✨│ai-help

[low shard]

colab no, but kaggle is more suggested

• Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
• Applio (ui)
• Mainline (UI)
• RVCDISCONNECTED (no ui)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
• Mainline (UI)

EVERYONE USING ILARIA RVC MAINLINE, CHECK #📰│dev-updates message

[tribal spade]

i get this error with the hugging face ai cover gen, is there any fix?

[tame mica]

use ilaria rvc instead

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

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

[low shard]

-hf

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by thestingerx Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces

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

[low shard]

seems like aicovergen zerogpu is broken too

[tame mica]

[low shard]

-hf

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by thestingerx Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces

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

[tame mica]

damn removed the aicovergen link

das crazy

[mighty fox]

Is "RVC V2 Disconnected" Colab still working?

I've tried several times, but I still encounter this error in Training cell:
"File "/content/Mangio-RVC-Fork/train/utils.py", line 206, in latest_checkpoint_path
x = f_list[-1]
IndexError: list index out of range"

I used a custom pretrained model, and I'm not sure if it has any impact. Can someone help me, pls?

[odd shale]

personally, i wouldn't suggest you using custom pretrains.

Various of these got issues that OG pretrain doesn't have

[mighty fox]

Thank you. I'm looking for a non-English model to improve accuracy, at least that's what I think.
Maybe I should try again with OG.

[odd shale]

Yep, go ahead and try.

[jade vault]

What is npy file and where can I find it?

[lilac beacon]

can someone send me the best settings for the voicechanger client demo?

[dusk breach]

Ihave RTX 3060 ti
what best voice changer version ishould used

[odd shale]

Use Deiteris' W-Okada fork.

-rt

[azure marsh]

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

[odd shale]

Check the first guide.

[odd shale]

.npy files are unnecessary for using RVC models

[jade vault]

[Voice Changer] Model Load Error! Check your model is valid. local variable 'embedder' referenced before assignment
Generated Strengths: for prev:(1024,), for cur:(1024,)
[Voice Changer] wait for loading model... -1 -1
[Voice Changer] [Exception] 'No model for model_common loaded. Please confirm the model uploaded.'
[Voice Changer] wait for loading model... -1 -1
[Voice Changer] [Exception] 'No model for model_common loaded. Please confirm the model uploaded.'
[Voice Changer] wait for loading model... -1 -1

[Voice Changer] Loading index...
Traceback (most recent call last):
File "voice_changer\VoiceChanger.py", line 301, in loadModel
File "voice_changer\RVC\RVC.py", line 84, in loadModel
File "voice_changer\RVC\RVC.py", line 153, in prepareModel
File "voice_changer\RVC\pipeline\PipelineGenerator.py", line 53, in createPipeline
UnboundLocalError: local variable 'embedder' referenced before assignment

why is it happening

[odd shale]

Which W-Okada version are you using?

Are you using RVC Realtime?

[jade vault]

v.1.5.2.9e

[rare goblet]

Ayo? @jade vault level 1 !!!

[jade vault]

yes

[odd shale]

I mean, are you using the voice changer built in on RVC?

[knotty moth]

I'd recommend this https://rentry.co/forkvoicechangerguide

[odd shale]

Exactly.

[jade vault]

thanks

[brisk quarry]

what is "best" webui

[oak widget]

Is there a way i can make someone of my choice or something sing an AI song only using a few audio clips of them?
I'm new to AI so any basic instructions will do

[brisk quarry]

Idk

[oak widget]

i don't know any good ways to train models

[weary fable]

This clearly was because Jameskii doesn't want his voice converted into AI.. A friend of his told me to remove it back on the old discord server or whatever cause why not.

[brisk quarry]

I do but i cant explain

[oak widget]

if anyone else has basic information do let me know

[brisk quarry]

dead ahhh server

[tame mica]

bro responded to a message from almost a year ago

[low shard]

i mean, the og pretrain is trained only on english

but its kinda weird u got that error ngl

[glossy raven]

error: /path/to/lxml/etree.so: No such file or directory
can anyone help with that?..trying to install locally

[brisk quarry]

Idl

Idk

[sudden lodge]

where i can download voice changer?

[low shard]

whats ur pc gpu?

[sudden lodge]

NVIDIA GeForce GTX 1080 Ti

[soft agate]

Any good Realtime AI voice changers?

[rare goblet]

Ayo? @soft agate level 1 !!!

[brisk quarry]

Idk

Voicemod

[brittle wing]

What's the fastest non-local model maker

idrc about the ui

[brisk quarry]

cmd

[elder hazel]

Hi there i need help getting rvc software on linux to my steam deck running steam os desktop mode

[brisk quarry]

NO

not on steam deck

NOOOOO

[copper sedge]

bad

[brisk quarry]

cry

[glossy raven]

hi has anyone had success installing RVC locally on an M3 pro Macbook? My old Intel one works fine but it just doesn't seem to wanna work with this new macbook

[low shard]

i'd suggest you to do it on cloud, as on mac you can only inference (use models) with mac, but cant train

[vale osprey]

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[low shard]

so u mean cloud

[low shard]

As you dont got a good PC, its better you use cloud for training an RVC Voice Model:

• Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
• Applio (ui)
• Mainline (UI)
• RVCDISCONNECTED (no ui)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
• Mainline (UI)

the best of the best would be kaggle, but its the hardest to use

[glossy raven]

i dont need to train right now tbh..plus I don't like using Colab..just wanna have it on my computer

[low shard]

it would be faster if you use Ilaria RVC Zero as its a ZeroGPU (A100 11 times faster than T4 of google colab) but alr, i dunno much about mac and there is no rvc mac guide in our docs, but maybe u could do it if u install python and follow the guide in https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/docs/en/README.en.md#preparing-the-environment

Maybe @proper shale knows more as a mac user

[glossy raven]

Thank you. Yeah there was a tutorial doc on here last year for Mac but it’s disappeared. Also I’m reading that the difference between the intel chips and the M series chips behave totally different

[still stone]

With Applio can I set it up so it generates a queue of quotes with a specific tts?

[odd shale]

Yep, i think so.

Applio has tts with a variety of languages anyway.

[still stone]

....Yeah i mean like

Is there a way I can set up like "Quote A, Quote B, Quote C" and get 3 seperate .wav's or am i gonna have to generate all 3 seperately?

[odd shale]

Ehh.. You'll must probably generate all 3 separately.

[still stone]

right thats what i needed to know

idk might see if i can convert rvc to bark and then see if bark can do that

[low shard]

convert rvc to bark ?

[still stone]

cant bark not load rvc files?

[low shard]

Applio TTS works by:

1. Generating a TTS output file via Microsoft Edge TTS API (which is good and mulitlingual but not emotional)
2. Using that tts output file as an input with the rvc model

[still stone]

.....I know that!

[low shard]

wdym with 'files'?

like the models?

[still stone]

yes the voice(.pth) and index(.index) files

[low shard]

i dont think thats possible with bark afaik

its 2 different programs

you can generate a tts with bark tts and use it as an input in rvc though

[still stone]

ugh yknow what

[still stone]

*Yes, Bark can load RVC models without the need to convert them for Bark. Extract the Model files to data/models/rvc_models in it's own folder using rsxdalv's TTS Generation WebUI to use it with Bark"

next time im gonna ask this channel for help, im gonna buy a rubber duck instead

[hot sierra]

Hi, i haven't been using rvc for like 1 year and as far as i can see, there is a new thing called pretrains. im planning to train models but how am i going to use pretrains? are there any guides for it?

[rare goblet]

Ayo? @viral island level 1 !!!

[viral island]

why do i have a robotic voice for no reason on every voice i have

I need a fix

i nneeeed ittt

[low shard]

yea, they mostly help for other languages as the og pretrain is english only https://docs.google.com/document/d/1j9J8A8Oop9bMOHmCs3jDXzPujuD6TQ0Q396rJ0MyuIc/edit

[low shard]

what? rsxdalv?

[vivid pewter]

-train

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[rare goblet]

Ayo? @vivid pewter level 2 !!!

[still stone]

https://github.com/rsxdalv/tts-generation-webui

load rvc -> send to bark

[rare goblet]

Ayo? @still stone level 3 !!!

[opaque glade]

i've got a question, is Ilaria_RVC downloadable or is it only accessable on huggingface

[low shard]

if it still just makes a tts audio with bark and uses that as an input for rvc model, that still means bark cant load the rvc model natively though

[still stone]

is it easier to put a 4 minutes audio into bark? get a voice that way?

[low shard]

the ilaria rvc on huggingface is the zero version, designed only to work with Hugging Face ZeroGPU (A100), so no, you can try downloading Ilaria RVC Mainline but there is no guide for that though

[deep pulsar]

where is colab to create model

[opaque glade]

i attempted to do this but i got stuck in the middle of downloading because of an error that isn't on my end

let me demonstrate really quick

[low shard]

• Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
• Applio (ui)
• Mainline (UI)
• RVCDISCONNECTED (no ui)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
• Mainline (UI)

[opaque glade]

oops

[rare goblet]

Ayo? @opaque glade level 1 !!!

[opaque glade]

it is infact not an error on either side, the script is linux based

how to get download aria2 binaries for Windows ?

[shadow sparrow]

anyone know why refreshing in the applio colab, it says "connection timed out"?

[opaque glade]

nevermind i installed everything, i just need a brief explaination on how to run it now

[opaque glade]

holy shit all im going to say is

please update which pythons and pips are being used becasue this is torture

im seeeing errors i have never witnessed in my life

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

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

💻 (Local) Realtime Voice Changer

• Guide for deiteris' modified W-Okada's Voice Changer Client (Fork)
• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[heady imp]

hi

i need help with this error code

operands could not be broadcast together with shapes (234000,2) (305280,2)

[random island]

did you manage to fix this? i tried installing this earlier and got same errors

[opaque glade]

i have

let me cook

im still cooking rn

[random island]

mimd sharing how you fixed it later 🙂

[opaque glade]

ofcourse

the issue at hand is that everything has been outdated since the code has been touched in 2022

so its a whole mess, u need to convert ur laptop into a time machine back to 2022 in order to get everything running correctly

[random island]

lmao

[rare goblet]

Ayo? @random island level 1 !!!

[opaque glade]

i hope admins can pin the soloution to this issue

[rare goblet]

Ayo? @opaque glade level 2 !!!

[opaque glade]

i will explain it only once

the installation requirements evolve around omegaconf and fairseq, the two intertwine, u cannot be using an newer version of omegaconf, only the version 2.0.6 works, here is where the series of problems begin

1. U cannot install omega through pip, continuing to do so will only result in constant errors and failures, the only way to get the old version is through the Github MANUALLY

https://github.com/omry/omegaconf/releases/tag/v2.0.6

add it to your folder where you are installating RVC mainline

then, you will ALSO need to manually redo the script, omegaconf is broken

here are all the changes you need to do:

1. setup.cfg:

[aliases]
test=pytest

[mypy]
python_version = 3.7
mypy_path=.stubs

2. setup.py:

and finally inside of the /requirements folder there is base.txt:

PyYAML>=5.1
# Use dataclasses backport for Python 3.6.
dataclasses; python_version<'3.7'
typing-extensions

once you are finished with all of these steps, run the following commands in order:

python setup.py sdist bdist_wheel
pip install dist/omegaconf-2.0.6-py3-none-any.whl
pip show omegaconf

[opaque glade]

IF it has been installed correctly it should show you:

Name: omegaconf
Version: 2.0.6
Summary: A flexible configuration library```

congratulations, you have successfully rebuild omegaconf

now you can hop on the second part of the equation, fairseq, luckily for us only one version has ever been released since 2022

pip install fairseq==0.12.2

and you're done

now im trying to figure out how to run this shit

its 4 am im so fucking lifeless

ill be back

[knotty moth]

I see no point of trying to do local port it, even the broken colab one

[opaque glade]

i think u might be right afterall

[rare goblet]

Ayo? @opaque glade level 3 !!!

[knotty moth]

my thought is that if you encounter sort of package conflict issue on the colab notebooks, try adding !pip install -U pip==23.1, idk about the rest tho

[opaque glade]

guys there is a whole fucking new chapter i just discovered just to make this run

ill keep u posted

yall istg this is hilarious ill keep u posted

everytime i try to run run.sh, i run into the following error for each individual module:

ModuleNotFoundError: No module named 'librosa' (example)

so i had to pip install each missing module one by one because i dont even know where they're located at, until i reached faiss

pip install nox pytest
pip install librosa
pip install torch
pip install soundfile
pip install numpy
pip install requests
pip install av

i dug as deep as i could to find faiss, github doesn't work

the original website that hosted the download for faiss shows this

[spare pivot]

does anyone know how to use colab I download the zip files

[sleek grail]

did u ever figure it out

[shadow sparrow]

i wish, im waiting till someone explains why its doing that

[opaque glade]

guys should i use cpu oriented faiss

or GPU oriented

i dont know which one is better

can someone help me out by experience, which one works best for them

[random island]

oh ya im running into this now XD

[opaque glade]

I finished it all but right now I don’t know which version of the faiss module do I use

I’m clueless

It’s either GPU or CPU based

I’m going to sleep GN

[lethal basin]

@acoustic scarab felt generous https://discord.com/billing/promotions/5fV2-GxZ3-mG7e-9MVz-tuGe-vYCZ

[nocturne mural]

it looks like the applio colab broke with version 3.2.5, how i hate you pydantic

🐈

[shadow sparrow]

bruh thats fucked

how long till its fixed?

[red stirrup]

Hi

Can someone help me

I have a audio file, but idk how to make it a voice

[nocturne mural]

f5

restart the env

[sleek grail]

how

[nocturne mural]

in other words, reset the colab environment and refresh the page, the problem was fixed 2 minutes after I found out about it.

[dusty sparrow]

so applio on colab is broken right now right? I can't DL any models or upload any audio.

[nocturne mural]

yes and no, I've fixed it, just disconnect the environment if you've already run it, then refresh the page and you're done.

LikeCat

I don't know if you understand me because I don't speak English.

[dusty sparrow]

I will try again right now! Thank you

[shadow sparrow]

i can understand you, hopefully it works

[dusty sparrow]

It worked! Thank you so much.

[red stirrup]

Dose anyone know how to make a model?

[opaque glade]

@nocturne mural I’ve got a question I hope u have an answer for it

Which preforms better to get tasks done

GPU or CPU

more specifically for the faiss module for rvc main

[brisk quarry]

Gpu @opaque glade

[mighty fox]

Actually I read your instructions about RVC V2 pretrained, I chose Rigel model and then I encountered the error. Maybe it's because the size of custom model is much larger than og model, I tried with og, it worked but that was not what I want :v

[rare goblet]

Ayo? @mighty fox level 1 !!!

[knotty moth]

Rigel is 32k, so you must choose that sample rate

[nocturne mural]

colab broken again and fixed ( refresh page )

precompiled fixed

[lusty garnet]

So if i play it through my desktop it sounds fine but if i try to use it in discord or in game Via vb autio virtual cable

its crackling and breaking up offten, is there any way i can possibly fix this?i have 2 graphics cards in my pc and i have the VC changer on its on GPU 1080TI

[hearty quail]

I know this may be unrelated but I don't know where else to ask is pi torch compatible with python 3.12

[rare goblet]

Ayo? @hearty quail level 1 !!!

[hearty quail]

Because I see things saying it is and also 3.8

[golden karma]

what's the best model for removing harmonies

[brisk quarry]

get rtx

[deft river]

hi
who understand in Midjourney?

[limpid sail]

Where can i find the beatzforge PTH G and D files ?

[tranquil blaze]

For some reason the file manager for applio on kaggle doesn't work anymore, I'm on the latest version of the notebook by Deiant (I think it's Vidal on here)

[old ice]

i downloaded rvc and how do i open it

the exe file wont do anything

[hearty quail]

I downloaded Cuda and installed it when I do the command to see if it's available it comes back as false help please

[odd shale]

You can try with Mel Karaoke or UVR BVE V2

[hot lagoon]

Where can I find mel roformer on uvr?

[marsh stag]

I did a ‏song savage by aespa and the sound is really messed up what do I do to fix it?

[odd shale]

Probably the model you used wasn't that good or you used an unproperly cleaned acapella.

[limpid sail]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/usr/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap
self.run()
File "/usr/lib/python3.10/multiprocessing/process.py", line 108, in run
self._target(*self._args, **self._kwargs)
File "/content/Mangio-RVC-Fork/train_nsf_sim_cache_sid_load_pretrain.py", line 230, in run
torch.load(hps.pretrainG, map_location="cpu")["model"]
File "/usr/local/lib/python3.10/dist-packages/torch/serialization.py", line 1114, in load
return _legacy_load(
File "/usr/local/lib/python3.10/dist-packages/torch/serialization.py", line 1338, in _legacy_load
magic_number = pickle_module.load(f, **pickle_load_args)
how do i fix this error @low shard

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

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

[odd shale]

It's not necessary to ping helpers.

[limpid sail]

sorry

[rare goblet]

Ayo? @limpid sail level 1 !!!

[limpid sail]

i was trying to fix a error

[odd shale]

Also, just use Applio Colab.

-applio

[azure marsh]

🖥️ Local stuff

• 🍏 Applio, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub
• GPT-SoVITS, by RVC-Boss GitHub

Credits to Faze Masta and Antasma for compiling these links.

More commands: -audio, -colabs, -overtrain, -realtime, -rvc, -help

[limpid sail]

where can i use the web ui ver?

[odd shale]

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

This colab version has Web ui.

[hot lagoon]

?

[odd shale]

Kimberley Jensen's Mel Roformer?

I think you must add it yourself manually to UVR.

But i don't have the download link to the model sadly.

[nocturne marten]

How would one go about making voice lines (using RVC for non-music purposes?)

https://drive.google.com/file/d/11D5wzFR0jz6p7Pws5irdXtq3cN4kXkyh I have this model of cave johnson from portal 2 i wanna use

like is there a hugging face space where you can just paste the link and upload the source audio file?

[dawn berry]

how big is the zip file for RVC?

[river eagle]

@odd shale

can u explain to me what epchos is

[odd shale]

Well, you can use Ilaria RVC Zero space on Huggingface.

[nocturne marten]

Thank you!

[rare goblet]

Ayo? @nocturne marten level 1 !!!

[odd shale]

In few words is this but replace "machine learning" with "model"

[river eagle]

thanks

[rare goblet]

Ayo? @river eagle level 1 !!!

[old ice]

i downloaded rvc and how do i open it
the exe file wont do anything

[odd shale]

Do you wanna use RVC for testing models or training models?

[old ice]

testing or using them

[odd shale]

You can use Applio Colab then.

[old ice]

oh

[odd shale]

Hola Lyery.

[old ice]

the link?

[analog obsidian]

or even simpler words: a snapshot of the model

[odd shale]

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

There you go.

[old ice]

im just gonna try clicking random buttons until im stuck

[old ice]

the download is done but now what do i do

[odd shale]

Have you finished executing the Install Applio cell?

[old ice]

wheres that agaiin

[old ice]

click the start one?

[odd shale]

No.

First, run the install applio one

[old ice]

i did

[odd shale]

It finished downloading?

[old ice]

i just said that-

[old ice]

this

[odd shale]

Alright then run the start applio one

And wait till a "Public URL" appears

And you click on it

[old ice]

alr

[random island]

anyone know where do i add custom models for rvc web ui for model inference

[odd shale]

Just use Applio Colab, it has a download tab where you can add your models with just an Huggingface/Drive/Pixeldrain link

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

[random island]

i thought colab doesnt work with the web ui anymore or is this different

[odd shale]

That one is encripted, so dw.

[odd shale]

Delete the spaces from the dataset name.

Also change the dataset's folder name deleting the spaces

[tranquil blaze]

Yeah, I think the notebook is broken, I tried multiple times but nothing

[rare goblet]

Ayo? @tranquil blaze level 4 !!!

[tame mica]

do you use Firefox

perchance

[tranquil blaze]

Nope, I use edge, so it should be fine since it's still chromium

[brittle wing]

Hi does anyone have a different good rvc client other than w odaka? It sounds alright when i listen to it myself, but when im using it on discord or whatsapp it cuts alot or echoes

[knotty moth]

if you think the ngrok is rate limited, delete the ngrok account then recreate it (you have to put the new token tho)

[tranquil blaze]

Idk, the ngrok is working fine for the applio ui and for the tensorboard, it doesn't work only with the file manager

And I still have like half a gig of bandwidth, so I should be fine from that perspective too

[azure marsh]

📚 Audio Guides

• Creating Datasets for RVC using iZotope RX11, by Cauthess
• Gathering and Isolating Audio, by SCRFilms :snowflake:
• Vocal Mixing Tutorial, by Roomie

🛠️ Tools

Audio Separation/Isolation

• https://mvsep.com/
• https://multimedia.easeus.com/vocal-remover/
• MDX-NET Colab

More commands: -colab, -uvr, -karafan, -overtrain, -help

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

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

[tranquil blaze]

Before you started training you should've enabled "save only the latest checkpoints", on top of the epochs it's saving the discriminator and the generator every time they update, and you don't need those

[rare goblet]

Ayo? @modern marlin level 4 !!!

[tranquil blaze]

I think you should've enabled this

And disabled this

*this

• 15 as a frequency is good
• you can bump up the epochs to like 500 or 1000, since you'll have to stop the training manually anyway by looking at the tensorboard and seeing if it's overtraining
• the first yes/no choice should be a "yes"
• if your dataset is less than 10 minutes you should put "yes" to the second choice as well
• the third one is fine as it is

I'm not sure it'll make any difference, I think you can leave it unchecked

No problem :)

[opaque glade]

mainstream is outdated and cannot be used guys,

i've spend

well over 24 hours

trying to make it work, im stuck at loops that have no soloution

conflicts because the modules used are outdated

and a big majority of the requirements's hosts are no longer even up, the modules are long lost

[low shard]

what rvc are u using

[low shard]

both applio and ilaria rvc zero (which is faster for infrence) have a tab where they can upload the model files btw

[random island]

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI

currently have it locally

[low shard]

you are doing it locally, whats ur pc gpu?

[random island]

do you need a gpu for model inference?

using i7 but rvc cant find so it said it was using cpu

[low shard]

well, you can do it on cpu too but its slow compared to a gpu

[random island]

i just assumed gpu was inly mandatory for trainjng

[quick jungle]

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

[random island]

i dont mind it being slow i just cant load models for some reason

[rare goblet]

Ayo? @random island level 2 !!!

[low shard]

if u want u can do it locally on cpu, but i would personally suggest you to use cloud methods like ilaria rvc zero (zerogpu hugging face space which runs on A100 gpu)

i shared you both links to do it

[low shard]

send the model link

[random island]

if i use cloud can i expose it to an api

kinda new to cloud

[low shard]

technically, but zerogpu spaces have a quota that recharges around every 2 hours, and google colabs work around for 4 hours a day with gpu

idk why u would want to use as an api though

[random island]

im trying to integrate the voice conversion into a project im working on

and need to save the output using code

[low shard]

free cloud ways have limits, on local u wont have limits except slow speed

anyways, u cant load the model? whats the model download link?

[random island]

oh sorry

i meant i have no idea where to put the model

for it to load on model inference web ui

[low shard]

i dont really do it locally but this should tell u https://docs.aihub.wtf/rvc/local/mainline/#1-upload-voice-model

[random island]

ah that was what i was looking for ty

[low shard]

yw

[lilac beacon]

Best quality Woman Voice? Someone have a link

[odd shale]

Nope.

But you can check the #1175430844685484042 channel

[lilac beacon]

I've already tested them all, but my colleagues say that they all sound very bad

[odd shale]

These are models made by random people, it's predictable various won't work properly on W-Okada.

[old ice]

how do you make a model

[lilac beacon]

If I bought one would it be better?

[odd shale]

Well, of course you can comission any model master on the #1191429836321849435 to make a model for you.

Nope.

[hushed spear]

whenever i try to download a model or search a model or anything on applio i get connection errored out

[lilac beacon]

I have a 4070ti super nvidia graphics card. Do you perhaps have good settings?

[odd shale]

Good settings on W-Okada?

[rare goblet]

Ayo? @lilac beacon level 2 !!!

[lilac beacon]

for this

This is also possible if you can tell me how I can add models there

[hushed spear]

when i use applio and search for a model or try to upload files i get "connection errored out", any help?

[rare goblet]

Ayo? @hushed spear level 1 !!!

[lilac beacon]

which of this both is wokada

[knotty moth]

I'd recommend the fork wokada https://rentry.co/forkvoicechangerguide

[lilac beacon]

can I also upload rvc models?

[hushed spear]

does anyone have any idea about my problem

[odd shale]

Yes

[hushed spear]

any help please

[astral perch]

My g/mel g/kl are going down but my g/total is going up. Does that mean its overtraining?

[knotty moth]

https://tenor.com/view/no-one-cares-nobody-cares-idc-i-dont-care-who-gives-a-shit-gif-16642477

[knotty moth]

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[tranquil blaze]

@nocturne mural looks like the kaggle notebook is broken, me and @low shard checked it and it seems like the file manager installation is failing

[lilac beacon]

this settings?

My voice is very strange at the beginning and at the end too

[ornate lava]

Hello everyone! Is there a voice to voice converter from RVC on some collab document?

[foggy ocean]

how to get the realtime voice changer pls ?

[fierce flax]

how do i install ai cover maker

[wintry torrent]

I need help installing aicovergen
im getting this error

[wintry torrent]

i tried so many things like installing the requirements one by one, changing the version of omegaconf to 2.0.5 (i couldnt succesfully do that it said it was corrupted), reinstalled 4 times, ran all the commands in administrator mode

nothing worked

[nocturne mural]

uh

[wintry torrent]

anyone?

[tranquil blaze]

Idk, the ngrok of the file manager was the only one erroring out, and when Nick checked the notebook it said this, hold on

[nocturne mural]

seems to be fixed by uninstalling and reinstalling the dependency, I guess I missed a path.

Fix :

!uv pip uninstall imjoy-elfinder
!uv pip install git+https://github.com/Vidalnt/imjoy-elfinder.git
    ```

well

I hope it will take me a short time to see what I have failed to change.

xd

[tranquil blaze]

Kinda like turning your pc off and back on when it isn't working HAHAHAHAHA

Thank you so much for fixing it :)

[fierce flax]

how do i install ai cover maker

[tranquil blaze]

Oh btw, if you can let me know when you'll figure it out and when you'll update the notebook, for now I'm gonna use the fix you gave me, thank you :)

[wintry torrent]

im getting the exact same error with RVC

@nocturne mural

[nocturne mural]

Maybe because you don't have version 3.9 installed on your device and you are trying to use another one?

[wintry torrent]

i have it

[rare goblet]

Ayo? @wintry torrent level 2 !!!

[nocturne mural]

maybe you have another python installed on your device, but maybe this has nothing to do with it because in the image you have sent it is a dependency conflict.

Also, what repository are you using?

[wintry torrent]

i deleted all other versions

[nocturne mural]

.

[wintry torrent]

im gonna be real wit u i have no idea what that means

[nocturne mural]

I mean how did you get the rvc you are trying to ‘install’?

[wintry torrent]

i git cloned it

i just searched rvc and cloned it

[nocturne mural]

link

[wintry torrent]

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI

idk if it has anything to do with that since i also got the same error with aicovergen

https://github.com/SociallyIneptWeeb/AICoverGen

[nocturne mural]

pip install "pip<24.1"

[leaden latch]

I saw applio has the ability to add custom embedders does anyone know if there are any other note worthy embedders to try out other than hubert and contentvec?

[nocturne mural]

run that in your console and try again to install the dependencies

[nocturne mural]

fixed for next updates

[nocturne mural]

I don't have an answer for you, I'm just saying this in case you think I'm ignoring you

[vale osprey]

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[wintry torrent]

i tried reinstalling python and adding python to path manually
still getting this

Btw can u ping me when u reply i dont get normal notifs

[tranquil blaze]

Thank you so much :)

[nocturne mural]

-m pip

@wintry torrent

[wintry torrent]

oh wait

1s

[nocturne mural]

now install the dependencies again

[wintry torrent]

yeah i did it but i didnt notice it worked

[wintry torrent]

alright

is there a way i can skip a dependency

theres one that is like 3gb and i have to install it every single time i try something

nvm it autoskips it

so far so good everything is working

[wintry torrent]

Now im getting a new error

  ERROR: Failed building wheel for fairseq
  Building wheel for antlr4-python3-runtime (pyproject.toml) ... done
  Created wheel for antlr4-python3-runtime: filename=antlr4_python3_runtime-4.8-py3-none-any.whl size=141217 sha256=92d9187cc31fffcb69896f7163234dee83722d8750ac479221241310efa1e46c
  Stored in directory: c:\users\ahmed\appdata\local\pip\cache\wheels\42\3c\ae\14db087e6018de74810afe32eb6ac890ef9c68ba19b00db97a
Successfully built antlr4-python3-runtime
Failed to build fairseq
ERROR: Could not build wheels for fairseq, which is required to install pyproject.toml-based projectsv```

[rare goblet]

Ayo? @wintry torrent level 3 !!!

[wintry torrent]

sorry im just gonna ping u again because idk if the reply pinged u @nocturne mural

[mild hornet]

Hey guys, can anyone tell how good will w-okada RVC client work on i5-11400H RTX 3050 4GB??

[gusty kestrel]

@hearty hatch, I have found 4 results that match your search!

mrsdog

Kai Cenat

• Uploaded: <t:1693180800:d>
• Server: AI Hub
• Likes: 0
• Lang: English

Epochs

400

Technology

RVC

Algorithm

Crepe

Requested by mbbruh · Sep 7, 2024, 5:21 AM

[lilac mortar]

Mine doesn't want to go, it just keeps giving an "error"

[violet heron]

What are you using?

[lavish escarp]

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by thestingerx Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces

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

[wintry torrent]

what the hell theres a 3040???

[mild hornet]

Bruh, I meant 3050*

[wintry torrent]

that makes sense

i was very confused

[mild hornet]

Lol my bad

So do you know the answer?

@wintry torrent

[rare goblet]

Ayo? @mild hornet level 4 !!!

[wintry torrent]

Idk i used it last year on my 3070 ryzen 7 5800H and it had like a 400ms delay

[mild hornet]

ohh i see

What "extra" did you use?

[wintry torrent]

Idk i dont really remember

it was a long time ago

it worked perfectly fine tho

[mild hornet]

Alright

Although the Beatrice v2 model in the latest alpha version doesn't even require a gpu

[proper shale]

Rough but it will run, just use crepe_tiny as your f0

[low shard]

@warped jungle i pinged u in the right channel, whats ur pc gpu?

[urban turret]

I'm having an issue with Applio inferences. It keeps cutting out or skipping for a split second and adds in blank time into the song. I've tried 3.2.4 and 3.2.5 and I've tried different voice models, they are all having this issue. I've also tried changing/toggling settings in the inference menu. GPU is RTX 3080.

[rare goblet]

Ayo? @urban turret level 1 !!!

[lavish escarp]

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[rare goblet]

Ayo? @lavish escarp level 1 !!!

[knotty moth]

did you enable Split Audio?

[lavish escarp]

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[wintry torrent]

anyone??????

[lavish escarp]

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[urban turret]

yeah

[urban turret]

I also tried disabling it, and it's still doing it.

[hybrid widget]

Where do I download it

[hybrid widget]

@brittle wing More people reccommended fork so I'm gonna download it

[brittle wing]

-realtime

[azure marsh]

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

[hybrid widget]

ty ty

[brittle wing]

What is good epoch recommendation for about 30 minutes of audio on Mangio Fork RVC ?

[low shard]

none, use the tensorboard, check https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/

@brittle wing there's no need to send me the youtube video of the voice in dms, just read the guide that explains you how to use the tensorboard

also, reminder that mangio fork is outdated, i'd personally suggest to use Mainline (original rvc) or applio (updated fork) instead

[wise minnow]

So I can't seem to get the new RVC colab to work

It downloads models and uploads audio but it won't actually convert

Just gives me an error when I try to convert it

not using NGROK

[low shard]

send the colab link

[wise minnow]

https://colab.research.google.com/github/ShiromiyaG/RVC-AI-Cover-Maker-UI/blob/main/assets/RVCAICoverMakerUI.ipynb#scrollTo=R0aXc4R1G8bV

[wise minnow]

its the new one in dev updates

[wise minnow]

ilaria is dead now ig

[humble sleet]

@low shard it just is not loading the google drive file perhaps

[wise minnow]

Kinda mad about that tbh I hate having to move to new colabs all the time

[humble sleet]

its the correct google account

its a .wav file

(inside the zip folder)

[low shard]

also for inference (using models like for ai covers), i'd personally suggest Ilaria RVC Zero which is a ZeroGPU HF SPace that runs on a better gpu than colab

[low shard]

oh the new one, ig ud need to show the error too but i never used iy

[low shard]

no-feature-todo = not feature extracting

is ur dataset format wav?

what program are u using to make the dataset

[low shard]

the old ilaria rvc mangio and ilaria rvc mainline yes (so ilaria rvc on colab), but on huggingface (the faster one) it works

[brisk quarry]

no

[brisk quarry]

all are dead

rvc is dead

[wise minnow]

huh

no shit

[low shard]

i mean they work, but the hype of rvc is ded true

but rvc does work

[humble sleet]

ok i just realized the issue, when i compressed the file, winrar decided to just not save anything inside the zip- my bad thanks anyway

[low shard]

oh lol, yw anyways

[wise minnow]

also btw is it possible to make a PC RVC version that doesn't require Nvidia cars (I use a 10 year old AMD GPU)

There was an early PC RVC program I used to use that worked

Took a while but it worked

[low shard]

applio technically has Zluda support (which is a cuda for amd u'd have to download) but idk how fast it would be

[low shard]

i think it was because it just used cpu so it was slow instead of using ur gpu

[wise minnow]

btw when I upload models to ilaria zero are they permanent or are they temporary?

There's a shit load of models already uploaded on there

[brisk quarry]

do you not see half of the server complaining

also major_brand : isom
minor_version : 512
compatible_brands: isomiso2avc1mp41
encoder : Lavf61.0.100
Stream #0:0: Video: h264 (Main) (avc1 / 0x31637661), yuv420p(tv, bt709, progressive), 3200x1800 [SAR 1:1 DAR 16:9], q=2-31, 2000 kb/s, 0.60 fps, 12288 tbn
Metadata:
encoder : Lavc61.2.100 h264_nvenc
Side data:
cpb: bitrate max/min/avg: 0/0/2000000 buffer size: 4000000 vbv_delay: N/A
Stream #0:1(eng): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 128 kb/s (default)
Metadata:
handler_name : SoundHandler
vendor_id : [0][0][0][0]
encoder : Lavc61.2.100 aac
frame= 2173 fps= 36 q=49.0 size= 138240KiB time=01:00:16.66 bitrate= 313.1kbits/s dup=2 drop=0 speed=59.4x

[humble sleet]

@low shard when resuming, what are you general recommendations, should i do about 5 songs at 500 epoch at a time, then resume and do the same for 5 new songs?

[brisk quarry]

do 300.

500 is too much

[humble sleet]

can i use the same songs to get an more advanced version on the next resume, or if i set it to 300 epoch, but i have to turn off my pc and it autosaves at 200 for example, can I use the exact same songs and would it just pick up where it left off?

[low shard]

i think after a while @proven hill resets the space so they get deleted

[low shard]

wdym

[proven hill]

every half hour its resetted

[obtuse galleon]

i can't get UVR5 UI colab to work, is anyone going thru the same problem?

[rare goblet]

Ayo? @obtuse galleon level 1 !!!

[obtuse galleon]

this is happening and in the UI is going "connection errored"

[humble sleet]

@low shard what voice model do i use in UVR to remove reverb and what do i select (and generally what do you recommend should i do with UVR)

[low shard]

the best models should be here https://docs.aihub.wtf/rvc/resources/vocal-isolation/#best-models

[brisk quarry]

ru dum

just

[low shard]

thats the loca version

[brisk quarry]

and

[low shard]

not everyone has a good pc

[brisk quarry]

then get job

lowkey

[low shard]

most people here are minors

just use cloud

[brisk quarry]

im also a minor and have one

[low shard]

yea and u got a job?

[brisk quarry]

https://tenor.com/view/accident-workplace-workplace-accident-gif-13655676427118804708

yuh

[low shard]

https://tenor.com/view/embed-no-no-embed-megamind-megamind-meme-gif-25261934

[brisk quarry]

u embed it

[low shard]

are u using the T4 GPU?

[obtuse galleon]

thanks lol

[astral perch]

Is a Index file still useful if you trained a model with different languages in the dataset?

[low shard]

yea i dont either

[nimble fractal]

hola alguien que me ayude?

[humble sleet]

thanks exactly what i needed

[errant plume]

guys why am i getting thsi error during conversion

[low shard]

Yw

[errant plume]

@low shard hey bro do u have any idea why its aint working

[low shard]

What's the model download link

[errant plume]

i didnt paste the link just uploaded the pth and index files

[low shard]

Could u share here a sec

Be sure is the added index btw not training index

[errant plume]

the pth and index files?

ok

[low shard]

Yes

[errant plume]

cant upload the index file its too large to upload

[low shard]

what if u try to zip the pth and index would it fit?

[errant plume]

its still too big if i compress the index file alone its like 120mb

[low shard]

what about uploading it in huggingface? is this your first model?

[errant plume]

no i have done a model before but this time its showing some error and not sure why

[knotty moth]

you should try uploading to huggingface and use its direct download link (make sure it's really direct download, not showing 404 page or nonsense)

[low shard]

its better u upload it on hf

and be sure its public so its possible to test

[errant plume]

yh here is the link https://huggingface.co/spaces/TheStinger/Ilaria_RVC

[errant plume]

oh you mean the url thats shown in the rvc disconnected after training?

[low shard]

no, i mean the link of the uploaded model, check https://docs.aihub.wtf/essentials/voice-models/#uploading-to-hugging-face

[errant plume]

uhm i found that but its not showing any option to upload my pth n index file in the upload model

[rare goblet]

Ayo? @errant plume level 4 !!!

[errant plume]

@knotty moth idk how to use the direct link lol

can you explain clearly?

@knotty moth can i dm you regarding this?

[low shard]

are u following the guide i sent? it shouldnt ask for those options

[errant plume]

@low shard yeah im following the guide you sent but i couldnt get to that section where it shows create new model below profile

[low shard]

it doesnt show that button for u?

[radiant bridge]

Uh

[brittle wing]

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[fading lodge]

my rvc keeps having like wierd crackles in my voice, i did the audiodg method and it still didnt work

[shadow sparrow]

does the format shifting on applio just not work? ive been trying to get it to work for the past like 6 minutes now and it doesnt load the audio

[errant plume]

only option is to increase the epoches if the dataset is clean and long

[indigo echo]

Helps why my converter take it so long

Is half 15 minutes

The clips only 10s

[errant plume]

how long is your dataset btw?

[indigo echo]

Where can i check that dataset

[low shard]

the pth depends on the epochs, the index training is different, but yea index depends on the dataset

[errant plume]

oh i see

[low shard]

are u training or making an ai cover?

[indigo echo]

Ai cover

[rare goblet]

Ayo? @indigo echo level 1 !!!

[low shard]

what are u using?

[indigo echo]

I do like the tutorial
Input audio file and model voice character

[low shard]

which tutorial?

there are different versions

if you are using a yt tut, its mostly outdated

[indigo echo]

https://youtu.be/CkHI4ZafF0w?si=cs73o0xdH9GoT59K

5 months ago

[low shard]

i never heard of that site lol

whats your pc gpu?

[indigo echo]

The GPU still 1%

[low shard]

that site is doing it on cloud, its not running on ur pc gpu..

Whats ur pc gpu name

[indigo echo]

Intel HD graphic

And nvdia geforce 820M

[low shard]

u have integrated graphics, which means u can inference (use models for ai covers) which means u can do it only on cpu so its slow

[indigo echo]

Maybe hardware issue but the clips so short

[low shard]

the site is nothing related to ur hardware, it runs on a remote pc, its not local on ur pc

u cant do it on ur pc (locally), its too weak for it

its better u do it on cloud on ilaria rvc zero

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

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

[low shard]

its a zerogpu (A100) huggingface space

it runs on a remote good pc, its the fastest for free on cloud

[indigo echo]

What?
So we can run RVC on cloud

[low shard]

yes.. ur pc is too slow to do it locally, that site u were using was not using ur pc, it was on cloud, so im suggesting u a better option which is faster

Also

You can search rvc ai voice model at:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/ (login required)
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models
• https://voice-models.com/
• https://thevoicemodels.com/ (for Turkish Models, login required with discord and level 2 on their server)

if there isnt one, you can:

• #1159289738314919936
• #1191429836321849435
• make it yourself with our docs guides https://docs.aihub.wtf/essentials/how-to-make-voice-models/

@indigo echo never use https://easyaivoice.com/#pricing, its basically rvc but paid, u can use rvc for free with the ways i told u

and i also sent u other ways to find models

Also about vocal isolation, u can check https://docs.aihub.wtf/rvc/resources/vocal-isolation/, ur pc won't be able to do it locally but u can use mvsep or uvr on cloud, theres also a zerogpu uvr huggingface space https://huggingface.co/spaces/TheStinger/UVR5_UI

but don't follow ever that youtube tut, u can do all those things for free

[grim bay]

Is 2 minutes of sample enough to create a model?

[indigo echo]

@low shard sir I upload the file .pth downloaded from discord to Model Loader but is said error

[rare goblet]

Ayo? @indigo echo level 2 !!!

[grim bay]

So how long is enough?

[indigo echo]

i uploaded the ai voice model file path but is just error

@low shard

Even the model downloader error too

[misty elk]

If i train an index on applio and a pth on mainline would that mess anything up

[indigo echo]

Oh i just fix that big thanks to @low shard

THANKS YOU SO MUCH

[low shard]

huh what happened i was eating

[low shard]

its fixed now after reading the guide?

[indigo echo]

yes jus some problem but it done now

BTW do any option can keep the music that will be perfect

the voice is ok but it always lost the music

[trim snow]

can somebody send me a settings for amd gpu

[half meadow]

I wanted to ask a question, I use RVC1006 (NVIDIA) locally and was wondering if there were better options to use?

[rare goblet]

Ayo? @half meadow level 3 !!!

[edgy marsh]

I need assistance

I don't know how to get my voice to sound like the files I downloaded.

[errant plume]

hey one small doubt.. to resume the training in rvc disconnected should i also upload the dataset that was uploaded in the previous training or can i just upload new data set and input the step number 2333333

[edgy marsh]

and I need guidance on what to use and how.

[errant plume]

@low shard

[edgy marsh]

Can Nick help me?

[errant plume]

yes ig

[errant plume]

you can also read the guide above if you not sure whats the process

[edgy marsh]

Where?

[errant plume]

-guides

[azure marsh]

Suggestions for @edgy marsh

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

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

[edgy marsh]

I'll check it out

[obtuse galleon]

and yes, im using but the problem still continues even with different accounts

[slim owl]

i am bored without making how to make ai voice model

[rare goblet]

Ayo? @slim owl level 1 !!!

[low shard]

Not really, i mean there is also applio that has some extra features like applio tts but its same quality

[low shard]

whats ur pc gpu?

and we use RVC

[errant plume]

@low shard to resume the training in rvc disconnected should i also upload the dataset that was uploaded in the previous training or can i just upload new data set and input the step number 2333333?

[low shard]

wait, do u mean like training your voice into a model, like making a model, or making ur voice sounds better with a model in realtime voice changing

[elder hazel]

@timid olive Hi need help setting up rvc voice changer on my windows 10 pc also i have a amd gpu on my hp laptop and its amd randeon r4 graphics gpu and cpu is amd a6series in a apu so i need help

[low shard]

previous dataset, i dont think u can with a new one

[edgy marsh]

I think im fine for now

I'm practicing

[low shard]

oh alr then

[low shard]

could u send the google colab link u are using?

does it happen at the start?

[low shard]

whats ur pc gpu?

[errant plume]

oh what! But in the tutorial they saif its just okay to input the stepcount without uploading the old dataset again?

[low shard]

ig ur using rvc disconnected, i thought what u meant is that u want to resume training with a whole new dataset lol, well i misread it then, to resume training u just need to be sure that last time u runned Save preprocessed dataset files to Google Drive & Export Model from Notebook to Drive so after u can follow the instructions https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/#instructions

[obtuse galleon]

this one! i tested with three different accounts and same problem https://colab.research.google.com/github/Eddycrack864/UVR5-UI/blob/main/UVR_UI.ipynb?authuser=2#scrollTo=oXYEyNjdWaMm

[low shard]

i mean wouldn't have changed by different accounts unless the gpu ended, lemme check

[low shard]

yea seems broken, for now u can use this zerogpu (A100) hugging face space if u want

[low shard]

@viscid moss the uvr ui colab seems broken

it seems to be doing that on all tabs

[viscid moss]

Does it disconnect when you use the gradio interface?

[low shard]

not disconnect

but throws a long ass error

i can't even send it as a file bc buggy cord

[viscid moss]

hmmm I'll test

and fix

[obtuse galleon]

thanks for the help nick!

[low shard]

alr

[low shard]

yw

[brittle wing]

i created an port of uvr5 ui lol: https://www.kaggle.com/code/ahaghjhj/uvr5-ui-kaggle/edit/run/195854974

tell me any errors you have using it

[brittle wing]

guess, a mistake with pydantic and gradio

[low shard]

ay thats good

[low shard]

bingo

[brittle wing]

https://github.com/Eddycrack864/UVR5-UI/issues/3

just add
pydantic==2.8.2
fastapi==0.112.0
starlette==0.37.2

[low shard]

@viscid moss

[viscid moss]

ya I'm testing that rn. Also updating gradio version and python-audio-separator version. I hope everything works correctly

[brittle wing]

eddy, plss silence the output in colab

[viscid moss]

The installation output?

I kept it because it's easier to find out if there is an error with a dependency.

[brittle wing]

yes xd

i mute the output in kaggle as the page is somewhat sensitive to receiving too many messages in the output.

[brittle wing]

works

[low shard]

🔥

[brittle wing]

fix:

!pip uninstall pydantic -y
!pip install pydantic==2.8.2
!pip install fastapi==0.112.0
!pip install starlette==0.37.2

[solemn sparrow]

Can someone give me the settings for an e-girl voice model and the model?

I just don't know how to adjust the models correctly.

[brittle wing]

high pitch

[solemn sparrow]

yes but it does not sound well

It's like you hear half of what I say and it cuts off

[brittle wing]

realtime or rvc?

[solemn sparrow]

im using rvc

[brittle wing]

adjust it according to your audio

[solemn sparrow]

i just want to have the best e girl model

where is it

[rare goblet]

Ayo? @solemn sparrow level 1 !!!

[brittle wing]

https://discord.com/channels/1159260121998827560/1182654791801917500

[dusky lion]

Hey guys quick question, I am being quota limited using huggingface RVC, i just bought Pro but still have the rate limit, where do i go to increase my quota?

[rare goblet]

Ayo? @dusky lion level 1 !!!

[low shard]

i guess you are talking about ilaria rvc zero, well as of rn you can't see the ZeroGPU quota, but on their ReadME they say

PRO users also get x5 more daily usage quota and highest priority in GPU queues when using any ZeroGPU Spaces.
ZeroGPU quota should reset every 2 hours

[dusky lion]

Okay awesome TY! Can i avoid the limit by duplicating to a private space and paying the per hour rate?

[low shard]

I can't really try but that should work

[quartz vigil]

using local install of Ilaria, I have the .wav file I chose to convert and the .pth model I chose to use. I click convert and I get this error everytime. thoughts? AttributeError: 'NoneType' object has no attribute 'dtype'

[rare goblet]

Ayo? @quartz vigil level 1 !!!

[brittle wing]

-local

[azure marsh]

Suggestions for @quartz vigil

🖥️ Local stuff

• 🍏 Applio, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub
• GPT-SoVITS, by RVC-Boss GitHub

Credits to Faze Masta and Antasma for compiling these links.

More commands: -audio, -colabs, -overtrain, -realtime, -rvc, -help

[brittle wing]

i recommend applio

[quartz vigil]

Thanks, I'll try it out.

[brittle wing]

use the precompiled version (less buggy) https://docs.applio.org/getting-started/installation#hugging-face

fix for gradio (use in case of error)

locate it in the applio folder
folder and run it

[brittle wing]

dont use covergen possibly broken

[quartz vigil]

now i'm getting this error using Applio

used the fix-env.bat as well before launching/converting

[low shard]

ye the zerogpu aicovergen hf space is broken, its removed from @azure marsh already

[azure marsh]

Wassup I'm Automaze!

• My prefix in this server is -
• Interested in how I'm built? I'm actually open source!
• Forgot a specific command? Try /help or -help

[quartz vigil]

damn. is there any other option to from the local list that has a good quality covergen, until it's fixed for applio?

[brittle wing]

-hf

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by thestingerx Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces

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

[low shard]

whats ur pc gpu btw

[quartz vigil]

7900xtx/7800x3d

32gb of ram

24gb vram in the gpu

[low shard]

yea AMD gpu, probably the reason

You could maybe have done it with ilaria rvc mainline with rocm in linux

With applio u should be able to do it in windows if u download Zluda

https://github.com/IAHispano/Applio/pull/513 you can find the readme on how to install it in the files changed section

idk much of it tho

[fading notch]

Nice to meet you. Can someone please teach me how to make an RVC model?

[brittle wing]

-hf

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by thestingerx Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces

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

[grim bay]

Which one is better, extracting the vocal then put it in together or vice versa? I'm trying to make a model from a sample from the Anime

[brittle wing]

Bs roformer

5 hp karaoke

[grim bay]

I mean I have this much sample
Should I extract one by one or combine them?

[brittle wing]

You can leave it in pieces or put it all together.

[grim bay]

Any difference of the result?

[brittle wing]

none

[errant plume]

hey guys can anyone explain me how the resume for rvc training works im s o confused rn

[brittle wing]

what rvc are you using?

What Batch Size should I apply for Ai Singing Training for 30+ mins of Datasets?

[errant plume]

rvc disconnected

[brittle wing]

depends on the vram of your gpu

[brittle wing]

I have an RTX 3090, 24 GB VRAM

[errant plume]

does it mean that only the audio files that was used in previous training should be used? i thought it can recognise it if its already saved in the rvc disconnected folder and just enter the step count after uploading a new dataset

[brittle wing]

put from 15 to 19 so as not to eat all the vram

[errant plume]

and whats this for!?

[brittle wing]

Export your model to drive

[errant plume]

my bad yeah just noticed that.. any commenst about my first qs?

[rare goblet]

Ayo? @errant plume level 5 !!!

[brittle wing]

Alright, Ill aim around 16 batch size. Thank you

[brittle wing]

use t de esser

and rx modules

[brittle wing]

-audio

[azure marsh]

Suggestions for @errant plume

📚 Audio Guides

• Creating Datasets for RVC using iZotope RX11, by Cauthess
• Gathering and Isolating Audio, by SCRFilms :snowflake:
• Vocal Mixing Tutorial, by Roomie

🛠️ Tools

Audio Separation/Isolation

• https://mvsep.com/
• https://multimedia.easeus.com/vocal-remover/
• MDX-NET Colab

More commands: -colab, -uvr, -karafan, -overtrain, -help

[errant plume]

this was my qs

[brittle wing]

just save the las D and G .pth

[errant plume]

so i dont need to upload the dataset that was used in the previous tra ning for resuming?

[errant plume]

there's nothing about it in the guide

[brittle wing]

Im so tired

sorry if i give wrong info, i am very tired after porting uvr5 ui to kaggle.

[errant plume]

ok lol

[errant plume]

can someone help me on this!

[analog obsidian]

8 for better generalization, 16 if you want a fast training
since is a large dataset the larger batch size is not going to degrade your model capabilities
ignore what the other person said, you don't choose batch size based in your vram, you choose it based in what your dataset needs

universally stick with 8 as is the most safe of all
tho in smaller datasets you can do 4
use 16 only in bigger datasets like 30 minutes and more

[brittle wing]

i dont use rvc disconected lol

old fork

[hardy orbit]

question im like brand new to this how do i download a AI voice for like game use?

[brittle wing]

realtime?

[knotty moth]

it's either 8 or 16 anyway (the former is more unstable), VRAM only limits the possibility without causing performance bottleneck

[hardy orbit]

yeah

[brittle wing]

-realtime

[azure marsh]

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

[hardy orbit]

wow!

[brittle wing]

Ok

[brittle wing]

Okay, ill keep that in mind. Thank you

[brittle wing]

sorry dude xd

[brittle wing]

No worries, you're were basically right. 16 batch is what I needed

[simple hill]

i have a question, for models that has a dataset with a time, what does that supposed to mean? how long it can be used for? or is it something else?

[strange hamlet]

i am confused is it possible to download rvc/w-okuda through windows on amd?

[dusky peak]

Can someone please explain briefly what RVC is, what it's for, specifically when using Okada? All the guides I've seen seem to skim over the basics

[surreal magnet]

basically a realtime voice changer, a program that could voice / mask your voice so when set up properly, you can speak like another model / person you wanted. W-Okada is most known tools for that specific task

[midnight fern]

Does anyone know the best ai tts voice

for like characters

[brittle wing]

Just a quick question, is making voicemodels easy and fast? I know its a relative answer but i am just wondering really.

[final sentinel]

Is there any way to create an rvc model on my phone?

[dusky peak]

Thank you. Is it necessary to use RVC if I don't care if the voice changes in real time? Like what if I just want to record my voice into a wav file or something and then have the voice model alter it later?

[knotty moth]

check the guide here:
https://docs.aihub.wtf/essentials/how-to-make-voice-models/
https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/

[knotty moth]

check the guide here:
https://rentry.co/VoiceChangerGuide
https://rentry.co/forkvoicechangerguide (recommended for better optimization)

[dusky peak]

ty

[rare goblet]

Ayo? @dusky peak level 1 !!!

[late juniper]

For training, what’s better extractions? RVMPE or Crepe

[knotty moth]

the former for most cases, though the latter with hop length 8-64 and high quality/clean datasets can also be good for some AI covers

[odd shale]

RMVPE can work for most cases.

But crepe can work for soft voices too.

[hushed steppe]

Inference on collab anyone 💀

[odd shale]

Applio Colab.

[hushed steppe]

All the ones I used do not work no more

[hushed steppe]

Do you have the link, I only have the no ui and that one doesn't work for inference

[odd shale]

There you go lil bud.

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

[hushed steppe]

Ty

[odd shale]

You're welcome bud.

[hushed steppe]

This one has risk of getting you banned right

Since it's with ui

I'll need to be careful

[tame mica]

nope

it's encrypted

[hushed steppe]

Oh damn

That's amazing

Ty

[late juniper]

Is 8 hop length better than 64? I have high quality clean dataset so I assume 8 better yeah?

[distant turtle]

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[low shard]

btw for faster inference i'd suggest u zerogpu huggingface spaces like Ilaria RVC Zero, it has a faster gpu for free than colab (A100 11 times faster than T4)
I mean it's your choice though

[hushed steppe]

My hugging face always tells me I run out

Aa

Or is that a bug?

Cuz it's like "log in for more gpu"

While I have an acc

[low shard]

that's the zerogpu quota, it recharges usually every 2 hour

[low shard]

its to prevent abuse as its free for everyone

[low shard]

nah its just the default message

[hushed steppe]

Ohhh okay

[low shard]

i was just suggesting u as its faster,but ofc if u use it very very much it runs out the zerogpu quota

[hushed steppe]

It's been running out faster these days

I figured this one was the fastest and least laggy https://huggingface.co/spaces/r3gm/rvc_zero

[low shard]

they are both rvc forks so the quality and time should be the same, the time depends on how fast it finds the zerogpu avaible before it starts inferencing as you may have noticed the lil warning at top right when u click the button to start, which depends on zerogpu

@proven hill did u ever compare ilaria rvc zero with the rvc zero of r3gm? like testing same settings, model and audio and compare the time of inferencing between the 2 after it finds the zerogpu

[knotty moth]

the thing is that rmvpe may be better for voice changer

[knotty moth]

no, it is a feature

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

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

💻 (Local) Realtime Voice Changer

• Guide for deiteris' modified W-Okada's Voice Changer Client (Fork)
• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[low shard]

https://tenor.com/view/linus-tech-tips-ltt-bug-gif-19900847

[late juniper]

My intention is not to use voice changer I have 1 hour and 30 minute dataset of someone singing and intend to use it for covers/original songs only so from what I see, crepe is better for that, but I don’t know what the best hop length would be for such

[rare goblet]

Ayo? @late juniper level 1 !!!

[odd shale]

I think you can test by using 64.

Also, using a full hour of audio isn't necessary.

[hushed steppe]

I saw somewhere something about hop length

Hold on

[knotty moth]

looks more like his choice for each of them, I dont see correlation between hop length and the pitch, since the former is temporal

[plush axle]

Is there any documentation on what each slider and setting actually does for aicovergen with any tips/suggestions for dialing it in? So far results have been kind of scuffed.

[knotty moth]

depend on the pitch range, but singing dataset tends to have less variation than talking ones with various intonations

[analog obsidian]

Crepe is less pitch accurate than rmvpe no matter the hop length
Edit: but usually is advised to use either hop 64 or 32 for training, below these numbers there are no audible difference at all

[astral perch]

While you're training a model on kaggle can you turn your pc to sleep mode?

[knotty moth]

#🧬│ai-chat message

[brittle wing]

eddy told me that aicovergen is possibly broken.

[brittle wing]

no

the session is adjourned

[plush axle]

Any suggestions on alternatives?

[rare goblet]

Ayo? @plush axle level 1 !!!

[brittle wing]

applio colab

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[plush axle]

Ah I was looking more so for local. Thanks for this though!

[brittle wing]

-local

[azure marsh]

Suggestions for @plush axle

🖥️ Local stuff

• 🍏 Applio, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub
• GPT-SoVITS, by RVC-Boss GitHub

Credits to Faze Masta and Antasma for compiling these links.

More commands: -audio, -colabs, -overtrain, -realtime, -rvc, -help

[plush axle]

Thanks 👍

[late juniper]

Ok well I don't really believe that because in some cases having a full hour of audio helped some vocal situations compared to half an hour, so the extra audio counts sometimes

[analog obsidian]

the larger the dataset, the better and realistic the model sounds

its basic ml stuff

there's not limit in rvc, you can use many hours/minutes as you want

[odd shale]

Alright then.

[odd shale]

Las papas eso es

🍟

[analog obsidian]

[dull crypt]

guys do u provide help on mangio rvc fork i try clicking on process data on step 2a on the train section and nothing works

[hushed steppe]

So I had this issue and it was because I didn't zip my dataset correctly

So it might be that?

[dull crypt]

i didnt need to zip anything i just put the directory of my mp3 file in the "enter path of the training folder"

i only used .zip file on model inference section but for train section in my youtube video im looking through they just put in mp3 file and it works

[rare goblet]

Ayo? @dull crypt level 1 !!!

[dull crypt]

it shouldnt be

[hushed steppe]

How old is the vid?

Personally it didn't work when I did that

So, get your mp3, go on a site to zip it or use ZArchiver

And put it in there

And see if that works

[dull crypt]

oooohhh

in my cmd its saying the error messages and its saying f = open("%s/preprocess.log" % exp_dir, "a+")
FileNotFoundError: [Errno 2] No such file or directory: 'C:\RVC\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0/logs/kaidou/preprocess.log'

[dull crypt]

not old 9 month i separated vocals, and are you sure you've zipped the mp3 before clicking on process data? but from the error message i just sent the problem seems to be the directory

[hushed steppe]

Yep it's cuz u didn't zip it, I think

[dull crypt]

ill try i guess if it works ur a genius

[hushed steppe]

Well that's what fixed it for me

[dull crypt]

what do u normally do to zip do u use zarchiver or on website

[hushed steppe]

I use ZArchiver

Cuz I can compress a single audio file lol

I always make my dataset a long file instead of multiple small

[dull crypt]

i used winrar and it worked fine i hope that works its a single audio file too

its sthe same no such file or directory error im gonna try figure this out on the github othre ppl had same error and same in yt vid i saw

[hushed steppe]

Hmm

Rvc disconnected works fine

I used it 3 times today to make 3 models

[dull crypt]

bro im so confused why is the error No such file or directory: 'C:\RVC\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0/logs/kaidou/preprocess.log'
im getting here not the same directory i provide im providing my downlaods directory wtf

[hushed steppe]

Honestly try RVC disconnected sobs

[dull crypt]

i dont want to give up i spent too much time watchign the vid 😭

caus i learned the model inference and batch stuff its an easy fix its not a error like thats random its just for some reason giving the wrong directoryy and different from the one i put into the training folder ;/

[rare goblet]

Ayo? @dull crypt level 2 !!!

[hot lagoon]

how do I run the RVC AI Cover Maker on pc?

[odd shale]

Which vid?

[dull crypt]

https://www.youtube.com/watch?v=ixB9oalT3cQ

im still trying

[odd shale]

RVC tutorials on youtube tend to be pretty outdated, for that reason we encourage instead to read the docs.

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

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

[dull crypt]

same error bruh :/

[odd shale]

Don't look for RVC tutorials on youtube.

[dull crypt]

ok i will read te doc

[odd shale]

🐢 👍

[dull crypt]

i jst feel that its such a simple step that there shouldn't be a problem why the error keeps arising but maybe it says why in the doc

i think i foudn it...

[dull crypt]

you are a genius IT WAS BECAUSE I HAD SPACE in the experiment NAME not the directory i thought that fix was only for the directory and the experiment name didnt effect it at all but apparently it does thank u sm this took an hour to fix

[odd shale]

🐢 👍

[azure marsh]

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

[blazing kite]

alguien que hable español y me ayude ?

[brittle wing]

Is Applio Overtrain Detector reliable?

Or just keep tracking with Tensorboard?

[odd shale]

Keep tracking with tensorboard.

i don't recommend using Applio Overtrain Detector since it's kinda innaccurate.

[brittle wing]

Thank you

[odd shale]

You're welcome buddy..

[shadow sparrow]

does anyone know how format shifting works for applio? ive tried using it a few times now and everytime i try to process the audio with format shifting, i dont see an audio to download

[knotty moth]

what error message in the console window?

[shadow sparrow]

should mention im doin this on the google colab

[knotty moth]

for inference, you can instead use local Applio or the online Ilaria RVC
https://docs.google.com/document/d/1YbXcLFPaGjhOdG5NFkK3QrucCEpHZBwFUxkeMO8aB18/edit

[brittle wing]

Mhm how do I make the ai vocals sound better with which Colab?

-colab

[azure marsh]

☁️ Google Colabs

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

ℹ️ Note

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

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

[knotty moth]

they should work the same with some features, just get better models if you think so
anyway Ilaria RVC is the easy way to do with the link already provided above

[brittle wing]

Tbh I think I'm really bad at applying reverb

I have to dm you some files.

[knotty moth]

use some plugins like FL studio's stock reverb, izotope Neoverb, etc

[brittle wing]

For mobile?

I use audioalter website but...

[knotty moth]

there's FL studio mobile

[brittle wing]

Any alternative

I can download pro apk but idk how to save file after processing

[knotty moth]

havent heard any mobile app that can use VSTs like for pc (not mentioning some possible workaround to use Winlator)

[misty elk]

Why is imjoy blank

[urban turret]

Can somebody please help me? I'm having an issue with Applio inferences. It keeps cutting out or skipping for a split second and adds in blank time into the song. I've tried 3.2.4 and 3.2.5 and different voice models. I've also tried changing/toggling settings in the inference menu. My GPU is RTX 3080.

[grave seal]

i cant even send imagea to tell what im having problem with

[proper shale]

Have you tried other RVC forks (or even the original RVC) to see if the problem persists?

[proper shale]

#1192011222023950368

[brittle wing]

Can I dm you some of the files and you tell me if they sound good?

[ionic crag]

while loading the .pth file i got an "Exception in Tkinter callback"

[odd shale]

Which colab are you using? Or are you using RVC locally?

[ionic crag]

RVC GUI locally i think

its downloaded i mean