✨│ai-help

[azure marsh]

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

[low shard]

first link

[violet heron]

-rt if your on AMD use the w-Okada fork

[azure marsh]

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

[low shard]

its !howtoask

!howtoask

[patent trellis]

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.

[violet heron]

No

[violet heron]

Oops

Another server uses !t so that’s why

[fickle lily]

Greetings, does anyone know how to make covers? I forgot how to make them?

And if so, can someone send me the link?

[low shard]

whats your pc GPU?

[oak mantle]

my voice changer is rly crackily when i talk

[fickle lily]

What do you mean PC? Now you have to make covers with a graphics card?

[low shard]

are you on mobile?

[fickle lily]

Yes

[low shard]

and yes there always have been a local way

you can just do it the cloud way

just use ilaria rvc zero which is the fastest for free on cloud

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

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

[tranquil raven]

Thanks

[obtuse forge]

How can I make my own model?

[low shard]

whats your pc gpu?

[obtuse forge]

3060 I think I don't remember

[low shard]

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

[prime grail]

hi does anybody know how to get the Kai Cenat voice in Okada??

[low shard]

You can search rvc ai voice model at:

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

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/

[obtuse forge]

I'm on phone but I checked a pic of my specs and yeah it's 3060

[low shard]

alr, rtx 3060 its good, u can train locally on that

check https://docs.aihub.wtf/ -> RVC -> Local -> Download either mainline (original rvc) or Applio (rvc fork with some extra features, same quality) on your pc

[obtuse forge]

Yep

Just tell me all the steps or a link to a guide and I'll note it down for tomorrow

Isn't this helpful?

https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[low shard]

ye its the general steps, the docs have everything u need to know

[low shard]

yea if u followed what i said ud find the applio and mainline guide link

[obtuse forge]

Oh mb

[rare goblet]

Ayo? @obtuse forge level 1 !!!

[obtuse forge]

Ok thank you 👍

I'll try it out tomorrow

[oak mantle]

my voice changer is rly crackily when i talk anyone know how to fix?

[terse oracle]

Im new can someone send real time voice changer link i have 1660 super and i5-11400 if can use local please send local version

[low shard]

-rt

[azure marsh]

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

[low shard]

1st is the wokada optimized fork

[obtuse forge]

@low shard can you quickly explain what rvc and w okada is and what's better for real time voice changing stuff

I wanna make my own model and I did like an hour of research and

I found out ab mangio

That should work right

[terse oracle]

Ty

[low shard]

@violet swift dont spam

[obtuse forge]

Oh actually nvm i think I've sort of figured things out

Yea I think I got it

[terse oracle]

Cpu or gpu version?

[rare goblet]

Ayo? @violet swift level 1 !!!

[low shard]

rvc (retrieval-based voice conversion) is mostly used for inference (use models on a prerecorded audio like ai covers), and training (making) models, i mean the original rvc (mainline) has a realtime version, but wokada is better for realtime stuff, especially the newest fork

[low shard]

mangio is an old fork of rvc, its not udpated since a year

[low shard]

gpu

💀

[obtuse forge]

Oh okay so

Do I use rvc to make my models then wokada to play them?

[low shard]

yes basically

[obtuse forge]

Oh okay

[obtuse forge]

Well it looks so much simpler to make a model with mangio

Mainline seems so complicated

[low shard]

if you are watching a yt video, its mostly outdated

i'd suggest u to use Mainline or Applio but alrigiht

[obtuse forge]

Yeah it's a few months outdated but I couldn't find a more up to date video

[obtuse forge]

Is applio simpler?

I saw that it's kind of the same as mangio

[low shard]

there is no up to date yt tuts, all youtube tuts are outdated

which is why all the guides here are written

[low shard]

the ui could be

[subtle cedar]

Yeah okay

[brittle wing]

´does anyone has girl model?

[rare goblet]

Ayo? @brittle wing level 1 !!!

[subtle cedar]

again I will give up to train models on Google Colabs right now until I had a laptop and because most of my favorite groups/bands/artists are popular/famous or from big companies because it's hard to be fan of underrated musicians/celebrities (especially underrated K-pop groups/artists ones because most of K-pop stans are only supported popular groups since K-pop nowadays only care about popularity) because only few people do models for underrated celebrities so I rarely do or train models for underrated musicians/celebrities and I will just train their models on Weights and Astra Labs for awhile, but even underrated western/American artists like Maggie Lindemann is easy to do models I mean because most of western artists are solo artists so it's easier to do datasets for them compare to Kpop groups because we should cut each K-pop group members' parts from their songs just to make datasets and models for them haha

[sour fern]

An error occurred extracting the index: need at least one array to concatenate
If you are running this code in a virtual environment, make sure you have enough GPU available to generate the Index file.

does anyone know how to fix this, it appeared when i tried to generate the index at applio

[brittle wing]

I have a question
I have a very clean dataset of 35 minutes.
I would like to know if there is any improvement by using the titan medium.

or a pre-trained model in general

[thin stump]

we came to the consensus that the original pretrain is better. Titan adds a bit of ringing noise at the end of each sentence. Still usable though for realtime

[brittle wing]

I have heard that it also distorts treble and was recommended to use klm.

[thin stump]

for really clean datasets. Shouldn't be an issue with yours then

[analog obsidian]

klm has gradient issues which in simple words means your training might become unstable

[brittle wing]

in my case I thoroughly cleaned my dataset xd

with uvr and rx 11 modules and T-De esser and Renegate

32k

also my audios are in .flac

[wheat sentinel]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

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

[olive bridge]

Does anyone know what voice changer clients support rvc 2

And what is a good rvc for a 3060 ti

[wheat sentinel]

Although not a helper or main IT staff, I'd assume you'd use the w-Okada Voice Changer. It should be good enough.

[wheat sentinel]

I'd check out #🔍│help-w-okada if you want to work with the Realtime Voice Changer.

[brittle wing]

-realtime

[azure marsh]

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

[subtle cedar]

Wait the 2nd of the list one is getting free training too? I thought it's only Streaks and I've been on Weights's server for a months but I still don't know there's another free training option besides Streaks I mean that use referal code or get 5 subscribers thingy

[tranquil raven]

Whats the best settings for rvc again?

[solemn turret]

i cant hear the voice in the voice changer i can hear mine but not the one i selected

[tranquil raven]

!help

[tawdry gull]

Great News!

Unlock the world of LunaBotPrime, absolutely free! Dive into the realm of premium music quality, enjoy it to the fullest. LunaBotPrime - where the music never stops, and the magic continues!

You can invite LunaBotPrime with this link

[tawdry gull]

LunaBot 🌙#9997

LunaBot 🌙 is the perfect music bot! Feature rich with high quality music! And Custom Playlist

You can start listening music by just joinning a voice channel and typing: /play [song name or link] (Remove brackets).
We support only Spotify, soundcloud, bandcamp and more!

To view more help on a specific command or category, run
/help <command> or /help <category>

Important Links:
Support
Premium
Invite

Command Categories:
🎶: Music
💰: Premium
⚙️: Utility
📕: Admin

Select A Page From Dropdown Menu Below

LunaBot · Sep 1, 2024, 6:29 AM

[tranquil raven]

oops

[tawdry gull]

No Information found for command rvc

[patent trellis]

<:error:865860636084142150> Command not found

There's no command called rvc

[tranquil raven]

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[dull orchid]

what is a good site or anything else for getting acapella,s ?

and is UVR still the best de reverb tool ?

for when you get vocals with alot of reverb and echo

[bronze zenith]

@timid olive from where i can download rvc

@crude bolt from where i can download rvc

@brittle wing from where i can download rvc

@steel forge from where i can download rvc

@hearty idol from where i can download rvc

@long forge from where i can download rvc

[timid olive]

do not spam

[bronze zenith]

bro please tell

[rare goblet]

Ayo? @bronze zenith level 1 !!!

[timid olive]

-rvc

[azure marsh]

Suggestions for @bronze zenith

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[bronze zenith]

okay then?

[meager quarry]

Is RVC Disconnected broken right now?
I keep getting "pickle.UnpicklingError: invalid load key, '<'," error when trying the training step

I was trying to use a custom pretrained model there

[wise roost]

I have a question, is it mandatory to provide index file path while running inference? I'm asking because for some of the models I couldn't find index files but have .pth files. (I'm using Applio).

[quick jungle]

https://cdn.discordapp.com/attachments/1160576561636966491/1165372018414342284/warning.gif

[rare goblet]

Ayo? @wise roost level 2 !!!

[unborn zodiac]

Is separating the sample voice lines into individual files a bad thing
(I hope not)

[knotty moth]

sounds like asking for a ban hammer

[obtuse forge]

Applio isnt able to generate an index file

[rare goblet]

Ayo? @obtuse forge level 2 !!!

[obtuse forge]

@flint geyser

[flint geyser]

Yes?

what are you using?

[obtuse forge]

applio colab

[flint geyser]

its actually the best practice

[obtuse forge]

i ran all the extras too

[flint geyser]

try another one for now

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

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

[flint geyser]

here are all the avaliable options

i recomment aicovergen for beginners

[obtuse forge]

can i atleast upload my pth file so i dont lose all my progress

@flint geyser

[flint geyser]

you were training?

[obtuse forge]

yeah

[flint geyser]

its recommended to first create the index file

[obtuse forge]

i did

it didnt work

[flint geyser]

before creating the pth

[obtuse forge]

An error occurred extracting the index: need at least one array to concatenate
If you are running this code in a virtual environment, make sure you have enough GPU available to generate the Index file.

i cant create the index file for some reason

so uh

[brisk quarry]

Idk what i should make a model of

[low shard]

It's written when u click the 'i' of referral codes

[subtle cedar]

Okay thanks

[low shard]

Yw

[subtle cedar]

But it's free or need to pay?

[low shard]

Free

Free premium job means training

[subtle cedar]

Thanks

[simple hatch]

can some one help me with fork wokada

[meager quarry]

how do I manually add models to applio? idk where to put the index or model file
I'm too used to using normal rvc

[pastel oak]

https://docs.aihub.wtf/rvc/local/applio/#1-upload-voice-model

[craggy brook]

How can I separate a track from double audio in Audacity?

[violet heron]

What do you need help with

[subtle cedar]

We should get 5 rewards to train models?

[low shard]

what? every 5 people that sign up using ur referall code get u a free premium job

[subtle cedar]

Oh okay thanks

[simple hatch]

How to start the web

[subtle cedar]

you mean this if we get people who joined using our invite code?

[low shard]

yea, u should be able to get 1 premium job for free if 5 people sign up using your code, u should be able to see the status clicking on your account

[subtle cedar]

Thanks

[subtle cedar]

But if we can train 5 models or more than just 1 model when we get 1 premium job unlike on Streaks that we can only train 1 model each/per 5 Streaks?

[low shard]

you get only 1 premium job free

[subtle cedar]

Oh okay it's the same with Streaks that we can only train 1 model right?

[low shard]

yes

[subtle cedar]

Okay thanks

[violet heron]

the exe if your using the fork

[carmine osprey]

how to run rvc locally

[low shard]

whats ur pc gpu?

[ocean pier]

hey , can we run rvc on node?

[brisk quarry]

no?

i mean you can try but you wont success

[ocean pier]

12c | 256GB | 1.8TB | 1Gbps

[brisk quarry]

what

[ocean pier]

oh

[low shard]

do you mean like as a comfyui node?

[brisk quarry]

i thinhk he mean nodejs

[ocean pier]

aio

[brisk quarry]

what

this guy is a npc

[ocean pier]

skidbi

[ocean pier]

i mean something like i can run it with console commands,

[rare goblet]

Ayo? @ocean pier level 1 !!!

[ocean pier]

oke

[low shard]

yea that wouldnt be possible

[brisk quarry]

also how to use realtime rvc

gvoice changer

[ocean pier]

sed :((

[low shard]

so you mean CLI/NO UI version? maybe https://github.com/blaisewf/rvc-cli

[ocean pier]

yeah something like that

but node doesn't have gpu

12 cores cpu only

[low shard]

are u talking about nodejs?

[ocean pier]

nodeaio

[brisk quarry]

@low shardwhat is voice changer rvc name

[ocean pier]

w-okada

[low shard]

-rt

[azure marsh]

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

[low shard]

wokada

use the first link which is a better fork

[brisk quarry]

real

[ocean pier]

okay if we can't train models , can we do voice covers?

[fleet cedar]

guys what is metadata.json for

[ocean pier]

On node aio

aio = all in one

[fleet cedar]

i downloaded the model from weights.gg

[brisk quarry]

https://tenor.com/view/im-dead-dead-bruh-skeleton-dead-bruh-skeleton-dead-im-dead-bruh-gif-26854866

[tight halo]

do u need to extract the zip file to use the voice model

[fleet cedar]

yes

[tight halo]

mines sounds like shit

[low shard]

i dont know much about that sorry

[low shard]

huh

[tight halo]

i cant send any pictures

[low shard]

its not used for anything dw, its just info about the model

[tight halo]

1.5.3.18a

[rare goblet]

Ayo? @tight halo level 1 !!!

[brisk quarry]

what

@low shardhow do i make my model match the voice its supposed to be over mine

[low shard]

like, it has some delay or it doesn't sound as it should?

[brisk quarry]

what are all those

[brisk quarry]

when it just doesnt match your voice and sounds different

like a male voice on a female voice

[fleet cedar]

also what does "Epochs" means

in models

it says like 200 Epochs

and stuff

[brisk quarry]

epochs do be epochs

how many times the dataset learning has been repeated

[flat bison]

Hello! Can I have help?

I don't know if this is the right chat or not

[fleet cedar]

so like what does 1000 Epochs mean

lot of detail put in the voice model?

[flat bison]

how do you set up the voice models?

[low shard]

its an unit of measurement of the training cycles of the ai model, there isn't a right amount

[flat bison]

I wanna try it out

[brisk quarry]

yes i totally have those monitors

[low shard]

ohh, i dunno much about wokada but u should change the pitch, like if its lower, the voice sounds deeper, if its higher, the voice sounds higher yk

[flat bison]

I downloaded the post malone thing and I wanna try it

[fleet cedar]

do u have another good girl voice?

not the one u sent

[low shard]

its just how many cycles it has been trained on the dataset, it doesnt mean its good quality, it varies alot

[brisk quarry]

@low shardWHAT ARE THOSE SETTINGS FOR

[low shard]

for realtime voice changing or ai covers and whats ur pc gpu

[low shard]

not really i dont even remember i sent u one

i think u confused me with someone else

[flat bison]

Like ai covers

[rare goblet]

Ayo? @flat bison level 1 !!!

[flat bison]

I make music and I wanna make my voice sound different

[flat bison]

Tbh idk the gpu stuff

[fleet cedar]

is this a good girl voice @low shard

[low shard]

there should be an explaination of every setting https://rentry.co/ForkVoiceChangerGuide#audio-setup

[flat bison]

can someone do it for me?

[low shard]

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

[flat bison]

I can send the song to you guys

[brisk quarry]

💀

what about super

[low shard]

if u got a good pc gpu you could do it locally which would be the best, else u could just use ilaria rvc zero

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

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

[fleet cedar]

idk what RVC i use

what should i do about it

[brisk quarry]

read the name

[fleet cedar]

it says voice changer client demo

[flat bison]

idk where to start haha im not smart

what do i download?

how do i start?

[wispy lodge]

Hover over labels with dotted underline and read

[low shard]

so basically, locally means u use rvc on your pc, cloud means u use it remotely on a good pc

[red kayak]

emojikage is a furry in disguise

[flat bison]

whats rvc?

what do i download

[low shard]

the whole program for this, retrevial-based voice conversion

[brisk quarry]

google

[low shard]

check your pc gpu first

[wispy lodge]

No, I'm an anime girl

[fleet cedar]

is nvidia rtx 4060 good for rvc

[rare goblet]

Ayo? @fleet cedar level 8 !!!

[wispy lodge]

With an army of neco arcs

[red kayak]

lie, i refuse to believe that

[flat bison]

I have nvidia geforce rtx 2060 or 70 i think

[low shard]

are you sure?

[flat bison]

I think so yeah

can someone do the post malone ai for this one song I'm about to send?

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

[low shard]

so you got 2 ways:

• Locally (no limits, but u have to download the program and could be a bit hard)
• check https://docs.aihub.wtf/ -> RVC -> Local -> Download either mainline (original rvc) or Applio (rvc fork with some extra features, same quality) on your pc
• Cloud (easy and fast, but limited by the zerogpu quota if you are gonna use it very very much)
• Ilaria RVC Zero (Guide with Link)

[low shard]

@flat bison

[brisk quarry]

the ms keeps going up

@low shard

its lagging hard af

[flat bison]

can someone do the post malone ai on this song?

[brisk quarry]

no

[flat bison]

i was mostly talking to Nick

[wispy lodge]

https://rentry.co/forkvoicechangerguide#i-am-lagging-what-to-do

[brisk quarry]

?

i already read that

the perf is green but tbe ping is high asf

[flat bison]

can someone help

[odd shale]

!howtoask

[patent trellis]

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.

[odd shale]

Can you explain your issue?

It's about RVC or W-Okada?

[brisk quarry]

@low shardit dont work on discord

[low shard]

u have a vac?

[brisk quarry]

yuh

but forget it

im not gonna use it anyways

[thin stump]

Headset = normal voice,
line 1 = voice changer

[brisk quarry]

what

i said forget it

[thin stump]

I have lots I want to say but you just cant say 'it don't work '

[brisk quarry]

i can

[austere path]

hey guys first time trying out the rvc tools, and okada and im having touble getting the voices to sound right, is simply just moving all the sliders a bunch all there is or is there some weird setting or software im missing

[brittle wing]

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[magic badge]

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[magic badge]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

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

[flat bison]

I'm not tech savy, can someone help me with my one track?

[stone ruin]

what are the minimum especifications for running rvc locally?

[odd shale]

You can use Applio Colab instead.

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

[dry kayak]

I'm trying to run a rvc voice of i show speed and kai cenat, but it doesnt work

[rare goblet]

Ayo? @dry kayak level 1 !!!

[dry kayak]

I cant hear anything

And it doesnt change my voice

[pastel oak]

install failed

Delete pretrain and model_dir folders then launch start_http again

[dry kayak]

Alright, I'll try it, thanks

[dapper spoke]

i can only hear myself with the voice chnager it doesnt go through my mic and other people cant hear it

[karmic plinth]

so...

why my CPU ussage so high even tho I am using GPU?

[pastel oak]

original wokada v1 is bugged with cpu, lower Extra to 8k or 16k

and f0 det: rmvpe / onnx

[karmic plinth]

actually, I am using this setting. it's okay I guess

[subtle gazelle]

How I can make a model?

[eternal sable]

my cpu isn't compatible or something

how fix

[low shard]

whats ur pc gpu?

[subtle gazelle]

NVIDIA GeForce RTX 4070

[low shard]

check https://docs.aihub.wtf/ -> RVC -> Local -> Download either mainline (original rvc) or Applio (rvc fork with some extra features, same quality) on your pc

[subtle gazelle]

thx

[low shard]

yw

[gusty kestrel]

@vagrant nebula, I have found 1 results that match your search!

nanowell

Rick Sanchez (English)

• Uploaded: <t:1691452800:d>
• Server: AI Hub
• Likes: 0
• Lang: English

Epochs

250

Technology

RVC

Algorithm

Rmvpe

Requested by moriiiiiiiiiiiiiiiiiiiiiiiiiiiii · Sep 1, 2024, 10:43 PM

@vagrant nebula, I have found 1 results that match your search!

kaythedon04

Benson (Regular Show)

• Uploaded: <t:1692921600:d>
• Server: AI Hub
• Likes: 0
• Lang: Unknown

Epochs

400

Technology

RVC

Algorithm

Rmvpe

Requested by moriiiiiiiiiiiiiiiiiiiiiiiiiiiii · Sep 1, 2024, 10:50 PM

@vagrant nebula, I have found 1 results that match your search!

kaythedon04

Benson (Regular Show)

• Uploaded: <t:1692921600:d>
• Server: AI Hub
• Likes: 0
• Lang: Unknown

Epochs

400

Technology

RVC

Algorithm

Rmvpe

Requested by moriiiiiiiiiiiiiiiiiiiiiiiiiiiii · Sep 1, 2024, 10:52 PM

[nocturne mural]

#🔍│find-models

I guess I should add the channel ignore system now.

*- traductor

[tame mica]

https://tenor.com/view/yippee-happy-celebration-joy-confetti-gif-25557730

[low shard]

do that to #1159514067187277865 too 🙏

people cant read channel names fr

@vagrant nebula use #🔍│find-models

[cerulean bison]

is this the channel to get help about the weights bot?

i do /create and it says "One files uploading failed"

[low shard]

i think #1212837348463607890 should be the one ig

[mortal plaza]

is there some type of post to make sure i get the best quality

i have a 4070 rn and it kinda sucks

[subtle cedar]

Is free premium job will be reset if we don't do things everyday like on Streaks?

[low shard]

if u dont do the 5 day streak u wont get a free premium job

[subtle cedar]

Oh okay thanks

[subtle cedar]

But I mean that premium job from 5 subscribers via invite code will be reset as well if we don't have 5 subscribers for few days?

[rare goblet]

Ayo? @subtle cedar level 15 !!!

[low shard]

i dont think it resets as long as u have 5 people who signed up, may be better to ask in weights.gg server but afaik it shouldnt

[subtle cedar]

Okay yeah

[knotty moth]

your skill issue doesn't fix it

[grizzled rapids]

Guys i have a question and i'm quite new to AI stuff. I downloaded some voice models for W-OKADA (rvc models) and i tried some. For example a female voice, it sounds cool but also unrealistic sometimes when i heighten my voice or try to sing or something it sounds completely off. how to get a PERFECT female voice or a perfect male deep voice for example. what can i do? do i need to train or something?

[mortal plaza]

help

[rare goblet]

Ayo? @mortal plaza level 1 !!!

[sour hawk]

First time attempting to load an rvc/tts model locally through hugging face, I keep getting a missing config.json error when using AutoModel

I see a lot of models actually don't have these files, but comments indicate that people can successfully use them such as: https://discord.com/channels/1159260121998827560/1279874091754721300

So I am not exactly sure what I am missing

[pure skiff]

i want a realstic model anyone know good one ?

[brittle wing]

anyone

[pure skiff]

why my voice dont change in discord ?

[gentle mountain]

I am trying to install RVC on docker with an RVC-Studio and same with chat , but it just wont Work. Is there any other open source alternatives ?

[odd shale]

Sorry but there's no such thing as a perfect female model for W-Okada. (Mostly due to hifigan/W-Okada limitations)

Probably the model you used wasn't trained on singing and for that reason it sounded off when you tried to sing.

[odd shale]

Check the #1175430844685484042 channel and try with the ones there.

[wraith coral]

Virtual Audio Cable suggestion?

[grizzled rapids]

What software has no limitations?

[odd shale]

No one if we refer to voice-changer type softwares.

There's no W-Okada alternative with no limitations.

[wraith coral]

whats the best VAC that wont cut off in VC

while voice changing

[grizzled rapids]

how do these g uys do it so realistic then. In a game there was some guy who sounded exactly like a female streamer

completely perfect

how did they do it

you could literally tell no difference at all

[grizzled rapids]

They sounded like 100% accurate

[low shard]

if u are on windows id suggest this

[low shard]

because it depends by the model quality

[grizzled rapids]

i tried like 10 different models, none of them were close to being somewhat realistic. you could always tell it's a Voicechanger

[odd shale]

Probably he had a pretty powerful PC and a girl model trained on more than 30 mins or audio or more.

[wraith coral]

that one cut offs

[grizzled rapids]

i've got a good PC also, 4080Super GPU

[low shard]

You can search rvc ai voice model at:

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

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/

But most GOOD female models are paid

[grizzled rapids]

ohh nice ;D

[odd shale]

Because public models made by random people won't always work with voice changer properly

[low shard]

what wokada are u using? original or fork and whats ur pc gpu?

[odd shale]

It's simple matter of testing.

[wraith coral]

im not using wokada

rvc realtime gui

mainline

[grizzled rapids]

What are some good ones? Can you send me some links for some very very good ones? It's okay if they cost some $

[wraith coral]

also my gpus 4050

[grizzled rapids]

thx for the link 🙂

[wraith coral]

before i reset my pc it was exactly like a woman

now it cuts off randomly

when i talk

[odd shale]

I don't have any, sorry.

[grizzled rapids]

What are some good ones? Can you send me some links for some very very good ones? It's okay if they cost some $

sorry Leo i wanted to answer this guy 😄

wrong quot

quote

[odd shale]

No worries.

[low shard]

oh thats not really suggested, i'd personally suggest using the wokada fork

[wraith coral]

wokada fork link?

[low shard]

i dunno tbh, its better u check out the model master shop

[odd shale]

Also, you can just personally DM a model master from the https://discord.com/channels/1159260121998827560/1191429836321849435 channel for a paid model or post a paid request on the #1159289738314919936 channel

[low shard]

-rt

[azure marsh]

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

[low shard]

1st link

[wraith coral]

alright ty

[low shard]

yw, u may wanna use #🔍│help-w-okada for any issues as this is channel is about rvc lol

[brittle wing]

are local model trainers hard to set up?

[odd shale]

Do you mean local RVC?

If you mean using RVC locally, nope, it's not that hard. All you need is a computer with a somewhat powerful GPU

[brittle wing]

is it faster than collab training

[odd shale]

The speed will depend on your GPU and your training settings if you use local.

[knotty moth]

colab uses T4, you can find out performance comparison against other Nvidia GPUs

[rare goblet]

Ayo? @knotty moth level 36 !!!

[wraith coral]

can anyone help

[thick dagger]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], by Eddy, Hina and Gdr Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer, Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

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

[lavish escarp]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], by Eddy, Hina and Gdr Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer, Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

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

[candid terrace]

how to upload the model file i download to the realtime?

[odd shale]

Are you using W-Okada or RVC Realtime?

If you're using W-Okada, on the GUI there's a button called "edit" you can click, there is a slot where you can add your model's pth.

[vocal forum]

Hi, is there a programmatic way to run models in quickwicks repo? no colab, no cloud hosting, i am looking for a way to run with either ( go/pytho/cpp ) or directly from a CLI app

#1159513888199540817 reads like unfiltered garbage mess, havent got any luck in there except finding garbage like deepfake (?????????) and other things

thanks for any help, smooch

[agile gale]

@odd shale could u upload the file for me via megauploads or mediafire github takes 2 hours just for 5gigs

[low shard]

there is a cli version https://github.com/blaisewf/rvc-cli

[low shard]

unfiltered garbage mess? whats wrong with deepfake or ai things like that ??

[vocal forum]

thanks, i didnt meant to sound aggressive

sorry for that

❤️

[low shard]

yw, its fine

[thin stump]

@jade marsh audacity > change your audio to mono and it'll show your true frequency in spectrogram view

[tough light]

What program is recommended for RVC local? I specifically want support for RVMPE and Titan. RVC GUI only has Crepe, crepe mini, dio, pm, and harvest

[agile gale]

could u upload the file for me via megauploads or mediafire github takes 2 hours just for 5gigs @low shard

[low shard]

what file are u talking about ??

[jade marsh]

My audio seems to reach 21.5k on linear

[analog obsidian]

mainline rvc (original, unedited rvc)

https://docs.aihub.wtf/rvc/local/mainline/
https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/releases/tag/2.2.231006

[thin stump]

strange. i would 32k anyways since thats what it wanted

[jade marsh]

ill just chalk it up to RVC being stupid and needy

[analog obsidian]

when that happens see where your audio is more detailed

so for example if most of your audio data is below 21k then is a 32k model

but if theres a lot of data in the 21k zone then it can work as 40k

simple words:
if the 21k audio is mostly blue/dark, then there's not much there

[thin stump]

yeah check the dm I sent you. It's an advanced view of a spectrogram

[agile gale]

rvc1006 nivida

[low shard]

what??

[agile gale]

@low shard

could u upload this

to mediafire or something like that

2hours for 5gigs it outrages

[jade marsh]

alright, sweet. hopefully RVC is fine with my new current settings for now.

[thin stump]

There may be some blue noise all over your audio if you dont check in Izotope first. Just saying that's what razer showed you

[jade marsh]

I know, but I was already training as soon as you sent that lol

worse case scenereo i gotta train again

[low shard]

i think that really depends by ur internet speed

[agile gale]

no bro the servers of github are just far away

[lethal basin]

@acoustic scarab

[brittle wing]

How does the 'save-frequency' work in RVC v2 Disconnected?
is it like a checkpoint so if it fails i can return to training at like 100 epochs instead of starting over?

if yes, how do i access the frequencies that i've saved

[unique blade]

help im using applio

[autumn dust]

i was reffering to a private payment

[half meadow]

I'm using the local version of RVC V2 (1006NVIDIA) and there's a bit where the voice model gets stuck on in a few songs that I'm trying to convert, specifically this part where it's supposed to say "I've built a little empire" but it keeps either getting too high pitcher or too low pitched

how would I go about fixing this?

[teal dome]

if a model doesn’t have an index file, does changing the search feature ratio make any difference?

[sharp vessel]

how do i use rvc webui

i really need help

[brazen osprey]

Is anyone else having issues using the real time voice models ? the client wont connect to the servers for me

[proper shale]

Seems like some background vocals might be interfering with the main vocals - did you already try filtering those out?

[proper shale]

If it's that wait web server stuff, just delete stored_setting.json and restart it.

[proper shale]

Explain your issue

[south nexus]

https://drive.google.com/file/d/12hwM7-nV1glcTBTfECyhSwS74CRcXThQ/view?usp=drivesdk

I have bad news I did it well I can’t do it I’m watching the video it’s a lie to all the people on discord said no

it doesn't work

[south nexus]

I need making rvc model voice humans

[ionic pewter]

how to use models located at hugging face? i downloaded freddie merkury 300k model, copy to weights dir of rvc, and if i load it "Inferencing voice" tab of web ui, i have an error

File "C:\Projects\RVC-WebUI\runtime\lib\site-packages\gradio\routes.py", line 321, in run_predict
output = await app.blocks.process_api(
File "C:\Projects\RVC-WebUI\runtime\lib\site-packages\gradio\blocks.py", line 1006, in process_api
result = await self.call_function(fn_index, inputs, iterator, request)
File "C:\Projects\RVC-WebUI\runtime\lib\site-packages\gradio\blocks.py", line 847, in call_function
prediction = await anyio.to_thread.run_sync(
File "C:\Projects\RVC-WebUI\runtime\lib\site-packages\anyio\to_thread.py", line 31, in run_sync
return await get_asynclib().run_sync_in_worker_thread(
File "C:\Projects\RVC-WebUI\runtime\lib\site-packages\anyio_backends_asyncio.py", line 937, in run_sync_in_worker_thread
return await future
File "C:\Projects\RVC-WebUI\runtime\lib\site-packages\anyio_backends_asyncio.py", line 867, in run
result = context.run(func, *args)
File "C:\Projects\RVC-WebUI\infer-web.py", line 439, in get_vc
tgt_sr = cpt["config"][-1]
KeyError: 'config'
Traceback (most recent call last):
File "C:\Projects\RVC-WebUI\infer-web.py", line 203, in vc_single
audio_opt = vc.pipeline(
NameError: name 'vc' is not defined

what to do with config.json that comes with the model?

[proper shale]

...I think that's a SoVITS model

Use another model

An RVC model

[rare goblet]

Ayo? @ionic pewter level 1 !!!

[ionic pewter]

ok. i know what to do with pth file, with index file. but what is npy file?

[surreal oxide]

im confused w
Picht raccometed +9 or +10 +11 +12 +13 for male

Dataset: 01:53
Rate 40k
Bucht size 8
Pretrein Original
Epochs 250

what are these supposed to mean

[wooden gate]

guys hows the rx6000 series work with rvc? Currently im using wokada and im curious which one works better in latency and conversion quality

btw it only works on linux with amd card or window is ok?

[low shard]

-rt

[azure marsh]

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

[low shard]

the wokada fork (1st link) is optimized for amd gpus and works good for windows too

[wooden gate]

yea im using it rn and im just wondering whether its possible to use rvc on window with amd card

[rare goblet]

Ayo? @marsh sonnet level 1 !!!

[azure marsh]

☁ (Online) W-Okada - Colab Version

• How to use Original W-Okada's Voice Changer Google Colab
• Modified W-Okada's Voice Changer Google Colab
• Original W-Okada's Voice Changer Kaggle
• Modified W-Okada's Voice Changer Kaggle

[marsh sonnet]

guys how to prevent outside noise

[tulip radish]

hello lads Rvc v2 disconnected aint working for me, i try to load the dataset and it finds all the wav files but at the end of it
it always just throws up a error and i cant figure out why

heres the error code if anybody knows how to fix it:
FileNotFoundError Traceback (most recent call last)
/usr/lib/python3.10/shutil.py in move(src, dst, copy_function)
815 try:
--> 816 os.rename(src, real_dst)
817 except OSError:

FileNotFoundError: [Errno 2] No such file or directory: '/content/temp_dataset/monkey/craftyz' -> '/content/dataset/craftyz'

During handling of the above exception, another exception occurred:

FileNotFoundError Traceback (most recent call last)
3 frames
/usr/lib/python3.10/shutil.py in copyfile(src, dst, follow_symlinks)
252 os.symlink(os.readlink(src), dst)
253 else:
--> 254 with open(src, 'rb') as fsrc:
255 try:
256 with open(dst, 'wb') as fdst:

FileNotFoundError: [Errno 2] No such file or directory: '/content/temp_dataset/monkey/craftyz'

[low shard]

the dataset is zipped and in the google drive rvcdisconnected folder right

[tulip radish]

yeah

[low shard]

could u tell everything inside the .zip?

[tulip radish]

monkey1.wav
monkey2.wav

[rare goblet]

Ayo? @tulip radish level 1 !!!

[tulip radish]

3 and 4 aswell

[knotty moth]

did preprocess actually work, and the feature part doesnt show nothing but "no-feature-todo"?

[low shard]

seems good, this happens when you load the dataset zip right?

[tulip radish]

it tells me no feature todo at the end yeah

[tulip radish]

yeah

[low shard]

wait if it didnt load the dataset how did it preprocess

[tulip radish]

lemme send screenshots of everything i see real quick

[knotty moth]

if you think the zip content structure is correct, re-export the wav files in audacity, or another audio editor with the include metadata option disabled

[tulip radish]

C:\audio\monkey.zip\monkey\craftyz this is how it looks

[knotty moth]

move the files to the folder "monkey", dont include any subfolders within it

[tulip radish]

oh my im silly

thank you so much sir

[errant plume]

hey guys im using kaggle to train my voice model and im keep getting thsis error in the screen shot so could someone help me how to fix this?

[brittle wing]

hi erm i made a acount but i cant log in bc it says csfr token mismatch can someone help me

[rare goblet]

Ayo? @brittle wing level 1 !!!

[drifting citrus]

the musics that i convert in separate youtube tracks are glitched
the voice is good but the instrumental sizzle
can i fix that or its common for the site?

[wicked apex]

how do i use voice models

[tranquil blaze]

Since it looks like the maximum frequency is 15-16k I should train at 32k, right?

And should I cut any frequency above 16k before training, or is it fine if I leave it as it is?

[analog obsidian]

rvc is going to resample it to 32k, so every frequency above 16k is going to be lost automatically so don't worry

[tranquil blaze]

Okie, nice, thank you so much :)

[analog obsidian]

rvc resampling is very bad and is not advised to allow rvc to do the process, instead use either rx10/11 or audacity to resample your audio

[tranquil blaze]

Oh I see, since I only have audacity how do I do it?

[analog obsidian]

then you export
(rvc uses mono audio, if it detects the samples are in stereo, is going to convert them to mono)

[vale osprey]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], by Eddy, Hina and Gdr Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer, Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

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

[tranquil blaze]

Okie, thank you so much :)

[rare goblet]

Ayo? @tranquil blaze level 3 !!!

[lavish escarp]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], by Eddy, Hina and Gdr Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer, Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

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

[broken lynx]

Which rvc would yall reccommend to have a dominant mommy voice

[lusty garnet]

is there aw ay so that my VC changer through discord and coms doesnt come through as robitic, is there a way to change that?

[rare goblet]

Ayo? @lusty garnet level 1 !!!

[lusty garnet]

I have 2 gpus

[odd shale]

Do you mean a rvc model?

Well, go and check the #1175430844685484042 channel.

[lusty garnet]

what does extra do? im assuming chunk just processes it so higher the better? but the longer it takes? to process it

[odd shale]

I think you can go and read the realtime docs.

-rt

[azure marsh]

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

[odd shale]

I'm not sure which version of W-Okada are you using tho.

[lusty garnet]

V1.5.3.17B i last downloaded this in 2023 of december , i tried to update it but i cant even download it just tells me link is broken on github and cant be downloaded

[odd shale]

I would suggest you to change to the Deiteris' W-Okada fork.

https://rentry.co/ForkVoiceChangerGuide

It got improvements over the OG version.

[lusty garnet]

oh nice

i gottta save my model real quick, since it doesnt exist in this discord anymorre for some reason

[broken lynx]

Yeah mb, i will

[rare goblet]

Ayo? @broken lynx level 1 !!!

[odd shale]

The original post of the model probably got deleted.

[lusty garnet]

T_T, wonder why it was decent enough, tried like 40 models before this one worked the best with my voice

[odd shale]

It's matter of testing models.

Some models won't work properly with your voice but other will

[lusty garnet]

i have a very Raspy voice, because of health reasons, so i was tired of being harassed online for not sounding Feminine enough,

so i used Ai voice change okada which helped ALOT

[odd shale]

I see.

[lusty garnet]

every time i try to download anything

[mortal plaza]

how do i get it realistic as possible

[odd shale]

Nope, not possible.

Unless you probably use a model trained on crystal clear audio.

[mortal plaza]

im talking about settings

like

which voice changer is the best, what settings to use for it to sound the best (idc about delay)

[bronze maple]

best voice extractor guys ??
to extract the acapella from a piece of music and make a model of it
I use UVR 5, except that we hear the voices in the background of the acapella, I don't know how to find the right setting

[odd shale]

Try with the Deiteris' w okada fork.

-rt

[azure marsh]

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

[odd shale]

There you got a guide for it too.

[mortal plaza]

also another question

is there a way to optimize it for discord or other apps

cause

it sounds really good using playback but people say it sounds like a voice changer in apps

if u know ofc

[lusty garnet]

is anyone else able to download this? ive tried 3 different browsers at this point

[icy python]

can anybody help me how to get the voices

i have download it but i cant play

with what app i can do that ?

[latent mica]

Dm

[still stone]

so im looking through the applio documentation. where did the different TTS Options come from? I wanna find a specific accent

[rare goblet]

Ayo? @still stone level 2 !!!

[still stone]

like are these microsoft ttses or

[sweet junco]

@lusty garnetcan you tell the client you use for voice changer

forgot the name

[lusty garnet]

wdym

[rare goblet]

Ayo? @lusty garnet level 2 !!!

[sweet junco]

this

[mortal plaza]

how do i make w okada better i have a 4070 and it sounds like shit

like what settings make the quality higher

[spare shale]

Friends of the helper Help, I haven’t worked with rvc for a long time How to make a data set better, also how to Remove noise (I mean how to make the Voice clearer so that it doesn’t sound like he’s in the bathroom speaking)

@mortal plaza -> #🔍│help-w-okada

[lusty garnet]

oh, i use a 1080TI for it

[lusty garnet]

Every time i try to download it it wont it just gets corrupted

[lusty garnet]

finally was able to download it this and install it

is it supposed to open up in the browser this time?

[knotty moth]

refer to the pinned guide in #🔍│help-w-okada and try another better model if necessary, dont blame on ur gpu tho

[surreal magnet]

yes forked w-okada uses browser (less ram usage vs desktop apps, at least on my PC)

[azure marsh]

📘 Kaggle Notebooks

• Applio Notebook, by Vidal Kaggle
• Applio Notebook, by Shirou Kaggle
• Music Source Separation, by Shirou Kaggle
• UVR5 NO UI, by Eddy Kaggle
• RVC Mainline, by Hina Kaggle
• Original W-Okada's Voice Changer, Kaggle
• Modified W-Okada's Voice Changer, Kaggle
• 📖 How to use RVC Mainline Kaggle by Cauthess

Note: Kaggle limits GPU usage to 30 hours per week.

[crude bolt]

seperating voices from each others is a bit harder.

the approach is to split instrumental and voice.

then use karaoke to split voice from background echo voice.

then de-echo / de-reverb on the background singers to get them sound dry.

This is not a 100% guarenteed method, especially like in cases you got duet voices and both singers are heavily on the foreground.

tweak / apply diffrent multiple steps depending your audio-source

sometimes you get also the background singers out it with just the de-reverb or de-echo.

[crude bolt]

it can depend on the browser, especially when you installed like bazilion extensions on your browser 😂

but yes lets assume we are on a clean vanilla browser 🙂

the rvc fork was for me a good incentive to get rid of extensions that i didnt actually need 😉

[knotty moth]

try in Edge or another browser if there's some issues in Firefox, though having uBlock origin is fine

[dapper scroll]

my rvc looks weird

it doesnt show a picture or any of the free models

does anyone know why

i got all the settings

just no models

[crude bolt]

the forked okada dont contain any

[knotty moth]

may depend on the accent compatibility with the voice to infer

[dapper scroll]

which one do i download

[crude bolt]

not sure if you particularly want the demo models (which imo are bad). So are these models what you really want?

if not, you can download models from here : #1175430844685484042

[knotty moth]

the zip downloads in #1175430844685484042 or weights.gg usually don't include the image, only .pth and .index files

[crude bolt]

you can edit the model in rvc and attach any picture you like

[knotty moth]

it may also depend on pretrain you're using, also have you tried different index rate values in inference?

[crude bolt]

the accent is stored particularly in the index. Do you generate an index?

there are many pretrains around depending on your target language you training for

i know theres a guide out there explaining which pretrain is trained for what use , but i dont know the link

[knotty moth]

https://docs.google.com/document/d/1j9J8A8Oop9bMOHmCs3jDXzPujuD6TQ0Q396rJ0MyuIc

[crude bolt]

😄

teamwork 👍

thanks @knotty moth

[knotty moth]

also take a look in #1235952130855010365

[crude bolt]

then do the try-approach

print the table out on a paper

and throw a dart on the paper

and pick that

or - choose something that looks good

there is no wrong or right here. Some sound better , some worste - its all subjective

beauty is in the ear of the beholder

❤️

[knotty moth]

pretrains have their pros and cons regarding quality and some others

[crude bolt]

they have 1 thing in common , they reduce trainingtime.

the whole point of pre-training

[knotty moth]

more like that training voice models will never be good without a pretrain as the foundation

[crude bolt]

not totally agree with it - but yea you easily will need to understand that training a model without a pretain will cost you like a week instead of a few hrs.

to get a descent (not perse good) model.

plus you need to do a big time of work to clean out the audio source, and have a lengthy big variety of spoken words of this person. ... All n all , do train with a pretrain rofl.

[knotty moth]

what I mean is the comparison of quality & how it generalizes, no matter what the optimal epoch is
for example, ~30m of assumingly clean voice dataset trained with og pretrain vs no pretrain

[crude bolt]

i was nitpicking on your statement that you cant make good models without pretrain. So yes you can do that for an unreasonable time with unreasonable effort compared if you actually train a model with pretrain. It's just not worth doing so. I would always recommend to use a pretrain.

Unless you really want to make your own pretrain.

using a pretrain will save you alot time and effort for making a good voicemodel.

there are veteran voicemodel makers out there totally unhappy with current pretrains and make their own pretrain that subjectively yields better results when creating their models with specific needs.

[knotty moth]

them who fallback to the og pretrain

[tranquil blaze]

If I'm training on kaggle and I'm using t4 x2, do I have to double the batch size, like, if I wanna use 8 do I have to put 16?

[knotty moth]

its batch size per GPU

[tranquil blaze]

Oh so it's still 8 right?

I stumbled upon the pretrains guide and on there it said to double the batch size if the gpus are two, but maybe that only applies to pretrains, idk

[knotty moth]

the actual batch size is doubled, so again I said the option is per GPU

[tranquil blaze]

Ohhh I see, so since I wanna use 8 I have to put 4 because it's per GPU, so 4 * 2, right?

[knotty moth]

the guide already says that

[tranquil blaze]

Yep, I just thought it was the opposite thing, thank you :P

[pastel oak]

say the same after u trained once

[fleet cedar]

why do i have echo voice

pls help

[pastel oak]

Mic picks up headphones, headsets and especially laptop speakers not recommended

U can try: lower volume, enable sup2, move s.threshold to the right

[strong flicker]

Hi everyone, I'm doing feature extraction for a model, Its been going for several hours, is that normal?
When I did my first model training some time ago it wasn't that slow

[knotty moth]

what are you using? mangio crepe in rvc disconnected?

[leaden ore]

Hi. I need stable and good rvc golab link\

[golden karma]

what's the best model in mvsep for isolating vocals

[gusty kestrel]

Translated to French

hello

Translated message

Bonjour

Requested by totoz8970 · Sep 4, 2024, 5:08 PM

[brave garnet]

Translation: Bonjour

[dark ginkgo]

How do I use RVC with XTTS? I get this error when trying.

[rare goblet]

Ayo? @dark ginkgo level 2 !!!

[viral island]

whenever im playing vr chat on vr i wannna use the voice changer but it doesnt work like the sound does not go to vrchat

[low shard]

what are you using exactly? like link of the fork

[cobalt oyster]

Any body give me link of rvc 3 one click install one

[low shard]

rvc 3 doesn't even exist

what's your pc gpu?

[cobalt oyster]

Rtx 4060

[low shard]

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

what? you just have to download a prezip and extract it

oh u deleted the message

[dark ginkgo]

https://www.youtube.com/watch?v=ds5LLIt5OLM&t=211s

I used the one click installer

[low shard]

its better you don't use those, youtube tutorials aren't maintained

I can tell you to use RVC or Applio that has TTS too

the problem shown in the screenshot is that it cant find the output file because it cant make it for some dependencies issues

[low shard]

whats your pc GPU btw?

[dark ginkgo]

RTX 3060

[vale osprey]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], by Eddy, Hina and Gdr Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• Modified W-Okada's Voice Changer, Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

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

[low shard]

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

About one with XTTS that is maintained, there isn't afaik

but applio uses Microsoft Edge TTS API to make tts and then uses that audio as an input to use it with rvc model

💀

[dark ginkgo]

Which is best for song covers and also manually controllable emotion in voice?

XTTS doesnt seem to have emotion control from looks of it.

Also this feature looks cool

[odd shale]

Neither RVC doesn't.

[dark ginkgo]

So which tool still works then?

because take away I'm getting is, XTTS-webui by
daswer123 is outdated and no longer supported

It does work if I dont use rvc so yeah, rvc setup seem to be the issue

[vale osprey]

why is it saying the llaria vc colab is a trojan virus?

i cant run it

[odd shale]

vc colab?

Do you mean Ilaria Mainline?

[vale osprey]

llaria rvc

yeah the mainline

did someone put malware in the downloads?

[low shard]

for tts? neither rvc (as its Speech to Speech natively) neither XTTS

gpt so vits is the best for emotions, but it works only for chinese, japanese and english

[low shard]

what ?? show a screenshot

[dark ginkgo]

so it's 2 serperate software?

One for normal talking and other for music and songs?

[low shard]

what antivirus is that?

[vale osprey]

AVG antivirus

[low shard]

google colab has no trojan, its owned by google

[vale osprey]

idk then cause i tested with avast antivirus and its saying the same thing

[low shard]

wdym? like one for ai covers, another for tts emotions?

[strong flicker]

yeah I was using an old rvc version, now I'm using RVC mainline with rmvpe gpu but it takes 1 minute per epoch, idk why its that slow

[dark ginkgo]

yeah, for one, I want to use it to make for ai cover or turnin humming into other instruments. I seen some ai tts do that.

It's for like instruments and stuff

And the other, being able to do voice dubs.

[low shard]

ig its a false positive, i have literally 3 antivirus (malware bytes, bitwarden and avira) and theres nothing wrong in it (im running it right now), its literally owned by google

also, its running on a remote google pc

its not even running on yours

[dark ginkgo]

One of the project I'm doing is converting bad voice acting on english dub with original japanese. that's testing of concept anyways

[vale osprey]

are you running the one that comes up with -colab?

or is there a newer version of it somewhere?

[low shard]

ig it gets false positive bc it connects to google's pc, ngl never heard of anyone else saying colab is a trojan, i use google colab much as i dont have a good pc

[low shard]

you are talking about https://colab.research.google.com/drive/1mHKTGH5e3SAyDSBss1KtiYRbDdQzwSMs#scrollTo=pz3lRmDHl_JG ?

or are you following a random youtube video?

[vale osprey]

nah i was talking about that one

let me try again

with that link

[low shard]

for ai covers, rvc is used, while for text to speech gpt so vits is the best

[dark ginkgo]

i'm downloading applio right now, does this cover both?

[rare goblet]

Ayo? @dark ginkgo level 3 !!!

[dark ginkgo]

or do I need to download something else?

also, how about the instument side of thigs?

[low shard]

applio uses edge tts, which isnt the best for emotions so no

[low shard]

i dunno about u but for me its all fine

[dark ginkgo]

which one you recommend for emotion then?

and again, which ones does song cover and music?

[low shard]

gpt so vits, its 2 different programs, the most updated guide should be https://docs.aihub.wtf/tts/gpt-sovits/

[low shard]

RVC for ai covers

[low shard]

from what i seen of your screenshot it seems an audio upscaler, it uses resemble enhance https://github.com/resemble-ai/resemble-enhance

💀

[vale osprey]

we are NOT buying that 😭

[low shard]

fr, banned that scammer

[vale osprey]

ima just use applio for now but thanks for the help

[low shard]

wait, u get that warning ONLY when visiting the ilaria rvc mainline colab ?

like, after u enter the site instantly u get that warning or is it for when you enter any google colab in general ?

[vale osprey]

just that one when i enter and when i try to install the rot13

[rare goblet]

Ayo? @vale osprey level 4 !!!

[vale osprey]

i went on applio and it was fine

[dark ginkgo]

so is this like a remaster of sort? If so, might be intresting to take some old classics and fix it to sound like it's recorded with modern instruments.

[low shard]

wtf that's very weird

[vale osprey]

yeah ikr

[low shard]

btw u can check even the code yourself

there is nothing malicious in it

its open source

[low shard]

its the same program that was used in the outdated ui you downloaded, it basically 'upscales' the audio quality of speech audios

[dark ginkgo]

by upscaling an audio, you mean like add detail to lost audio content? Like fake audiophile? Adding missing infromation to mp3 and convert them into flac sort of things?

If it's just that, it's not all that useful...

[vale osprey]

yeah im honeslty too lazy to go through it all so ima just turn off the antivirus next time lol

[low shard]

i mean that it enhances the quality of the audio, and resemble-enhance specifically denoises it before enhancing

like the example from their github is

i just sent it as a file as the link was too long

[dark ginkgo]

if it's just for speech, why is it under instrument tab

but yeah, if it's just for speech, it's only really useful to use it to enhance something you struggled hearing due to distortion then

but if thats the case, subtitle is enough

[low shard]

i think it was just meant as like 'other tools', because in the official resemble enhance github repository its said its just for speech

[dark ginkgo]

different question, you said GPT was best for emotion in speech right? Aka making character talk. I dont see an option on controlling the emotion in the screenshots. Is it prompt based?

[low shard]

i mean,its good for low quality audios, another audio upscaler that can work also for music is AudioSR, which of i have made a fork that fixs a bug in the webui version

[dark ginkgo]

also what is ApplioV good for if it's not good for emotion or songs?

Like it defeat both purpose of ai voice tts doesn't it?

[low shard]

i dont use gpt so vits locally myself as i dont have a good pc, but it depends by the reference audio as seen in https://docs.aihub.wtf/tts/gpt-sovits/#inference-

[low shard]

Applio is not made for TTS especially, Applio is just a fork of RVC, which is Speech To Speech natively, RVC is not for TTS, what applio does is make a tts audio with edge api, and uses it as an input with the rvc voice model, it has good quality but no emotions. About songs, rvc is used for ai covers, or are you talking about making songs like suno?

[peak pebble]

When is where I should enter a link to make the cover?

[rare goblet]

Ayo? @peak pebble level 1 !!!

[dark ginkgo]

So Applio is for song covers then?

[low shard]

the public url

[peak pebble]

okii!

[low shard]

Applio is for AI Covers and multilingual TTS

so yea mostly for song covers

[dark ginkgo]

ok, thanks, so I now know what model to use for what.

[low shard]

yw

[low shard]

alr, yw

[peak pebble]

How do I find the voice I want to use? (already downloaded it)

i forgot how to do it

[low shard]

click 'Refresh' and you should find it in the 'Voice' drop down menu

Btw, i see you are using Ilaria RVC Mainline right? if you want i can suggest a faster ilaria rvc version

[dark ginkgo]

where to I place the rvc folders for ApplioV?

[odd shale]

RVC folders?

[peak pebble]

please!

[low shard]

there is this guide if you want https://docs.aihub.wtf/rvc/local/applio/#inference-

[peak pebble]

there is nothing when i refresh

[dark ginkgo]

where in the rvc folder? there is ton of sub folder

[low shard]

if you want the faster one, there is Ilaria RVC Zero

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

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

[low shard]

whats the model download link?

[low shard]

dont worry you can upload it directly in the webui as seen in the guide

[peak pebble]

Uhhh, Mephiles the Dark

[odd shale]

Tell me what you want to do.

[low shard]

the link?

[odd shale]

He said the Huggingface link to the model.

[peak pebble]

aaahh

srry!

hugginface

[low shard]

yea you sent the right one

it seems like that the model files are right

[low shard]

you gave it a name without spaces right

[peak pebble]

OOOOOHHH!!!

IM SO DUMB!! THANKS!!

[low shard]

your welcome

[peak pebble]

well

[low shard]

show the google colab output

so i can see whats the error

[rare goblet]

Ayo? @peak pebble level 2 !!!

[low shard]

could u try to stop running the cell, run it again and accept the google drive thing (if u want to backup) or else uncheck the drive backup checkbox and run the cell?

[peak pebble]

on my way

[dark ginkgo]

Hey, how do you change GPT-SoVITS to English? It comes out as Chinese by default.

[peak pebble]

ready

[low shard]

could u retry to download again the model and re test if it works?

[low shard]

did u download it from https://docs.aihub.wtf/tts/gpt-sovits/ ? the prezip should be english

[dark ginkgo]

yeah this link

[low shard]

u downloaded the latest which is this one?

[peak pebble]

It tells me to select a file from the audio folder, but I don't know what it means

[low shard]

ah thought it was all english, unfortunately the helper who made that guide isnt here anymore, but the guide i sent should help you to use it all

[low shard]

could you show a screenshot ?

[dark ginkgo]

also, I dont see an option to make text to speech anywhere

[peak pebble]

btw, I already did everything again

[dark ginkgo]

Hey @low shard how about Bark TTS?

[low shard]

dw you shouldn't need to do that, but i just tested myself and it seems like that colab isn't working as i tested with different models and audios unfortunately

[low shard]

@proven hill Ilaria RVC Mainline colab seems broken, no matter what model, audio, or if u used the ilaria tts, it seems to give always:

[peak pebble]

ow :(

[proven hill]

colab is not mantained by me

[low shard]

who maintains it then?

[low shard]

its better you use ilaria rvc zero which i sfaster

[proven hill]

theres the credit section

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

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

[peak pebble]

Haven't I been doing something wrong then?

[low shard]

This colab was created by

Angetyde
Poopmaster
l3af
Using ilaria mainline by thestingerx

but the program is made by you though

i guess i can either ask angetyde or leaf

[proven hill]

yea but it always worked so i assume its a colab error

[low shard]

i will tell them about the error and if there is no fix ig i have to say that its broken in #📰│dev-updates if none is able to use it

[low shard]

@flint geyser seems like the Ilaria RVC Mainline Colab is broken, could you check it out and tell me if its fixable?

let me know because if its not fixable i will just tell users to use ilaria rvc zero instead

[peak pebble]

The last time I used this program was more than a year ago and it also gave me an error.

[low shard]

I think what you used was the old ilaria rvc mangio rather than mainline

its better to use the zero version tho

[low shard]

.

[peak pebble]

ooh!! i see, can i use the zero instead?

[low shard]

its in the inference part,

[low shard]

it was an old version of suno, its emotional but not that good quality

[low shard]

Yes

[peak pebble]

thanks man!

[low shard]

i sent you the links

your welcome

[peak pebble]

alr!

[tranquil cliff]

Nick talking about me? Now he hates me, even though I was hoping this moment would come

if you want to clarify you can dm me, since I can't, after you blocked me

[dark ginkgo]

So GPT-SoVITS is still best for emotion then?

[rare goblet]

Ayo? @dark ginkgo level 4 !!!

[dark ginkgo]

Like I want it so I can maybe make the character yell, or laugh.

ect

[low shard]

yea good quality and emotions

[low shard]

i just asked u and leaf to check whats wrong in the colab

[dark ginkgo]

where?

[tranquil cliff]

yes, but I'm taking advantage of the situation to clarify

[low shard]

1c-inference

[tranquil cliff]

answer this

Nick088

[low shard]

no need, asked only to check whats wrong, because if its not fixable i have to say it in #📰│dev-updates

[dark ginkgo]

That's a little demanding

[tranquil cliff]

Yes, I'll check now, but for me I need to clarify

[tranquil cliff]

?

[tranquil cliff]

Yo?

Nick è morto

[dark ginkgo]

Nice, Applio was the first one that sorta works. But yeah, no emotion I guess. Unless maybe...

[dark ginkgo]

so just for confirmation, applio doesn't support emotion correct? Like gasping for suprise, happy noise, angry, etc.

[brisk quarry]

Idk

[dark ginkgo]

@low shard

[low shard]

yes

[dark ginkgo]

so for tts, I'm kinda back to square 1 since I cant figure out how to get the other one to work

[low shard]

i dont know if there is any other helper that knows about gpt so vits sorry, there used to be @dawn nebula but he left, hopefully an helper can help you there

[sharp glacier]

how do I use this now they recently updated it to a new dumb llayout and now no audio plays even when i set it up correctly

I wish people would stop updateing layouts i get use to

or how do I go back to the previous version.

that i knew how to use

well now it's being even stupider

nvm it's solved... sorry

[drifting sand]

I dunno if this is the right place, but every time I try to use the RVC it just spits out 'error' after convert

[rare goblet]

Ayo? @drifting sand level 1 !!!

[low shard]

which rvc

be sure u arent watching a youtube video

[drifting sand]

im going of f memory

[drifting sand]

https://colab.research.google.com/drive/1mHKTGH5e3SAyDSBss1KtiYRbDdQzwSMs#scrollTo=8RwuGj0LmSB_

this one

[low shard]

do you get an error similar to this in the colab ? #✨│ai-help message

(btw u can check the actual error code in the google colab output, in the gradio ui it will just say error)

[drifting sand]

im only really familiar with the front end, where would I find an error like tha- oh

uhh

would that be under the 'run' cell?

[low shard]

from what i know that colab is broken very recently, just asking if its the same error to be sure that its broken

[low shard]

yea, like after the public url thing

[drifting sand]

2024-09-04 22:30:34 | WARNING | infer.modules.vc.modules | Traceback (most recent call last):
  File "/content/Ilaria-RVC-Mainline/infer/lib/audio.py", line 63, in load_audio
    audio2(f, out, "f32le", sr)
  File "/content/Ilaria-RVC-Mainline/infer/lib/audio.py", line 41, in audio2
    ostream = out.add_stream(format, channels=1)
  File "av/container/output.pyx", line 132, in av.container.output.OutputContainer.add_stream
  File "av/stream.pyx", line 111, in av.stream.Stream.__setattr__
AttributeError: attribute 'channels' of 'av.audio.codeccontext.AudioCodecContext' objects is not writable

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/content/Ilaria-RVC-Mainline/infer/modules/vc/modules.py", line 188, in vc_single
    audio = load_audio(input_audio_path.name, 16000)
  File "/content/Ilaria-RVC-Mainline/infer/lib/audio.py", line 67, in load_audio
    audio = file[1] / 32768.0
TypeError: unsupported operand type(s) for /: 'str' and 'float'

Traceback (most recent call last):
  File "/usr/local/lib/python3.10/dist-packages/gradio/routes.py", line 437, in run_predict
    output = await app.get_blocks().process_api(
  File "/usr/local/lib/python3.10/dist-packages/gradio/blocks.py", line 1349, in process_api
    data = self.postprocess_data(fn_index, result["prediction"], state)
  File "/usr/local/lib/python3.10/dist-packages/gradio/blocks.py", line 1283, in postprocess_data
    prediction_value = block.postprocess(prediction_value)
  File "/usr/local/lib/python3.10/dist-packages/gradio/components.py", line 2586, in postprocess
    file_path = self.audio_to_temp_file(
  File "/usr/local/lib/python3.10/dist-packages/gradio/components.py", line 360, in audio_to_temp_file
    temp_dir = Path(dir) / self.hash_bytes(data.tobytes())
AttributeError: 'NoneType' object has no attribute 'tobytes'```

I got this-?

[low shard]

yea its broken

use ilaria rvc zero instead

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

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

[drifting sand]

thank you, I will give it a try

[knotty moth]

stop using either of those two AVs

[knotty moth]

how about the latest one? it couldnt be so slow as mangio crepe, even with low end gpu like 1660 super

[half meadow]

Ah very sorry about the late reply, had to take a break due to personal reasons, I use the built in vocal and instrumental divider of RVC, how would I go about filtering out the background vocals?

[knotty moth]

UVR BVE or mel roformer karaoke

[strong flicker]

I managed to get 30s per epoch using RVC Mainline rmvpe gpu (running on a 3070), but I feel like it shouldnt be that slow

[rare goblet]

Ayo? @strong flicker level 1 !!!

[strong flicker]

When I trained a model some months ago on an older version of RVC i was getting 10s per epoch or something

[knotty moth]

I was talking about feature extraction stage

[half meadow]

What does the BVE stand for? I assume you're talking about Ultimate Vocal Remover?

[strong flicker]

so you suggest to change rmvpe to mangio-crepe in the feature extraction?

[knotty moth]

stick on rmvpe/gpu if that's what you intend, only that it shouldnt be as slow as mangio crepe

[strong flicker]

Is rmvpe better quality than mangio-crepe?

when I trained months ago I didnt have rmvpe so I never tried it

[knotty moth]

rmvpe is generally good and pitch accurate, especially for voice changer use, but it's also fine for making covers

[strong flicker]

aight, ill stick to it then, thanks

thats where I am atm

[knotty moth]

get BVE in VR arch category

[half meadow]

4B_SN

right?

Sorry if I'm asking too many questions I just don't want to misunderstand you

OH

OH HOLY SHIT

NO

I AM SO STUPID

THANK YOU ALYA AND MJ

[strong flicker]

What gpu batch size is recommended to use?

[knotty moth]

8 which fits on your 3070

[strong flicker]

Thanks!

[low shard]

#📰│dev-updates message

[low shard]

Whats ur PC GPU

[sweet harbor]

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[odd shale]

Please don't post acapellas on the server.

These are copyrighted.

[half meadow]

My apologies

[desert ledge]

anyone have the public url?

[lapis tinsel]

Hi everyone

I wanted to create a model, so how much batch size do I get?

Please help me how to make the batch size better

[glossy raven]

what happened to the RVC docu for training locally?

if anyone has that please send it to me

[torpid wasp]

Ilaria RVC on maintenance?

[proper shale]

https://docs.aihub.wtf/rvc/local/mainline/ ?

[low shard]

the ilaria rvc mainline google colab is broken #📰│dev-updates message

But you can use Ilaria RVC Zero which is better

[low shard]

the public url is different for everyone using the google colab

what are you looking for?