✨│ai-help

[rapid echo]

yeh happens whenever i run the bat

[viral mason]

oh I dunno how to use it locally, NeuralNoob may know more

[rapid echo]

i followed the website for the AMD section perfectly but whenever i launch run-applio-amd.bat i just get

`HIP Library Path: C:\WINDOWS\SYSTEM32\amdhip64_7.dll
Failed to launch on port 6969, trying again on port 6968...
Failed to launch on port 6968, trying again on port 6967...
Failed to launch on port 6967, trying again on port 6966...
Failed to launch on port 6966, trying again on port 6965...
Failed to launch on port 6965, trying again on port 6964...
Failed to launch on port 6964, trying again on port 6963...
Failed to launch on port 6963, trying again on port 6962...
Failed to launch on port 6962, trying again on port 6961...
Failed to launch on port 6961, trying again on port 6960...
Failed to launch on port 6960, trying again on port 6959...

Press any key to continue . . `

nvm fixed it changed it to the same port Wokada uses in app.py

[viral mason]

applio and wokada are very different programs tho

how did that fix it?

[rapid echo]

DEFAULT_SERVER_NAME = "127.0.0.1" DEFAULT_PORT = 18888 MAX_PORT_ATTEMPTS = 10

in app.py

all i did was change the port

from 6969

[viral mason]

o

good job wizard

I don't understand code

[silent grove]

Hey I uh

[viral mason]

Hi Kinger

[silent grove]

downloaded Replay

and well

frankly I'm pretty excited

got the voices downloaded and all

now I just gotta wait for Replay to download these files it needs

It's been on this D-f048k-TITAN-Medium.pth thing for a while

[viral mason]

do u want my Digital circus models that I don't have public in this server? they're not the best but they're all decent

[silent grove]

that's normal right?

[silent grove]

Agh erm... nah... I hate TADC

sorry

[viral mason]

probably, it may be a large file

[silent grove]

ah oki!

[viral mason]

https://tenor.com/view/pomni-the-amazing-digital-circus-tadc-pomni-smirk-smirk-gif-16561828646995621947

[silent grove]

Yep we're good it's almost done

[silent grove]

I said it

and I'LL FUCKING SAY IT AGAIN!

I-

awwwww :((

that would've been a good one

[viral mason]

just gotta talk more to get image perms

I saw the gif tho it was very silly

[silent grove]

yeah ikkk

wanna be friends?

[viral mason]

yea!

could you elaborate what program you're having issues with

what gpu do u have(Nvidia or AMD) and what are you using it for?

like playing as Darth Vader or Goku, something cool right?

no, but I can give u a better thing

vonovox

does the same thing but it's the current best for voice chaning realtime rvc stuff

way better quality

ect

wdym? 👀

[devout dome]

lol

[viral mason]

have you read the rules?

anything at all related to catfishing isn't allowed

mmmmh

<@&1159293140440723499>

[viral mason]

told ya, weirdos all the time come here

https://media.discordapp.net/attachments/869395754052124723/997702870835138700/523BE0F5-45E8-4EF5-9138-7827EBAE09F5.gif

[proven hill]

catfish

sadly it isnt still allowed :(

[viral mason]

ok?

[proven hill]

yea thats catfishing and not actually solving the problem

pedos as you stated

[viral mason]

just get rid of preds the normal way, find them and eliminate them, remove their breathing tube

preds are not a stable source of income

I don't understand it

that's

mhm

[proven hill]

you should report to the police not taking advance

[viral mason]

just don't bring it here annnnd use literally anything but egirl models

who

who is jidion

never heard of him

[silent grove]

yikes we got a problem

Error creating song: CUDA error: no kernel image is available for execution on the device Search for cudaErrorNoKernelImageForDevice' in https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__TYPES.html for more information. Compile with TORCH_USE_CUDA_DSA` to enable device-side assertions.

says this when I try to start the conversion

[proven hill]

what are you trying to use

[silent grove]

Replay

[viral mason]

this is my setup

[silent grove]

Hey uh

where did you go?

I'm still waiting man

[proven hill]

use applio

[silent grove]

wha-

[devout dome]

ohhh i see damn thats a lot of stuff

[silent grove]

I JUST SPENT ALL THIS TIME DOWNLOADING REPLAY ARE YOU SERIOUS???

ughh

fine

[proven hill]

replay is no longer supported

[viral mason]

it works fine for me tho

👀

[devout dome]

which are the parts that cost money

[silent grove]

was literally about to ping you

[devout dome]

and could u send a link for where to download

[viral mason]

nothing, I pirated bought fl studio, everything else is free

[devout dome]

i see i see i might know a guy

[silent grove]

Great

so now I gotta uninstall Replay

AND

download a new thing

[proven hill]

what gpu do you have

[silent grove]

lemme look...

NVIDIA GeForce GTX 1050

Yea I have a gaming laptop

biggest regret ever

[proven hill]

my fourth pc have that it will be sloooooow

[silent grove]

Ight well fuck this ai bullshit

was nice being in the server

[proven hill]

😭

[silent grove]

happy holidays and good fucking night

[viral mason]

oh

[proven hill]

[bronze pier]

I finished training in Applio and i got the index file along with D_333333.pth and G_333333.pth what do i do with these?

[viral mason]

you need the pth file of the model not the d and g

index and pth

[bronze pier]

it didnt generate any pth file with the same name as the index file, in the output it had the index file along with G_333333 and D_333333

[viral mason]

that's not good

[bronze pier]

what folder should I find it in once it trains? is it ApplioExported?

[viral mason]

I'm unsure, I don't use it locally

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[viral mason]

check the applio docs for the best support for it

or ping a helper

[bronze pier]

<@&1159293204038955078>
I'm using Applio interface with public URL https://colab.research.google.com/github/iahispano/applio/blob/main/assets/Applio.ipynb and once i train the model, i dont see a .pth file with the same name as my index. The index is there, but its there with D_3333333.pth and G_3333333.pth in the logs folder. am i looking in the right place?

[devout dome]

im learning fl studio now XD

[devout dome]

@viral mason what is line 1? i only have the VB-Cable input and output

[viral mason]

line 1 comes from vac lite

[devout dome]

https://vac.muzychenko.net/en/download.htm here or another place?

[gusty socket]

Gang, is anyone running okada on AMD gpus? I'm trying everything and it's not working

[viral mason]

have you tried wokada tg fork?

[gusty socket]

I even tried a fork of tg fork of okada fork

forkception

[viral mason]

what?

that's not a thing

[gusty socket]

https://github.com/gajae1/voice-changer#prerequisites

[viral mason]

that link isn't familiar at all

[devout dome]

mine or his?

[proven hill]

this is a scam and malware and its uses by scammers.
i suggest you NOT using that

[gusty socket]

WHAT

[devout dome]

jeepers

[proven hill]

im joking but that fork is useless prob

[gusty socket]

I'm trying tg now

[viral mason]

try this one, and not whatever that is https://github.com/tg-develop/voice-changer/releases/download/b2397/voice-changer-windows-amd64-dml.zip

this is wokada tg fork for AMD

not that thing u just had

[gusty socket]

I will!

[viral mason]

:3

[gusty socket]

is dat windows version?

[viral mason]

mhm

[gusty socket]

am on linus

[viral mason]

oh

idk how to help anymore

<@&1159293204038955078> give this kind soul the wokada tg fork download for linux amd

i dunno what one it is

[graceful scaffold]

guys what are gpt sovits in #1175430844685484042

[viral mason]

tts models

[stable basalt]

How do i make it sound less roboty

[graceful scaffold]

im not sure, maybe try with more epochs? :)

[gusty socket]

beep boop boop beep

[stable basalt]

Whars epochs

[viral mason]

they're talking about in the voice changer I assume, epochs cannot be changed

[graceful scaffold]

how many "cycles" the model is trained on to learn the voice better, less = worse quality, higher = better quality

[graceful scaffold]

ohh ok

[stable basalt]

How do i train it then

[viral mason]

u cannot train an existing model

you can only create new ones with audio you collect or public datasets

[gusty socket]

model, sit!

model, paw!

[viral mason]

so true

[gusty socket]

model, start running on linux!

[viral mason]

if only I knew anything about it

[stable basalt]

https://voice-models.com/model/1m8tmDQ5aEe

this good?

[viral mason]

I dunno

test the model

the best models can be found here https://discord.com/channels/1159260121998827560/1175430844685484042

just look around

[graceful scaffold]

model, start production of quantum computers that cost 40 cents and can run deepseek 4 pro at over 300 tokens per second

[gusty socket]

also make me a hot choco

ts is stuck on downloading gang

[proven hill]

linux?

[gusty socket]

It's downloaded, just doesn't run anyway

[viral mason]

Hatsune miku? :D

[gusty socket]

yes

Getting this on tg's fork on RDNA2 6800XT, can't copy text somehow

[proven hill]

<@&1159293204038955078>

[gusty socket]

I was able to fix it with

sudo pacman -S patchelf binutils

target="./_internal/onnxruntime/capi/onnxruntime_pybind11_state.so"

cp -a "$target" "$target.bak"

readelf -lW "$target" | grep GNU_STACK

patchelf --clear-execstack "$target"

readelf -lW "$target" | grep GNU_STACK```

but it runs only on CPU and not GPU ![pepocoffee](https://cdn.discordapp.com/emojis/1164623357384327188.webp?size=128 "pepocoffee")

[viral mason]

that's no good

it should run on gpu

[proven hill]

sorry i dont use linux and gemini wont help me :(

[gusty socket]

Gemini is the dumbest AI I've ever used

I remember asking it some technical question about spotify and it was like uhh cant help you with that

[proven hill]

it helped me tons with my discord bot

[gusty socket]

I was like WDYM YOU CANT HELP ME and then it proceeded to answering (got the asnwer wrong)

[low shard]

fuck all those, the e girl tag has been deleted a long time ago (why have it? no reason compared to like the technical tags we have), the commissions in fact caused a lot of issues especially with scams, female models themselves aren't a problem, the issue is people misusing

all of those tutorials aren't affiliated to us and the youtubers are banned

[low shard]

do you have a slow internet connection usually?

check https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/#failed-to-download-or-verify

[gusty socket]

no, I got a gigabit channel

[low shard]

that's really weird, could you try downloading the fix in that link?

[gusty socket]

it's not downloading issue, it's not running

[low shard]

on the first run, the program downloads internal pretrained models, it might be that they didn't successfully download, it happened just earlier to another user where the program runned but the program didn't download correctly so it didn't work, but they were on windows

btw are you sure your amd drivers and os are up to date?

[gusty socket]

yeah, am on a rolling release, fresh as a bunny

[gusty socket]

oh my god, fixed this with docker

[viral mason]

what is easy effects sink

literally never seen that before

[gusty socket]

dont mind it

[viral mason]

should be using line 1 there and is the hyperx bla bla is your normal mic then that's fine

[gusty socket]

I havent setup output thingy, was just trying to make this whole thing work on linux first

This is how I managed to make tg fork of okada work on Linux with AMD GPU wtih Docker

OS: CachyOS
GPU: AMD Radeon RX 6800 XT / gfx1030
Repo: tg-develop/voice-changer

1. Problem summary:
The prebuilt MMVCServerSIO bundle may say:
Voice changer version: AMD-ROCm
but still only show CPU conversion.

In our case, the bundled ONNX Runtime was ROCm-capable:
ROCMExecutionProvider MIGraphXExecutionProvider CPUExecutionProvider

but bundled PyTorch was CUDA-only:
libtorch_cpu.so libtorch_cuda.so

and not ROCm/HIP:
missing libtorch_hip.so missing libc10_hip.so

So PyTorch reported:
torch hip: None cuda available: False device count: 0

That means the UI only exposes CPU.
The fix is to run the source version of tg-develop/voice-changer with proper ROCm PyTorch.

2. Host requirements
Install Docker, your user should be in video, render, docker groups (sudo usermod -aG video,render,docker "$USER" if not)

sudo systemctl enable --now docker
sudo usermod -aG docker "$USER"
docker run --rm hello-world```
**3. Pull ROCm ONNX Runtime base image**
We use AMD’s ROCm 6.4 + ONNX Runtime 1.21 image because tg-develop/voice-changer ROCm requirements are built around onnxruntime-rocm 1.21.0. ONNX Runtime documents ROCm and MIGraphX execution providers for AMD GPU acceleration.
```docker pull rocm/onnxruntime:rocm6.4.1_ub24.04_ort1.21_torch2.6.0```
Test that the container sees the GPU:
```docker run --rm -it \
  --device=/dev/kfd \
  --device=/dev/dri \
  --group-add video \
  --group-add render \
  --ipc=host \
  --security-opt seccomp=unconfined \
  -e HSA_OVERRIDE_GFX_VERSION=10.3.0 \
  rocm/onnxruntime:rocm6.4.1_ub24.04_ort1.21_torch2.6.0 \
  bash -lc 'rocminfo | grep -E "Name:|gfx" | head -30'```

[viral mason]

I have been intimidated

[gusty socket]

4. Clone tg-develop/voice-changer
git clone https://github.com/tg-develop/voice-changer.git

5. Create Dockerfile
   Inside of your voice-changer folder
   nano Dockerfile.rocm64
   and paste this (2 separate posts because discord wont let me post in one, you need to copy from both)

FROM rocm/onnxruntime:rocm6.4.1_ub24.04_ort1.21_torch2.6.0

ENV DEBIAN_FRONTEND=noninteractive
ENV ROCM_PATH=/opt/rocm
ENV HIP_PATH=/opt/rocm
ENV HSA_OVERRIDE_GFX_VERSION=10.3.0
ENV HIP_VISIBLE_DEVICES=0
ENV ROCR_VISIBLE_DEVICES=0

RUN apt update && apt install -y \
    git \
    wget \
    curl \
    gcc \
    g++ \
    make \
    unzip \
    ffmpeg \
    libportaudio2 \
    portaudio19-dev \
    libasound2t64 \
    libasound2-dev \
    libsndfile1 \
    alsa-utils \
    python3-pip \
    && rm -rf /var/lib/apt/lists/*

WORKDIR /tmp/build

COPY server/requirements-common.txt /tmp/requirements-common.txt
COPY server/requirements-rocm.txt /tmp/requirements-rocm.txt

RUN python3 -m pip install --break-system-packages --upgrade pip wheel setuptools

# Common voice-changer deps.
RUN python3 -m pip install --break-system-packages -r /tmp/requirements-common.txt

# Install ROCm requirements, but skip torch/torchaudio/onnxruntime lines.
# We install ROCm torch manually later from the PyTorch ROCm index.
RUN awk '\
    /^[[:space:]]*($|#)/ { next } \
    /--index-url|--extra-index-url|--find-links/ { next } \
    /onnxruntime/ { next } \
    /^torch==/ { next } \
    /^torchaudio==/ { next } \
    /^torchvision==/ { next } \
    { print }' /tmp/requirements-rocm.txt > /tmp/requirements-rocm-filtered.txt && \
    cat /tmp/requirements-rocm-filtered.txt && \
    python3 -m pip install --break-system-packages -r /tmp/requirements-rocm-filtered.txt

# Some deps may accidentally pull CUDA torch. Remove it.
RUN python3 -m pip uninstall -y torch torchaudio torchvision triton pytorch-triton pytorch-triton-rocm || true

# Install ROCm PyTorch.
# If 2.9.1+rocm6.4 is unavailable later, use the newest version pip says is available.
RUN python3 -m pip install --break-system-packages \
    --index-url https://download.pytorch.org/whl/rocm6.4 \
    'torch==2.9.1+rocm6.4' \
    'torchaudio==2.9.1+rocm6.4'

# Sanity check.
RUN python3 - <<'PY'
import torch
import onnxruntime as ort

print("torch:", torch.__version__)
print("torch hip:", torch.version.hip)
print("torch cuda available:", torch.cuda.is_available())
print("torch device count:", torch.cuda.device_count())
print("ort:", ort.__version__)
print("ort providers:", ort.get_available_providers())
PY

CMD ["bash"]```

6. Build image
from your voice-changer folder:
docker build -f Dockerfile.rocm64 -t tg-vc-rocm64 .
7. Run container
Create persistent data folders, for me it was

mkdir -p /mnt/data/random/VC/tg2/rvc-data/sound_dir```
Run:
```Bash
docker run --rm -it \
  --device=/dev/kfd \
  --device=/dev/dri \
  --device=/dev/snd \
  --group-add video \
  --group-add render \
  --group-add audio \
  --ipc=host \
  --network=host \
  --security-opt seccomp=unconfined \
  -e HSA_OVERRIDE_GFX_VERSION=10.3.0 \
  -e HIP_VISIBLE_DEVICES=0 \
  -e ROCR_VISIBLE_DEVICES=0 \
  -v /mnt/data/random/VC/tg2/voice-changer:/workspace/voice-changer \
  -v /mnt/data/random/VC/tg2/rvc-data:/data \
  tg-vc-rocm64```
8. Verify ROCm Torch inside container
```Bash
python3 - <<'PY'
import torch
import onnxruntime as ort

print("torch:", torch.__version__)
print("torch hip:", torch.version.hip)
print("cuda available:", torch.cuda.is_available())
print("device count:", torch.cuda.device_count())
if torch.cuda.is_available():
    print("device:", torch.cuda.get_device_name(0))

print("ort:", ort.__version__)
print("providers:", ort.get_available_providers())
PY```
Should look like
`torch: 2.9.1+rocm6.4
torch hip: 6.4.x <--------------- you're good if it recognizes torch hip
cuda available: True
device count: 1
device: AMD Radeon RX 6800 XT
ort: 1.21.0
providers: ['MIGraphXExecutionProvider', 'ROCMExecutionProvider', 'CPUExecutionProvider']`

9. Start VC server

python3 main.py --log-level debug```

if you get any errors on your system, just feed these instructions into any AI assistant

[proven hill]

pants? pissed.

[viral mason]

https://tenor.com/view/baldi-baldi's-basics-aaaaa-magicdippyegg-gif-11683388191454623599

[gusty socket]

and dont forget to install portaudio inside of the container
pulseaudio-utils libpulse0

[bronze pier]

I just trained a voice model with this https://colab.research.google.com/github/JackismyShephard/ultimate-rvc/blob/main/notebooks/ultimate_rvc_colab.ipynb where can I put it over an audio file?

i have RVC GUI locally but i think its too old

[viral mason]

install what

what gpu do u have (Nvidia or AMD) and what are u going to use it for?

why are u using it on roblox? that platform already has enough weird ppl on it

<@&1159293140440723499> if u read the rules that isn't allowed sorry

[bronze pier]

Are there any ways to train ai voice models? I been looking for about a week and all of them are outdated

[viral mason]

try Kaggle, it has Applio on it and works fine for me

I even recorded how to use it

[plucky crown]

You guys train the voices in collab/kaggle which is like cloud pc what is the maximum gpu collab has to train voices best?

[viral mason]

I train in specifically kaggle using applio

[plucky crown]

Whats the cpu or gpu specs for kaggle do u use for training

[bronze pier]

Could you send me that video

[civic kelp]

Idk if this may be the right channel to post this, but I want to upload 2 voice models and would like the model maker role.

I read the info on the website but there's no model-maker-role channel no more (or at least I can't see it?)

What's next?

[amber niche]

how do I post some of my latest ai images in the ai-images channel? do I need a certain role for that?

[viral mason]

first go here https://discord.com/channels/1159260121998827560/1453581938538315847

[civic kelp]

Oh thank you so much

[viral mason]

you'll need image perms if u cannot post images at the moment, just talk

no problem, if u have questions on models or anything I'm available for a bit

[civic kelp]

Thanks!

[amber niche]

just talk?

[viral mason]

sure

[viral mason]

yea you need a specific level to send images or gifs, not sure what level that is tho

[pulsar token]

but how many poems do one have to write to get perms =_=

5 level i think

[viral mason]

I dunno tbh

[civic kelp]

Actually now that I think of it... what tool would you recommend that can separate the backing vocals and the main vocals from a song?

Lalala.ai site works the best, but I'm looking for a free alternative with the same quality as them

[viral mason]

I use mvsep and I use this specific model

[civic kelp]

Holy gold, thank you! Will definitely try it later

[viral mason]

no problem!

u can add me if you like so u can message me if u have questions or anything

[civic kelp]

Sure thing lemme do it

[viral mason]

https://tenor.com/view/meowl-egypt-egyptian-egyptianmeowl-meowlegypt-gif-11055941431584742372

[hallow thistle]

Most online cloud services generally use Intel Xeon CPU, including Kaggle and Google Colab. Kaggle offers dual NVIDIA T4 and single P100. Kaggle's GPU quota has a fixed 30 hours in a week.

You're banned.

[eager wing]

hey how yall brainstorm saas ideas ?

[hallow thistle]

What is SaaS?

[eager wing]

software

as a service

[plucky crown]

I7 13700
3090 gigabyte oc
32gb 3600mhz ram
Msi mag mortar b760m ddr4 motherboard

Is this okay my pc i got compared to the cloud kaggle/collab since what you mentioned sounds like those nvidia a100s datacenter gpus with like 40gb or 80gb vram

[hallow thistle]

Bro.

[eager wing]

?

[plucky crown]

So what does it mean then, i havent tried training yet so , and also i just found out training text ai for a 3090 they said it cant handle much and is slow local ai text response compared to the latest models
So thats why im asking

[plucky crown]

The ones i asked about that does ai with a 3090 said they use dual gpu with 3090 just so its alot faster with just text ai

[hallow thistle]

To be honest, your PC specs and GPU are actually faster than typical Kaggle specs, even though most GeForce RTX GPUs have lower VRAM than those top-tier RTX Pro and data center GPUs. T4 is an older entry-level one.

[plucky crown]

Really? The ai group im in is like 4 gpu 3090 for ai and they want more, and i feel like shouldnt expect much with just one

[hallow thistle]

Let move on.

[viral mason]

<@&1159293140440723499> don't forget about this thing

[plucky crown]

Im about to train with the ai you guys have ,probably soon , hopefully things go well aaaa

[viral mason]

on kaggle?

[plucky crown]

Locally with my pc

[viral mason]

ah applio then

I cannot help with local applio since I've never used it

[hallow thistle]

Applio RVC.

[torpid oriole]

hey is that possible to make a v1 pth model to v2 ???

[hardy yew]

No

Better to just train a new one

[torpid oriole]

ahhh but we cant merge v1 and v2 right?

[uneven flame]

why whenever i talk it takes like 10 seconds for it to say it.

its so delayed

[cosmic holly]

hi

@hushed perch hi?

[plucky crown]

Hello

Question, how much tb do i need for storage for training the ai software here

[low shard]

Linux Voice Effects

[low shard]

you run it through docker?

[plucky crown]

@low shard also can i run the softwares on an external ssd and its training data to external hdd

[low shard]

this is weird, we usually don't have linux users but i don't remember them needing such a workaround to make it run, maybe its related to your specific linux distro or gpu drivers, I could try adding this in the docs

[low shard]

what's your pc gpu and os?

[low shard]

you sure you don't want to try training locally first?

[low shard]

they're executed

[low shard]

what ai software? there are multiple ones, elaborate using the help template

!help-template

[patent trellis]

AI HUB | Technical Support Desk

📋 Required Help Template

To receive assistance, you must provide your system details. Copy and paste the block below into your reply and fill it out.

⚠️ NO INFO = NO HELP

👉 Fill this form:

- Goal (e.g., TTS, AI Covers, Roleplay):
- Specific Issue:
- Full GPU Name:
- Operating System:
- Tutorial Link used:

📍 Quick Checklist Before Asking

• Check Docs: Many fixes are in the AI Hub Docs.
• Be Specific: Say "RTX 3060 12GB", not just "NVIDIA".
• English Only: Keep all discussions in English.

⚖️ Community Guidelines

• No assistance for NSFW/Porn or ANY Illegal Activities.
• Read the [Full Guidelines](#1402790586028789830 message).

Helpers are volunteers. Be patient and don't ask to ask.

[plucky crown]

The ones you have in these server ill try all of them

[low shard]

this is a general ai discord server, we talk about any ai programs lol

[plucky crown]

This one

Will its softwares training data be enough on a 2tb ssd or no get the training data to another hdd

[low shard]

elaborate your specs using this:

• Goal (e.g., TTS, AI Covers, Roleplay):
• Specific Issue:
• Full GPU Name:
• Operating System:
• Tutorial Link used:

[plucky crown]

I already answered that though

[plucky crown]

Full gpu: gigabyte 24gb oc
Operating: windows 11

[full sky]

Does anyone know how to train voice models? I honestly wanna make one myself

[proven hill]

theres the guide

[proven hill]

wich model of the gpu, also theres missing the goal

[full sky]

ah thaks mate 👍

[lilac moss]

hi @simple ore sorry if you don't appreciate pings but is changing the LR for pretrains recommended? or is sticking to 1e-4 better? i was thinking cause im doing a weird hacky way of embeddings that might need 5e-5 to prevent mode from collapse

[lilac moss]

wtf hi ilaria!!! its me frail/helena if you remember heh

[icy fjord]

So how do you guys make the AI vocals apply to the backing vocals in a song as well?
I'm using Replay and it doesn't really do that.

[proven hill]

sorry i had a problem and my memory is wiped almost completely (not a joke) but helena tickles me something

[bronze pier]

windows 11 omen laptop 7.8 GB of NVIDIA GeForce RTX 3050 Ti Laptop GPU

[lilac moss]

ouh it's okay! np! i was one of the helpers in the server far before this server was acquired / bought out / whatever, back when it was kalo tea snoop menh running things around here

[proven hill]

jeez thats 2023 stuff

[proven hill]

that tells us nothing

[bronze pier]

what do u mean

[proven hill]

whats the gpu

[lilac moss]

yeah wayyy back then

[proven hill]

sadly the server is mostly dead

[lilac moss]

yep i miss those days they were really fun i learnt alot and got into all this

[lilac moss]

^ would you happen to know anything abt this btw ilaria

[bronze pier]

7.8 Gb and its NVIDIA GeForce RTX 3050 Ti Laptop GPU

[proven hill]

ask me anything in dms

[proven hill]

download applio

[lilac moss]

dmed

[bronze pier]

i did and it does not work for me, i click on run-applio.bat and it looks like this

[proven hill]

did you download the precompiled version?

[proven hill]

use applio

[icy fjord]

What settings would I put on Applio for that?

[proven hill]

no special settings just take the backing vocals and convert

[icy fjord]

I'll give it a shot.

[low shard]

check https://docs.aihub.gg/rvc/local/applio/

[low shard]

what's your pc gpu and os?

[low shard]

rtx 3050 ti has 4gb vram, not 8gb, 4gb vram might not be good for local rvc training

[full sky]

err I cannot check it rn, I'm currently having it cleaned up a bit. I'll check it by tomorrow

[low shard]

it might be best you tell us later because that can help a lot directing you to the right guide and telling you if its good enough

or i could just give you the Applio guide but you risk wasting time if it's not good enough

[full sky]

I'll just check it myself if it can run good or not

[low shard]

your choice: https://docs.aihub.gg/rvc/local/applio/, https://docs.aihub.gg/essentials/how-to-make-voice-models/

[bronze pier]

yes, i run the install and it says this

but then run-applio.bat does nothing

[icy fjord]

Ah wait, I have to extract the backing vocals separately, is it?
And then convert that and the main vocals as separate files?

[low shard]

delete everything, download the latest precompiled version, unzip it and just run run-applio.bat, show what it does

[hallow thistle]

I once tried the single 53-stem BS Roformer model within an MSST notebook on Google Colab with NVIDIA T4 GPU; the GPU suddenly ran out of memory and process crashed. I just think because MSST attempted to load up 53 audio stems at once, which in my theory it would of course. The model itself is meant to be run elsewhere. Funny observation.

[simple ore]

there were some experiments, but nothing particuarly good came out of them... maybe half of current one for very small datasets

[viral mason]

What software are you using? And what for?

[gusty socket]

yesh, other option would be following og okada guide and going thru venv + replacing torch packages

I was surprised this worked worse on linux, you would assume rocm and ai stuff works better

[gusty socket]

I think someone more knowledgeable needs to come up with a proper guide for this

[hallow thistle]

Nick is one of staff members who manages AI Hub's docs.

What is your PC GPU? Did you follow any tutorial or guide before? And what would you use the voice changer for?

[low shard]

wait this looks interesting, might check that out later, that didn't work either?

[still current]

Hey, I have an RTX 2060, what's the best real-time option for it?

I was using w-okada, but I feel it's outdated.

[hardy yew]

Vonovox

[still current]

Maybe I'm just being dumb, but I'm testing it out and it's lagging a lot

[viral mason]

have you tested the current beta https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b779b/Vonovox_beta_17_11.zip

[viral mason]

ngl

the voicemodels.com server is ass and has no helpers

we are much more organized

[proven hill]

i remember working with a site like that

it was weights before weights

[viral mason]

were the staff also dumb?

[proven hill]

i mean i dont care.
they paid me

[viral mason]

oh

I wonder if they'll hire me

I have no life soo

[proven hill]

i think they closed down

dms

[viral mason]

aw man

[still current]

I'm on this version, and what I'm feeling is like the Vonovox isn't picking up my voice, because when I speak right up to the microphone, it picks it up, but the volume is at maximum.

I'm using the same template I use in W-Okada, the template doesn't look bad, I really feel like I have some wrong setting.

[viral mason]

try having it like this

turn off smart sine and rnnnoise reduction if that's on

[still current]

Believe it or not, I copied your settings. lol

[viral mason]

you did?

[still current]

yeah, but I feel the same problem

Maybe my RTX 2060 can't handle Vonovox.

[viral mason]

the lower gpus tend to struggle somewhat yea, have you tried wokada tg fork?

[proven hill]

the 2060 is very old too

[low shard]

@gusty socket could you try my experimental fork without any type of workarounds just following the AI Hub Docs guide normally?: https://github.com/Nick088Official/voice-changer/releases/tag/b2401

[viral mason]

what is this?

[low shard]

It should fix the Linux ROCm issue, there was a typo in the requirements-rocm.txt and needed to remove legacy flag from libraries when building

[viral mason]

that's neat

is it a fork of wokada tg?

[low shard]

yeah I just forked it to try to fix that, nothing else changes and I can't really test it since I'm not on AMD GPU nor Linux

[viral mason]

is this for the AMD linux version?

making sure since I could send this to a friend as well who's been having issues

[low shard]

Yeah this should fix his issue on Linux with AMD GPU, that uses ROCm

[low shard]

I'm not sure if it works and if he had the same issue as minn, but let me know

[viral mason]

def not the same issue but it might help it run as he got errrors before with it not working

[gusty socket]

sure, I'm done with work anyway

[gusty socket]

does it work only on Ubuntu?

[hasty iris]

hello. how can i use ai voice changer pls? mine is outdated

[paper bloom]

hey

im using the b2332 nvidia-cuda version of the owakada voicechanger is that an old version?

[low shard]

It's not working for you on CachyOS? It uses Ubuntu 22.04 to build the release, what issue do you get?

[low shard]

!help-template

[patent trellis]

AI HUB | Technical Support Desk

📋 Required Help Template

To receive assistance, you must provide your system details. Copy and paste the block below into your reply and fill it out.

⚠️ NO INFO = NO HELP

👉 Fill this form:

- Goal (e.g., TTS, AI Covers, Roleplay):
- Specific Issue:
- Full GPU Name:
- Operating System:
- Tutorial Link used:

📍 Quick Checklist Before Asking

• Check Docs: Many fixes are in the AI Hub Docs.
• Be Specific: Say "RTX 3060 12GB", not just "NVIDIA".
• English Only: Keep all discussions in English.

⚖️ Community Guidelines

• No assistance for NSFW/Porn or ANY Illegal Activities.
• Read the [Full Guidelines](#1402790586028789830 message).

Helpers are volunteers. Be patient and don't ask to ask.

[low shard]

that seems to be wokada deiteris fork last update which is old and not suggested, elaborate more

!help-template

[patent trellis]

AI HUB | Technical Support Desk

📋 Required Help Template

To receive assistance, you must provide your system details. Copy and paste the block below into your reply and fill it out.

⚠️ NO INFO = NO HELP

👉 Fill this form:

- Goal (e.g., TTS, AI Covers, Roleplay):
- Specific Issue:
- Full GPU Name:
- Operating System:
- Tutorial Link used:

📍 Quick Checklist Before Asking

• Check Docs: Many fixes are in the AI Hub Docs.
• Be Specific: Say "RTX 3060 12GB", not just "NVIDIA".
• English Only: Keep all discussions in English.

⚖️ Community Guidelines

• No assistance for NSFW/Porn or ANY Illegal Activities.
• Read the [Full Guidelines](#1402790586028789830 message).

Helpers are volunteers. Be patient and don't ask to ask.

[hasty iris]

gpu: msi rx 570 8gb
os: win 10
link used: i used it back in mid 2025

[paper bloom]

is there a better version?

[hasty iris]

im using site version of voice changer

[low shard]

this is a general ai discord server, you need to elaborate more, if you're trying to do AI Covers, E Girl Trolling / E boy Trolling / Catfishing or Roleplay

and also the screenshot of the program or program files if you don't have the link, there are many programs, I can't know whatever you're using

[low shard]

you need to first elaborate what I asked, read up the template

[hasty iris]

okay let me show you real quick

[gusty socket]

Just was going through the commits, I'm installing it now

[hasty iris]

its called MMVCServerSIO

[paper bloom]

tbh idk those answers im rlybad at stuf like this

[gusty socket]

Gives an error

[low shard]

open the program and show a screenshot to check the version

[paper bloom]

is there no tutorial how i can install a newer version?

[low shard]

those are needed, what can't you answer? there isn't just a single better voice changer, this is a general ai discord server with multiple programs, elaborate everything

[hasty iris]

is this old?

b2309

[low shard]

that seems to be an old version of the wokada deiteris fork https://github.com/deiteris/voice-changer/releases/tag/b2309, which is outdated

now there are multiple replacements, but you need to elaborate your goal, are you trying to do Realtime TTS, AI Covers, E Girl Trolling / E boy Trolling / Catfishing or Roleplay?

[hasty iris]

yes i want to real time TTS. i have got voice models from #1175430844685484042

or like trolling yk

[viral mason]

what trolling?

[gusty socket]

aha, so catfishing

[hasty iris]

no not like that i already got my lesson when i did that

[viral mason]

like Darth Vader or Goku?

[paper bloom]

Goal (e.g., TTS, AI Covers, Roleplay): roleplay

• Specific Issue: i wanna know if my voice changer is outdated and if there is a better version, and what makes the btter version better?
• Full GPU Name: NVIDIA GeForce RTX 4060
• Operating System: Windows i think
• Tutorial Link used: idk that

like this?

[hasty iris]

yeah, movie characters and etc.

[viral mason]

thank you for being normal

Nick, this one is good do not ban them for their previous answer

[viral mason]

what kinda roleplay? like playing as a movie character or something?

[low shard]

then you're using those wrong tools from the start, most models in #1175430844685484042 are RVC, which is STS not TTS

You could check https://docs.aihub.gg/tts/realtime-tts/ for realtime tts

[low shard]

will check it up rq

[hasty iris]

no no i dont need tts

like i need to mimic their voice in real time

idk what its called

It was RVC i believe

[paper bloom]

ehm like i want to use anime voices from jjk megumi espicially

ive been using the wikada one but so far i couldnt find any models and want to know fi mine is outdated

[viral mason]

https://discord.com/channels/1159260121998827560/1175430844685484042

[paper bloom]

do u also know how to get a newer version of the wokada one

[paper bloom]

and do u know why those voice models sound so good like u cant even tell its an ai and when i use it it sound so rovbotic

[viral mason]

what gpu do you have? nvidia or AMD?

[paper bloom]

nvidia

[viral mason]

is it like 3070, 1060 ect

if u know

[paper bloom]

Specific Issue: i wanna know if my voice changer is outdated and if there is a better version, and what makes the btter version better?
Full GPU Name: NVIDIA GeForce RTX 4060
Operating System: Windows i think
Tutorial Link used: idk that

[viral mason]

do u know the link u got the voice changer from?

this is the newest one for Nvidia https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b779b/Vonovox_beta_17_11.zip

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

[gusty socket]

try latest forks from the wiki

[paper bloom]

idk i saw it in this channel before

[paper bloom]

where can i find it? back then when i got mine there was tutorial soemwhere in this discord

[paper bloom]

is there also a tutorial for it?

[gusty socket]

https://discord.com/channels/1159260121998827560/1393389200862089240

[viral mason]

Kinda, but it's very simple to install both, for the second link just extract it and run setup64 and then install drivers, for the first one extract and run Start

[paper bloom]

there is a message with a security alert

whats that about im scared now

[viral mason]

?

Are you able to send a screenshot

[viral mason]

Ignore that, that was about a discord server that copied our server

This server is completely safe now

[viral mason]

Did you download both files btw

[paper bloom]

oh and wich one should i use it says there is 3 versions of the voice changer

[viral mason]

I sent you the files you need

[low shard]

did you download the source code instead of the release and tried to build it manually ? You seem to be running the source code vc_install instead of the MMVCServerSIO built file

You need to follow exactly https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/, but instead of the original github links you have to just download the tar gz aa and ab from: https://github.com/Nick088Official/voice-changer/releases/tag/b2401

[paper bloom]

not yet is thjere no tutorial where i can do it step by step

[viral mason]

No, I already said what to do, I would record how to use it but I'm not home rn at my pc

[viral mason]

Just follow these instructions and if you're confused I can elaborate more

[low shard]

your gpu is old, but you could try https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/

[low shard]

check the actual guide, it might be easier than using manually the files provided by local worm: https://docs.aihub.gg/realtime-voice-changer/local/vonovox/

[paper bloom]

do i have to uninstall the wokada i got now?

[viral mason]

I would recommend that yes, just to save space

[low shard]

you got an old version, it's not needed anymore, it's better to just use vonovox

[viral mason]

What Nick said lol

[paper bloom]

does this one run better?

[low shard]

i mean you could just give the guide link and tell them to use the beta download link if you wanna give them the beta, some people might need the guide to understand better

[low shard]

vonovox is more updated

[paper bloom]

does both work the same? like both uce rvc models?

[viral mason]

Some people struggle with the guide not being adjusted to the beta version and ask about stuff that isn't included or was removed/changed

[low shard]

both use RVC models

[viral mason]

Yes they both work the same, Vonovox just has better quality and is much more optimized

They all use rvc models

[low shard]

well because the beta isn't considerated as a stable release, and the guide can't be updated for every single beta since they come out faster than stable releases, it wants to be as stable as possible rather than some finding some rare issue

[viral mason]

That's fair but the beta release is the most recommended version of Vonovox at least from my view

Much better than the current full release 1.6.9

[paper bloom]

ty very much another question what makes this one better then the version i got? are the voice better?

or just liek the speed is betteR?

[low shard]

performance and feature-wise

[viral mason]

There's a lot that makes it better, it's much more optimized so it shouldn't be laggy in games or on discord, the quality is noticeably higher because of the upscaler that enhances quality, it's very fast and sounds good overall

There is only 64 slots though which is a downside to the massive over 100 models slots that other okada programs have

That's the only negative thing about it, other than that it's amazing

[paper bloom]

wdym by slots?

[viral mason]

Like how many models you can have on the app

[paper bloom]

is this how i start it? and it also says post-installation do i have to do it with my gpu

[paper bloom]

nice ty can i aslo make like breathing sounds more real with it or like can ppl hear when i use my keyboard? cause mine makes weird sounds when it hears the keyboard

[viral mason]

It automatically uses your gpu, to start it run the file called "Start"

[viral mason]

That's dependent on the model still and it does pick up stuff mostly depending on your microphone but also there are features to suppress it picking up such noises

[paper bloom]

tyy what do i click on first on my screenshot?

sryy im rly newbei i stuff like this

[viral mason]

Wdym

[paper bloom]

like do i click on ''github repository'' to download that

[viral mason]

these links https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b779b/Vonovox_beta_17_11.zip

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

[paper bloom]

ahh the second one isa virtual cble?

[viral mason]

Yep!

[paper bloom]

do i need to download it i fi already haveo?

[viral mason]

It'll allow you to use it in games, discord, ect

[paper bloom]

i think is use the lte one s

[viral mason]

Nah you only need one

[paper bloom]

alrigty

[low shard]

@gusty socket please try using the built build of my fork and let me know, don't build yourself with the source code

were you perhaps using vc_install instead of the built release for the upstream repo too?

[quasi pagoda]

hi i lost my rights idk what happened can i have a star back or just the right to post gifs or whatever or maybe not thats cool too

[low shard]

that's because you left and rejoined, gave you the verified role back but please use the modmail by dming @hushed perch next time

[quasi pagoda]

ty

[teal void]

Hello, I need help with a workflow in ComfyUI for LoRA

[paper bloom]

hey i downlaoded it what now?

[viral mason]

Extract the zip files and then run the file called Start

[gusty socket]

yes, I git cloned it, lemme try

[proven hill]

i love your pfp

[paper bloom]

hpw do i extract it

[paper bloom]

haha ty

[proven hill]

right click, unzip

[paper bloom]

which one?

[viral mason]

Start

[gusty socket]

ImportError: /mnt/data/random/VC/vc-nick088/MMVCServerSIO/_internal/onnxruntime/capi/onnxruntime_pybind11_state.so: cannot enable executable stack as shared object requires: Invalid argument```

[sharp thunder]

anyone know how to use the okada voice changer on mac

[paper bloom]

start adn then extract?

[viral mason]

No no, extract the zip file then run start

[paper bloom]

but liek how do i extract it how do i exract it

[low shard]

Seems like the same execstack security policy issue, I will modify the github workflows to try to patch that again

for now you could run this in your terminal:

cd /mnt/data/random/VC/vc-nick088/MMVCServerSIO
find . -type f -name "*.so*" -exec patchelf --clear-execstack {} \; 2>/dev/null

This temp fix will make it run, and make us understand if my fork atleast fixed the GPU detection in that build

Or you could wait for my next build in like 20 minutes that should fix both GPU detection and the security policy

[gusty socket]

it runs, but no GPU is detected

[low shard]

could you send logs?

[paper bloom]

ehm is it normal that it has over 40k elements?

[viral mason]

I'm not sure

I downloaded it a while aho

[low shard]

first try this to show me if you actually have the files or there's a pyinstaller bug that somehow removed the rocm libraries:

ls -l _internal/torch/lib/ | grep hip

then try to run it with the flag you talked about earlier:

cd /mnt/data/random/VC/vc-nick088/MMVCServerSIO
HSA_OVERRIDE_GFX_VERSION=10.3.0 ./MMVCServerSIO

[gusty socket]

no files, grep returns nothing

[viral mason]

mmm grapes

[low shard]

weird, could you try that run command with the flag?

[paper bloom]

so i extracted it but where do i find the extraxted file

it still looks like this

[viral mason]

check your downloads folder

[paper bloom]

cant seem to find it what happens if i extract again

[viral mason]

try this

instead of clicking and dragging with left click use right click instead

[paper bloom]

wont it download it again?

like i extractedit but idk wher eit is but ill try this one out

[gusty socket]

I did, same thing

[viral mason]

?

[paper bloom]

does the folder has the same name after extracting it?

[viral mason]

yep it should

[paper bloom]

is it this one

[viral mason]

this is it!

[paper bloom]

perfect so how do i hear my self?

[viral mason]

just like this

[paper bloom]

omg ur so awesome

one more question whats the diffrence between ouztput and input?

does both effect how loud i am ?

[viral mason]

this one controls overall loudness and this one u can change for each model

[paper bloom]

adn what is output?

[viral mason]

I'm not really sure, but the audio settings should be
Input: normal headset/headphones mic
Output: line 1

[paper bloom]

yes i men like u can swipe to rght and elft at the output thign does it change the way how lod ppl hear em or is input for it?

[viral mason]

oh yea it should

[hardy yew]

Input volume is applied directly on the mic input, before passing to RVC. If the mic level is low, it's worth increasing input volume for better conversion. Other than that, output volume should be used, because that's what directly controls the volume of converted voice without affecting feature extraction.

[paper bloom]

oh so input is for my voice get sinto the rvc and output controls how loud the other person can ehar me?

[hardy yew]

technically, yeah. Obviously vonovox also handles the volume, so input volume indirectly will affect output volume too, but it's a side-effect, not the primary purpose

For example, If you set input volume very low, then output will be mumbly because feature extraction from low signal will fail. But if you keep input volume normal and set output volume low, the voice will be quiet, but clear

[viral mason]

what are u looking for?

[paper bloom]

thank you guys verymuch ❤️

[paper bloom]

u 2 thanjk youu ❤️

[viral mason]

you're welcome!

[viral mason]

@coarse laurel what are u looking for

[low shard]

the issue is still missing files excluded from the pyinstaller

https://github.com/Nick088Official/voice-changer/releases/tag/b2404 try this executable, try it both normally and with the flag, it shouldn't give policy issues anymore nor not detect the GPU

If it doesn't work again, give logs and I might give it another shot but I'm not really a Linux ROCm expert

[gusty socket]

had to use a flag to get it working, but not GPU still

[low shard]

what flag? are you talking about the policy issue still?

I might try with another pytorch version with rocm6.4, i just checked and pytorch 2.7.1 doesn't have rocm6.4, i checked your message but in their pypi there's no 2.6.0 with rocm6.4, so hopefully a last try with pytorch 2.8.0+rocm6.4 doesn't break anything

Will try to make it soon, else I will just add the docker method to the docs and explain the issue

[gusty socket]

oh yeah, that's what I've been doing with docker, had to downgrade rocm as well

[feral crater]

Any good Pre-trainer for female singers?

I remembered KLM was a thing, has the DEV stopped updating it?

Has the DEV for KLM made any statements?

[viral mason]

use Legacy core 1.6 as of now, klm is ok but idk what the best newest one of his pretrains is

[idle cedar]

Anyone knows how to use gemini without a google account?

[green birch]

ok im unsure this is the right room im trying to get Mangio-RVC-Fork-main to pick up audio from another pc how would i do this from pc to pc suggestions welcome ping or reply if you have done a multi audio setups

object pc a to pc B and have pc B edit audio live

i am sort of stuck it picks up normal audio in the teleport (obs to obs) but but it does not seem like the i think its mangio picks up or edits it and i think it should ping me if you know / ty im brain broken.

[proven hill]

whats your goal

[green birch]

have pc (b ) edit audio

from pc a live

[proven hill]

okay for what porpuse

[green birch]

thats the porpuse and the goal pc a has the mic pc b is the stream box

[proven hill]

no.
what are you going to do? catfish or something?

[green birch]

i just need to know why or possibly how to set up the voice thing to pick up something other than "desktop"

[green birch]

ooo no i just need some lag

so if i crash my main pc it wont crash my stream pc or my record / lab stuff

[green birch]

example before update

[proven hill]

so youre going to use it to change your voice in real rime for streaming

[green birch]

its a yes or no question: can it pick up from a to b ?

[proven hill]

yes

[viral mason]

this conversation is very confusing

[green birch]

the point is audio separation and lag

[green birch]

nah i got my answer so i know its my fault now just unsure how its sending from a to b and why B can not see A's audio 2 it only picks up desktop so it starts doing a bzzzzzz and its terrible

[proven hill]

honestly that graph is scaring me

[viral mason]

don't they just want to use a voice changer or smth?

[green birch]

im a professonal hacker / redteam / purple team

[proven hill]

yes

[proven hill]

5 bucks youre trans

[viral mason]

then all they need is vonovox on nvidia or TG fork on AMD

[viral mason]

they could be a femboy too

[proven hill]

i placed my bet

[green birch]

no i just like to break things and im testing something thats about the end of it

[proven hill]

https://tenor.com/view/serial-experiments-lain-lain-stare-into-the-abyss-y2k-2000s-gif-24846261

[low shard]

i'm still trying to fix it, but is the same issue with https://docs.aihub.gg/realtime-voice-changer/local/deiteris-w-okada-fork/ ? This is what tg develop fork is based on

[gusty socket]

I think so, I've tried both before trying docker

[green birch]

@proven hill ty for info while for w/e reason you are simi hostile knowing it is possible will let me know that its my issue not audio to audio

[proven hill]

im hostile because this informations is usually asked by people who catfish people

[green birch]

i suppose thats fair: i get the "trust me bro" its my account daily

im just trying to clean a audio scream / hiss and add lag for dc fail over

[gusty socket]

speaking of catfishing, does anyone have a model for a dommy mommy voice? For uuh reasons

[green birch]

im using 4 - 6 computers at the same time for stuff so i have to isolate everything

[viral mason]

no

ask again and you won't be here very long

❤️

[gusty socket]

I need to call my fren a good boy once he gets me back to celestial in marvel rivals

[proven hill]

me

[gusty socket]

im in my female loki phase

[night cave]

can someone give me the best voice mod and the settings to go with it to sound like a girl please?

[proven hill]

do you need it in real time?

[night cave]

yeah

[viral mason]

why?

[proven hill]

why girl voices specificallu

[night cave]

to create content, as i've been doxxed for my voice before and i don't want it to happen again

[viral mason]

why girl voices tho, there's so much cooler stuff like Spongebob or Darth Vader or Goku

[night cave]

i'm known as a girl in all the servers i've already been doxxed in, and yes i'm actually female, so if i ever interact there again, the majority might think that the doxxing was false

[night cave]

@viral mason

[viral mason]

I dunno man, just look through here https://discord.com/channels/1159260121998827560/1175430844685484042

[night cave]

i tried, but they all sound really bad

[viral mason]

what are u looking for

[night cave]

and whenever i speak a full sentence it starts breaking

[proven hill]

im not allowed to speak

[night cave]

also there's this kind of breathing sound in the background constantly

[proven hill]

but trust theres place to find

[viral mason]

my models sound great

[night cave]

@proven hill help please?

[proven hill]

you smell like a catfisher

[night cave]

wallahi

i can swear i'm not

[viral mason]

yea I don't trust this

I'll let someone else deal with this

if u have Nvidia gpu download vonovox, if u have AMD get wokada tg fork, that's all I'll say

[proven hill]

what are you searching for anyways

[night cave]

just a traditional girl voice

[proven hill]

theres plenty of options

[night cave]

could you please help me with them?

[viral mason]

just look around

it's not hard

try this one smh

https://discord.com/channels/1159260121998827560/1501708538420924436

[proven hill]

no because i suspect youre a catfisher

[night cave]

im not tho?

but alright

[torpid dawn]

heyy everyonee quick question about mic setups. i’m a trans girl and after 8 months of voice training not really working out for me i’ve switched to using a voice changer to help me with my dysphoria and i actually found a model that sounds realistic and feels like me. i’m just trying to make the audio setup sound as natural as possible over discord etc. does anyone know if adding a tiny bit of background noise or static through voicemeeter helps mask any weird artifacts and make it sound all more natural and convincing? any tips for a natural sounding setup would be amazing 💕 thankssss

[viral mason]

Hi! what gpu do you have (Nvidia or AMD)?

I'll gladly help wit has much info as u can give ^^

[patent trellis]

<:error:865860636084142150> Message error

The message template:help-template could not be displayed. More details can be found in the error log.
Please report this to a server admin.

[proven hill]

@low shard

[delicate cradle]

i gotchu king

https://discord.com/channels/1159260121998827560/1341216399372062823

[night cave]

say wallahi bro

[delicate cradle]

why

who / what even is wallahi

[torpid dawn]

hiii thank u so much for offering to helpppp. i actually have an nvidia gpuu. what other info do u need from me? i’m a little new to the deeper audio routing stuff so any advice you have would be appreciateddd

[low shard]

rip, last try: https://github.com/Nick088Official/voice-changer/releases/tag/b2418

if this doesn't work in any way, I'm just going to use your Docker fix in the docs, I tried helping myself with Gemini 3.1 Pro Preview and claude opus 4.6 thinking but seems to be a rare bug, i can't really test it myself and its niche so i dont want to waste neither mine or your time, docker might just have a lil bit more of overhead but atleast it works guaranteed

[viral mason]

if you could say exactly what gpu that might help me decide which voice changer you can use like RTX 3060 or RTX 1060 ect

[low shard]

-# GTX 1060*

[viral mason]

what is a GTX

[low shard]

you didn't do a typo? the RTX serie came after the GTX 16 serie

[viral mason]

I pretty much only know RTX is Nvidia and AMD is something else

I dislike confusion

[low shard]

don't you have a gtx 1660 super yourself

[viral mason]

haven't had that for a year

forgot about it tbh

what voice changer, there is more than one

what gpu do you have(Nvidia or AMD) and what are you using the voice changer for?

and what are you wanting to use it for?

mh

that's not allowed here

<@&1159293140440723499>

[wide oasis]

eyo

https://tenor.com/view/i-have-been-summoned-gif-19881033

[gusty socket]

kk, let me try it

[wide oasis]

i have been summoned

[low shard]

summoned 2 seconds earlier

[wide oasis]

fuck

i became predictable

https://tenor.com/view/masters-of-the-universe-skeletor-mot-u-gif-5430301619065962680

[gusty socket]

unsummon yourself rq, they wnot expect it

[viral mason]

this is the quickest I've seen mods actually do something

wonderful :3

[wide oasis]

no iwas there but im slow typer and i need to use the gif as my intro

[wide oasis]

it depends on if im online or not

[low shard]

jokes aside all that matters is that the case has been taken care of

[gusty socket]

is the archive corrupted?

[low shard]

hopefully not, im downloading it, it doesn't work for u?

[gusty socket]

can't open with Arc either

[low shard]

I'm starting to think that this has been an issue since the original wokada

[gusty socket]

[low shard]

Because there's barely any linux amd gpu user using the program, majority of users are atleast on windows

[gusty socket]

Jensen Huang transferring 21 million dollars to original okada developer to not include ROCm libraries

[low shard]

You could try opening an issue about it but I wouldn't expect much about it, tg-develop and deiteris are both using windows iirc

[low shard]

they are included, the issue is PyInstaller that discards them

[gusty socket]

Jensen Huang hijacking pytorch repo to replace
if amd then pyinstaller_amd
to
#if amd then pyinstaller_amd

[low shard]

btw sorry for wasting our time about this, I think it's just better to give up and use the docker way atp,

I will link to your message and give credit

Btw which AI Did you use to help you out for the docker way?

[gusty socket]

no worries, happy to help, you can ping or dm me if you want to tinker more

[gusty socket]

gpt plus

[low shard]

I will add you to keep that in mind, but for now I will give up, if you ever find out a fix please let me know, I might delete my fork in like a week or smt if I totally give up or retry later

You might also try the original wokada and let me know if that works for you

[gusty socket]

I can make a better instruction that also includes portaudio inside of docker as well

or just send it to you, you can format it however you need

[low shard]

I added already a formatted section: https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/#amd-gpu-on-linux-only-showing-cpu-docker-fix

You can let me know if there's anything wrong and could dm me or PR it

[torn orbit]

what do epochs mean im a new gen i keep seeing it on voice models and dont know what it means

[safe oracle]

how do i fix lag in audio while gaming only

[viral mason]

it's just how many times the ai saw the audio used to train that model before it sounded good to the person making it

more or less doesn't equal better or worse model

[torn orbit]

i see

thank you

[viral mason]

what are u using? and what gpu do u have? (Nvidia or AMD)

[low shard]

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

[low shard]

!help-template

[patent trellis]

AI HUB | Technical Support Desk

📋 Required Help Template

To receive assistance, you must provide your system details. Copy and paste the block below into your reply and fill it out.

⚠️ NO INFO = NO HELP

👉 Fill this form:

- Goal (e.g., TTS, AI Covers, Roleplay):
- Specific Issue:
- Full GPU Name:
- Operating System:
- Tutorial Link used:

📍 Quick Checklist Before Asking

• Check Docs: Many fixes are in the AI Hub Docs.
• Be Specific: Say "RTX 3060 12GB", not just "NVIDIA".
• English Only: Keep all discussions in English.

⚖️ Community Guidelines

• No assistance for NSFW/Porn or ANY Illegal Activities.
• Read the [Full Guidelines](#1402790586028789830 message).

Helpers are volunteers. Be patient and don't ask to ask.

[torn orbit]

thank you my question was already answered though

appreciate you

[buoyant cave]

Help, on my end I can't choose any models at all for inference in original RVC. It just doesn't show a list of models. I've installed pretrained models btw

• GPU: RTX 3060 12gb
• Operating System: Linux Mint
• Tutorial Link used: https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/docs/en/README.en.md

[viral mason]

are you wanting to make covers or something?

[buoyant cave]

I actually have a goal for realtime voice changer

[viral mason]

I can't help much with Linux but I can point you in the right direction

[buoyant cave]

Gotcha

[viral mason]

try Vonovox, I am unsure if it works with Linux but why not try

[buoyant cave]

sure, I think original RVC just dead or smthing

[viral mason]

https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b779b/Vonovox_beta_17_11.zip

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

[viral mason]

rvc doesn't stand for realtime voice changer btw:3

[long agate]

if i have a mac w a m4 chip can i use this 😅

[buoyant cave]

o h

[viral mason]

wha

[long agate]

the voice changer

[viral mason]

no what's a m4 chip

[long agate]

uh

[viral mason]

ye, basically just download those two links, install vac lite (second one) by running setup64 then install driver, after than for vonovox run start

make sure to extract them

[buoyant cave]

wait, is it only for windows?

[long agate]

um like the chip that my macbook uses

[viral mason]

unsure tbh, but if it does only work with windows Wokada TG fork has many versions

[buoyant cave]

yep

it has only setup.exe

[viral mason]

unfortunate

<@&1159293204038955078> when u can this kind fellow needs help with the Nvidia linux wokada tg build (idk what download version it is)

while u wait tho the quickest way to look is in the guide for it

-rt

[patent trellis]

🔊 Realtime RVC

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

• Vonovox

A Realtime Voice Changer with similar performance to Wokada Tg-Develop Fork, with extra features, but it supports only Nvidia GPUs on Windows 10/11 unlike other options that have wider support. and without cloud options

• Local 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.

• Local Guide
• Cloud Guide(s)

• Applio Realtime

A Realtime Voice Changer with similar performance to Vonovox & Wokada Tg-Develop Fork, with extra features.

• Local Guide

• Wokada Deiteris Fork

Deiteris' fork (modified version) of wokada that doesn't get updates anymore.

• Local Guide
• Cloud Guide(s)

⛔ 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]

https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/

ignore the warning at the top btw that's been solved

[buoyant cave]

yea, I'm reading it rn

oooh I hope it does work

[viral mason]

:o

that's

did I send the wrong thing

that isn't tg fork that is deiteris it looks like

much older than tg

[buoyant cave]

hm?

I just followed this guide: /home/lf/Documents/VOICEMODELS/GeneralGrievouscvlv1.5V1_235e_9635s

[viral mason]

whuh

that's not a guide that is a voice model

[buoyant cave]

fuck

I made a mistake when copied things from place to place

https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/

ths one

[viral mason]

yea

tg fork looks like this tho

[buoyant cave]

🥀

[viral mason]

tho I've never used anything but the windows version of anything so maybe for whatever reason it's different

prob not tho

[buoyant cave]

yes, I just deided to try again voice changer

I used only on windows previously

before ai hub was reborn

bruuuh
2026-05-08 05:34:20,880 INFO [OnnxLoader] Converting model to FP16...
2026-05-08 05:34:24,854 WARNING [symbolic_shape_infer] Unable to determine if n_samples <= ConvTranspose_308_o0__d2, treat as equal
2026-05-08 05:34:25,183 INFO [OnnxLoader] Done!
2026-05-08 05:34:25.690178271 [W:onnxruntime:, transformer_memcpy.cc:74 ApplyImpl] 2 Memcpy nodes are added to the graph main_graph for CUDAExecutionProvider. It might have negative impact on performance (including unable to run CUDA graph). Set session_options.log_severity_level=1 to see the detail logs before this message.
2026-05-08 05:34:25,736 INFO [PipelineGenerator] Loading index...
2026-05-08 05:34:25,737 INFO [PipelineGenerator] Try loading "model_dir/0/GeneralGrievouscvlv1.5V1.index"...

no nvidia?

[viral mason]

I think you should prob switch back when u can

[buoyant cave]

I won't

I hate windows because of it's artifacts, glitches, blue screens (now black screens) and so on

[viral mason]

my windows pc on my 5070ti works completely find without any bugs or errors

[buoyant cave]

Oh damn lucky you

In my case I had crashes and kernel errors

just at random

[viral mason]

damn

[buoyant cave]

sometimes parts of windows weren't drawn and win10 just fell into idk kernel panic? Looped last few ms of audio from yt and went into black screen spinning my fans fast

I just better not touch win except for pc clubs

that's inevitable

[viral mason]

weirdd

I use the evil windows 11 and it seems fine

[buoyant cave]

YEAH fuck both of this versions

when I build a pc with a bro we decided to install win11. Well, it was a pain in the ass because after downloading first updates it just started lagging

we lost about a fucking hour, ruining feeling of high end pc being right there

[viral mason]

[buoyant cave]

sometimes I wish I never learned anything digital

[idle fossil]

goal = oprate open notebook on my pc ,,issue = desktop doker run into errore ,,,gpu=rtx 5080 ventus ,,,op=win 10 link =https://www.open-notebook.ai/get-started.html

[buoyant cave]

holy shit it's working

[idle fossil]

please help

[buoyant cave]

@viral mason yoo it's working just requires to reload page a few times xd. Built in protections I guess ruin everything

[idle fossil]

cant open open notebook

[buoyant cave]

now I guess I have to figure out how to make a virtual cable, but it's probably easy to do\

idk what magic they did but few years ago I couldn't use models with delay less than 128. Now I can

[obsidian crater]

What u supposed to do to fix choppy audio / voice cracks with the voice changer

Just sounds kinda unnatural

[viral mason]

what gpu do u have (Nvidia or AMD) and what are u using the voice changer for?

[obsidian crater]

Nvidia

Using it for fun

[viral mason]

like Spongebob or Goku or somethin

[obsidian crater]

Yea

[viral mason]

ok good good, you should switch to Vonovox

https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b779b/Vonovox_beta_17_11.zip

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

I have the links right here :D

[obsidian crater]

Ok

Ty bro

Does this work with all the voice file things that ppl post in this server?

[viral mason]

ye works the same as the thing u already had but wayy better

[obsidian crater]

Ok thx 🙏

Preciate it

[viral mason]

btw the second link is a virtual audio cable like VB cable but it causes less issues on windows

[obsidian crater]

Ah

So i should replace the one im using rn with that one?

This the thing that i was using before btw https://youtu.be/dZ_2HELnWJU?si=qCBjcHX1AykTUrJL

[viral mason]

o yea that's super old

[obsidian crater]

Ok lol

Explains why i was having issues

[viral mason]

yea delete the folder for the old vc and cable completely

[obsidian crater]

Ok 👍

Im boutta go to bed rn but ill get on that tmrw

Thx for the help

[viral mason]

alr night!

[obsidian crater]

https://cdn.discordapp.com/attachments/1114334579424231494/1471912195989307404/image0.gif

Oh

[viral mason]

https://tenor.com/view/dab-me-up-dab-me-up-but-something-beautiful-is-missing-gif-8902167673945806588

[feral crater]

Thanks, I'll give legacy core 1.6 a try.

I have AMD, more specific I have RX 9070 XTX.

I would love to train locally, which one would you recommend for AMD?

[viral mason]

Locally you can use Applio, tho I have zero idea how to install it locally the guide should be very helpful for that!

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[viral mason]

Just look into the applio docs

[gusty socket]

@low shard I saw there's a new build, but it seems to be also corrupted, can't extract with cat or arc

[viral mason]

Meow

[magic depot]

can someone help me when using the voice changer egirl some words cut out

[low shard]

Are you looking to do ai covers, training models, e girl trolling / Catfishing or roleplay?

[low shard]

!help-template

[patent trellis]

AI HUB | Technical Support Desk

📋 Required Help Template

To receive assistance, you must provide your system details. Copy and paste the block below into your reply and fill it out.

⚠️ NO INFO = NO HELP

👉 Fill this form:

- Goal (e.g., TTS, AI Covers, Roleplay):
- Specific Issue:
- Full GPU Name:
- Operating System:
- Tutorial Link used:

📍 Quick Checklist Before Asking

• Check Docs: Many fixes are in the AI Hub Docs.
• Be Specific: Say "RTX 3060 12GB", not just "NVIDIA".
• English Only: Keep all discussions in English.

⚖️ Community Guidelines

• No assistance for NSFW/Porn or ANY Illegal Activities.
• Read the [Full Guidelines](#1402790586028789830 message).

Helpers are volunteers. Be patient and don't ask to ask.

[low shard]

!help-template

[patent trellis]

AI HUB | Technical Support Desk

📋 Required Help Template

To receive assistance, you must provide your system details. Copy and paste the block below into your reply and fill it out.

⚠️ NO INFO = NO HELP

👉 Fill this form:

- Goal (e.g., TTS, AI Covers, Roleplay):
- Specific Issue:
- Full GPU Name:
- Operating System:
- Tutorial Link used:

📍 Quick Checklist Before Asking

• Check Docs: Many fixes are in the AI Hub Docs.
• Be Specific: Say "RTX 3060 12GB", not just "NVIDIA".
• English Only: Keep all discussions in English.

⚖️ Community Guidelines

• No assistance for NSFW/Porn or ANY Illegal Activities.
• Read the [Full Guidelines](#1402790586028789830 message).

Helpers are volunteers. Be patient and don't ask to ask.

[feral crater]

Thank you

[low shard]

!help-template

[patent trellis]

AI HUB | Technical Support Desk

📋 Required Help Template

To receive assistance, you must provide your system details. Copy and paste the block below into your reply and fill it out.

⚠️ NO INFO = NO HELP

👉 Fill this form:

- Goal (e.g., TTS, AI Covers, Roleplay):
- Specific Issue:
- Full GPU Name:
- Operating System:
- Tutorial Link used:

📍 Quick Checklist Before Asking

• Check Docs: Many fixes are in the AI Hub Docs.
• Be Specific: Say "RTX 3060 12GB", not just "NVIDIA".
• English Only: Keep all discussions in English.

⚖️ Community Guidelines

• No assistance for NSFW/Porn or ANY Illegal Activities.
• Read the [Full Guidelines](#1402790586028789830 message).

Helpers are volunteers. Be patient and don't ask to ask.

[low shard]

Are you trying to do e girl trolling? What's your PC GPU and OS?

[low shard]

It's better we use DMS since this channel is full,

[valid cosmos]

hello im trying to create ai content for tiktok/youtube mainly short form and i dont know how to do it correctly. im trying to make a video similar to your life as a twich streamer for example but i try multiple promts and the answer is not what im asking anyone having a tutorial to help me do it correctly ?

btw i dont want to pay for any service or anything just to know

[kindred nacelle]

which one do i download for an amd gpu

[low shard]

!help-template

[patent trellis]

AI HUB | Technical Support Desk

📋 Required Help Template

To receive assistance, you must provide your system details. Copy and paste the block below into your reply and fill it out.

⚠️ NO INFO = NO HELP

👉 Fill this form:

- Goal (e.g., TTS, AI Covers, Roleplay):
- Specific Issue:
- Full GPU Name:
- Operating System:
- Tutorial Link used:

📍 Quick Checklist Before Asking

• Check Docs: Many fixes are in the AI Hub Docs.
• Be Specific: Say "RTX 3060 12GB", not just "NVIDIA".
• English Only: Keep all discussions in English.

⚖️ Community Guidelines

• No assistance for NSFW/Porn or ANY Illegal Activities.
• Read the [Full Guidelines](#1402790586028789830 message).

Helpers are volunteers. Be patient and don't ask to ask.

[low shard]

You need to elaborate more, you might be using old versions of a program, this isn't just a voice server, this is a general ai discord server

[buoyant cave]

Problem solved

[fringe crag]

windows & linux, radeon gt 9060 xt oc 16GB: i use thoth, LMstudio, ollama, alpaca (on linux) but not one program is using my GPU, they all stay on the CPU

toth (qwen 3; 14b is already thinking for 5' about what to say after i said hello

[simple ore]

LM studio has options to use vulcan, etc

on linux it should not even be an issue as you should be using rocm torch

[hasty iris]

yo, i installed the latest okada but whenever i put .pth and .index it does not sound like the model i'm trying to use

[hasty iris]

it sounds like peter griffin instead of butcher

[simple ore]

you probably want to activate another model in the UI

[fringe crag]

I walked away from my PC right now, but in win for lm studio there are indeed options to choose how much my GPU should be used... It is set at 100%

But still, not even a whiff

[simple ore]

make sure it is selected

[hasty iris]

so right now im trying to use travis scott

input device is my mic, output is VAC

maybe its the audio driver issue?

[simple ore]

no

audio driver does not change the voice

wokada has model slots

you add a model and it takes a slot

[hasty iris]

yeah

[simple ore]

you can stop the changer, select a different model and start it again

[hasty iris]

i know, right, i just installed the newer version

and has only 1 slot selected

the voice model is RVCv2

maybe its the cause?

[simple ore]

you may have just selected a wrong model

replace it and pick the right one

[hasty iris]

let me try real quick, i'll let you know what happens next

[simple ore]

when you're trying to read War and Peace and instead you're reading Pinoccio, it does ot mean your glasses are bad, it means someone swapped the covers of the books

be logical

[balmy sluice]

can someone help me with RVC voice models? All I can find here is weird fictional characters

[hasty iris]

one question

how can i use TTS

like peter griffin tts

or make the model sing a song

[viral mason]

Tts or singing can be done easier on applio which is separate from the voice changer

-rvc

[patent trellis]

📚 RVC Documentations

AI HUB Docs

https://docs.aihub.gg

🍏 Applio Docs

https://docs.applio.org

[viral mason]

Why would you need a real person voice?

Most fictional characters have normal voices like my Poppy model

[fringe crag]

@simple ore this is my Linux setting:

[viral mason]

I've never seen a Linux computer before, that's interesting

[fringe crag]

i am autistic, so most of the time i don't get if people are serious or joking? :/ You really haven't messed about with linux?

[viral mason]

I'm autistic as well, and I'm being serious I've only used windows growing up and it's cool seeing a different operating system

[fringe crag]

i really like linux (i have a dual boot win11/Bazzite), but only use win for the few games that not run in windows...

linux helped me reviving a lot of old laptops

however, i am totally not tech intelligent, like i don't understand not even a tenth of what i should to use linux

[viral mason]

That's fair, I really only use Windows for playing games via steam and as well as it being the most commonly used is for projects like rvc or other stuff

[fringe crag]

rvc?

[plain cloud]

Hello there fellow clankers and clankettes

I am in dire need of assistance

[viral mason]

It stands for retrieval-based voice conversion
Basically ai voice cloning as well as voice changer software

[viral mason]

What with?

[plain cloud]

So long story short I wanted a voice changer and Claude code brought me here for the rvc and apploi and w-okawada and stuff

Now I did try rvc it wasn’t anything spectacular might be that the ui and all were done by Claude so nothing much to change to adjust the voice

It then suggested me to delete it and the downloaded a fork which then it said to delete and get appolio and w/okawada

[viral mason]

Alrighty so applio is separate from okada, as applio mainly trains models and does ai covers but can be used for real-time but okada/Vonovox does real-time as its main thing

[plain cloud]

What distro is that ?

[viral mason]

What's the main thing you're lookin for?

[fringe crag]

bazzite

[plain cloud]

Oh cool

[plain cloud]

I want real time voice changing for vrc and Roblox

🫪🫪

[viral mason]

And very important question what is your GPU (Nvidia or AMD) also are u using it for sounding like a cartoon character or something like Darth Vader or Optimus prime ect

[plain cloud]

It first did suggest me w okawada but it was in Japanese so I scraped that

[fringe crag]

talking about voices, do you guys know a really good site/app to put a whole pdf book into an audiobook?

[hasty iris]

thanks a lot

[plain cloud]

Nvidia I want to sound like Sunday he sounds like a choir boy who is 28

[plain cloud]

You could always just tts a pdf no ?

[fringe crag]

i tried the build in app from Edge, in windows, but they read everything, like the page numbers, the repeated title at the top of the pages,...

[fringe crag]

that is possible, but i want to know if it possible to do whole books

[plain cloud]

Heard there is a better ai for singing the hmm what was its name ace xl something ?

[plain cloud]

True true btw might I ask what brings you to this server ?

[viral mason]

Never heard of it, I only trust things I know are commonly used here us unless it can use rvc models it's useless to me

[plain cloud]

Gotcha mb

[viral mason]

Do you know exactly what Nvidia gpu you're using? Like RTX 3070 for example

It'll better help me pick out which of the two voice changers would work better

[plain cloud]

Rtx 5070ti ryzen 9700x and 64 GB 6000mhz cl30 ram

[viral mason]

Yooo I have the same setup

W

[plain cloud]

W

[viral mason]

One second while I get the links

[plain cloud]

K gotcha

[fringe crag]

i want to learn more about AI, and experiment with it. Like i finally succeeded in running a local LLM (however, i don't think they are as good (yet?) as the online ones)

[plain cloud]

True I mean the online ai are whole corporates with billions in hardware it’s honestly astonishing for me that we can even run some llms in our own hardware

[viral mason]

Here's the two downloads you'll need, I recommend installing the second one first as it's required so ppl can hear u in game with the voice changer
https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b779b/Vonovox_beta_17_11.zip

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

I gotta piss brb

[plain cloud]

Nature call lmao

[plain cloud]

Could you like umm tell me what is what and do I need to remove anything also have you used it ?

[fringe crag]

but for my studies, i really find it more helpful to create audio files, but instead of creating many small files, it would be nice to be able to create one large

[viral mason]

I'd delete any older okada software tbh so it doesn't interfere
First link is Vonovox which is the voice changer
Second link is a virtual audio cable called vac lite which connects the voice changer to games or discord ect

[plain cloud]

I think notebook LLM does that I’m not sure tho because I don’t study :p

[plain cloud]

I think I have rvc and appolio

And virtual cable and vb cable

And voice mod

[plain cloud]

Ah gotcha I already have the second link done and dusted now this vonovox thing is it a fork of w okawada

[plain cloud]

You here to do any projects ?

[viral mason]

U can keep those last two and applio

[fringe crag]

not yet

[viral mason]

Voicemod can be helpful with adding voice effects to models that may need them to sound better

Tho I use fl studio for that

[viral mason]

It's basically okada but at its best

[plain cloud]

I was thinking of doing an automated YouTube short uploader and creating anything but now that YouTube has changed its policies and algo I am switching to ph

[plain cloud]

My my thanks imma set it up and let ya know

[plain cloud]

You use it for sound board ?

[viral mason]

For voice effects on models live in vrchat, as well as for sound design

[fringe crag]

i have an idea about a (story based) game that i want to create, and i would need AI to interact within certain parameters

[viral mason]

Only use voicemod for the soundboard lol

[plain cloud]

Uh you mean like how the models speak and stuff when they join the world tbh I just got my pc a week ago and new to pc vrc

[plain cloud]

Gotcha

So I do t need fl thing rn

Btw you do assets in vrc ?

My my I yap a lot

[viral mason]

Not at all sadly, just into sound design like making creature noises for games ect or make a custom effect on a guy's voice to make it sound like he's speaking over the radio

That also goes hand in hand sorta with making voice models

[plain cloud]

can i add you

[viral mason]

Of course

[viral mason]

Nah not unless you want the complicated setup I use

[balmy basalt]

hihi lowkey i wanna create a rvc myself, but idk where to start, beginner here, i'm used to asking people but now i wanna make my own

[plain cloud]

imma first make things work then go to the big legues

[plain cloud]

like training your own model ?

[balmy basalt]

Yeah

[plain cloud]

mind if i ask what you are trying to train it for

[balmy basalt]

for tts or vtv

vtv is what i'm aiming for

[plain cloud]

umm whats vtv ?

[balmy basalt]

voice to voice

[plain cloud]

i mean whose voice mb i should have framed it better

[viral mason]

You got it!

[viral mason]

Ye like the voice changer?

[viral mason]

If you wanna learn how to train a model I got u