✨│ai-help

[analog obsidian]

https://huggingface.co/GaboxR67/MelBandRoformers/resolve/main/melbandroformers/vocals/voc_fv4.ckpt?download=true
https://huggingface.co/GaboxR67/MelBandRoformers/blob/main/melbandroformers/vocals/voc_gabox.yaml

gabox fv4

https://huggingface.co/becruily/mel-band-roformer-karaoke/tree/main

becruily's mel karaoke, for vocals that have harmonies

[stiff idol]

I found it a bit baffling that CN instruments were forgotten

[analog obsidian]

u need to use those in the original audio with the instrumentals + vocals tho

[stiff idol]

You mean to load the original audio? No problem. lol I'm trying to learn audio editing

just so I cna make a good dataset because I will never have audio without SE or music and other stuff

[analog obsidian]

yea, if your original audio has harmonies you use becruily's, if not, you use fv4

[stiff idol]

I think I'll have to use an excel sheet like I had for LoRa models

too many models are coming my way

but it doesn't take long and it's quick even on my GTX 1650

I wanted to try Reaper instead of Audacity but it looks like a Spanish village to me (or idk how people say it 🤣 )

now only to make it a bit sharper and that'll be it, I don't think I can make the quality better 🧐

[fossil sage]

yes

i noticed anything above 90 epochs it has more noise

xtts

but its to hard to install

i was told xtts was the best at reading out loud

thats what im looking for

[teal ferry]

thats what my whispering model is but its outdated model

and has dumb quirks

i wouldnt use it

[fossil sage]

????????

sound good to me

[teal ferry]

yeah but the effort to get to that was significant

[fossil sage]

yes

😭

now how do i install xtts

[teal ferry]

if you installed the right version of cuda it should install for you

[fossil sage]

?????????????

should i just use the google collab at this point

[teal ferry]

no colab will most likely kick you before the thing finishes

just try to install it again, pussy

[waxen wigeon]

Anyone know the AI that lets you combine rvc voice models to merge voices and where I can download it?

[fossil sage]

THATS thats im trying to do

[fossil sage]

@royal kettle is it posssible to change the embedder for wokada

@teal ferry E:\xtts\alltalk_tts\alltalk_environment\env\Lib\site-packages\torch\cuda_init_.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]
[AllTalk Startup] AllTalk startup Mode : Standalone mode
[AllTalk Startup] WAV file deletion : Disabled
[AllTalk Startup] DeepSpeed version : Not Detected
[AllTalk Startup] Model is available : Checking
[AllTalk Startup] Model is available : Checked
E:\xtts\alltalk_tts\alltalk_environment\env\Lib\site-packages\torch\cuda_init_.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]
[AllTalk Startup] Current Python Version : 3.11.13
[AllTalk Startup] Current PyTorch Version: 2.8.0+cu128
[AllTalk Startup] Current CUDA Version : 12.8
[AllTalk Startup] Current TTS Version : 0.22.0
[AllTalk Startup] Current TTS Version is : Up to date
[AllTalk Startup] AllTalk Github updated : 19th March 2025 at 20:50
[AllTalk Startup] TTS Subprocess : Starting up
[AllTalk Startup]
[AllTalk Startup] AllTalk Settings & Documentation: http://127.0.0.1:7851
[AllTalk Startup]
E:\xtts\alltalk_tts\alltalk_environment\env\Lib\site-packages\torch\cuda_init_.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[

[teal ferry]

You need to grab the location of python.exe again. In the env folder. Paste that location into a terminal then -m pip uninstall pynvml then paste it again -m pip install nvidia-ml-py

[fossil sage]

E:\xtts\alltalk_tts\alltalk_environment\env\Lib\site-packages\torch\cuda_init_.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]

didn't work

@teal ferry IT FINALLY LAUNCHED

FINALLLLLLLLLLLLLY

[royal kettle]

no

Use vonovox

[viral mason]

If u have Nvidia use Vonovox

[fossil sage]

o man who was the guy who just left the vc

[teal ferry]

Nice

[fossil sage]

time to read the documentation

[stiff idol]

I'll try these models in a jiffy. I also found what was "that" problem. The problem is that I want to vocals only, but there is this instrument called "Erhu" that sounds like violin combined with harmonics. The models don't know this instrument.

[fringe snow]

Is Kaggle broken? The urls are all the same so Applio takes me to Tensorboard a few times before it takes me to Applio, then I try to use it and then I get an error that reads "Connection Errored Out"

[fossil sage]

this is horrible

and it told me to use 22050Hz, Mono, 16 bit to

the documentation said i can use a model

and imported it

but idk how to use it

[teal ferry]

id use 48khz for the reference

and stereo

[fossil sage]

ok imma try that

[teal ferry]

its 24khz to train

[fossil sage]

[teal ferry]

youd want to test on out of distribution data

[fossil sage]

whats that

[teal ferry]

or write some words that arent an exact match to the reference audio

[teal ferry]

sounds like it could be the same as reference audio

[fossil sage]

mman this is horrible

[teal ferry]

then make a better one

[fossil sage]

the reference audio is long enough and its high quality

i dont understand what else i could do

[teal ferry]

make a better model

[fossil sage]

how do you import models into xtts

i looked in the docu

[teal ferry]

do you have a model to import

[fossil sage]

yes i alredy imported it

[teal ferry]

are you implying that you intend to use an rvc model where an xtts model goes?

[fossil sage]

yes i did

[teal ferry]

you cant

[fossil sage]

OH

THAT MUST BE WHY

[teal ferry]

models are not cross swappable

[fossil sage]

so a xtts model is just a .wav file

[teal ferry]

no

[fossil sage]

how do i tune it

[teal ferry]

follow the instructions

[fossil sage]

doing it rn

[fringe snow]

Kaggle is cooked

[fossil sage]

when creating a model why do i have to use low quality audio

22050Hz

[teal ferry]

several reasons. mostly has to do with the size/shape of the tensor

[fossil sage]

imma train this one

[fringe snow]

bruh now it's just taking me to my files

https://tenor.com/view/tense-table-smash-mad-gif-13656077

[stiff idol]

Don't models expect mono?

[stiff idol]

🧐

[teal ferry]

if you can in some cases you cant

but xttsv2 has no problem with it

it depends on model type, i believe its the diffusion archetecture because of the latent space

[stiff idol]

I've just found a good model for vocals that removes a chinese instrument from vocals. finally (Mel-Band-Roformer-Karaoke-Gabox)

[teal ferry]

so youd be mapping the empty space with 48khz or a higher resolution

[stiff idol]

it can't deal with multiple vocals but that's fine, ... it's an experimental model, after all

[teal ferry]

if its just one thing of audio davinci resolve is kind of a hidden gem

its ai audio stuff is really good

[stiff idol]

Which AI stuff? I may not have caught on that.

I have yet to explore other types. For now only RVC and Stable Diffusion.

[teal ferry]

audio seperation from music and noise

[stiff idol]

I have a question, which model is the best for removing reverb and echo form audio?

[teal ferry]

i dont work with music often. when i denoise i just use mel band roformer

the original one by kim iirc is her name

[stiff idol]

Could you write the exact name of hte model?

Kim has many models

[teal ferry]

https://github.com/IIEleven11/Automatic-Audio-Dataset-Maker/tree/main/tools/denoiser/Mel-Band-Roformer-Vocal-Model

here i have it in my repo

[stiff idol]

thank you

[fossil sage]

Could not locate cudnn_ops_infer64_8.dll. Please make sure it is in your library path! its not in my libary path what do i do

[pure obsidian]

yo can anyone help

how to set up

some voice changer

[tight halo]

HI GUYS

IM AI

[stiff idol]

ok, so my test was: Mel Band Roformer Karaoke V2 (experimental) by Gabox is better for the audio where I tried separating vocals and isntruments

[languid kernel]

Ok

[young halo]

i just opened the damn kaggle

[low shard]

check #📰│dev-updates message , it's not a bug related to Applio nor Kaggle

Get your own Ngrok account, also check #📰│dev-updates message

Cloud shouldn't be your first option

[low shard]

It's not a Kaggle nor Applio issue, the Ngrok Tunnel used to make the Public URLs just got a stricter free tier compared to before, read: #📰│dev-updates message

[fringe snow]

Damn rip

[graceful igloo]

Is there an alternative to weights.com? I want to train a voice model, but now weights.com is charging people to train voice models and I like to use ai for free.

[young halo]

is there any way to fix this? my screen turned off and then it was like this, it still training but not showing up on applio/kaggle

[knotty moth]

he was just showing his application repo that only pre-includes the single og Kim model. you should try this colab that lists many more models (including newer roformer models being better than the og Kim) or clone the MSST repo for local use: #📰│dev-updates message

[knotty moth]

you can also try becruily/frazer model (try search it if not sure)

[young halo]

How can i resume training on kaggle?

[split warren]

i dont understand why i can hear myself and is like 3 sec delay?
can someone help?

[fringe snow]

Unfortunately Ngrok just got that much stricter for free users they dropped the number of requests from 4k to 120 making kaggle extremely buggy and unusable

Besides for paying for more Ngrok benefits, there is a workaround you can do in #📰│dev-updates

[young halo]

i want to stop the training but when i press that button nothing happens and the training goes on

[meager lichen]

how do i uninstall okada voice changer?

i installed the original one but i want deitaris version

[young halo]

help, how do i stop it? the button wont work

[sonic agate]

who pinged

[nocturne mural]

Sometimes the interface just disconnects, and to stop it, you’d have to manually stop the cell in the notebook.

[knotty moth]

@sonic agate sorry was about to summon you but it turns to be a rehosted og Kim model instead of his own trained model

[teal ferry]

correct

[sonic agate]

xd

[deft finch]

I am trying to install Automatic1111 with an AMD graphics card. I followed all of the install instructions, but whenever I run the webui-user.bat file, I keep getting this error and I don't know how to resolve it. "RuntimeError: Torch is not able to use GPU; add --skip-torch-cuda-test to COMMANDLINE_ARGS variable to disable this check"

[viral mason]

never heard of it

[deft finch]

Wdym, it's one of the most popular Stable diffusion uis. Or so I have heard idk. This ai stuff confusing fr. I don't blame you.

[simple ore]

automatic1111 is old and dead

use SD.Next

[hard forge]

Hi everyone,

I wanted to bring up something that’s been on my mind regarding Alignerr. I’ve been finding it difficult to actually secure projects on the platform, and I was wondering if others here are facing the same challenge.

If you’ve managed to get projects through Alignerr, could you share what worked for you? like how do you usually approach projects there?

I think it would be helpful if we could share our experiences and strategies so we can all figure out the best way to make Alignerr more fruitful for us.

Looking forward to hearing your thoughts!

[simple ore]

also if you got like 7900 or 9000 series card you can try manually installing rocm windows wheels for pytorch

[elfin cypress]

Yo I need help

[knotty moth]

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[viral mason]

I've never heard of this

[chilly furnace]

literally what this server is about dawg

you can find someone linking resources every minute or so

[tall merlin]

index file not found

[tame yacht]

Hi
So i wanna develop a chatbot type ai which will be modeled after me and will communicate with my social media friends on my behalf and they won't notice whether it is me or ai

So
For that
I will need to learn python

But there are really many varieties in python
So i ask for you guys to help me

I would need to know which are the topics i need to learn in python to be able to start the project

[tame yacht]

Nah let it some times

[brittle wing]

i need help applio installation

[stray pier]

hello, is it possible to request the creation of a voice model that I need?

[stiff idol]

use ComfyUI

haven't heard of people using Automatic111, I tried it at the beginning too, but too many problems with it and it'ss not versatile

[brittle wing]

i need help

[viscid topaz]

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[oak adder]

hei i wondeer is any open source ai like warp or cursor exists thank you .

[low shard]

if you mean like coding agents

https://github.com/Mote-Software/nanocoder this is terminal only

https://github.com/openai/codex this is also terminal only

https://github.com/block/goose?tab=readme-ov-file this has a UI

[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)
• https://voice-models.com/ (just scrapes the huggingface link from the voice-models channel)
• 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/1159260121998827560/1159289738314919936
• Be aware that we don't allow any paid comms in the server

[earnest musk]

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

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

[low shard]

please elaborate

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[subtle cedar]

We can train voice models locally on Chromebook or ChromeOS?

[low shard]

chromebooks are made mostly for just surfing the web and being cheap, most dont even have a dedicated gpu, so no they wouldn't have enough power for any local ai program decently

[lethal shale]

heyy

i didnt use okada for 4 months and now I am back and it doesnt work

can someone pls help

[low shard]

please elaborate

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[lethal shale]

my mic works but it doesnt transfer my voice to line1 i guess

just when I speak I dont get a feedback

[low shard]

that's not enough info for us to define the clear problem, please read above the bot message to understand how to ask for help :)

[river dune]

How can i use Stable Diffusion on AMD GPU?

[lethal shale]

i dont know why it doesnt work either it just doesnt work when I speak the app does nothing

[low shard]

please elaborate your pc gpu, operating system and show a screenshot of your full program

[low shard]

I'm guessing you're on windows 10 or 11
What's your full gpu name? is there a specific ai image model you're looking for?

[river dune]

I have RX 6600 8 GB models.. idk

[lethal shale]

sorry for hdr

[low shard]

this is very low quality, but you seem to be on windows 11 with the rtx 4060 ti 16gb desktop on wokada deiteris fork b2332?

[lethal shale]

oh it works now

echo was cutting all my voice

[low shard]

Vonovox would be more suggested for windows nvidia, it still gets more updates and about performance too

[lethal shale]

omg lol

[low shard]

wokada deiteris fork latest update was b2332, the 7th december 2024

[lethal shale]

im windows 10

which okada should i use? i think that one is outdated right?

[hallow thistle]

"Chrome OS" refers to an operating system that made for Chromebook or any PC that supports it, while Chromebook refers to a type of a laptop that runs Chrome OS especially, which typically manufactured and sold by other PC brands. Unlike traditional laptops/desktops, there's no known Chromebook that is made for training AI locally, but for "online" training it sure can.

[low shard]

windows 10 support is ending 14th october 2025, you're going to need to update soon prob btw, just saying

[lethal shale]

well time to update then hahaha

does it work better with 11 alreadyt

[low shard]

wokada is just one RVC AI Realtime Voice Changer program, which has 3 main current versions:

• original: not suggested anymore as it doesn't have proper rvc performance updates
• forks:
  • deiteris: a modified version which added some performance things, last update was in december 2024
  • tg-develop: a modified version off deiteris, added some quality of life updates but isn't considerated a maintained fork, last update was in august 2025

for windows nvidia, vonovox would be more suggested, which is another RVC AI Realtime Voice Changer for calls

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[low shard]

here are the guides for each

I would ofcourse suggest Vonovox, it's your choice, you can read the pros and cons of each in each guides

[lethal shale]

Okay so I use it realtime in vr should I use vonovox and do I have to delete mine to download it

and is mine original?

[low shard]

I was suggesting you that for your Operating System security and features and support in the future, it's not related about performance

just like you can't use windows 7 forever, as an example

[low shard]

the one you got is wokada deiteirs fork b2332

vonovox gets more frequent updates and migh have better performance for your setup

[subtle cedar]

Oh okay thanks

[low shard]

you can read their specific pros&cons, that would be better if you want to understand more the difference, it's all ur choice tho

[lethal shale]

thanks bro do I have to delete my VC I have to use Vonovox

or can I keep both to compare

[subtle cedar]

Okay thanks

[lethal shale]

do they clash with each other I mean

like different versions of same app so one deletes when other gets in

[viral mason]

You can't use both at the same time, delete the current one you have and then download vonovox

[lethal shale]

Last question, Vonovox is better with performance and convert quality too right?

[low shard]

you can do both

they are 2 different programs that do a similar goal with the same type of models

[lethal shale]

oh really? thats wonderful now

[low shard]

they aren't "updates" of the same program

[viral mason]

Vonovox is cool and has an auto update feature

[lethal shale]

Thanks Nick for helping me for 2 years lol

I always like people named Nick

[lethal shale]

Niceee I will give it a shot now

[low shard]

oh lmao i forgot i talked to you before 😭

You can guess I helped a lot of people here lol

[lethal shale]

thanks to you I made an amazing model and 2 great ones

you also helped me with voice training

2 years ago lol

If I make a website of voice model market I will send you gift lol

[low shard]

I'm glad :)

[lethal shale]

I am just sad I deleted my american accent models thinking they were bad

it is just that I got German accent

so they didnt work lol

[low shard]

lol you aren't forced to even tho that's nice to say, but just be aware that promotions / paid commissions aren't allowed here, just saying in case you were going to promote it in the server

[lethal shale]

I know that hahaha

Thanks for reminding

Well I didnt know but I could guess I think 😄

[low shard]

I mean you can use them, but you'd have to lower or not use at all the index (which is the trained accent)

[lethal shale]

should I use index? I never use them

I feel like it causes inconsistency

[low shard]

Well yeah, we actually did allow people to promote here before some months ago (I think almost half a year ago) but there were just too many scams, scummy services and chaos that it wasn't worth it, even bc not many people were actually checking the promo channel

[lethal shale]

with my models I can sing, laugh, cry

they all work somehow

[low shard]

the index is the trained accent of the voice model, you shouldn't use them especially if the models are in a language you don't speak

[lethal shale]

so my model is german and when I speak english I shouldnt use it

right?

Vonovox is realtime vc right?

[viral mason]

Yea

[river dune]

can somebody help me install stable diffusion for amd gpu ?

[lethal shale]

okay I think vonovox is better

but i got few problems lol

where is echo cancellation

it is so complicated to me right now

[lethal shale]

nickkk

okay I use Steel Series Sonar smart noise reduction and it works well so far

[simple ore]

if you got nvidia, their broadcast app is the best

[lethal shale]

I got Nvidia yeah

but I dont know that one

Ill check now

[simple ore]

https://www.nvidia.com/en-us/geforce/broadcasting/broadcast-app/

[lethal shale]

does it create another microphone like steel series

or just enhance the one you use

[simple ore]

it create an output, yes

you configure it with your real mic input

[pure obsidian]

yo

any expert in here that can help?

[lethal shale]

I will give it a shot now

[pure obsidian]

trying to download the vonovox thingy

[simple ore]

and then you pick nvidia broadcast output as an intput in the other app

[lethal shale]

i am launching

[low shard]

if you mean the index, yes

[pure obsidian]

nick bro

can you please reply to my aihelpforum?

[lethal shale]

I downloaded vonovox

but I dont speak that language

[low shard]

realtime voice changer, ai that uses RVC (Retrieval-based-Voice-Conversion) models like Wokada, yes

[lethal shale]

there is no echo cancellation and sup 1 sup 2 like okada

and I hear myself 3 times lol

Steel Series Sonar made it better but now I will try Nvidia

[simple ore]

you may need to fix the audio path

[pure obsidian]

yo bro can you help me rq?

[lethal shale]

I mean I hear myself with echo

even echo echoes

[simple ore]

means your have some other inputs mixed in

[lethal shale]

I Use Voicemeter

[simple ore]

you dont need to route your actual mic thru voicemeeter

[lethal shale]

My Mic -> Line 1 -> Voicemeter

[simple ore]

stop that

[pure obsidian]

little help please?

[simple ore]

why dont you ask instead of asking to ask

[pure obsidian]

i cant send a photo

[lethal shale]

man you can see we are talking about something

[pure obsidian]

can i tag u in the place where i posted the photo

[lethal shale]

can you stop interrupting and ask when this one is done

[pure obsidian]

yeah bro but u recently started i been asking since 2 hours

[low shard]

I replied

[pure obsidian]

ik but i already downloaded that one

[low shard]

@pure obsidian @lethal shale please don't argue about who gets helped first, there aren't many helpers (not all staffers are helpers, just the ones with the role "helper" are official ones), there are different timezones and sometimes might be multiple people asking at the time time, I hope yall understand we are volunteers :)

[low shard]

let's continue there

[pure obsidian]

okay tysm brother

[pure obsidian]

sorry @lethal shale

[low shard]

show a screenshot of your program settings

!give-media-perms 1h @lethal shale

[lethal shale]

I can already send pics ig

[low shard]

why are you using voicemeeter?

you seem to be using 2 VACs at the same time

you only need Vac Lite

your audio setup should look like:

• vonovox:

  • input: microphone
  • output: line 1

• other programs:

  • input: line 1
  • output: headphones

[simple ore]

you really need to figure out the purpose of a1, b1 buses

[lethal shale]

I will try now

how will I listen to myslef

in vonovox

[low shard]

by using this workaround https://docs.aihub.gg/realtime-voice-changer/local/vonovox/#how-can-i-hear-myself

[lethal shale]

okay I will now try nvidia smart noise + Line 1 only

I will tell difference

then my own mic

and line 1

[simple ore]

A1, A2, A3 buses are hardware outputs, you send them to physical headphones, speakers, etc

B1, B2 are virtual outputs, you used the mix audio, like discord voice chat + game audio to push it to OBS

if you're using a noise canceling software like Nvidia broadcast, you do not select a real mic as any of the inputs

and you generally do not need to mix it to any bus either

you send the real mic to broadcast app, then from broadcast app to voice change, then from voice changer to virtual cable line 1, then you take that as input for discord

and output for discord is voicemeeter's virtual input

technically with banana you dont need virtual audio cable

you can use b1 to route audio from voice changer to discord

[lethal shale]

Nvidia broadcast made my voice like I speak in a cave

[simple ore]

change the effects there

[lethal shale]

I used both now and I can say steel series is much better with it

[simple ore]

it should not add any effects

[lethal shale]

wym

[simple ore]

the broadcast app

[simple ore]

https://x.com/thegunrun/status/1252789873699745792

theGunrun (@theGunrun)

Nvidia's RTX Voice tech looks AMAZING. I can finally use a desk mic again with my mechanical keyboard!

@Barnacules's test with the fan and hammer is so impressive

https://t.co/ES7ELMiMQS

X · Apr 22, 2020, 2:42 AM

[dusty delta]

model isnt loading

and i cant hear myself

[viral mason]

what is nvidia broadcast?

[simple ore]

noise canceling

[viral mason]

ah

only added fx I usually use is de-esser and extra stuff to add fullness in fl studio

[dusty delta]

I NEED HELPP

[viral mason]

like added bass

https://cdn.discordapp.com/attachments/1159290139609137264/1419738795833819269/Jaxcostume.gif

[simple ore]

it is not for recorded audio

[viral mason]

fl studio? it is but I figured a way to use it live and with okada deiteris and any other realtime vc

pretty reasy actually

[iron maple]

Hi, not been in the space for a while but is RVC working with RTX 50 series cards yet?

[low shard]

with RVC, do you mean Retrieval-based-Voice-Conversion, so a fork (modified version) like Applio to train rvc models and use it on pre-recorded audios?
Because RVC doesn't mean Realtime Voice Changer

[iron maple]

apoliges yes that

like applio

[viral mason]

-# told u nvm

[iron maple]

give an audio sample and use a weight to change it locally

[low shard]

if you're talking about applio, yes it works on rtx 5000 series

[iron maple]

last time I tried it, there was issues with pytorch or something on the 50 series

is that fixed?

[low shard]

-# told u people don't think ALWAYS that rvc is realtime voice changer lol

[low shard]

yeah, it got fixed in applio 3.3.0 and newer versions, the latest rn is 3.5.0: https://docs.aihub.gg/rvc/local/applio/#nvidia-on-windows-precompiled

[iron maple]

fantastic, thanks very much I appreciate it

[dusty delta]

dude

i cant bro

[low shard]

please elaborate what you're talking about

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[dusty delta]

okay so my full gpu name is NVIDA GeForce GT 1030 im windiows 11 im trying to start it by running start http bat but it keeps opening then closing

[low shard]

to start what? we help with multiple programs

[dusty delta]

multiple?

[low shard]

Yeah, this is a general AI server, it was previously RVC-Only but we expanded to multiple things

[low shard]

I see you're talking about start http, Did you perhaps come from a youtube video about realtime voice changing for like roleplaying in games/calls from 2023?

[dusty delta]

yes

i think so yes

[low shard]

Yeah they are all outdated, you can uninstall that 2 year old version of original wokada and vb audio cable

[dusty delta]

kk

now what do i do

[low shard]

GT 1030? That's pretty old, I could suggest you Vonovox but im not sure if it supports such old GPUs, you want to try?

Don't expect it to work in intensive games like gta5 FiveM max graphics

[dusty delta]

sure

nah im using it for roblox

[low shard]

-realtime

[patent trellis]

💻 Local Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[low shard]

Read up the 1st guide

[viral mason]

https://tenor.com/view/david-baszucki-dave-roblox-wave-roblox-connect-gif-7510206570543806539

[dusty delta]

i just have to ask a question

i have a intel r hd graphics 530

is that okay?

or no

[low shard]

Absolutely not that's 10 times worse, I'm guessing you have both the integrated GPU and the dedicated Nvidia GPU, the integrated GPU would be absolutely worse and 100% surely not recognized

AI needs power

[dusty delta]

so what do i do?

[low shard]

You can try to see if the gt 1030 works

And let me know if it does or not

[dusty delta]

man i feel embarassed

i have an unc intel card

[low shard]

You said you have a gt 1030, or were you just joking ?

[dusty delta]

no im not

im not joking

[low shard]

I'm guessing you have both the Intel integrated GPU and the Nvidia dedicated GPU, right?

[dusty delta]

correct

[low shard]

If you want, instead of running it locally on your PC, there's cloud (remote good PC) with limited free GPU time, so depending on a remote GPU, and ofc you'd need good wifi for low delay

Would u want to at least try locally, or try cloud?

Hopefully you have a decent wifi connection

[dusty delta]

i do

full bars all the time

[low shard]

Did you do a speed test?

[dusty delta]

not yet

im about to

rocking a 30-35 mb

33 mb and 10.33 for upload

[low shard]

MB or Mb/Mbps? Because MB is 8 times faster than Mb

[dusty delta]

mb\s

[low shard]

MB = MegaByte
Mb = Megabit
Byte = 8 bit

[dusty delta]

this good?

[low shard]

Eh I mean, should be decent for cloud

Soo, would you rather try Vonovox locally to see if it can work? Or do you want to directly try cloud?

[dusty delta]

ill try locally

[low shard]

Alr let me know

[dusty delta]

just gotta download it correct?

[low shard]

You gotta first of all setup vac lite, then download the precompiled Vonovox and extract it

Be sure to not miss steps on the guide

[dusty delta]

vac lite kk

i downloaded vac right?

ran 64setup

and when i opened it

said it had to make changes

so i accepted

and it didnt work

[viral mason]

wait vonovox works on intel???

[low shard]

what?

[low shard]

no

[viral mason]

didn't he say he has intel? or am I missing something

[dusty delta]

SORRY BRO I WAS OUT

but uh

it didnt work

wait its working

nick

alright what now

nick i got it set up

[mellow ermine]

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[mellow ermine]

-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

[elfin cypress]

So my voice changer is going high pitched and will cut out a lot is there a fix?

[dusty delta]

NICK

BROO

I GOT VONX AND DOWNLOADED ALL THE STUFF

[coral kraken]

Can someone please tell me if I can uplod a pth file model to Kits AI to convert my voice to one model I have in my computer? I got the starter upgrade version in kits AI but I can't fine where I can Upload the pth file like I use to do a year ago.

[viral mason]

do not use kits it's horrible

if anything use weights

[last rivet]

it echos for me

no fix it seems

[subtle cedar]

Yeah but I rarely do train models and do an AI Covers nowadays because many K-pop fans using AI Covers for shipping opposite gender Kpop artists (well I dislike hetero ships because most of people who ship opposite gender celebrities are toxic that don't respect anyone's actual partner) and I can't train models too often anymore because of my Dad for some reasons so it's okay if my future laptop is Chromebook or ChromeOS or don't use a local training or unlimited GPU

[knotty moth]

it's okay I can run borderlands 4 on a sub $300 phone that is even slower than Switch 1

[subtle cedar]

Yep but I can't afford to pay anything

[knotty moth]

not sure if you think my joke is really true

[subtle cedar]

Oh okay I don't know that was a joke haha sorry

[knotty moth]

why did you easily believe that without using common sense?

it's like believing that my chicken can fly because I have trained him so hard, till you realize I was joking on it

[hallow thistle]

You can try Weights.com as an alternative website to the super expensive Kits.ai. I don't know what to guarantee if you have paid for Kits.ai just for this.

[robust drum]

so if i were to download this juice preset thing can it work for pro tools?

[low shard]

please elaborate what ur talking about

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[low shard]

it seems like you replied to a convo that wasn’t related to you, please elaborate

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[robust drum]

where do i drag the file i downloaded? like the voice ai

do i drag it into models

in the RVC

[low shard]

it won’t train ever locally on a chromebook or chrome os pc, it’s too weak

ur only option is using a cloud (remote good pc) service with limited free gpu time

[low shard]

voice.ai is a different service that we don’t suggest

please elaborate as i asked

[low shard]

^

rvc doesn’t mean realtime voice changer, it measn Retrieval-based-Voice-Conversion

[robust drum]

No i mean the voice changers in the #1175430844685484042

do i drag the files into the RVC?

[low shard]

Well even tho that’s a joke, you can’t run borderlands 4 but you can run borderlands 3 https://www.reddit.com/r/OdinHandheld/comments/1m843on/borderlands_3_trough_winlator_on_odin_2_portal/

[low shard]

elaborate your pc gpu, operating system, the tutorial link you’re using, what program you’re using, be sure to not use youtube/video tutorials, there are different RVC versions

[low shard]

any issues?

[low shard]

we don’t suggest kits ai
What’s ur pc gpu and operating system?

[low shard]

pleaß elaborate

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[robust drum]

I downloaded the RVC off a yt video with a discord link to this discord am i cooked nick

[low shard]

yeah you are, all youtube tutorials are outdated

[knotty moth]

any handhelds are more capable than switch and average smartphones

[low shard]

^ please elaborate the rest of the things i asked

[robust drum]

by cooked did i just download a virus?

[low shard]

that’s an android handheld, it’s the same as buying a phone just with an integrated gamepad, it runs windows emulation

[low shard]

i’m not even sure what tutorial or download link you used so i’m not sure, but just know that theres no updated video tutorial

i can’t help you if you don’t elaborate

[robust drum]

Ok so then whats the right RVC?

[knotty moth]

not with snapdragon 6xx or weaker

[low shard]

there are many different versions, if you don’t elaborate I can’t know which one is right for you

open source ai isn’t easy to use

[hallow thistle]

Read help guidelines before start asking.

[robust drum]

Ok so i wanna use the an ai preset and I dont know how to start people are saying use a RVC? i dont even know what that is, is there something i gotta download

im finding my AI presets in #1175430844685484042

[low shard]

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

[hallow thistle]

Are you looking for W-Okada or Applio RVC? If this helps giving you some ideas.

[robust drum]

nah i just wanna know how to make my voice sound like a specific artist

[low shard]

I do understand that, but you need to elaborate:

• your pc gpu
• your pc operating system
• if you want to use models in realtime vc or pre-recorded audios

[hallow thistle]

That's not the answer.

[low shard]

there isn’t a right program and version for everyone, it depends on your hardware and needs

[robust drum]

pre recorded, GE FORCE RTX, Windows 11 64

[low shard]

Thanks, but which rtx gpu?

[hallow thistle]

RVC stands for Retrieval-based Voice Conversion, an AI program where you make AI cover and training voice models; it doesn't always mean realtime voice changer.

[quick pine]

is the voice changer supposed to be delayed

[hallow thistle]

What number is your NVIDIA GeForce RTX?

[quick pine]

cause its delayed by like 20 secs

[low shard]

Please elaborate, there are many voice changers programs

!howtoask

[patent trellis]

❓ How to Ask for Help

✅ Before You Ask!

1. Check Docs & Guides: Your answer may already be in the AI Hub Docs or the https://discord.com/channels/1159260121998827560/1159513888199540817 channel.
2. Search the https://discord.com/channels/1159260121998827560/1192011222023950368 : Look for existing posts that solve your issue. Do not invade someone else's post.

📝 How to ask?

Tell your:

• Full GPU Name: (e.g., NVIDIA RTX 3060)
• Operating System: (e.g., Windows 11)
• Detailed Description: What were you trying to do and what went wrong?
• Tutorial Used: Link to the guide you were following.
• Screenshot: A picture of the full error message is very helpful.

🚫 Prohibited Topics (We Will NOT Help With These)

To maintain a lega, safe & ethical community, we will NOT provide help for:

• (E girl, as an example) catfishing/trolling, scamming, impersonation.
• NSFW/Porn.
• Any illegal activities.
  Requests for these topics will be ignored and may result in moderation action.

<:matsuripray:1159685390156967936> Community Expectations

• Be Polite & Patient: Our helpers are volunteers. You may ping the Helpers role once.
• English Only: Please keep all conversations in English.

[low shard]

be sure to not use youtube or video tutorials for ai realtime voice changers for roleplay in calls/games

[hallow thistle]

To check your PC GPU name, open Task Manager, go to Performance tab, and spot GPU 0 or GPU 1 in the left panel.

[robust drum]

GPU 1

[hallow thistle]

Read help guidelines before start asking anything. By providing more information, I may be able to identify your issue.

[hallow thistle]

Click GPU 1 to reveal its GPU name on right panel.

[robust drum]

4050

RTX 4050

[low shard]

great, that’s enough for local rvc conversions

please read the applio guide https://docs.aihub.gg/rvc/local/applio/ :)

[robust drum]

so then i drag the file i want into this

[low shard]

you need to first download the precompiled windows nvidia version, then extract it, run the run-applio.bat, then it will open and you can upload your model

[hallow thistle]

Don't skip steps, we aren't on that part yet. You'll need to download the zip file of the program first.

[hallow thistle]

NVIDIA GeForce RTX 4050 is a mobile GPU for laptops. Applio will work with this GPU for inferencing (doing AI cover) but not very much for training.

[robust drum]

which one

[low shard]

the latest, 3.5.0

[hallow thistle]

ApplioV3.5.0.zip

[robust drum]

this aint gonna hack me right

[hallow thistle]

Sometimes, it's too much to ask every step, I can focus on known settings and how to use the program.

[robust drum]

sometimes?

[hallow thistle]

Don't worry, this program won't gonna hack your system. If one of your antivirus programs detected it, try allow it.

[robust drum]

ok its downloaded

now extract all?

[low shard]

nope, it’s open source, meaning the source code is visible to you and you can even edit the code, just to make you understand that it’s safe

[hallow thistle]

If I tell you on "how to extract the zip", would you still want me to do it?

[robust drum]

do i extract it or no

[hallow thistle]

This step is so simple. You open WinRAR or 7-Zip, you use the zip program to extract the internal files of "Applio" zip into somewhere like D:\Applio. No need to ask me for confirm like this.

[low shard]

yup

[hallow thistle]

Yeah, I know you want me to tell you every step, even on how to run the program, so do I.

[robust drum]

Ok im doing it

[robust drum]

do u want me to extract it uhhh 6 to 7 times ?

[hallow thistle]

After you extract the Applio zip, go inside Applio folder, and double click on run-applio.bat batch file.

[hallow thistle]

I was being sarcastic.

[robust drum]

is it normal for the file to tell me my full name?

[hallow thistle]

After you double click on the batch file, wait for it until the program opens your browser.

[robust drum]

Ok so im on this right now

[knotty moth]

isnt it PS2 emu?
btw my point was talking about the spec demand level of borderlands 4 or cyberpunk 2077

[robust drum]

its just stuck on that black screen

why is it stuck on this black screen

[hallow thistle]

You can send your screenshot to here huh.

[robust drum]

it wont let me lmao

its stuck on that

[hallow thistle]

This is my screenshot, not yours.

[robust drum]

Ik i was saying its stuck on that

[hallow thistle]

Launching a Python-related AI program can take some time to load up files and resources, so don't expect it to launch within seconds.

[robust drum]

Ok

[robust drum]

so its says press any ket to continue

i hit a key

and it does nothing

is that normal

[low shard]

Nope, ps vita, anyways ik i was just joking around about how well is android emulation

[hallow thistle]

This likely means that Applio failed to loading.

[robust drum]

so now what

[hallow thistle]

Run the batch file again.

[robust drum]

Ok

[hallow thistle]

This is what Applio might look like if launched successfully, but I'm using the older version.

[robust drum]

its saying file in one drive please move to a different file

[hallow thistle]

Don't extract anything on your desktop.

[robust drum]

Ok

you know @hallow thistle ur a legend did u know thatg

Ok got it working

now how do i add this vocal ai

???

dont die on me

[robust drum]

i move it to logs?

[hallow thistle]

This is how you upload voice model files into \Applio\logs\ folder, so these voice models will appear on Applio GUI.

[robust drum]

good?

i just moved the two bottom files

[hallow thistle]

I won't be repeating the same step for another time.

[robust drum]

i love u

thank u

u deserve a raise

ur a goat

i love u

[hallow thistle]

I explain things in the most simplest I can do, and you're here to ask me for little more when it's too much for me.

[subtle cedar]

Yeah haha

[hallow thistle]

On Applio, click "Refresh" to load models.

[robust drum]

YESSS

THANK YOU SO MUCH

HOLY

[hallow thistle]

Now you got those models.

[robust drum]

YESS

i love u

[hallow thistle]

This is how you upload an audio file into Applio.

[robust drum]

😆

😩

[hallow thistle]

If you act like this to some other helpers, you might not only be getting mocked, by the way.

[robust drum]

wdym

im just saying thank u

[elder coral]

how do i make my models sound clean not robotic or artifactual creature than your bald friend's hair

[fluid horizon]

My Applio stopped working today. Can anyone help? I'm using Pinokio as my "launcher"
It can't do its RVC anymore today, so I figured to update it to the latest (which is a simple button Pinokio has)
But its the same problem. So I can't even see the Web UI. Had to open the local url to my browser to see (it is indeed running)
Any clues why? Also before it had no problem reading .m4a files and converting it. What is going on?

[simple ore]

why do you need pinokio when you can just download the compiled verison, unzip and run?

[fluid horizon]

ah it was an older setup

like way back 2024

[simple ore]

the only things I've heard from pinokio is that it breaks shit for no reason

[fluid horizon]

that is true 🤣

i will migrate stuff out, i have my voice models in here

also im troubleshooting too! i think i found the new folder where the models go in

so yeah, pinokio broke it 🤣

may i have the link to that compiled version? 😄

i might just nuke pinokio entirely

[low shard]

you can find the guide in https://docs.aihub.gg/rvc/local/applio/

[fluid horizon]

thanks!

[low shard]

I was actually thinking about adding pinokio as an unofficial way to use applio in the ai hub docs, seems like its best to not do that for now lol

[simple ore]

i mean, even Dione had f'd up the install

and it is from the same guys lol

[fluid horizon]

i did initially run applio in a google colab hahaha back when my rig was shittier

then i learned about pinokio lol they were gassing it up on youtube (last year)

right now, i cant even see it in their Web UI view. i think it started when i saw the themes ahha i wanted to try it

so yeah, im nuking this applio copy lol i just need to move out my trained models

like wow, pinokio must have broken it soooo much it cant even infer

thanks again for the guidance!

[fluid horizon]

thanks again! i just got the compiled version working. amazing!

[stiff idol]

Hi. I'm creating a dataset from videos on bilibili and other stuff. I've got a Chinese singer and wanted to ask someone with good hearing if 135Hz high-pass is fine or not. I think it's ok, 150Hz was too much, but there might be people with better hearing.

1. HPF sample: high pass filter -> truncate silence -> loudness normalization
2. no HPF: truncate silence -> loudness normalization

I'd like to ask if you hear any difference and whether it's good or bad, thank you.

[stiff idol]

yup, 135Hz was a bit too much, I needed to listen several times to notice when her voice goes lower, this one should be an okayish starting point

[low shard]

You're welcome :)

[dusty delta]

nick

i have vonox

but my file explorer wont open

[knotty moth]

explorer.exe has stopped working

[thorn plinth]

RuntimeError: The expanded size of the tensor (3850) must match the existing size (8192) at non-singleton dimension 0. Target sizes: [3850]. Tensor sizes: [8192]
//how to changer the size of the tensor

[pure obsidian]

guys

question

how to hear myself

from the vonovox

[low shard]

https://docs.aihub.gg/realtime-voice-changer/local/vonovox/#how-can-i-hear-myself

[pure obsidian]

bro my college wifi is embarassing as fck

its not opening

okay worked

[oak edge]

yooooo

what's the new spin-v2 @low shard

[low shard]

why are u using college wifi 😭

are u deadass running ai on college's pcs?

[pure obsidian]

no no i live in a dorm lol

[low shard]

idk how it works there but they might like track what u do

oh lol

[pure obsidian]

using my own laptop

[pure obsidian]

btw bro i did the steps but theres no "listen" inside the properties of line 1

if u want i can send photos

of what i see

[low shard]

Spin v2 is a version of Spin made by @shy spruce which should improve it for RVC, he can deffo give you more technical details,
The pretrains for it were made by @simple ore

[low shard]

are you sure you're checking in the recording tab?

[pure obsidian]

yes

wait

can i open the forum

and show u

cuz i dont have perm photos to send here

[analog obsidian]

what lmao spin v2 was made by dr87

is not a pitch extraction either

its an embedder

[knotty moth]

bet if it's not as bad as public mall wifi (unless you're talking about the security paranoia)

[analog obsidian]

[pure obsidian]

my uni security is so good bruh

like i need to use vpn to use steam and epic games and discord and so on

idk how to by pass it

so i use a vpn but they make wifi slower by 10 times

[low shard]

I confused it up with it's pretrains lol https://huggingface.co/Aznamir/spin/tree/main/spin-v2, thanks for correcting it up, the message has been edited pinging both noobies and dr87

[pure obsidian]

nick bro see the forum i sent u the photos

[low shard]

that can vary, but other than that, they can check all you do and have blocks

[pure obsidian]

yeah they have so many blocks idk how to bypass

[knotty moth]

just for browsing something, I'd rather use custom dns (cloudflare, google. etc) if it works

[pure obsidian]

i just need to use for like steam and discord

[knotty moth]

in case of region block, I'd agree vpn would work

[pure obsidian]

yeah but like my uni wifi does like 350mbs

but when i use vpn

that shi drops to 20-30

[low shard]

what vpn do u use btw?

[pure obsidian]

tunnelbear but u get like 2gb a month / hotspot shield good one but free one takes u to america and fucks ur wifi / mullvad vpn its paid but a friend gifted me a sub for a month

[low shard]

ProtonVPN is also nice

Mullvad and ProtonVPN are prob the best privacy wise iirc

[pure obsidian]

idk i saw on their website theres no free theres only 3 plans

and all are paid

[low shard]

it's unlimited, you just can't choose the region

[pure obsidian]

imma give it a try

thanks man

[meager lichen]

my voice changer was working well for abou 3 days now it stopped producing any output, any help?

[pure obsidian]

do like a forum

post

[meager lichen]

i didd

https://discord.com/channels/1159260121998827560/1420761216770637895

[pure obsidian]

oh alrr wait for some help mods

[pure obsidian]

go private msg

[meager lichen]

ok

[pure obsidian]

bro lowkey nick is goated

promote that guy asap

[viral mason]

He is so cool

[whole sky]

My Play AI Hub account still shows the “Link Discord” button even though my Discord is already connected. When I press it, I get the error “Discord already linked” and the console shows multiple 401/500 errors from the /auth/discord and /user/discord/roles endpoints.

I already tried:

Removing Play AI Hub from Discord → Authorized Apps

Clearing cache and cookies

But the issue remains. It looks like my Discord is linked on the backend, but the UI still thinks it isn’t. Could someone manually reset or unlink my Discord connection from oficail site?

My Play AI Hub account Crypto Phan

[pure obsidian]

ong

[viral mason]

Maybe it's because crypto is stinky

[low shard]

We aren't affiliated with any discord bot called "Play AI Hub" nor crypto

[hallow thistle]

This server here is not your "My Play AI Hub" nor any crypto-type server. This server is mostly about AIs and codings.

[oak edge]

@shy spruce @simple ore can you guys elaborate on the languages used for spinv2 (i'm thinking about using it to train tamil voice of mine)

[meager lichen]

sm1 plz help if they can in https://discord.com/channels/1159260121998827560/1420761216770637895

[viral mason]

There's no spinv2 pretrains that I know of so I don't think u can even use it yet

[oak edge]

oopsies

@viral mason alsooo if i were to build a voice model from ground up without pretrains how long would it take

[latent yacht]

can i send an image here?

[oak edge]

and how the quality would be like? (because I'm training tamil voice-for which I don't see a pretrain anywhere)

[simple ore]

there's my retune

[simple ore]

english

[oak edge]

oops then not gonna work I think

[oak edge]

can you guide me on how to build a model from ground up without pretrain?

[viral mason]

I've got no experience with that so I wouldn't be sure, I always use pretrains on my models

[oak edge]

ahhh alrighty

[viral mason]

You'd have to have a lotttttttttt of data tho

[analog obsidian]

24 hours minimum

[oak edge]

how many hours are we talking about, voice is mine so I can record 1 hour per day everyday

[oak edge]

i've already got around 2 hours, so I think I need 22 more hours

[simple ore]

you can't

[oak edge]

[latent yacht]

can one guys tell me if i should download these vcclients or just use the 1.5.3? like whats the difference are they better, new or what

[simple ore]

it should work with another language

[viral mason]

What is this...

[knotty moth]

neither of them

[viral mason]

Outdated technology

[oak edge]

the pronounciation is way way awful than contentvec (at least that was how it was for spin - v1)

[analog obsidian]

if you train from scratch then you gotta find a way to fight rvc degradation

[oak edge]

what's that

[analog obsidian]

coz rvc stock just blows up

[latent yacht]

oh damn, so what should i do?

[knotty moth]

2.2.2 beta only if you want to use beatrice model

[oak edge]

it's goin into uncharted territory for me but I'm willing to learn more

[latent yacht]

im gonna be using custom models...soooooo?

[analog obsidian]

after training for maybe 50e, the model will start to degrade

[viral mason]

Download one of the three that are actually good

-rt

[patent trellis]

💻 Local Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[analog obsidian]

forgetting stuff it learned prior and overfitting on low frequencies

[simple ore]

use spin-v2

[analog obsidian]

so it will never clean the high ends

[oak edge]

as in the epoch's following that will get worser and worser in terms of quality?

[viral mason]

What gpu do u have, that is the important part

[analog obsidian]

everything will get worse

[oak edge]

holy

[oak edge]

[latent yacht]

3060Ti 8GB

[analog obsidian]

50e is not enough for a pretrain either

im saying that it will improve up to that point then start to degrade

[oak edge]

alright I'll try training with v2 for 10 epochs and tell you the results

[oak edge]

then how did these people manage to create pretrains?

like how are these spinv2 contentvec models were made?

[viral mason]

Is that Nvidia, AMD, or intel

I'm kinda slow

[latent yacht]

Nvidia brother

[analog obsidian]

retunes

they finetune og pretrain

[oak edge]

ohhhh

then how were the og pretrains were made?

[analog obsidian]

retunes also degrade anyway

i noticed degradation in noobies model

[latent yacht]

which one of these would you prefer personally?

[oak edge]

ohhh

[analog obsidian]

og was trained from scratch, rvc-boss most likely used layer freezing in order to train og

[oak edge]

@analog obsidian can you tell me what degree are you doing/did?

[analog obsidian]

i trained almost 1m steps of a failed pretrain

then did tons of tests

so i kinda know stock rvc cant train a pretrain

[latent yacht]

i checked vonovox guide , and it says that its features are paid

[oak edge]

damnnnn

you really really are a pro in this @analog obsidian

[viral mason]

Ty, u can download Vonovox as it's the best of the three just scroll up and download it from the guide but I personally use W-Okada Deiteris as it has over 100 slots and I use okada in vrchat

[oak edge]

i wonder how many years of studies it's gonna take me to learn properly U_U

[analog obsidian]

why so interested in a new pretrain? the og one is great

[oak edge]

bcz all are made from languages that's not in tamil

[latent yacht]

does vonovox have slot limits?

[oak edge]

which results as some specific kind of pronounciations never getting right

[analog obsidian]

spin v2 is also a english embedder

[oak edge]

[analog obsidian]

im sure spin and cvec were trained using the same dataset

[oak edge]

i'm just tryin out for @simple ore

[viral mason]

It has only 8 slots currently but eventually is getting an update to have them build up over time as you add models in future updates

[analog obsidian]

you'll need to train a spin embedder in tamil then train a pretrain with it

[oak edge]

but i tested for same data set for 50 epochs, the content vec had less electronic noise and better pronounciations opposed to spin v1

[viral mason]

Besides that small flaw of being so limited model slot wise it beats all other okada forks in every way

[analog obsidian]

yes spin v1 had some problems that got fixed with v2

[oak edge]

how do I do that?

[oak edge]

lemme try rn

[analog obsidian]

idk

[oak edge]

[analog obsidian]

you need a spin v2 pretrain

[oak edge]

i think i can download custom pretrain in applio download menu?

i think it will be available in it?

[analog obsidian]

these are so outdated, i told noobies to remove them

[oak edge]

bruh

[latent yacht]

I'll go with w-okada

[oak edge]

then how can I train with spin v2 Y_Y

[analog obsidian]

download noobies retune i guess

[oak edge]

and upload it manually to my colab?

also I train in google colab for free btw Y_Y

kaggle was uhh... tough Y_Y

[analog obsidian]

https://huggingface.co/Aznamir/spin/tree/main/spin-v2

[oak edge]

ohh the sizes are relative low

i'd go with 40Khz i assume

[hallow thistle]

[oak edge]

but can you confirm where to put this in colab ? @analog obsidian

[latent yacht]

@viral mason which one of these

[viral mason]

Alr download w-Okada Deiteris then, use the guide and download vac lite and the voice changer itself in the guide

-rt

[patent trellis]

💻 Local Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Deiteris Fork, with extra features, but supported only for Nvidia GPUs on Windows. and without cloud options GUIDE

• Wokada Tg-Develop Fork

A personal fork (modified version) of Wokada Deiteris Fork, it just adds some Quality of Life improvements to it like supporting Spin Embedder and Audio Effects. Don't expect too much about it since the creator made it originally as a personal project. GUIDE

• Wokada Deiteris Fork

Most suggested WebUI with the best general support for many platforms. GUIDE

⚔️ Wokada Deiteris Fork vs Vonovox

For Windows Nvidia, Both Wokada Deiteris fork and Vonovox have similar performance & quality. Users should read the pros and cons for both and choose based on their differences, such as UI and Vonovox's paid effects.
Read Wokada Deiteris Fork Pros&Cons & Vonovox Pros&Cons

⛔ Outdated/Discouraged

These options are not recommended for use.

• Original Wokada

Not suggested, older versions in youtube tuts are even way worse. GUIDE

• RVC GUI Mainline Realtime

The program is worse compared to the ones above, and much less updated. GUIDE

[viral mason]

No

[analog obsidian]

rvc > models > pretraineds > custom

[latent yacht]

can you vc?

[viral mason]

@latent yacht download from the correct guide

[oak edge]

okie lemme download and use em

[hallow thistle]

This may be your answer. https://huggingface.co/Shadicti/deiteris-Fork/blob/main/voice-changer-windows-nvidia-b2332.zip

[oak edge]

wait

[viral mason]

Nope I'm in the car with my grandfather rn so I can't, and I've got work soon in a few hours

[oak edge]

is there any ways to directly import them to colab rather than downloading and re uploading? as in uploading from a link ? @analog obsidian

[analog obsidian]

create a colab cell that downloads the pretrains and place them in the correct place

[oak edge]

bruh I'm on free version how do I create a one

wait can chat gpt 5 work in writing this code?

[analog obsidian]

im sure a llm can help you with this

lol

[oak edge]

i can create a cell with chat gpt

[viral mason]

Ew collab

[oak edge]

yup lemme go to my good ol chat gpt

[viral mason]

Worst option for training

[oak edge]

vrother colab Y_Y

[oak edge]

i can't risk my 4060

[viral mason]

I forget it has 1 l

[latent yacht]

I got the link from a another kind fellow, thank you for the help so far. I genuinely appreciate it

[analog obsidian]

[latent yacht]

it was right there i just needed to scroll a lil

lmao

[oak edge]

wym i like colab gpu better than 4060

[viral mason]

Np man! I would have copied the link from the guide if I had my pc on hand but it's fucked rn

[oak edge]

[analog obsidian]

the t4 is 20 times slower

[oak edge]

uhm idk how to use my 4060 Y_Y tbh i'm so used to training in colab

now idk how to do it locally

[analog obsidian]

[viral mason]

Kaggle and lightning ai better, imagine having more than a random number of hours between 1 and 4

[oak edge]

i opened kaggle yeah but in kaggle i had this weird freakin issue where it deletes my backup

and no way to save the epochs to drive

or any safe splace

[viral mason]

Wdym?

Just turn on the save options variables and files

[hallow thistle]

The download links from GitHub can be confusing, especially those zips that are named identically, looking duplicated but are actually splitted.

[oak edge]

like it literally deleted all my training data after training for 15 hours

[oak edge]

THAT WAS ONN

that was on

[viral mason]

Uhmmm you got unlucky

[oak edge]

it just disappeared

[oak edge]

that's why i'd rather go colab

it saves my data properly in my drive

[viral mason]

Sometimes u just get unlucky but it's rare

[oak edge]

and i share my drive folder between 3 accounts

so 12 hours training

[oak edge]

i guess my bad day i lost 15 hours of data

[viral mason]

I will never ever ever use Google colab again for training because I don't wanna go back to switching accounts transferring, I'd rather have errors that make it impossible to train a model over the torture of 2 hour training sessions

It sucks ass

[hallow thistle]

You can try train a voice model on NVIDIA GeForce RTX 4060, but the VRAM won't be that much of it, so the best bet is to go for online services instead for much faster speed.

[viral mason]

Joker when Batman or something idk

[oak edge]

💀 gang I've been training for 15 days with colab

talk about patience

[analog obsidian]

eh it's enough for batch size 8 and fp16

[viral mason]

I'd rather kill myself

[analog obsidian]

i can do batch 8 and fp32 too

[oak edge]

i mean i got lot of patience though

[oak edge]

exactly my training data is 2 hours Y_Y

[viral mason]

One clean 15 hour training session over literally putting yourself through horrors beyond the mind is better for me

[oak edge]

what's the dataset's time?

[oak edge]

mind it I trained for 3 days with kaggle

[analog obsidian]

why train for 15 hours lol a model can take 2 hours or 3

[oak edge]

i mean i love slow progression though

[oak edge]

uhm i forgot to mention i'm currently at 370 epochs with a 2hour and 30 minutes something data set and i think it even exploded at some point

[analog obsidian]

my safest recommendation will be always 40-100 epochs lols

[viral mason]

The most I've had to take is like a day if I train a model later in the day and run out of time to be allowed to be awake (I'm an adult please help me)

[oak edge]

40-100? but it still had some electronic noise between that range

i think the electronic noises slowly subsided after 200

[oak edge]

[hallow thistle]

"100 compute units" or Colab Pro monthly option on Google Colab both listed for $9.99, kind of exact to the price of Weights basic plan which is also $9.99.

[analog obsidian]

v2 is naturally more metallic, i asked to bring back v1 but as always i was not taken seriously lol

[oak edge]

YOOO IT WORKED

i mean chat gpt code worked to download stuff

URLS = [
    "https://huggingface.co/Aznamir/spin/resolve/main/spin-v2/f0G40k_spin-v2.pth",
    "https://huggingface.co/Aznamir/spin/resolve/main/spin-v2/f0D40k_spin-v2.pth",
]

DEST_DIR = "/content/Applio/rvc/models/pretraineds/custom"

import os, pathlib, requests

# Ensure destination exists
pathlib.Path(DEST_DIR).mkdir(parents=True, exist_ok=True)

def download_to(url: str, dest_path: str, chunk=1024 * 1024):
    with requests.get(url, stream=True, allow_redirects=True) as r:
        r.raise_for_status()
        with open(dest_path, "wb") as f:
            for part in r.iter_content(chunk_size=chunk):
                if part:
                    f.write(part)

saved = []
for url in URLS:
    filename = url.rsplit("/", 1)[-1]
    out_path = os.path.join(DEST_DIR, filename)
    print(f"→ Downloading {filename} ...")
    download_to(url, out_path)
    size = os.path.getsize(out_path)
    print(f"   Saved to: {out_path} ({size:,} bytes)")
    saved.append(out_path)

print("\n✅ Done. Files now in:")
for p in saved:
    print(" -", p)

W gpt Y_Y

[analog obsidian]

[oak edge]

im kinda new to this

[hallow thistle]

Colab Pro+ and "500 compute units", while both more expensive than Colab Pro, I think they typically not needed for most people unless someone needs to train something big all day and night or just wanna use A100 for something longer.

[oak edge]

can't spend a penny fr Y_Y

@analog obsidian are there any other custom pretrains like the one you've mentioned?

holy hell nvm

[analog obsidian]

for spinv2? nope

[oak edge]

colab noped out at preprocessing 2.4gb

imma go to good old kaggle Y_Y

HOLY SHIT

mine never got deleted in kaggle

i just thought mine got deleted but it never was @viral mason u

VROOOO @analog obsidian i need help

[viral mason]

Told uuuuuu

[oak edge]

gang help me with kaggle

i don't see the spin v2

[viral mason]

I don't believe it's added yet

[oak edge]

also this is the current environment btw

[oak edge]

BUT it shown in colab

!!!

like spin v2 was there in colab applio

[viral mason]

Keep it to the original environment

[oak edge]

okie how do I go custom

how do I add custom embedder for spin v2 @viral mason

[viral mason]

@analog obsidian uhhh ^

Help

[oak edge]

i mean there must be a way

[analog obsidian]

i did not made the notebook lol

most likely is downloading a stable version and not the main branch

[viral mason]

Who did make it?

[analog obsidian]

vidal

[viral mason]

Ah

[oak edge]

how do i import spin v2 here as custom embedder

[analog obsidian]

i dont even like applio

[oak edge]

or should I click always use the latest environment?

[oak edge]

😭 gangggggggggg

what if I press the use latest environment? bcz im damn sure the applio colab had spin v2

that's why i even asked in the first place

[analog obsidian]

just remove the line that downloads a specific branch

[oak edge]

what you mean specific branch?

[analog obsidian]

can you please at least send me the link of the thing youre using

[oak edge]

gosh yea sorry let me

https://www.kaggle.com/code/arkeshan/notebooka66a3f0eeb/edit/run/258706189

[analog obsidian]

you're downloading an old version of applio

just git clone the main branch

[oak edge]

wait tell me your kaggle name

[analog obsidian]

but you have to encrypt everything because kaggle bans rvc usage

[oak edge]

oops wait let me share you properly

[analog obsidian]

show me a screenshot of your first cell

most likely says --branch 3.6.0 or whatever

[oak edge]

HOLY SHIT

IT'S 3.3.0

[analog obsidian]

change it to 3.5.0

[oak edge]

@analog obsidian is that the latest?

[analog obsidian]

yes

[oak edge]

okiee

wait what will happen to my training data

it didn't change anything ? @analog obsidian

[analog obsidian]

bro just reinstall everything im not the tech support of this server ask gpt idk 😭

[oak edge]

i'm sorry vroskieee

imma try somethin out

[shell pewter]

anyone got fixes for electrical background noises when speaking with a model? im using W-Okada

[pure obsidian]

put the echo or go task manager and put the voice changer priority to high

i think thatll work