✨│ai-help

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[jaunty bloom]

Gtx 1650
Windows
Trying to make a voice model. The process "finishes" in a few seconds and gives me an index file but not a pth file. Im just using the applio link that shows up with the -colab command. I dont have a screenshot right now because I'm not home.

Windows 11 for operating system

[simple ore]

read what the console windows shows you

training on 4GB GPU is pointless

[lusty ore]

4070 super ti
win 11 pro
wanna troll in valo
https://www.youtube.com/watch?v=SxdnGxicJOg

[low shard]

that tutorial is outdated asf, uninstall everything

wanna troll in valo
are you trying to catfish in valo as an egirl?

[lusty ore]

no

im trying to troll with like changing voices and stuff

like charakter voices

[lusty ore]

ok

[low shard]

so just characters like peter griffin, not e girl stuff right?

[lusty ore]

yea

[low shard]

good then, catfishing isn't allowed

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[lusty ore]

so basically i looked in ai help forum and i downloaded it from a source you gave and npw ts shpws

[low shard]

1st link, wokada deiteris fork

[lusty ore]

how can i send pic

i downloaded the first thing

[low shard]

what thing? wokada deiteris fork?

did it open your browser?

[lusty ore]

yea

[low shard]

what browser do you use?

[lusty ore]

chrome

[low shard]

alr, show a screenshot of the wokada

did you also uninstall the old original wokada and vb audio cable?

[warm cedar]

yo my voice changer is SUPER delayed when i open fortnite, like 15 second delay.

[lusty ore]

ye

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

show a screenshot of the error

!give-media-perms 1h @jaunty bloom

[warm cedar]

dude you helped me set up yesterday

this guy

[low shard]

extra: 2.7
chunk: 80
f0: rmvpe without onnx
output: line 1
monitor: optionally to headphones to hear urself

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Disable JIT compilation: off for faster loading speed of the program, on for slightly better performance (10-15 ms) for Nvidia only)
• Reduce the delay via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

[low shard]

oh lmfao i forgot you

set fortnite's settings to the lowest, 1080p, 60 fps capped, literally the most dogshit quality you can get 😭

then set extra to 2.0, and chunk higher than perf value

[warm cedar]

dang i cant run my crisp graphics

i’ll set it on performance mode and see if that helps

but it’ll go to like 40000ms

[low shard]

yeah ofc, ai is an intentive task, don't expect to play any games on good graphics while using the program

[low shard]

you got 8gb vram, which are pretty low for intensive tasks like high gaming, and ai btw

[warm cedar]

dude i am on performance mode and running 2700ms

nah i got like 16 or 32 i forgot

16 with 15.7 usable

[low shard]

that's ram, not vram 😭

vram is the virtual ram of your GPU

the rtx 4060 got only 8gb vram

[warm cedar]

damn dude maybe i just can’t run fortnite and use ai i paid $1600 for my computer i thought it could withstand it but i’m STRUGGLING, fortnite looks like a nintendo switch game

it’s just weird though because yesterday everything was running perfectly fine while using the ai on my normal high graphics

i feel like i’m doing something wrong because why would it work perfectly fine yesterday and just completely stop working today

[low shard]

did you use it yesterday on fortnite with the same settings?

[simple ore]

remember to turn off this shite

[warm cedar]

yeah, my complete usual graphics i run without the ai

do you think that would help?

[warm cedar]

okay i turned off that setting

i’ll see if it makes a difference

[warm cedar]

are all my settings right? this is how delayed it was sitting idol in the fortnite lobby

[low shard]

simply put chunk higher than perf value until perf value is green lol

[low shard]

this should help the voice changer at the cost of less gaming performance right? I dont remember about it

[simple ore]

it barely helps gaming and it messes up AI stuff

[dull canyon]

does anyone knwo why ai voice changer keeps picking up discord audio?

like im speaking into it and the other person speaking, his voice is also going through the voice changer and going thru the mic

[foggy karma]

hey, can someone help me?
i want to use ai voice models but w okada does not support intel macos, is there any alternative app?

[sour venture]

Don't anyone know how to use w-okada?

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using

be sure to not use yt tuts

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[sour venture]

nvm I solved my issue

[low shard]

just be sure to not use Youtube VIdeos tutorials, they are ALL OUTDATED

@sour venture i hope you're using wokada deiteris fork b2332, right?

[sour venture]

this is honestly my first in a while using this so idk

[foggy karma]

2019 intel mac, mac os 13, i want to find an alternative of w okada or someway to use the models without the app, i could not find any

[low shard]

show a screenshot, i can help you up with the program and settings

!give-media-perms 1h @sour venture

i can help you for lowest delay and best quality

[sour venture]

this is what I'm using

can't remember all this

[low shard]

2019 intel mac
that's kinda old,
the issue is that you used youtube tuts, they use an over year old original version who doesn't support mac

-realtime

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

1st link, wokada deiteris fork is better

[low shard]

that's almost 2 years old

[foggy karma]

bro what are you talking about

[low shard]

it's outdated asf

[sour venture]

r u serious?

[low shard]

RVC = Retrieval-based-Voice-Conversion, the best Few Shots Speech To Speech AI Models (on v2), Inferences (use models) pre-recorded audio (ai covers) and train (make) models. Technically, Mainline RVC does have a go-realtime.bat (aka RVC-GUI), but it's pretty messy and outdated so it's extremely not suggested for realtime.

Wokada = uses RVC for realtime inference. There's 2 main versions, Original made by Wok, and the most suggested one is Deiteris Fork (modified version)

[foggy karma]

is Deiteris Fork avaible for intel mac?

[low shard]

did u check when downloading?

[sour venture]

a friend actually helped me with it

[low shard]

yes, tho macs aren't the best for ais

[low shard]

you and your friend are COOKED 😭

[foggy karma]

IK

[low shard]

also, the settings were wrong, and you prob used vb audio cable which causes issues on windows

[foggy karma]

but is there any way?

or could you help me with kaggle

i dont understand the tutorial

[low shard]

yes, read up the wokada deiteirs fork link https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

try local first

[sour venture]

ok before we move on should I switch to a more updated version of wokada?

[low shard]

just to see if its good enough

[low shard]

you should uninstall and literally forget everything your friend said

it got way worse quality and performance

[sour venture]

I don't mean to sound to like a jerk but could you tone it down please

[foggy karma]

damn

tysm

[sour venture]

anyway I'll take out all the files related to the client

[low shard]

Sorry, but to make it easy its better you uninstall everything

then you can follow

-realtime

[sour venture]

yeah I'm on it rn

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

1st link, wokada deiteris fork

this will help you a lot

if you got any issues, please let me know

[low shard]

yw, lmk for any issues

[low shard]

btw, for uninstalling vb audio cable, you gotta do that via the windows app settings

also you should prob telll ur friend its outdated lol

[foggy karma]

BRO

iTS WORKING

you goated

[low shard]

want to share a screenshot of ur settings to let me help u set them up ?

[foggy karma]

its still downloading

give me a sec

[low shard]

oh thought u said its working 😭

[sour venture]

srry late reply the recycle bin won't help?

[low shard]

recycle bin will help for the old original wokada, but vb audio cable is a program installed directly on your pc

any programs with a setup, need to be uninstalled from windows app settings

[foggy karma]

i mean

its downloading

[sour venture]

ok so how would i do that?

[low shard]

https://support.microsoft.com/en-us/windows/uninstall-or-remove-apps-and-programs-in-windows-4b55f974-2cc6-2d2b-d092-5905080eaf98

just this basically

you may want to check if you also had other apps you thought you uninstalled but never really did too btw

[sour venture]

alr done now what?

[low shard]

now read up the wokada deiteris fork

and let me know for any issues

[sour venture]

this one?

[low shard]

yep

[foggy karma]

hey

there in no output i can select

[sour venture]

I'm downloading the version as of dec. 7th 2024

my pc is running a GeForce 3060 so it's the only viable download

[low shard]

what browser are u using?

did you give it microphone perms?

[low shard]

be sure its the normal nvidia on windows one, and not the 50 seire since you have the 30 serie

[sour venture]

ok soooo my pc is having problems

my pc is keeps ghost disconnecting the HDMI

[sour venture]

ok I'm back hopefully nothing goes wrong I have the file and it's the one before the NVIDIA one

[dense cairn]

i need some help decidig weather to scarap this model or not its suposed to copy all of juices sound dataset is 5 hours long

[brittle wing]

Can anyone tell me how reliable AI chatbots are, especially if I'm doing research on something?

[abstract rain]

anyone have an issue where game audio disappears but voice monitoring and other people's voice in the game is present. (Default device is Cable Input) E.g valorant GAME audio is gone but I can hear people's voices in the game and my own ai voice. If I set the default audio to my HD audio driver I hear the game but the AI voice changer doesn't work. Any tips?

[livid surge]

peoples

ive solved my proble

[ivory lynx]

AI chatbots (I'm assuming LLMs) are generally quite reliable, but it definitely depends on the prompting. If you can prompt well or instruct an AI chatbot well, then yes, they can be exceptionally useful for research-tasks. You should pick a chatbot you think is reliable. Bear in mind that an AI still doesn't demonstrate the analytical, expressive and cognitive abilities of a human, so make sure to keep an eye on the AI chatbot's work. All in all, yes, they are reliable, especially for your task of 'research on something'. I hope that helped! 🙂

[brittle wing]

First of all, thanks for the reply, brother. Secondly, yes, I think I can agree with what you just said. Yes, it helped a lot. Thanks again and may God bless you.

[ivory lynx]

No problem. 🙂

[ashen spear]

is there anyone here who generates ai videos using google collab? I tried it always happens run out of vram

I want to use wan 2.1 but it seems impossible if it's in the free google collab.

[flint vapor]

If you mean voice changer with AI Yes

-wokada

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[flint vapor]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/ There is here a guide

But first of all if you want run it locally you should understand the limits of your gpu, which gpu do you have?

[low shard]

Any updates?

What's your PC GPU and what do you want to do?

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[brittle wing]

oh, i need a help, this codes is not working on cmd

[sweet notch]

how to download

@oak plank

[oak plank]

Click the three dots on the top of voice model name and click download :)

[sweet notch]

thnks

[oak plank]

[brittle wing]

C:\Users\djvey\Desktop\ApplioV3.2.9\env\lib\site-packages\torch\cuda_init_.py:209: UserWarning:
NVIDIA GeForce RTX 5060 with CUDA capability sm_120 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_50 sm_60 sm_61 sm_70 sm_75 sm_80 sm_86 sm_90.
If you want to use the NVIDIA GeForce RTX 5060 GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/

😔 help

this codes are not working

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

u sure u followed the same steps

[brittle wing]

yes

[low shard]

what error do u see?

!give-media-perms 1h @brittle wing

[brittle wing]

frstly I followed instructions on github and then I got the error "system cannot find the specified path" while writing the cmd codes

[low shard]

on github?

the instructions arent on github

u just need to read the ai hub docs guide

[brittle wing]

oh

i clicked this

,

oh sorry

i mean, aihub docs

its not github

[low shard]

clicked what

[brittle wing]

i clicked this article

also i can access the Applio localhost site, but it gives this error on CMD

I think I need to run the codes you shared here to fix this error, but I get the error "system cannot find the specified path"

[low shard]

can u uninstall everything, and share a screenrecord of what did u do?

[brittle wing]

Do you want me to reinstall everything

when i unistall everything

Applio, etc.

[low shard]

yeah

like a screenrecord

of u reinstalling

[brittle wing]

oh, sure

[simple ore]

https://github.com/IAHispano/Applio/issues/1020

[viral mason]

Don't use yt tuts 😭

[brittle wing]

🥺

Thanks so much!!! its working!!! @low shard@simple ore thanks again <3

[low shard]

i'm guessing you want a realtime voice changer for calls right?

yeah yt tuts are old

[abstract rain]

Radeon rx 590 Fatboy
windows 10
im trying to fix game audio not working : anyone have an issue where game audio disappears but voice monitoring and other people's voice in the game is present. (Default device is Cable Input) E.g valorant GAME audio is gone but I can hear people's voices in the game and my own ai voice. If I set the default audio to my HD audio driver I hear the game but the AI voice changer doesn't work. Any tips?

[simple ore]

show the screenshot of the voice changer

[low shard]

share the tutorial link you used, and a screenshot of the program

!give-media-perms 1h @abstract rain

[simple ore]

read the guide

-rt

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[simple ore]

500e for a model with 163MB index

probably overtrained af

anyway, index uses a lot of cpu so things may be lagging

that is why it is recommended not to use any for rt changer

[low shard]

that's outdated

it's original wokada

every yt tut is outdated

what's your pc gpu? what do you want to do?

[azure patio]

is it possible to continue training my model?

[blazing carbon]

do ctrl + shift + esc

[low shard]

You can check your pc gpu on Windows via:

ctrl+shift+esc (task manager) -> Performance tab -> GPU

your pc pgu is the most crucial part

open source ai is complex and intensive

it can't run on weak hardware

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

@nova sundial any updates?

[azure patio]

i am training a voice model on kaggle applio, and already have done 150 epochs but want to continue for more. My GPU i dont remember exactly but it was not good enough so im instead just using the T4

[low shard]

that's not a gpu name

exactly

it should be there, as an example from google

[blazing carbon]

i think they got some 8gb vram gpu

it says 4 from 7.9 not sure tho

[low shard]

are you sure? did you check up in task manager?

[azure patio]

yeah i got nvidia quadro p620

[pseudo swift]

changing the volume envelope on local applio seems to make breath sounds really loud, is there a way to fix this?

[viral mason]

why the random russian

[low shard]

This server is english only

[viral mason]

don't feel like translating that

[pseudo swift]

so these are supposed to go down right? is this model doomed?

[viral mason]

up=bad

[analog obsidian]

they're supposed to go down and have low values yeah, they usually fluctuate and randomly go up but as long they're not over 500 it's fine

iirc if its over 1k things gets weird

idk what causes them to go up and have high values honestly

[pseudo swift]

it reached 2.3k at its highest >>

[pseudo swift]

so this would be the ideal scenario?

[analog obsidian]

the highest value i got here in those two trainings i remember was maybe 500 or 600 (in the non avg50 gradss)

[analog obsidian]

yes

those two graphs are from different datasets

[pseudo swift]

ah damn ok. would trying a different batch size make a difference? or maybe a different pretrain?

[lusty ore]

i need help when i only talk like 2 words it sounds well but if i talk longer it starts to get glitchy

[analog obsidian]

dataset size? if you got 10 mins or more, you can go up to 64 max

[lusty ore]

ts my settings

rtx4070tisuper

[pseudo swift]

16 minutes

[analog obsidian]

reduce extra to 2.7 and raise chunk to maybe 150ms

[pseudo swift]

but wow i only have 6vram but i can use the colab, its just that it bugs out a lot

[analog obsidian]

i used batch 64 in two 17 minute datasets, they were fine

go the highest ur gpu can handle

[pseudo swift]

okay thank you for the help! ill try it again

[analog obsidian]

if your grads are still high then its something your dataset has... im not really sure what causes that instability in both grads, could be the quality (maybe the speaker is using two different mics?), mp3 compression, noise, silence, etc

[pseudo swift]

oh right, i took lossy songs from different sources but i did think some were a bit muffled

i did truncate silence so that should be alright

[analog obsidian]

yea may be the quality

[pseudo swift]

btw does putting in speech affect a singing model?

[analog obsidian]

yes

[pseudo swift]

ahhhh okay i see

[analog obsidian]

in a negative way

[pseudo swift]

shouldve known

ill rework my dataset then, thanks again

[analog obsidian]

good luck

[cold shoal]

Hello how is everyone doing

[red dune]

hi guys
I need a little help
I'm not updated with the newest rvc colabs and I need to make a model
can someone please send me a link for the colab?
is there any guides too? how many epochs to choose?

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

ALL YOUTUBE/VIDEO TUTORIALS ARE OUTDATED, don't use them

[low shard]

hello, do you need any help? this is an AI help channel

[cold shoal]

Hello, I’m trying to be an AI rapper so I guess so

[low shard]

could you elaborate more? like your pc gpu, operative system and what do you want to do exactly

[low shard]

also, there's no specific epochs amount to choose, epochs dont mmeant anything at all in quality

epochs are a unit of measuring the training cycles of the AI model

basically the amount of times the model went over its dataset and learned from it

they don't mean how good is the model, it's just an info provided on how they trained the model by the model maker
More ≠ better
Less ≠ better

There's no way to determinate how good the RVC model is until you try it out or listen to the audio samples if there are

[red dune]

@low shard I want something online my laptop is clinicaly dead

[low shard]

and google colab is a service only for people with bad pcs for cloud computing with limited free time

[cold shoal]

Well, I’m trying to find some good AI models, but I can’t seem to find any

[low shard]

your laptop stopped working?

[low shard]

You can search rvc ai voice models at:

• https://discord.com/channels/1159260121998827560/1175430844685484042
• In https://discord.com/channels/1159260121998827560/1163592055830880266 , Do /find with @earnest musk
• https://weights.com/ (login required)
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• Suggested Models for Realtime Voice Changing (Wokada)
• https://voice-models.com/
• https://thevoicemodels.com/ (for Turkish Models, login required with discord and level 2 on their server)

if there isnt one, you can:

• make it yourself with our docs guides
• Ask a free request in https://discord.com/channels/1159260121998827560/1159290139609137264
• Be aware that we don't allow any paid comms

[earnest musk]

:wave: @low shard, How can I help?

Available Commands:
• @weights find <query> or /find <query> - Search for RVC Voice Models

[red dune]

it works but the gpu isn't suitable for any ai usage, I bought in 2019

[low shard]

may you tell me the name of the gpu?

[red dune]

amd it barely works too

how to check it?

[low shard]

You can check your pc gpu on Windows via:

ctrl+shift+esc (task manager) -> Performance tab -> GPU

I mean just to check

we had some users saying their pc was good, then found out it was fro the early 2000s

then we had others who said their pc was bad, and found out it was actually good

[red dune]

there gpu 0 and gpu 1
gpu 0 is intel uhd graphics 620
gpu 1 is NVIDIA GeForce MX 130

[low shard]

yeah you were right 😭

[low shard]

Train (make) RVC Models on cloud:

1. Prepare the Dataset
2. Setup RVC:
   Choose a cloud way to use RVC,

• Google Colabs (max 4 hours of daily T4 16gb gpu not granted for free, not much hours for training, but easy to use, there's a paid tier):
  • Applio (UI)
  • Mainline (U, broken right now)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus, either T4x2 16gb each or P100 16gb, only free):
  • Applio (UI)
  • Mainline (UI, broken right now)
• Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly, Free Studios run 24/7 but require restart every 4 hours. There's a paid tier):
  • Applio (UI)
  • Mainline (UI, No guide as of right now)

3. Be sure to know about the tensorboard

Google Colab = Easier but risk of getting disconnected
Kaggle = Harder but way more gpu time
If you are looking for the easiest way and for free, is using https://weights.com/ which ofc uses RVC

RVC Inference (use models) on pre-recorded audio on Cloud

You can use either:

• Weights.com: Easiest Possible Ever Automatic
• Ilaria RVC Zero: Fastest free on cloud
• Applio UI Colab: RVC Fork with some extra features like TTS
• RVC AI Cover Maker UI: Automatically Separates the Vocals and Instrumentals, converts the voice and mixes them back

[low shard]

also id really suggest getting a better one

[red dune]

😭😭
I want to buy a new one, but I need to save a lot
laptops here are so expensive
not only laptops anything is expensive technically 💀😭

I'm saving already and waiting for the black friday or a miracle

[low shard]

goodluck 😭

the mx130 dedicated gpu was released back in 2017

[red dune]

it's way older than what I thought 💀
but 2019 from 2017 isn't a long time
I never thought that I was going to use AI back then, I was just a freshman + in Egypt it felt like I bought a new model trust me 💀

[red dune]

thanks for the help anyway ❤️

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

be sure to not use video tutorials, they are outdated

[red dune]

for future when I have enough money to buy a better laptop what gpus are good?

[low shard]

what tasks are you going to use it for?

also, laptop gpus are weaker btw

I can help you up if you elaborate what I asked

[red dune]

making ai models, text to image, 3d modelling, and running engineering programs like matlab, solidworks..those stuff

[simple ore]

those are normals, no, but they should not go over 1k

[low shard]

making ai models
im guessing rvc models

I suggest an rtx 20+ serie gpu card with minimum 8gb vram, but its better to have more
and would also be better to be a desktop tbh

[pseudo swift]

is that for the smoothed value or exact value? its still continuing to go up 😦

[simple ore]

you need to look at the loss charts

normal can be up and down, as long as they do not shoot into stratoshpere it is fine

[low shard]

elaborate

[low shard]

@coarse tinsel I need to know those infos to help you up

[pseudo swift]

my grad/norm_g reached 2.3k so this model should be trashed yeah? the loss charts look like this

[low shard]

what m chip is it? what operative system? and what are you trying to do?

Open Source AI is an intensive and complex task

[simple ore]

you can only check the result to see if anything is wrong... usually norms going into thousands is not good

[pseudo swift]

yeah i was waiting until epoch 200 to stop training so i could test. thank you for the advice

[flint vapor]

Is outdated because it use rvc gui and the video is also old

[low shard]

That video is extremely outdated
There's no updated tutorial for realtime voice changing nor RVC (RVC doesnt mean realtime voice changing)

If you elaborate what you want to do, pc gpu, operative system etc i can help you

Also, may I ask why are you trying to pass as the opposite gender? are you trying to catfish/troll online as an e girl, or are you trans? I need to know that info to help

that tutorial is outdated asf, forget everything in it

catfishing also includes trick people into thinking you're an e girl, are you trying to trick your friends?

I'm just asking to know how to help you since that tutorial is pretty old, that's why im asking if you're trying to catfish/troll as a girl

not sure if shady posted it, but it is over a year old and pretty old 😭

alright good then, we ban catfishers

you should uninstall everything from that video

what's your pc gpu? i'm guessing you're on windows

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/ ?

[brazen ore]

is it possible to pause a uvr inference? like pause it and resume it later?

[low shard]

yeah that's great then, the info in it is better, and there's also https://docs.aihub.gg/rvc-voice-changer/realism/ that can help, want me to check your settings?

[stray iris]

hello, does anyone have a good guide for uvr splitting voice and instrumental?

[low shard]

alright, let me know fro any issues

[low shard]

check out https://docs.aihub.gg/rvc/resources/dataset-isolation/#vocal-isolation--cleaning

[brazen ore]

@low shard you know if is it possible to pause a uvr inference? like pause it and resume it later?

[low shard]

Nope

[stray iris]

oh, thanks a lot, i was still using docs.aihub.wtf

[low shard]

docs.ai-hub.wtf is the old domain, docs.aihub.gg is the new one

[ancient sleet]

someone help me emulating wokanda?

[brazen ore]

im saying this cuz its taking way too long, i left it infering overnight and its still not done (its one hour audio) i have a gtx1650 do you think on my macbook it will run faster?

[low shard]

emulating wokanda?

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

nah macbook has worse ai support, you could try also other cloud methods like mvsep and the uvr huggingface zero space https://docs.aihub.gg/rvc/resources/dataset-isolation/#vocal-isolation--cleaning, or split the audio
or upgrade to a better nvidia gpu

[brazen ore]

the thing is that im using a custon model, will the cloud method let me import my model?

[stray iris]

and one more thing - i dont have any of the listed in guide models(e.g. Gabox's voc_fv4 or Anvuew mel dereverb v2) in my uvr's download center. It's latest version of app(v5.6.0 from official github). I should download them from the outside somewhere?

[simple ore]

it is in this https://github.com/Eddycrack864/UVR5-UI/releases/tag/v1.8.4

[stray iris]

got it, thanks!

[ancient sleet]

RX 580 2048sp
WIN10 pro
im using the aihub

also i cant use voice and play at same time

i tried using my own gpu but i can only voice without playing

cuz its crashes

[low shard]

well you could try to check if the model is available first, else i'm not sure if you could import it in like @rare loom 's UVR UI HF Space

[low shard]

aihub
aihub isn't a program at all, it's just the name of this server, and this isn't a voice server, this is a general ai server, i'm guessing you mean a realtime voice changer

[low shard]

your gpu is kinda old, play at lowest settings ever 1080p 60fps, show a screenshot of your program, and tell the game name

!give-media-perms 1h @ancient sleet

[ancient sleet]

Okay

@low shard

[flint vapor]

I recommend you to use the deiteris fork, this version of okada is a bit outdated

-wokada

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[flint vapor]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

The guide 😄

[ancient sleet]

what it is fork

[flint vapor]

is a modified version of the original one

But is more stable

[ancient sleet]

i have an amd

[flint vapor]

There is amd version

[ancient sleet]

i download this?

[flint vapor]

Yes 😄

[ancient sleet]

okay

@flint vapor i have a ryzen 5 7600

its better than my gpu i think

i downloaded @flint vapor

i open this .exe

?:

[flint vapor]

Yes

Sorry for the late

[flint vapor]

Don't run it with cpu it will be bad and slower

[ancient sleet]

okay

[ancient sleet]

dw

[ancient sleet]

where the models?

[flint vapor]

You have to go on edit and upload

Btw there is all in the guide and is explain you how to add models 😄

[ancient sleet]

i can upload the default models?

[low shard]

that version is an over year old of original wokada
and vb audio cable causes issues on windows
uninstall everything

[low shard]

RVC = Retrieval-based-Voice-Conversion, the best Few Shots Speech To Speech AI Models (on v2), Inferences (use models) pre-recorded audio (ai covers) and train (make) models. Technically, Mainline RVC does have a go-realtime.bat (aka RVC-GUI), but it's pretty messy and outdated so it's extremely not suggested for realtime.

Wokada = uses RVC for realtime inference. There's 2 main versions, Original made by Wok, and the most suggested one is Deiteris Fork (modified version)

[ancient sleet]

i downloaded the fork one

[low shard]

what?

there's no default models

if you mean the original wokada ones, they suck, its better to not use them, they aren't even rvc v2

[ancient sleet]

what means 60 ms chunk?

im using like this

[low shard]

great, may I ask on what you're going to use the program to suggest you voice models? are you going to use it to roleplay as a character like homer simpsons, or catfish/troll as an e girl?

[ancient sleet]

just roleplaying being egirl

not troll cuz im using this with my friends only xD

[low shard]

so like just use roleply as an e girl with a girl right? nothing else like catfish or troll online?

[ancient sleet]

yes nothing else

look my profile theres my pronoun

[low shard]

great then, catfishing is horrible

[ancient sleet]

im a good person

i dont do that to people

[low shard]

you're on amd, you have to always check the triangle

ofc non nvidia gpus dont have as good ai support, and yours is also old

[ancient sleet]

perf number + 60 ms?

[low shard]

the chunk
you have to put it as the perf value at the top left while in game, + 60

[ancient sleet]

i fixed

oh

triangle again -

[low shard]

input: microphone
output: line 1
monitor: headphones optionally to hear urself

[ancient sleet]

fixed again

[ancient sleet]

i dont need to use virtual cable?

[low shard]

everytime you change settings, you have to check the triangle, that's because you're on AMD

[ancient sleet]

ohh

[low shard]

you need to

A VAC (Virtual Audio Cable) makes a fake audio device, used to re-route the audio of different programs

In Wokada context, it's used to get the output of wokada as the input in other programs

Line 1 is the cable of VAC Lite, if you have vb audio cable, uninstall it

you got vac lite, right?

vb audio cable can cause crackling and randomly stops working on windows

[ancient sleet]

nah i dont

i will unistall vb audio cable

[low shard]

never ever trust a video tutorial or voice changers again, even in the future, always check the server

[low shard]

do it from the windows app settings

[ancient sleet]

im new at this i saw a tutorial they saying to use crepe in the f0 idk lol xd

[low shard]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#virtual-audio-cable

[low shard]

the last time that was suggested was 2 years ago here lol

[ancient sleet]

i unistall?

[low shard]

yeah uninstall it from there

[ancient sleet]

okay

sorry if im being a npc

[ancient sleet]

where i find vaclite?

[low shard]

nah dw

[low shard]

read https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#virtual-audio-cable

[ancient sleet]

thanks

[low shard]

yw, read everything, be careful, missing a step can fuck up your whole audio system

[ancient sleet]

ohh

@low shard

i change the default device to my mic?

[low shard]

yes

that is the exact step that can fuck up ur audio system if u dont do it correctly lmao

[ancient sleet]

now i think its right

[low shard]

yep

[low shard]

set extra to 0.5

on wokada deiteris fork, you can **optionally **use more advanced settings for benefits:

• Force FP32 mode: on (THIS IS OFF BY DEFAULT! Turning this on improves stability. Increases VRAM usage by 200 MB)
• Reduce the delay via: https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#reduce-more-delay

[ancient sleet]

what method you recommend? in the reduce delay

[low shard]

wasapi is easier, flexasio is better, your choice tbh

[ancient sleet]

i will flexasio then

i need the .6x?

[low shard]

yes

[ancient sleet]

flexasio gui didnt open

i downloaded the 6x here

@low shard i can join support vc for u help? i can screenshare

[low shard]

I cant VC

[ancient sleet]

Okay

@low shard i installed

and i still cant open

[low shard]

you need to just download the .exe and install it thats all

[ancient sleet]

i mean the GUI

when i open nothing happens

this was installed

wait forgot it

this not the runtime :c

now it works yay!

[low shard]

🔥

[ancient sleet]

good?

[cerulean warren]

hey i need some help it isnt working

cant submit images too

i followed the guide but i cant put it to work

might be something stupid but i dont understand anything thats in front of me lol

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

What isn't working? There's thousands of AI programs

we aren't just a voice server

[cerulean warren]

ah yeah

its the voice

ai thing

i put the screenshot in the thread

[low shard]

there's thousands of programs who could be classified like that, but yes will check #1192011222023950368

[ancient sleet]

but in the tutorial

its saying to use flexasio

[low shard]

sorry was meant to reply in that other guys thread lol

[ancient sleet]

oh

okayy

[low shard]

seems good, just be sure to keep chunk higher than perf value

[ancient sleet]

my game crashed ?c

:c

[low shard]

what game are u trying

[ancient sleet]

minecraft

[low shard]

set lowest graphics ever, 30/60fps capped, 720 or 1080p

close everything else in the background

dont expect too much from that old gpu, is bare minimum

[ancient sleet]

okay

[ancient sleet]

should i use index?

[low shard]

in rvc context:

• pth files: contain the voice
• added index files: contain the accent
• metadata.sjon file: it's just some extra info about the model download link if you downloaded it off weights.com, it's not needed and won't impact the actual model at all

using the index, will result in being more similar to the person's accent, but may cause autotune on some cases

[summer ibex]

guys i have a question

the voice changer everything kinda works only problem is when someone talks to me they hear theire own voice delayed in the changed pitch how do i fix that, the voice changer dont change the voice of whoever speaks to me

[ancient sleet]

thx

[ancient sleet]

i think its echo cancelation or noise supression should be on

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

i have a feeling they are using an outdated one from yt tuts ngl

[ancient sleet]

XD

[summer ibex]

yeah maybe that

[summer ibex]

yeah i did that

[low shard]

can you please elaborate the info I asked you so I can help you out ?

be sure you are using wokada deiteris fork and not video tuts

there's no updated video tut for realtime voice changers currently, and prob will never

[low shard]

yw, need any other help?

[ancient sleet]

i cant play using the voice, i think i tried minecraft and roblox "just when i open the game it crashes"

[summer ibex]

rtx3060 windows 10 i want a voice changer who i can speak to people without them hearing themself in a weird pitched voice

[low shard]

try to close the program, open the games, set lowest shittiest settings, then open the program and use fcpe as the f0

[ancient sleet]

fcpe?

[low shard]

im guessing realtime voice changer for calls
and that issue seems to be because you also used an old program

could i ask if you are going to use it for like fun and rp like "peter griffin", or catfish/troll as an e girl?

[ancient sleet]

i should put a higher extra?

[low shard]

yep, sacrifices the quality a bit but will work atleast

[summer ibex]

cant send screenshoot

[summer ibex]

i tried both

[low shard]

extra controls the quality, the higher, the more its heavy and can cause cutoff issues over 2.7, its better to keep it as 0.5 for ur gpu

[summer ibex]

and yes thats seems to be an older model

[low shard]

wdym with both? like you are going to use it for catfish/troll as an e girl too?

[ancient sleet]

oh ok

so should i llower it for gaming?

[summer ibex]

bro idk i just download a few one until now i just pranked my friends and wanted to use it on rp

but they all sound like fkin batman

[low shard]

keep it at 0.3 or 0.5,
try to firstly set the game settings to the lowest

then let me know

[low shard]

yeah i just needed to know so i can help you the best for models and settings, that's why im asking if you're going to use it just for rp, or catfishing/troll e girling too

[cerulean warren]

when i tried opening valorant it gave me a grapics error, let me see if i can show it to you

[summer ibex]

for now rp

or on discord with friends

[low shard]

could you please open a new #1192011222023950368 as it might be a long convo?

and you can share screenshots here

[cerulean warren]

yep no problem

im just restarting everything

[low shard]

alright, so no catfishing/troll stuff right?

[low shard]

alright, let me know!

[summer ibex]

no

[low shard]

great, as catfishing is bad

[ancient sleet]

cant open the game XD

[low shard]

!give-media-perms 1h @summer ibex

show a screenshot of your program

[low shard]

does it open when the voice changer is closed?

[ancient sleet]

yes it does

also i play with like 2k+ fps

[young halo]

i'm trying to resume a training on applio colab and i see this instead of the links

[ancient sleet]

dont use this

[cerulean warren]

wait now it worked

hold on lol

[low shard]

it's minecraft and roblox so its normal anything can play them fine, but actually speaking that gpu is bare minimum for AI tasks, its 8 years old

anyways, you tried to first close the program, set both games graphics to lowest, then open the program?

[low shard]

wokada and rvc/applio are 2 different type of tasks and programs

[ancient sleet]

oh

ok

minecraft its cpu boundgame @low shard

its supposed to doesnt affect

[young halo]

why?

[ancient sleet]

my bad dw

[low shard]

can you try what i said?

[young halo]

can someone pls help with this

[ancient sleet]

but i think i know why @low shard my gpu its bottlenecking cuz my CPU its way better than my gpu

so it is using 100%

[summer ibex]

this one

[low shard]

that one is over a year old version of original wokada, and vb audio cable cause issues on windows

yep you used a video tutorial

crepe is not suggested since like 2 years here

the settings are wrong asf

uninstall everything, forget every piece of info from video tuts

-realtime

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[summer ibex]

okey haha

[low shard]

read up the 1st guide, which is wokada deiteris fork

[summer ibex]

okey thanks man

[low shard]

be sure to delete the old original wokada folder, and to uninstall vb audio cable from the windows app settings

[low shard]

yw

[summer ibex]

aight bro thanks

[young halo]

python3: can't open file '/content/Applio/app.py': [Errno 2] No such file or directory

how can i fix that?

[ancient sleet]

@low shard how i adjust sens?

[low shard]

wdym?

[ancient sleet]

sensitivity microphone

its getting background noises

[low shard]

are you doing it from a new session? you can just do install, then in extra do mount drive, then load a backup

[young halo]

yeah that's what i did

[low shard]

set sup2 on, echo on, and set "in. sens" more to the right

[ancient sleet]

im using server

still works?

[low shard]

oh nvm ur on server

[ancient sleet]

ye

[low shard]

forget the first 2, in sens should still work tho

[ancient sleet]

in sens its in the max -60

[low shard]

the only other way is using another 3rd party software for suppressing noises, which the only one im aware of is nvidia broadcast

but i dont think you have that since you're on AMD

[ancient sleet]

oh

[low shard]

so you either have to find a 3rd party program for noise suppression or have less delay or have more delay

[ancient sleet]

@low shard alright thx

[summer ibex]

daamm im struggling setting up python 3.10

[high basalt]

did they remove e girl tab in voice models

[analog obsidian]

yep

[high basalt]

do you know why?

[analog obsidian]

attempt at reducing catfishing

[viral mason]

yea it's getting replaced by "regular woman" and "hot eguy" trust

[rare loom]

What

[low shard]

oops wrong person, i meant to ping @viscid moss

discord ping moment

[viscid moss]

lmao

[viscid moss]

u have ur own model? What arch?

Depending on arch and config files u can import that model to UVR5 UI/UVR5

But not that cloud version, must be locally

I'm kinda interested 👀

[red dune]

thx for your suggestion
I'll see what I can afford and buy it

[fleet cedar]

why is this happenign to my friend

[unique tangle]

anybody can do pain 93 carti?

[low shard]

#✦│chat message

[fleet cedar]

?

im trying to help him download it

nvm

he fixed

[simple ore]

@low shard ^ scam

[low shard]

his connection is slow https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#help

[fleet cedar]

he fixed it dw

DUDE??

IM TRYING TO START MY OWN OKADA

AND THIS HAPPENED

[low shard]

are yall on the same wifi?

[fleet cedar]

no

we are way far

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do

[low shard]

wtf first time i see that issue, could you try restarting?

[fleet cedar]

no lol

its not a issue nvm

its just cause i had my cmd open

for okada that i opened before

and closed

i fixed it

[fleet cedar]

oh yeah, why is this okada working fine on app of discord

but not website

[low shard]

lol

[low shard]

wdym

[fleet cedar]

the voice changer is working fine smoothly the audio on my discord application

but not website discord

[low shard]

is it having crackling issues on the website discord?

[fleet cedar]

no

just shows no audio

[low shard]

can you show a screenshot of that?

[fleet cedar]

no dude

it just shows no audio

when i try to talk

its not okada

its discord

idk why

[low shard]

yeah i meant a screenshot of discord web settings

[fleet cedar]

i had to zoom out lol

[low shard]

you sure this might not be a vb audio cable issue? did you also try with vac lite on the discord web?

[fleet cedar]

ill try vac lite ig

[low shard]

not sure if this might be a discord web issue, but it's worth a try to check

also because you don't have anything else to test 😭

[fleet cedar]

??

im trying to install vac lite

[low shard]

setup64, not setup64a

[fleet cedar]

mb

[fleet cedar]

vac lite is ass bro

it doesnt show any audio

[low shard]

did you do this step?

[fleet cedar]

yes

[low shard]

could i see your discord web, wokada, playback&recording settings?

[fleet cedar]

wtf

why does it show a echo

[low shard]

what echo?

[fleet cedar]

when i talk

theres echo

[low shard]

"We are not horses"

[fleet cedar]

😭

[low shard]

btw set echo on

then, be sure that line 1 is the input in discord web

[fleet cedar]

still

theres echo

[low shard]

and be sure that it has microphone perms on

[fleet cedar]

it doesnt show any green thing

[low shard]

s3 ending was so sad

[low shard]

so like, what browser are you using? does discord web work when you don't use the voice changer? you sure the perms are on?

@analog obsidian hello lurker

i have summoned you

[analog obsidian]

[fleet cedar]

google chrome

[fleet cedar]

yes

[low shard]

yes to which ?

[low shard]

how's your lurking going, dear lurker

[fleet cedar]

to whichj what

[analog obsidian]

boring

[low shard]

does discord web work when you don't use the voice changer?
or
you sure the perms are on?

[low shard]

help bom up 🔥

[fleet cedar]

i am sure both of them are

[unique tangle]

nun on weights or voice model ai thar has pain 93 in its name

[abstract rain]

can i still send it?

this was the tut

https://www.youtube.com/watch?v=hvPoxsA4lKo

[simple ore]

it is an older version of wokada

[low shard]

It might be better to use the discord app, maybe @analog obsidian got an idea why this happens specifically on the discord web only

[low shard]

Try similar names, or if there's Nun, make it yourself or do a free request

[low shard]

It's an old version of wokada deiteris fork

Uninstall it, keep vac lite tho

-realtime

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

1st link for the updated guide

[wild sierra]

when i run the start_http it opens and closes instantly

[knotty moth]

it's recommended to get the better version here

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/

[grizzled sapphire]

what are virtual audio cables used for? or canble output or wtv

[sterile yacht]

I started the start_http script and it didn't show anything, my cmd opened and then closed, is it normal?

[sterile yacht]

Same problem

[abstract rain]

how to get new version?

nvm i see

[stiff goblet]

Which one is better at making voice models ? Mainline RVC or Applio ?

[latent kettle]

I use applio and it is pretty fine

[livid atlas]

Hey, quick question, I am trying to copy a comfyui workflow from a youtube Video but the guy uses an nvidia gpu, I got an amd gpu, is it possible that I cannot use his workflow because of that? Comfyui in itself works fine, I just cant seem to get the workflow to .. well work or flow 😅

[knotty moth]

idk but try googling some comfyui for rocm or zluda

[finite dew]

I'm trying to test out the voice and it cuts out a lot, anyway to fix this?

[livid atlas]

Thank you for your answer, googling that just results in mostly installation advice in general, I am more interested in the compatibility of workflows when different gpus are used

[knotty moth]

that means it's your own task to troubleshoot your installations

[livid atlas]

that is true indeed 🫡

[knotty moth]

Claude, Gemini pro, etc. could help you tho

and linux (ROCm) is usually better for AMD gpu users

[toxic ermine]

whats the latest google colab link can anyone help?

[livid atlas]

I heard that about Linux yeah.. but I am just too comfortable on windows I think 😅 thank you tho 🙏

[knotty moth]

soon you'd decide to move to linux

linux mint is usually what ppl recommend

[unique tangle]

aight twn i made my own

[brittle wing]

After the recent incident, can we still use Grok?

[rugged mirage]

https://cdn.discordapp.com/attachments/1392473621531988029/1393894169449074698/live_insaan_animatin_3.gif?ex=6874d49b&is=6873831b&hm=4fdc5239b0da8faeea2291398f7405b223a384418afd9920cecff5eaf33930e2&

[rugged mirage]

i need editor to add live insaan text on this gif

[simple ore]

You just need to get comfyui-zluda

[low shard]

that was old original wokada, all video tuts are outdated for realtime voice changer, uninstall everything

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do

[low shard]

A VAC (Virtual Audio Cable) makes a fake audio device, used to re-route the audio of different programs

In Wokada context, it's used to get the output of wokada as the input in other programs

be sure to be using vac lite and not things in yt tuts like vb audio cable

[low shard]

that was old original wokada, all video tuts are outdated for realtime voice changer, uninstall everything

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do

[low shard]

read the guide

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

google colab is just a cloud computing service for people with a bad pc that need good remote power

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

hopefully you aren't using yt tuts

did you unzip the folder first?

[proper forum]

why is my mic not picking up my voice when I put the vc audio virtual cable as input in discord

[proper forum]

same thing for me too but also not working on discord application

[hidden crane]

can anyone help me with my problem?

I am training a model which takes a lot of time but is showing promise. Unfortunately, colab is hitting its gpu usage limit before i can do anything significant

any free alternatives?

[pseudo swift]

i think you can backup and resume training on a different account

[hidden crane]

i know i actually changed it 3 times. 😦 A faster way would be better.

Did anyone use azure ml? I heard its free

[pseudo swift]

oh wow you must have a really long dataset then

[hidden crane]

yeah but the model is simpler and the accuracy and the loss are both improving without showing overfitting. Trained it for 3 days and i am tired now 😦

[pseudo swift]

mb you can try kaggle altho they have limits too. sorry i cant help more but good luck

guys what audios do you use to test your models?

[knotty moth]

next time you should predict the colab session time, it's usually 4 hours unless you have used it often recently

also you should download saved model checkpoints (the pth and G & D files) or move them to the google drive while the training is ongoing

[low shard]

what's your pc gpu?

cloud has limited free time

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[hidden crane]

gtx 1060

[twin moth]

for some reason when trying to upload voices to w-okada the progress bar just stays at 0 and im not sure what im doing wrong, could anyone help with this?

[hidden crane]

see, the training is easy because the model is simple, but the dataset is huge

wait i have heard of a fix from my uncle. Lets see what happens

i will ask later

btw, this isn't anything related to llms, just an experiment for a new way to train any models in general.

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

damn it could technically train locally, but not a good option

Train (make) RVC Models on cloud:

1. Prepare the Dataset
2. Setup RVC:
   Choose a cloud way to use RVC,

• Google Colabs (max 4 hours of daily T4 16gb gpu not granted for free, not much hours for training, but easy to use, there's a paid tier):
  • Applio (UI)
  • Mainline (U, broken right now)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus, either T4x2 16gb each or P100 16gb, only free):
  • Applio (UI)
  • Mainline (UI, broken right now)
• Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly, Free Studios run 24/7 but require restart every 4 hours. There's a paid tier):
  • Applio (UI)
  • Mainline (UI, No guide as of right now)

3. Be sure to know about the tensorboard

Google Colab = Easier but risk of getting disconnected
Kaggle = Harder but way more gpu time
If you are looking for the easiest way and for free, is using https://weights.com/ which ofc uses RVC

RVC Inference (use models) on pre-recorded audio on Cloud

You can use either:

• Weights.com: Easiest Possible Ever Automatic
• Ilaria RVC Zero: Fastest free on cloud
• Applio UI Colab: RVC Fork with some extra features like TTS
• RVC AI Cover Maker UI: Automatically Separates the Vocals and Instrumentals, converts the voice and mixes them back

[twin moth]

i'm using windows, my gpu is a nvidia geforce rtx 4070 super, i want to upload an ai thing to w-okada to change my voice, i'm using this as a tutorial https://www.youtube.com/watch?v=_JXbvSTGPoo and i don't seem to have image perms

[low shard]

that tutorial is outdated

dont trust ever video tuts for realtime voice changers, they all get outdated in sonic speed, because ai is too fast

so yeah, you're using an over year old version of origina wokada, and vb audio cable which causes issues on windows

plus the settings used are pretty messed too

@twin moth delete the old original wokada folder, uninstall vb audio cable from windows app settings, forget everything said in it

[twin moth]

i'm using version 2.1.4 alpha cuda

i don't think that's the one in the video

[low shard]

that's still original wokada, the only difference is the UI lol

[low shard]

also, may i ask what are you trying to do to help, like roleplay for fun as characters as "peter griffin", or e girl troll/catfish?

[knotty moth]

we don't support that version, unless you want to use the beatrice model

[low shard]

beatrice models are lightweight, but lower quality compared to rvc lol, no one uses them nor trains them infact, original wokada doesnt have the best performance and quality

[twin moth]

i just want a funny voice

[oak plank]

Nick, you should get the bot role

[knotty moth]

fr

[twin moth]

where do i download the right version of wokada?

[low shard]

alright then, that's great as catfishing isn't allowed lol

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

1st link, wokada deiteris fork

[low shard]

[oak plank]

😭

[twin moth]

I'm assuming I need to click wokada deiteris fork

[low shard]

Hello — how may I help you?

[twin moth]

is that right?

[hidden crane]

Epoch 1: 100%|██████████| 3907/3907 [03:12<00:00, 20.32it/s]

Epoch 1 | Train Loss: 0.9888
🔸 New best model saved.

Epoch 2: 100%|██████████| 3907/3907 [03:13<00:00, 20.21it/s]

Epoch 2 | Train Loss: 0.7292

from 15 i have got the loss down to this 😎

And its still not overfitting. I think i will be reaching the end of the training soon

[low shard]

you have to read the guide, if you're talking about this, this is just an introduction lol

the download link is below

[pseudo swift]

what is train loss about...? arent you supposed to look at the graphs instead to determine best model

[hidden crane]

i have a callback but it hasn't activated yet because the model was always improving. And i trained it for 3 days so couldn't make a graph

[oak plank]

😂

[pseudo swift]

dont really get it tbh but ok

[hollow crater]

when i generate video on veo 3 it says "something went wrong" what can i do?

[brittle wing]

hey can anybody call me and see if the ai voice changer working or not

[quartz moon]

hi
is there a free tts service that i can run on my pc?
and if yes how
but like i want to download the mp3 file for it i dont want it js to read for me

[simple ore]

what language?

[quartz moon]

english

[simple ore]

do you care about voice cloning? what's your gpu?

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

GeForce 256 /j

[viral mason]

I don't understand all that but good job for getting it to 2 epochs

[viral mason]

The graph and listening to how the model sounds

Doing both will be good :3

[analog obsidian]

they're not training a rvc model

[viral mason]

Are they training a pretrain?

[analog obsidian]

no, its another ai

in other ai architectures the losses do matter a lot

in rvc our only good metric are our ears

[simple ore]

answer

[quartz moon]

answer what

[simple ore]

do you care about voice cloning? what's your gpu?

[quartz moon]

rx 570
idk what that even is

[simple ore]

then you can use something like edge_tts, the microsoft edge screen reader

[quartz moon]

i found one called kokoro but idk where to put the pip install stuff

[quartz moon]

those only read tho

right?

[simple ore]

ask chatgpt

[quartz moon]

k

[simple ore]

kokoro requires gpu

edge tts does not

https://github.com/rany2/edge-tts?tab=readme-ov-file

it is online

[quartz moon]

oo ok ty

[simple ore]

you need to install Python

[quartz moon]

i think i have python

[simple ore]

https://www.python.org/ftp/python/3.11.9/python-3.11.9-amd64.exe

make sure you check [x] add to Path at the bottom when you install

[quartz moon]

ye i have it

[simple ore]

run cmd.exe, then type 'where python'

[quartz moon]

then after that i paste the lines in there right?

[simple ore]

before

just to test

[quartz moon]

nothing happened

oh im on terminal right

ye it told me where it is

[simple ore]

C:\Users\user\AppData\Local\Programs\Python\Python311\python.exe which python it is?

311 or 312 or 313

anyway, run pip install edge_tts

here's the script you can use.. save it somewhere, make tts.txt file with the text to read, click the script to run it

it makes tts.wav file

[pseudo swift]

oh interesting wonder if its better than rvc

[viral mason]

Voice cloning is like taking someone's vocal cords out and putting them onto yours without all the violence

And through a machine

[quartz moon]

kk tyy

[quartz moon]

oh ok

so mimicing human voice

[simple ore]

mimicking someone else's voice

if you want the text read to you by Bill Gates

instead of a random nobody called Ryan

[quartz moon]

oh ok

ty

[viral mason]

Yeah fuck that Ryan guy

[low shard]

We all prefer that Ryan guy 💔

[simple ore]

anyway,with edge tts, you can dump an entire book into a text file and let it read

takes basically no time and steals resources from Microsoft. win-win

[viscid moss]

@marble folio

-colab

[patent trellis]

📒 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

[marble folio]

anyone have a good google collab

[viscid moss]

Try the Applio one

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do

[low shard]

are your windows and gpu drivers up to date?

[viral mason]

Kaggle is better

[digital cairn]

Any ideas what might have gone wrong? My wokada deiteris fork was working fine week ago, I tried today and the voices are super glitchy (robotic and breathy), I tried fresh reinstall of deiteris but no help. I run it on RX 6750 XT 12GB on Fedora Linux

[low shard]

please show a screenshot of ur wokada deiteris fork settings

!give-media-perms 1h @digital cairn

[digital cairn]

sure thing, 1 moment

I haven't changed anything since the last time i used it about week or so ago

[low shard]

can you also show advanced settings?

[low shard]

turn force fp32 mode on, a bit more delay but helps with stability and model's quality

[low shard]

also, what model are you using? are you using wokada for some fun roleplay like as "peter griffin", or to troll/catfish as an e girl in games/vc?

[pseudo swift]

is there a local version of https://huggingface.co/spaces/TheStinger/Ilaria_RVC ?

or is it better to just use mainline?

[digital cairn]

Using it for experimental work (audio research), no usage in online environments

[low shard]

What's your PC GPU & Operative System? What are you trying to do?

[digital cairn]

Doesn't seem to fix the problem

[low shard]

for experimental work about audio research? you're talking about your models sounding glitchy, it's a realtime voice changer

[analog obsidian]

yeah, its named rvc project and applio

[pseudo swift]

rtx 2060 windows 11, i just want to infer. i was using applio but it has issues with volume envelope

[digital cairn]

No, I understand that, I've been using it for past 3 months so I do have knowledge of how it should sound normally, but this is completely off

[pseudo swift]

when i use local and colab applio, adjusting the volume envelope makes breathing sounds really loud for some reason. but when i use that hugging face its fine?

[analog obsidian]

yes because volume envelope is broken in applio

[pseudo swift]

i see

[analog obsidian]

ilaria rvc is mainline inference

so it works there

[pseudo swift]

ahh okay thank you

[analog obsidian]

[pseudo swift]

😂 hope it gets fixed one day...

[analog obsidian]

actually noobies gave up and just removed it

[pseudo swift]

nyooooooooooooo

[low shard]

there is technically an Ilaria RVC local fork #📰│dev-updates message tho i'm not sure on if it is even still maintained, so that's why it's not much suggested

it would be better to use Applio and wait for a fix on volume envelope which I'm sure @simple ore is aware of, or use mainline (the original rvc, which is the same inference used in ilaria rvc)

[analog obsidian]

its weird coz the applio inference code is the same exact used in mainline

but for some reason volume envelope never worked properly in applio

[simple ore]

i dont fking know why volume envelope is messed up

I did remove it in the exp branch

[low shard]

I mean, can't you just revert the code related to it?

[simple ore]

because whatever it is it uses frames to calculate loudness

[analog obsidian]

applio inference code is the same as the one used in mainline
volume envelope not working there is an actual mystery

[simple ore]

so during silence it just boosts that to stratosphere

[low shard]

if they are the exact same.. why the fuck is there a difference 😭

[analog obsidian]

exactly lmao
i mean there is no difference, is just that volume envelope broke for no reason

[low shard]

this can't be possible lmfao

[mental spruce]

where do i get the voice changer?

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do

I'm guessing you didnt follow any tuts yet

[mental spruce]

nope.

[low shard]

alright, elaborate what I asked so I can help you out

[mental spruce]

GPU Nvidia GeForce GTX 1660 super

windows 10

[simple ore]

i tried to figure that out but after Blaize's changes who the fuck knows

[low shard]

what models are you using exactly? have you tried other ones?

[mental spruce]

and i want to test it

[simple ore]

I've checked the code and after fixing the variables it is exactly the same

[low shard]

are you in the need of a realtime voice changer for fun roleplay characters like "peter griffin", or for trolling/catfishing as an e girl?

[analog obsidian]

maybe try a different browser

[mental spruce]

for fun

[low shard]

so, will you do e girl trolling in games(like valorant)/vc? I need to know this to know how to help u out with the best models and settings

[mental spruce]

just trying to test it...

i aint going to be using it for games

[low shard]

yeah but I just need to know if you're going to use it for e girl trolling/catfishing too or not

[mental spruce]

nope

[low shard]

great then, catfishing isn't allowed

[digital cairn]

Sorry can I have mp3 files permission? I just downloaded the "Peter Griffin" voice model from #1175430844685484042 released on 21th June and used the same male-to-male pitch as with my other male voice models (-1) and would like to send the recording for you to see what it outputs

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

Guides for Programs that use RVC Models in Realtime for Calls/Games

• Wokada Deiteris Fork

Most suggested. GUIDE

• Original Wokada

ONLY the latest alpha comes close to the Deiteris Fork performance, older versions in youtube tuts are way worse. GUIDE

• **RVC GUI Mainline Realtime**

Unavailable, the guide is outdated and the program is worse compared to the ones above, and much less updated

[low shard]

1st link

wokada deiteris fork

read it up

be sure to not mess steps

[mental spruce]

thanks

[low shard]

!give-media-perms 1h @digital cairn

[low shard]

yw and lmk

[low shard]

could you please try other models?

also know that rvc got limitations for things like screaming and laughing

[digital cairn]

Here is with a female voice model

[digital cairn]

That I am aware of, didn't plan on using that anyways

[mental spruce]

Um where is the windows link to download it

there is only for Mac

[digital cairn]

depending on your GPU you have windows downloads above the Mac ones

[low shard]

https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#download-nvidia-on-windows

be sure to not miss https://docs.aihub.gg/rvc-voice-changer/local/deiteris-w-okada-fork/#virtual-audio-cable

[low shard]

could you try searching for your distro updates, gpu drivers update, and maybe try for a reinstall of the program? be also sure to try many models

[digital cairn]

I will try upgrading to higher version of distro, though I doubt this will have positive effect on this, will keep you updated, thanks for the help so far

[drowsy bronze]

how do i change from rvc to rmvpe

like

how do I use an rmvpe model

because I'm kinda dumb

[digital cairn]

Yeah so as I exepected, update only causes more problems, now on running RVC i get this error:
_internal/onnxruntime/capi/onnxruntime_pybind11_state.so: cannot enable executable stack as shared object requires: Invalid argument

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

wtf? could you pease check if your drivers are up to date and maybe try reinstalling?

sorry but we don't really have many linux helpers as windows is the OS chosen by the majority of users, maybe @simple ore got a better knowledge on this

[digital cairn]

It's the case with linux that you usually don't want to update anything because it tends to break python related projects, this is the case as well

trying one more fresh reinstall

yep no help, oh well probably will switch to NVIDIA with the amount of issues i've been having lately related to ML

[unique tangle]

my rvc doesnt catch the index of models any idea why or an solutions?

[viral mason]

?

could you explain what you mean by that

[low shard]

This is a General AI Server, AI has many fields

Elaborate:

• your PC GPU
• your operative system
• what you want to do
• what tutorial link are you using
• a screenshot of the program

[low shard]

I mean, Nvidia has better ML support, which is unfortunate for non nvidia users

[digital cairn]

no doubt, but also worse linux support, no ideal gpu nowadays ugh

[viral mason]

love being an nvidia user

are the prices high for the gpus?

I might upgrade mine someday

[low shard]

aren't the official closed source driver support great? since every cloud computing service uses linux + nvidia

[low shard]

what's ur pc gpu