to try the 48k is so you dont lose the 40k data and keep some of the high sr clearness
i dont recommend the 40k pretrain because the sibilants are veeery bad

if i had to guess maybe the model is going to sound like a 32k model tho lol but idk i cant predict that, experimentation is key

[rare totem]

Gottcha!

[glad nebula]

but there's a chance that it might sound like a 40k model tho

[rare totem]

So to sum up: I'll try your 1.5 32k and 48k. And your 2.5 right? (2.5 doesn't have sr options)

[glad nebula]

2.5 is only for 32k

[rare totem]

Yep

[glad nebula]

ye just experiment with all of them and see which one you like the most

if you dont want to experiment, just use 1.5 32k

[rare totem]

I'll do that first! Pragmatic choice!

[rare totem]

I like experimenting tho

[glad nebula]

me too

[minor blade]

*we

[rare totem]

So as soon as you have the new model, shout it out loud

[glad nebula]

[rare totem]

Leo, I think it's you who always posts turtles. (I'm kinda new). Love it. Anyway, bye everybody

[trail heath]

guys why cant i change anything in settings?

[gentle trench]

<@&1159293140440723499> throw this user to the flesh eating crabs

[solar torrent]

Can it create Skibidi Toilet though?

[solar torrent]

More like a Discord server's tag, like AI Hub by Weights used to have the tag "AI", which would appear on your username.

[loud sorrel]

-colab

[pine acorn]

📒 Google Colab Notebooks

Google Colab is a Cloud (Remote Good PC) Service. While the Free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

• **Applio**

by IA Hispano
Google Colab

• **RVC Mainline**

by Hina
Google Colab

• **UVR5 NO UI**

by Eddy
Google Colab

• **UVR5 UI**

by Eddy
Google Colab

• **Wokada Deiteris Fork**

by Deiteris & Hina
Google Colab

• **Hina's Modified Original Wokada**

Google Colab

• **RVC-AI-Cover-Maker-WebUI**

by Shiro & Eddy
Google Colab

• **FaceFusion UI**

by Nick088
Google Colab

• **FaceFusion NO UI**

by Nick088
Google Colab

• **Music Source Separation Training (Inference)**

by Jarredou & Makidanye
Google Colab

[solar torrent]

Is there something you would like to get help on?

[loud sorrel]

no hoe

[limber nexus]

ay, wtf

[solar torrent]

Well, this server might not be where you find women getting Skibidi'd.

[loud sorrel]

ok

I just need to find some hot gay furry top

[gentle trench]

Yeah that's enough Internet for me tonight

[solar torrent]

Bless you.

[solar torrent]

Since December, 2025 to this day, the bro in #✨│ai-help still pretends his laptop is powerful and usable for RVC, when in reality his laptop has NVIDIA GeForce MX, which Applio RVC actually detects but struggles with it, the online options are available but then he refuses to do so.

[quick slate]

slm

[sullen dagger]

yo does anyone have a voice model that sounds really high quality is english and male over 16yo?

[restive bluff]

does anyone know why if sometimes when there is a loud voice that isnt me the ai voice repeats it? but faintly

[sullen dagger]

if ur headsets topo loud or ur in a call they speak and if ur headphones are tooloud they get picked up by the ai

[vocal flume]

yall

do you know any ai that can write long scripts

long youtube scripts specifically,

and is actually creative with it, not blanky

[elder goblet]

anyone else doesnt work on rvc?

[stark maple]

hey i wanna troll un fortnite with realistic voice girl but i have to many delai or the voice deson't work good

[azure ledge]

same

my delay is crazy

[inner anchor]

please make heyseluck voice model

[solar torrent]

To request for a voice model, there's #1159289738314919936.

[gentle trench]

Why did you word it like that..

[gentle trench]

Ew

[final nimbus]

Hello 👋

[sullen dagger]

18yo*

[gentle trench]

Kinda weird tho

Why not just ask for a human voice

[bold lily]

https://tenor.com/view/heyselcuk-hey-bradar-labubu-sabubu-lex-sus-gif-12299152818219392172

[cold hatch]

Yo random question but did they remove tts for weights

[fossil pumice]

I hate AI

[gentle trench]

yea, they said it was outdated even tho it worked fine, I think they're stupid lol

[gentle trench]

https://cdn.discordapp.com/attachments/1303206967308713985/1456732618577150124/HenryEmilysuicide.gif

[cold hatch]

Ohh gulp are they gon add it back

[gentle trench]

No clue yet:(

[inner anchor]

how do people create their own voice models? Please explain where you learned this. I would greatly appreciate it

[gentle trench]

Ai voice models are created using applio

I can't really explain the process because I'm not good with words or sounding smart

[inner anchor]

can you send tutorial?

[gentle trench]

doesn't explain everything but it's ok

also got this one that's kinda old, ignore everything about tensorboard

[hasty dagger]

My guys i know you guys are AI voice in here primarily.

would you guys be interested in image to video? i just built something but need some feedback on trying it

[river gust]

and by "i just built something", you mean you let an AI build something

is that correct

[hasty dagger]

Well ngl. Front end is my strong suit but the engine is Ali Baba just trained on my workflow tbh

by AliBaba i mean "WAN"

[river gust]

well what's the deal now, rich

[hasty dagger]

Its free btw i opened up usages so im not doing this as a cash grab. the USP is it being uncensored i was pretty mad about all the other tools having woke filters etc

[cold hatch]

Ohh gotchu

[gentle trench]

sora ai exists lol

[hasty dagger]

It ain't uncensored bro

[gentle trench]

https://tenor.com/view/murder-drones-drones-serial-designation-j-j-falling-gif-15437868468999170116

[harsh dome]

Where is Options text to speech

[zinc oasis]

yo anybody knows free ai models api access unlimited and free and unrestricted ?

[soft beacon]

it is on SeaArt

as long as you use a PC / Laptop to do generations

[fossil pumice]

claude is so much better than chatgpt for code

I might be saying something akin to "the sky is blue", but I'm new to this stuff haha

[gentle trench]

Is this you?

[fossil pumice]

Haha, yes

[gentle trench]

That's silly

[fossil pumice]

It's insincere

[shrewd venture]

whats the best free voicechanger

[naive lance]

ay

i need help

i took the gta san andreas 10th anniversary trailer's audio bc i wanted the shotgun audio sound

can someone isolate it?

[late crater]

why is there no e girl filter anymore

[gentle trench]

ew

[elder willow]

whats the best was to use realtimevoicechangerclient with vb virtual audio cable on a laptop that only has an i3 and intergrated graphics?

[gentle trench]

you don't

[safe umbra]

guys how to make models

how to make voice models

[still sigil]

@violet zealot no advertising

[solar torrent]

Because you can only expect getting mocked asking for E-girl here.

This is not where you share your IP grabber, poopy.

[solar torrent]

You won't do that. Instead, better go for an online cloud option. For help about W-Okada voice changerm better ask in #✨│ai-help.

[solar torrent]

Irony.

[solar torrent]

Gooner.

[late crater]

i really dont care

[solar torrent]

You cared enough to respond.

[fossil pumice]

I have a lot of perfectly isolated wavs of a voice I wish to create (or fine tune) a voice model with. Are there any competent tools for automating transcription/segmentation/metadata creation end to end? I found some stuff on github but I thought to ask before I start downloading and trying random gits.

[lethal pine]

If you need a high-quality, natural-sounding voice clone, open-source tools won’t be sufficient; you’ll have to rely on cloud-based services

[fossil pumice]

Thanks for the reply. Any recommendations?

Thanks, but I chose this channel intentionally. I don't need help setting something up. I had a general question about AI tools. I wanted to #🧬│ai-chat .

I think I'll try TTSizer.

[fair trail]

i hired someone i wanted to ask if these models are ai made

lmk if u can help

check #🏙│ai-images

nvm i cant send image

dms

is there a pic i can send image here?

[slender mist]

hey pookie

[slender mist]

pd me on type soul rn

[fossil pumice]

Dataset tool question. Not a model question.

[fair trail]

sadly i dont play type soul just rogue lineage

[slender mist]

pd me on demon hunter mate

[fair trail]

check dms rq

[forest wyvern]

how do I use the voice models for a voice changer I downloaded, it only lets me use it for musics and stuff

[fossil pumice]

Thanks for making it clear this is not the discord server for me.

[solar torrent]

If you ever feel useless, remember.

And then he left the server because he was embarrassed to challenge further.

[proper goblet]

HELLO

[solar torrent]

For help about W-Okada voice changer or Applio RVC, better go to #✨│ai-help.

[solar torrent]

https://tenor.com/view/konuri-maki-maki-blue-archive-blue-archive-maki-wave-gif-8825291380717028696

[elder hearth]

Hi

[small sparrow]

Hello there👋

[silk hare]

Hi, I'm looking for a French female voice so I can modify my voice with high quality

[rotund trench]

ya un serv fr normalement

mais le link est invalid

[silk hare]

ah merde

et toi tu la pas du tout

[silk hare]

tu en connait toi ou pas du tout ?

[rotund trench]

nn jai join aujrd

[silk hare]

ah merdee

[sleek flume]

Hey guys you know a free AI tool where I can humanize some text

[kind minnow]

hwlp