✨│ai-help

[vagrant marsh]

aah

[low shard]

@viscid moss maybe you got an idea? (sorry for ping but i noticed it was your colab)

do you need wokada (realtime voice changer for calls) or rvc (ai covers and train models) ? ilaria sent that wokada guide here bc also the user asked in the wrong channel

[cedar cobalt]

why does it sound patchy when I'm in a Discord call ?

[heady gorge]

someone tell me how i can make my own rvc model and upload it? i tried looking for a JP Aigami Duel Links voice but they dont seem to have it yet and i wanna learn how because there's a video of all of his JP voicelines on YT found here -> https://youtu.be/PB6vckdb1EA?si=EUZKDUXuR2VhUVl4

[viscid moss]

what were the settings?

That denoise is different from the denoise model

[oblique trout]

how do i do the ai voices?

[subtle phoenix]

any idea why I'm getting this error when trying to train index using the Applio colab (with gui)? 😢

I've already done preprocess and feature extraction.

[low shard]

check https://docs.aihub.wtf, if you got a good pc do it locally else do it on cloud

[low shard]

do what?

• ai cover / use them
• make them
• use them for calls

be specific and also say if you got a good pc

[oblique trout]

nah its aight i figured it out

[subtle phoenix]

I'm trying to finetune a pretrained model to a specific language..
I have gtx 1070, and was using colab paid..

but I keep running into issues using applio in colab, I was using mangio locally before I read that it was abandonded.
edit: my pc has 1070, so I think colab gpu is significantly better..

would you recommend a specific notebook for finetuning in colab?

[rare goblet]

Ayo? @subtle phoenix level 2 !!!

[low shard]

yea gtx 1070 isn't the best for training, you are trying to finetune a pretrain and not train a model right?

[subtle phoenix]

yep.

[low shard]

you got the paid gpu right? cus a pretrain requires alot of resources as seen in the pretrain guide

[subtle phoenix]

yep, I'm using paid

[low shard]

Also, i seen your error before, but technically you don't need an index when your making a pretrain, you need only the G & D files, the index and pth are a thing of models, so unless you wanna use the pretrain as a model instead of a pretrain (for some reason ) you don't need an .index & .pth

[subtle phoenix]

hmm.. yes. but I was afraid something wrong was going on becauses it wasn't able to create an index.
thought I should fix it before consuming resources.. 🤔

[low shard]

Ngl kinda weird, i'd try to restart and disconnect the runtime via the runtime tab and then refresh the page, and restart it all, also be sure that the dataset is ALOT OF HOURS, and that it doesn't contain weird characters in its name

[subtle phoenix]

yep... I can't seem to figure out what the problem is .

I have 14 hours of high quality audio in the target language.

it only contains _,numbers, and english letters (is the underscore problematic?)

one thing I am doing which I'm not sure of is preprocessing and feature extraction on a cpu runtime since these steps don't appear to be using gpu and are taking a looong time.

and I'm backing up using the auto backup inside the colab and also loading using it in the extras section.

[low shard]

i think if you change the colab runtime mid using it, it deletes everything did before

like if you start the colab on CPU, then do everything till feature extraction, then go to colab and change the runtime type to gpu, it should delete everything technically

I don't know if its different for Google Colab Paid tho

but i mean like

[subtle phoenix]

yeah it does but doesn't using the backup save the logs in drive?

[vagrant marsh]

Idk 100%

[vagrant marsh]

🤔

I have to learn how to use this AI denoiser

[low shard]

tbh haven't used that extra section, check in your drive tho and see if everything is backed up, also while doing this be sure that you don't get any 'no feature todo' error while feature extraction

[subtle phoenix]

yeah I'm getting a about 9gb backup in drive.
I'll try again being more cautious and see if it works out.
thx!

[low shard]

Tbh im not really sure if you should be doing the preprocess and feature extraction in CPU, maybe i'd try to do it all on GPU

[subtle phoenix]

yeah, I guess I'll try to do it on gpu I guess goodbye to those credits 😂
I was wondering if there was a better notebook for this sort of stuff but this seems to be the best one I guess...

one thing that is also annoying is that backups take a very long time to bacukup and load, the logs for the model seem to be around 50k files

[vagrant marsh]

Ig the settings were like this:

@viscid moss

[thin stump]

Is this for Ai covers or model cleaning? Just Use BS Roformer 1296 then UVR denoise

[vagrant marsh]

But there's no option to enable denoise

[thin stump]

It's a music and vocal seperation model so it's not a denoiser

[vagrant marsh]

This notebook has any kinda uvrs

[thin stump]

but its the best one compared to Instc hq you were using

[vagrant marsh]

I'm using this: https://colab.research.google.com/github/Eddycrack864/UVR5-NO-UI/blob/main/UVR5_NO_UI.ipynb#scrollTo=gmjUWmz8iecd

[low shard]

well there's also the No UI RVC-Disconnected Google Colab, i dont think theres really a best one tho

[thin stump]

Ik, so just use BS Roformer 1297 then UVR denoise separately. Keep in mind that stacking seperator models will slowly damage the quality of your model so the least audio processing the better

[vagrant marsh]

Fine

I'll use my dataset for it

[subtle phoenix]

will try it out.
thanks for the help, I've also found the pretrained guide very useful thanks for your efforts

[low shard]

i usually always used rvc disconnected for personal preference, the guide for it is https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/, anyways you can use whatever you want dw
Also thank you i appreciate you like the guide

[spiral osprey]

Ilaria mainline still works? I added the voice model and it downloaded but when refreshing it gave an error.

[viscid moss]

Use Roformers and then De-Noise

Roformers is located on BS/Mel Roformer cell

And De-Noise in VR Arch cell

there are 2 models

That Denoise check is not a model

[vagrant marsh]

Okay, a user already recommended me to use it, but thx anyway

[viscid moss]

It's just an audio cleaning but not an AI audio cleaning

[vagrant marsh]

Ah I get it

I'll do it. Thx

[vagrant marsh]

We can choose any kinda separator for example, mdx-net?

[spiral osprey]

?

[thin stump]

It has a low SDR score. The finetuned bs roformer is technically the best one currently because it took bunch of rented GPU's to make

[vagrant marsh]

Hmm

[thin stump]

it does work so restart your runtime and all that

[thin stump]

oh meant to show this one https://mvsep.com/en/algorithms

[vagrant marsh]

Ig I found this ones

Mdx23c v2

And v1

[thin stump]

Yup thats the second best one for vocal and instrumental separation

less aggressive than BS roformer

[vagrant marsh]

The website says that remove noises

But first I'll send the audio to the reformer

Roformer*

Am I doing right?

What are these settings?

[thin stump]

check your colab folder and copy and paste your file path for audio_input

[vagrant marsh]

I already did

It

I uploaded on the models folder

[thin stump]

k and just leave it as default ig

[vagrant marsh]

Ready

On the mdx23c we choose the path of the converted file?

From the bs roformer?

[spiral osprey]

I put my voice model in and the audio doesn't come out, can anyone help me?

[vagrant marsh]

Didn't work

Still the same

[thin stump]

No you either do MDX23C or BS Roformer. BS Roformer is the best one

to reiterate

[vagrant marsh]

Ah...

Okay

[thin stump]

then after that UVR Denoise and done

[spiral osprey]

I just made my model with 440 437 epouchs and it doesn't sing the dataset is ok and it doesn't sing. Can someone help me?I just made my model with 440 437 epouchs and it doesn't sing the dataset is ok and it doesn't sing. Can someone help me?

[vagrant marsh]

Ah, yes.

You meant this?

[thin stump]

skill issue https://docs.google.com/document/d/1YbXcLFPaGjhOdG5NFkK3QrucCEpHZBwFUxkeMO8aB18/edit

[thin stump]

ye

[vagrant marsh]

Okay, with the path of the converted file?

Okay, time to follow the steps

[rare goblet]

Ayo? @vagrant marsh level 8 !!!

[vagrant marsh]

Let's convert the file to the BS Roformer

And use the UVR-Denoiser.pth with the vr arch only

Tab

[spiral osprey]

Even so, it still doesn't work, other models work, but mine doesn't work. Since I just trained.

[thin stump]

It's suppose to be the added_index file and not a regular index file

[vagrant marsh]

The result was that:

[vagrant marsh]

Sorry, I sent the instrumental version

[viscid moss]

try to increase the aggression of the DeNoise model

[vagrant marsh]

Stills the same

[vagrant marsh]

To 10?

[viscid moss]

ya

[vagrant marsh]

K

[spiral osprey]

What should I do in this case?

[viscid moss]

show me ur model files

[spiral osprey]

Can I send it here?

[viscid moss]

sure

[spiral osprey]

I just trained him.

[viscid moss]

Files are fine

lemme test rq

[vagrant marsh]

Still the same. This notebook does not work

I changed the uvr denoiser to lite

And I increased the agressive to 100

And it keeps the same

I give up

I can see that I'll never fix my model

[viscid moss]

model works but.... there is no output

hmmmmm

[spiral osprey]

yeah.

[vagrant marsh]

I think I'll look for a better one from the internet

Colab doesn't work for nothing

[spiral osprey]

what could be happening?

[viscid moss]

That's not noise, that's sound effects

[vagrant marsh]

Ah

Yeah

But does it affect the voice of the model?

[viscid moss]

yep

that's more difficult to remove

try Bandit Plus on MVSEP

lemme send u the link

[vagrant marsh]

K

You're gonna save my day lol

[viscid moss]

https://mvsep.com

[vagrant marsh]

Okay, thx

[viscid moss]

tbh idk

ur train was succesfull?

lemme check the dataset

wait a min

[vagrant marsh]

Thank you again

I have to wait for the queue?

[thin stump]

create an account to skip the queques btw

[vagrant marsh]

This is a joke

[thin stump]

no u

[vagrant marsh]

Okay

[viscid moss]

I'll try to add that models to my notebooks

[spiral osprey]

I just trained him today, meaning he finished his training.

[viscid moss]

My theory is that something went wrong in the training and 'cause that the model is broken or something

because no output

[spiral osprey]

Worse, if I want to train I'll have to wait 2 days or so because I've just finished training so 48 hours will be necessary. Only if you want to give me a hard time then you will help me lol.

[thin stump]

kaggle exists besides colab. More hours of training than before

-kaggle

[azure marsh]

Kaggle Notebooks

• 🆕 UVR5 NO UI by Eddy
• How to use RVC Mainline Kaggle by Cauthess
• 🍏 Applio RVC by IA Hispano
• ✨ RVC Mainline by Hina

More commands: -colabs, -hf, -local, -rvc, -help

[vagrant marsh]

I keep in the same position

[thin stump]

It's been reduced from 200 people to 20 so not that bad

[vagrant marsh]

Hmm

[spiral osprey]

My notebook doesn't support it locally because it doesn't have an Nvidia GTX video card. My notebook has an integrated Intel card, so there's no point locally unless you can train and keep your voice clear. And it doesn't require much of a video card, then that's fine.

[thin stump]

no kaggle is on the cloud like colab. Click the mainline guide specifically

[viscid moss]

u can also use another google account

[vagrant marsh]

Well, I used the mvsep

And the sound effects still there

I'll update my model anyway

[thin stump]

Most lines will get removed from the dataset because quality > quantity. So if it sounds bad then it will reflect in the model. It's hard to remove sfx noises

manually

[heady gorge]

there's a lot a steps to just making your own rvc voice 😭 all i wanted to do was create my own JP DL Aigami voice loll

[vagrant marsh]

Okay

Tomorrow I'll update the model anyway

Ig it will not affect 100% the voice

[amber snow]

hi, well, i've got a problem with the google collab rvc, like, when i talk, i can only hear whispers. With the past ver I didn't have any problem. Pls help me 😦

[simple mango]

It wont let me download voice models on the ilaria RVC it just keeps erroring anybody know what causes this?

[violet heron]

What’s the link you were using?

[simple mango]

The google collab

[violet heron]

Use the HuggingFace version instead

-spaces

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• 🆕 UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by Ilaria Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[limber verge]

Hello,

For training my model I've used RMVPE + One-click training. Total Epochs = 250. It sounded quite weird? I can't really explain it but very artifact-y. I'm now training using the "Train model" button instead of "One-click training".

My question is will the "Train Model" button get rid of the artifact sounding model, or is there something else I should be looking into? Thanks!

[rare goblet]

Ayo? @limber verge level 1 !!!

[vague dew]

I tried downloading RVC but the files seem to be a py file

[violet heron]

Is there no bat file?

[vague dew]

oh so I open the bat file

thx that worked ._. im a newb lol

[vague dew]

how do I import the model into RVC ?

[violet heron]

Put the .pth file in the weights folder (weights folder might in assets)

And the .index in the logs folder

[vague dew]

lemme see

[rare goblet]

Ayo? @vague dew level 1 !!!

[vague dew]

@violet heron sorry to bother u but not sure if im doing it right can I please Dm u ?

[violet heron]

Sure but why not here

[vague dew]

show screen shots and what not

cuz I cant send images

of what im doing

plus I tried converting it sounds horrible T-T

@violet heron

[violet heron]

Make sure your using rmvpe

And you are converting audio that has vocals only, no music

[vague dew]

its only vocals

[vague dew]

how do I make sure its rmvpe?

[violet heron]

It’s a pitch/f0 extraction method

[vague dew]

I dont have that option

I only have dio, pm, harvest, crepe, crepe-tiny

@violet heron

[violet heron]

Your probably using RVC-gui which is outdated

-local

[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

[vague dew]

which link do i click ?

[vague dew]

.

@violet heron

[violet heron]

I personally use original RVC

But applio has some extra QOL

[vague dew]

alr

[vague dew]

Im looking for the bat file and I cat find it in this version

[violet heron]

It doesn’t come with it in some installations for some reason

[vague dew]

think I found it

[vague dew]

so do I paste that into cmd ?

[rare goblet]

Ayo? @vague dew level 2 !!!

[violet heron]

No you download the file or run it

You can paste it into cmd in that directory tho but it’s easier with the file

[vague dew]

alr

[crimson wraith]

is requesting isolated vocal datasets for artists a thing on here, or is that prohibited?

[modern surge]

-uvr

[azure marsh]

Ultimate Vocal Remover

One of the best free and open source vocal and instrumental isolation tool.

Tip: Use /guides UVR to learn how to use it.

[modern surge]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[misty elk]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[crimson wraith]

Yeah, I get that UVR is a thing but it's redundant if a bunch of people are processing the same request. I'm not lazy, just efficeint. 🙂

[open stag]

• colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[crimson wraith]

Is kim vocal still the model to use for isolating a singer's vocals from the music ?

[limber verge]

Where is the .index folder location within RVC?

[violet heron]

logs

[limber verge]

Thanks!

[limber verge]

It doesn't matter what voice model I use the audio always has a lot of artifacts. Is there some settings that I need to follow correctly? Or am I not using the correct audio files?

[violet heron]

Try turning down the index ratio

The index ratio controls accent and can add artifacting

[limber verge]

Can I post an example of some of the audio?

[violet heron]

No because copyright

[limber verge]

Ahhh, can I dm you then?

[violet heron]

I guess

[limber verge]

Do you use the RVC Model Interface, RVC GUI, or another software? I'm not sure where Index ratio is on the RVC Model interface.

[violet heron]

It could be under “Index Rate” “Index Influence” “Accent Strength” “Search Feature Ratio”

If your using RVC-GUI that’s outdated

[thorny vapor]

hello

anyone know whats the best singing juice wrld model?

[limber verge]

Ahh it was Search Feature, and I lowered the Transpose to better match. Thank you so much! I'm currently using RVC WebUI

[violet heron]

Good version

[woeful glacier]

hello

can some 1 help me with the download

i cant download my pc says

Windows protected your PC
Microsoft Defender SmartScreen prevented an unrecognized app from starting. Running this app might put your PC at risk.
More info

[glossy fjord]

That sounds like a firewall issue. Turn off your windows defense (firewall) and try again. Windows does this with all unfamiliar apps. Of course do it at yuor own discretion tho cuz idk what you're downloading. Make sure it's a trustworthy dev/source

🙂 Hey folks so in the guide it says Cloud Applio has Mangio Crepe but I don't see it. The applio wiki also mentions the removal of Mangio Crepe and Mangio Crepe-tiny in the latest version. Is there a reason why? Although the Crepe in applio also offers hop length adjustment. Kind of confused because I heard mangio is good but then i can't try it?

[subtle cedar]

How to clean dataset for better or "perfect/flawless" models and there's alternative for MVSEP and Audacity for cleaning datasets? I mean because MVSEP's queue nowadays is too long and I don't have a laptop or PC for Audacity ✌️

[left rampart]

Hello everyone, there was a problem with the voice changer, after launching in discord I installed a virtual cable and everything was as it should, but in the end the voice just doesn’t work, any voice at all, the message pipeline is not installediazed appears in the console, help solve the problem

[pastel oak]

mvsep queues are pretty short if you have an account which is for free, takes 3 minutes max

UVR local is the same as mvsep ig, you can use rx izotope for cleaning theres a lot of options

[little jungle]

-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

[rare goblet]

Ayo? @little jungle level 1 !!!

[subtle cedar]

but even when I logged in my account it's still too long but I will try tho if it's not too long anymore when we logged in our accounts there

[open stag]

• colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[subtle cedar]

Yes you're right it's shorter when I logged in my account so I will use MVSEP again for clean my datasets

[vagrant marsh]

start preprocess
['infer/modules/train/preprocess.py', 'datasets', '32000', '1', '/content/training/logs/Tater', 'False', '3.0']
datasets/20240722232530-a75c893172-1-2xna8ozp0loz-vocals-vocals-model-bs-roformer-ep-317-sdr-12_bandit_v2_model_speech_[mvsep.com] (1).mp3->Suc.
datasets/place-audio-here.txt->Traceback (most recent call last):
  File "/content/training/infer/lib/audio.py", line 37, in load_audio
    ffmpeg.input(file, threads=0)
  File "/content/training/.venv/lib/python3.10/site-packages/ffmpeg/_run.py", line 325, in run
    raise Error('ffmpeg', out, err)
ffmpeg._run.Error: ffmpeg error (see stderr output for detail)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/content/training/infer/modules/train/preprocess.py", line 87, in pipeline
    audio = load_audio(path, self.sr)
  File "/content/training/infer/lib/audio.py", line 42, in load_audio
    raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: ffmpeg error (see stderr output for detail)

end preprocess

[tribal thicket]

`Updating and installing system packages...
Installing build-essential...
Installing python3-dev...
Installing ffmpeg...
Installing aria2...
Updating and installing pip packages...

CalledProcessError Traceback (most recent call last)
<ipython-input-3-cfc2dfe3fcb2> in <cell line: 12>()
10
11 print("Updating and installing pip packages...")
---> 12 subprocess.check_call(['pip', 'install', '--upgrade'] + pip_packages)
13
14 print('Packages up to date.')

/usr/lib/python3.10/subprocess.py in check_call(*popenargs, **kwargs)
367 if cmd is None:
368 cmd = popenargs[0]
--> 369 raise CalledProcessError(retcode, cmd)
370 return 0
371

CalledProcessError: Command '['pip', 'install', '--upgrade', 'pip', 'setuptools', 'wheel', 'faiss-gpu', 'fairseq', 'ffmpeg', 'ffmpeg-python', 'praat-parselmouth', 'pyworld', 'numpy==1.23.5', 'numba==0.56.4', 'librosa==0.9.2']' returned non-zero exit status 2.`

got this error

what do i do

[twin pagoda]

should i use applio or rvc disconnected?

[low shard]

Doesn't really matter, applio is UI, RVC disconnected is no ui

[low shard]

You can also use the better MDX23C, which requires more resources tho

[low shard]

Be sure you downloaded RVC from https://docs.aihub.wtf and not a random YouTube video

[tribal thicket]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[stable crater]

what happened to this colab? it wont download voice models https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=F6zbsuNs6xZF

[rare goblet]

Ayo? @stable crater level 1 !!!

[proper shale]

that's outdated

use stuff like ilaria rvc zero

https://huggingface.co/spaces/TheStinger/Ilaria_RVC

for other colabs go to https://docs.aihub.wtf/

[proven hill]

guide: https://docs.google.com/document/d/1YbXcLFPaGjhOdG5NFkK3QrucCEpHZBwFUxkeMO8aB18/edit

[stable crater]

the hugging face worked, thank you!

[proper shale]

you're welcome :)

[tribal thicket]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[tribal thicket]

what do you reccomend me for the best realistic voice?

[heady gorge]

i tried extracting the folder and it took 15 mins or so and it failed bc something went wrong over and over so i'll probably need someone to make my desired model for me..it's in the request-model channel listed yu-gi-oh

[spiral whale]

Hi friends. I have all my vocals on an audacity project and am now following one of the guides for processing. I'm sorry for the dumb question, but can anyone explain this step to me?

I spent a while deleting silences and pauses previously because I was following a different guide before

[low shard]

i don't know exactly which you download nor error, also be sure you have a good pc (good gpu) for doing it locally

[low shard]

doesnt really matter, just a different ui tbh

[fading stump]

I'm confused about "Step 4"

[rare goblet]

Ayo? @fading stump level 1 !!!

[fading stump]

Why TensorBoard said "No dashboards are active for the current data set."

[heady gorge]

i mean, i have an omen laptop with nvidia geforce rtx and i went to the link you sent to extract the folder but failed

[fading stump]

I read in google docs it told me to install Python so I did and put the command "install tensorboardX" and "install torchcrepe" and I did.

[low shard]

your using an outdated colab, check https://docs.aihub.wtf

[low shard]

its weird it failed, maybe low storage?

[tribal thicket]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[heady gorge]

maybe..ill just be patient and wait until i get a reply notif that someone creates it for me then : (

[spiral whale]

When processing a data set do I need to generate silence? Or can I skip that if I've went through it deleting the gaps

[fading stump]

The new one, can I use my voice? I have prepared my Dataset.

[low shard]

i mean if u really wanna just wait

[low shard]

which new one? in the docs there are different ones, do you need to do inference (use models) or training (making models, which is related to the tensorboard) ?

[fading stump]

I pressed download from this one.

[low shard]

Do you got a good pc?

You were using google colab, which is a cloud computing service for people who dont got a good pc

locally is only for people who got a good pc

[fading stump]

oh

[low shard]

i guess you dont right?

[fading stump]

I pass the "Inference" criteria.

[low shard]

alright, to do inference (use models) on cloud for free, you can try Ilaria RVC Zero which is the fastes free cloud way

Else, you can use applio google colab (with a daily free limit T4 gpu, which is slower than ilaria rvc zero's ZeroGPU that is an A100)

[fading stump]

okay

[low shard]

also you don't have to worry about the tensorboard, its a training thing, if you got any issues ask here

[fading stump]

ok, understood

[potent brook]

I can't extract the mp3 file. The option isn't showing up and ive looked up every article and yt vid

Can I have someone extract the file for me

[heady gorge]

i totally can.

[low shard]

extracting the mp3 file ? wdym ? also dont watch yt vids, they are outdated, use our docs https://docs.aihub.wtf

[potent brook]

ok

[potent brook]

I found one of the voice models from the discord and it was in a mp3 format

[rare goblet]

Ayo? @potent brook level 1 !!!

[brittle wing]

@viscid moss can I ask you something, when will you add the ensemble cell to your UVR NO UI COLAB?

It's been a month

[spiral whale]

Hi guys. Does this mean I can't do the voice model?

[low shard]

thats not an rvc voice model then

voice models aren't in .mp3

where did u find that

rvc voice models are made of .pth and added.index

[potent brook]

i found it in the voice-models tab

[potent brook]

I typed female and it was at the bottom

[pastel oak]

if you have AMD gpu you cant unless you use Linux ROCM

only nvidia gpus on windows

[fading stump]

i don't understand

[rare goblet]

Ayo? @fading stump level 2 !!!

[fading stump]

which name

Or did I miss something?

[rotund prawn]

experiment name is not defined

[fading stump]

yes, but i don't understand where is it

did I skip something?

[low shard]

send the link of that model

[knotty moth]

"Set training variables"

[low shard]

like

[fading stump]

oh ok i see

[potent brook]

idk how to do a link

[low shard]

click the link icon show in the icon of the voice model u found

bc that should be deleted if its .mp3

[potent brook]

https://cdn.discordapp.com/attachments/1186431241176764517/1186431242158231592/normalp.mp3?ex=66a0d3f2&is=669f8272&hm=2fe9b566a67b70e000411e6e823fb3e9cee8888427f45c560abf85a053cc0ac0&

[tribal thicket]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[rare goblet]

Ayo? @tribal thicket level 3 !!!

[twin pagoda]

i tried using applio and when i press preprocess my whole pc crashes is there a way i can fix this?

[low shard]

?

i think you are using the demo .mp3, you need to use the voice model download link in the post

[low shard]

do you got a good pc?

[potent brook]

yea u right mb

[low shard]

yea, be sure you are using the download link of the voice model, tell me if its all fine now

[potent brook]

i downloaded it with a few other ones but im having trouble setting it up

[brittle wing]

How do i fix robotic audio in rvc?

anyone notice my message?

[twin pagoda]

i got a rtx 4080 with 16gb ram

and a i7 13700hx

[shrewd nebula]

someone else having problems with applio inference?

-kaggle

[azure marsh]

Kaggle Notebooks

• 🆕 UVR5 NO UI by Eddy
• How to use RVC Mainline Kaggle by Cauthess
• 🍏 Applio RVC by IA Hispano
• ✨ RVC Mainline by Hina

More commands: -colabs, -hf, -local, -rvc, -help

[low shard]

Damn

Does it ENTIRELY crash?

Maybe @nocturne mural could help you

[twin pagoda]

my screen goes black all apps crash and my windows goes into lock for some reason

[low shard]

Holy shit, I don't really know but maybe that applio staff could help

[twin pagoda]

im thinking it might be a memory issue the chrome tab says out of memory before the screen going black

[low shard]

How much VRAM does your GPU have?

Also do you have other things in the background?

[twin pagoda]

i have 12gb of vram i also tried it with everything closed on the background didn't work

[nocturne mural]

this problem has happened before with a user, this problem was fixed by updating the graphics card drivers.

[olive hill]

Which is best collab for model creation, Suggestions please.
THIS IS FOR MODEL TRAINING and which is BEST RVMPE OR MANGIO_FORK OR ANY SUGESSTIONS

[low shard]

there isn't a best one

it depends, you can either use applio colab that has ui, or rvc disconnected that has no ui

Btw, also kaggle (another free cloud computing with more gpu time and better ones) is a good option

but its harder to use

if you wanna use that

check below

-kaggle

[azure marsh]

Kaggle Notebooks

• 🆕 UVR5 NO UI by Eddy
• How to use RVC Mainline Kaggle by Cauthess
• 🍏 Applio RVC by IA Hispano
• ✨ RVC Mainline by Hina

More commands: -colabs, -hf, -local, -rvc, -help

[olive hill]

How many epochs is good for a 10 mins dataset?

[low shard]

There isn't a perfect amount

you need to use the tensorboard

https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/

[olive hill]

Thank you so much. Kaggle works like Collab?

[vagrant marsh]

I already ran the first cell

why is notebook asking me to run the first cell?

[low shard]

kinda, its a bit harder to use and you need a phone number for it, but you get 30 hours weekly (compared to the 4 daily ones of google colab) and u can use P100 or T4x2 (better than T4 google colab gpu)

[olive hill]

Thats interesting thank you 🙂

[low shard]

yw

[vagrant marsh]

can someone help me pls?

[thin stump]

What is this notebook? Colab?

[vagrant marsh]

https://colab.research.google.com/github/hinabl/RVC-Online/blob/main/Mainline_Colab_Full.ipynb#scrollTo=BFb9nDEM10-i

[thin stump]

-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

[thin stump]

Though if you're already learning the colab one, you might as well try Kaggle. It's the same steps pretty much

[vagrant marsh]

Kaggle is complicated

there is no run button in other users' notebooks

[thin stump]

I made it dummy proof on the guide since ppl keep asking about the cells . It's a requirement to run !wget. You don't have to run the pretrain cell

idk when you last checked but there's new images

[vagrant marsh]

🤔

I'm gonna try the mainline from Kaggle

[arctic cargo]

From an understanding perspective why is only 45s audio recommended in most cases? Isn't having more audio better to clone a voice?

[vagrant marsh]

It takes so long

[vagrant marsh]

okay, plan b

RVC Disconnected

[thin stump]

You didn't verify your number did you?

ok noted

[vagrant marsh]

what? verify my phone number?

[thin stump]

yeah I even said this part

[vagrant marsh]

aah

I'm gonna check this correctly

[vagrant marsh]

Is this from I'm using? https://www.kaggle.com/code/hinabl/mainline

[thin stump]

yes, i just turned on dark mode

[brittle wing]

how can i get the voice changer?

[vagrant marsh]

Fine

[thin stump]

because the white images weren't great for the guide and made it look cluttered than it actually is

[proven hill]

https://rentry.co/VoiceChangerGuide

[brittle wing]

sus

why is there that gif?

[proven hill]

idk for fun but thats the guide

[vagrant marsh]

Okay

[brittle wing]

i can't read japanese or idk..

[proven hill]

its english

[brittle wing]

on hugging face found it

[rare goblet]

Ayo? @brittle wing level 1 !!!

[brittle wing]

YOOO

[proven hill]

make sure to download the right version

[brittle wing]

wich would you recommend?

[proven hill]

what gpu do you have?

[brittle wing]

idk

...

[proven hill]

check it

[brittle wing]

how do you do that sry i forgot

[proven hill]

task manager -> performance

[brittle wing]

ok one sec

16GB

[proven hill]

thats the ram

im asking the GPU

[brittle wing]

i am stupid

[violet heron]

It should be GPU 1

If it’s not there it’s GPU 0

[proven hill]

GPU 0

[violet heron]

GPU 0 is usually the integrated graphics if the CPU has them

[brittle wing]

amd ryzen 5 5600h with radeon graphics?

[thin stump]

I'm updating a few things so if you refresh you'll see the changes

[proven hill]

so you dont have a graphic card

[brittle wing]

i do

rtx 3050

nvidia

[proven hill]

okay good

download this version: https://huggingface.co/wok000/vcclient000/blob/main/MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a.zip

[brittle wing]

ty

[vagrant marsh]

Fine...

[rare goblet]

Ayo? @vagrant marsh level 9 !!!

[brittle wing]

sry for the trouble i am lost sometimes

[proven hill]

nah its fine

[vagrant marsh]

@thin stump where are the updates?

[thin stump]

i don't own the notebook https://rentry.co/RVC-Mainline-Kaggle

I made this guide

[vagrant marsh]

ah

Okay

I'll read everything

thx

I'm gonna do it on my phone

[woeful glacier]

hello guys

can someone help me how to make me voice in discoard

i just hear my normal voice

[proven hill]

https://rentry.co/VoiceChangerGuide

[tribal thicket]

how much epochs do you reccomend?

[pastel oak]

theres no epoch recommendation, it all depends on tensorboard graphs that determine when its trained ideally, read more

https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/

[woeful glacier]

TY

[paper hinge]

Hi everyone,

I would like to use a Boots the Monkey voice RVC in particular like I saw used on some YouTube videos used to cover popular songs,

But can't seem to find it on the official page for voices in the search bar. Does anyone know where it is?

[pastel oak]

https://www.weights.gg/models/clm72lgmf04agcctcvjazqhzs

[woeful glacier]

hi i cant talk i put my pic in input and audio cable in output and i cant talk in the voice changer client

[rare goblet]

Ayo? @woeful glacier level 1 !!!

[brittle wing]

Does anyone know how to split vocal parts? like in k-pop songs that have multiple parts including some overlapping and backups? That way with AI covers, it's not constantly sounding like a pile of rubbish?

[potent brook]

does anyone know how to setup the voice changer with fortnite?

[rare goblet]

Ayo? @potent brook level 2 !!!

[vagrant marsh]

start preprocess
['infer/modules/train/preprocess.py', 'datasets', '32000', '1', '/content/training/logs/Tater', 'False', '3.0']
datasets/.ipynb_checkpoints->Traceback (most recent call last):
  File "/content/training/infer/lib/audio.py", line 37, in load_audio
    ffmpeg.input(file, threads=0)
  File "/content/training/.venv/lib/python3.10/site-packages/ffmpeg/_run.py", line 325, in run
    raise Error('ffmpeg', out, err)
ffmpeg._run.Error: ffmpeg error (see stderr output for detail)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/content/training/infer/modules/train/preprocess.py", line 87, in pipeline
    audio = load_audio(path, self.sr)
  File "/content/training/infer/lib/audio.py", line 42, in load_audio
    raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: ffmpeg error (see stderr output for detail)

Well I got this error

How do I solve this?

[thin stump]

it's dataset/Tater. Make a folder called dataset in the imjoy

[violet heron]

What’s your GPU

[potent brook]

NvIdia something

[vagrant marsh]

Imjoy?

[potent brook]

do you need the specific type?

[violet heron]

Which one

[thin stump]

the imjoy gui is the FILE URL

[violet heron]

You can check in task manager under preformace

[vagrant marsh]

Ah

[potent brook]

Nvidia GeForce GTX 1070

[thin stump]

ye put your tater folder inside the dataset folder. I don't know why the /datasets/ folder is locked so again, just make a new folder

[violet heron]

https://github.com/deiteris/voice-changer you will need to limit settings and fps on Fortnite

[vagrant marsh]

I didn't get

[thin stump]

It's in the guide

[vagrant marsh]

I copied the path to the audio and it gives the same error

[vagrant marsh]

I'll read

It

[potent brook]

So i just follow the steps on there and it will work?

[thin stump]

https://rentry.co/RVC-Mainline-Kaggle#formatting-the-dataset

[olive shale]

hi can anyone help me

[violet heron]

Yes you will need to set your input on Fortnite to the audio cable tho

[violet heron]

Don’t ask to ask, just ask

[olive shale]

it says
Looks like there’s a a bit of a problem

[potent brook]

ok

[vagrant marsh]

Ig it's this one

I'll read it

[potent brook]

what about the output

[thin stump]

i linked the section for you to read. It'll automatically take you there

[vagrant marsh]

Yes!!! God!!! Thx @thin stump

[cloud jewel]

Hi there, I'm using MangioRVC to create a voice model of Ranni from the Elden Ring but the voice model turned out to sound very whisper-ish. Is the problem of the voice samples I used?

[vagrant marsh]

Is a GPU necessary to do this?

[near vigil]

Hi ! Im not able to select any voice model provided like its not loading

[vagrant marsh]

The problem is my GPU use limit was reached

I can't use GPU anymore

On colab

[thin stump]

oh, then use kaggle

I told you it's the same steps :/

[brittle wing]

I'mma need serious help with this because I've seen this be possible in IZ*ONE's panorama, where the high note was separated from the lyrics but I do not know how to do something like that. It could be very good and enhance ai covers when I need to do this.

[rare goblet]

Ayo? @brittle wing level 1 !!!

[versed cairn]

-spaces

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• 🆕 UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by Ilaria Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[lilac canopy]

cant open "C:\Users\User\Desktop\RVC-GUI-pkg"

cant open RVC-GUI.bat

to start the programming

i've got all requierements

[olive shale]

The voices don't work for me i but my mic in input and audio cable on output and still the voice are not working

[lilac canopy]

it says it cant find the path

[proven hill]

dont use it its old

[proven hill]

show okada screen

[frosty lichen]

yall whats the newest ilaria

[violet heron]

@proven hill what’s the newest of you

[proven hill]

https://docs.google.com/document/d/1YbXcLFPaGjhOdG5NFkK3QrucCEpHZBwFUxkeMO8aB18/edit

[agile apex]

does anyone know why this is happening

[woeful glacier]

hi . my voice is so laggy i talk and after 10-20s i talk in discoard and i cant hear my self in the app how can i hear my self in the voice changer?

[violet heron]

Use RMVPE instead of crepe or harvest

[violet heron]

Set monitor to your headphones/speakers

What’s your GPU?

[woeful glacier]

OKAY IT WORK THANKS

[agile apex]

im using weights

[brittle wing]

i need it for calls sorry for not responding

[low shard]

alright dw

-rt

[azure marsh]

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

[low shard]

Use wokada locally if you got a good pc, else use the google colab (u still need a pc)

[trim sphinx]

Which collab should I use for finetuning?

[brittle wing]

okay thank you

[low shard]

you mean for training rvc voice models?

[low shard]

yw

[trim sphinx]

Yes

oh and it should be able to make German rvc voices

[vagrant marsh]

Invalid syntax

Also Kaggle stops running automatically after some seconds

[low shard]

you can choose whatever you want:

• RVC Disconnected (has no ui)
• Applio Colab (has ui)

[tidal trail]

@low shard

[low shard]

don't send ai covers as files here for copyright issues. also you may need to clean it more

[low shard]

didnt see, ty

[low shard]

I mean, you can do that, maybe the rigel pretrain in the pretrains guide could help you as its actually the only pretrain that has a german dataset too

[vagrant marsh]

How can I fix it?

[tidal trail]

good job, ty

[low shard]

why are you trying to run an html code line ?

[vagrant marsh]

I thought it was necessary

Also Kaggle stops running automatically after some seconds

And idk why

[low shard]

i dont see that line on top of the title tho

also be sure to read https://rentry.co/RVC-Mainline-Kaggle

[low shard]

you are phone verified on kaggle right

[vagrant marsh]

Yep, I've verified my phone number

[low shard]

and when does it automatically stop

when running that line?

[vagrant marsh]

Yep

[low shard]

did you add it yourself ? also you dont need to run that

[vagrant marsh]

Yep, I clicked on copy & edit

So how do I do it without running?

[low shard]

don't add that line, idk where u seen that, read from https://rentry.co/RVC-Mainline-Kaggle#cloning-the-notebook-and-configurations

[vagrant marsh]

I followed the example of the guide

[thin stump]

Please don't train on mobile for your sanity . I never said to add a line of code anywhere

kaggle stops running because you keep hitting your head in a wall trying to run that syntax for some reason

[vagrant marsh]

Can't we train anywhere like on the PC, on the phone?

[thin stump]

bruh unless you have a 18 inch phone screen and can see two browser windows at once, then you can have that luxury

[vagrant marsh]

Mine is a Xiaomi Redmi Note 12

[spiral osprey]

@viscid moss I trained my voice again today, and even so, when I converted the music, the audio didn't come out.

[glass ember]

does it matter if u dont put the index file and just put the pth file? will it make the model sound any different

[spiral osprey]

Does this have anything to do with it?

[viscid moss]

no way!

[viscid moss]

what hd space is that?

[spiral osprey]

i don't know.

[viscid moss]

use ilaia

-hf

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• 🆕 UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by Ilaria Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[spiral osprey]

Just trying and nothing.

[viscid moss]

holy

u train locally or cloud?

[spiral osprey]

cloud

[spiral osprey]

I don't have a GPU to do it locally

[viscid moss]

using...?

[spiral osprey]

Unfortunately, mine is integrated from Intel.

[viscid moss]

i mean what cloud service and using what program

[lilac canopy]

whcih python version
i should use guys
the latest?

[brittle wing]

Anyone know how to separate parts in terms of vocals so that the cover doesn't sound garbage, but that way if someone is hitting a high note, you can have them hit it instead of using just the full acapella? It's something that would help a lot.

[lilac canopy]

should i let it like this?

[viscid moss]

for what?

[lilac canopy]

to use the app

of voice cloning

-'

[viscid moss]

u can use UVR to separate a song into stems

-uvr

[azure marsh]

Ultimate Vocal Remover

One of the best free and open source vocal and instrumental isolation tool.

Tip: Use /guides UVR to learn how to use it.

[viscid moss]

u don't need to install python, the program comes with python

[lilac canopy]

ok.

only the requierements tab?

[brittle wing]

so I'mma be new to this

[lilac canopy]

i install just the requierement?

and open it?

thats it?

i have a 4080

2 4080

[viscid moss]

what program are u tring to install?

[brittle wing]

not new to making AI covers but :P

[lilac canopy]

Applio

and the

RVC

does rvc clone voices?

[rare goblet]

Ayo? @lilac canopy level 4 !!!

[viscid moss]

Applio is RVC so is the same

[brittle wing]

you can train voices and ai covers.

[lilac canopy]

i see

ty

i will use this to apply dark stuff

thanks for giving me forbidden knowledge

to make money

JK

[viscid moss]

https://docs.aihub.wtf/rvc/local/applio/

check our guides

[brittle wing]

-rt

[azure marsh]

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

[brittle wing]

so I downloaded UVR so..how do I work it?

I'm using Shooting Star by XG since there's tons of overlaps in it, so how do i work this thing?

Permission to dm a moderator about it or send screenshots??

[lilac canopy]

'env\python.exe' is not recognized as a built-in command
or external, an operable program or a batch file.

cant open applio with that

i get this error

i dont know how to fix it anymore

i'm going crazy

3 hours now

trying to fix this error

[tidal trail]

move applio to another place that not contain spaces or rare characteres

[lilac canopy]

can u give me an example

please?

i just unziped it.

and i'm following the guide

[tidal trail]

C/RVC/Applio V3.2.1, also do not open as admin

[brittle wing]

this is what I see when I open, I am choosing a song and the same folder the vocals are in but as there is overlapping, Idk what i'm doing

[tidal trail]

https://docs.aihub.wtf/rvc/resources/vocal-isolation/#local-uvr

[lilac canopy]

if i dont open them as adm it shows i dont have permission

to open it

i really dont know what else to do...

tried to use RVC or applio and it showed the same error

[tidal trail]

put it in another place

[lilac canopy]

which one

exacllt

my desktop?

[tidal trail]

don't put it in a folder with privileged access and ensure the path does not contain any spaces or special characters.

[tidal trail]

maybe

[brittle wing]

I need to remove overlap, so which vr model would work?

[tidal trail]

i think should work for applio and rvc

[brittle wing]

@viscid moss I asked something

[lilac canopy]

ok

[brittle wing]

what do I do to remove overlapping vocals? in "shooting star, there's an overlapping with "baby if i give it my all will it pay off" in the main vocal, so I want to remove the overlaps.

[rare goblet]

Ayo? @brittle wing level 2 !!!

[lilac canopy]

@tidal trail assim?

like this?

[tidal trail]

hablas español

[lilac canopy]

portugues

pero entiendo un poco

[tidal trail]

a

[lilac canopy]

de espanol

jajaa

[tidal trail]

no lo pongas en tu one drive

[lilac canopy]

they are not

on my one drive

they are on a folder called clonar

and i've unzipped the archive

and now what?

[tidal trail]

iniciar backup/clonar?

[lilac canopy]

just open it?

[lilac canopy]

Desktop > new folder > folder name > clonar > Applio unzipped on that folder

is that correct?

@tidal trail

its on my desktop inside this folder

and unzipped

[tidal trail]

intenstaste con lo que te dije C/clonar/Applio

[lilac canopy]

i did that before

didnt't worked

i received that error

[tidal trail]

probaste con el otro programa

[lilac canopy]

same error

😦

for both

[tidal trail]

[lilac canopy]

this error 😦

[spiral osprey]

cloud:ilaria main line
program:uvr to extract the vocals.

[lilac canopy]

guys please

can someone make a miracle

i'm trying to install this program

for like 4 hours

[tidal trail]

lo mismo con el otro verdad?

[brittle wing]

it says I have to select a VR model but idk which one removes overlaps..pls send help

[lilac canopy]

si

mismo error

[tidal trail]

weird for both programs

[tidal trail]

try this https://docs.aihub.wtf/rvc/local/mainline/

[brittle wing]

Hey uhhh, meow?

[latent crag]

how do I get pic perms

[brittle wing]

i need to select a vr model but idk which one allows me to stop overlapping vocals

Could I dm you for advice and to show what I mean?

[latent crag]

is the gpu option better on nividia gefore gtx 1160 supor or cpu

[lilac canopy]

nah

i want to fix the problem

i want to use my GPU

its built for that shit

and i can install the programa right

[viscid moss]

I tried but got a lot of errors, so I think I'll wait for the original author of python audio separator to add it.

[brittle wing]

Thanks for the effort but how long can that take? Any idea?

[viscid moss]

tbh idk, for him python audio separator it's just a hobby

[brittle wing]

Oh...

I shouldn't wait then

[violet heron]

Always GPU

[knotty moth]

your GTX 1660 super > threadripper 7995WX alone

[proud inlet]

hi how can i clone my voice?

llevo un tiempo intentando en el rcv del collab pero en la parte de descargar siempre se traba y no me deja descargar

[rare goblet]

Ayo? @proud inlet level 1 !!!

[violet heron]

-docs

[azure marsh]

Suggestions for @proud inlet

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

🎙️ Realtime Voice Changer (Local)

💻 (Local) Realtime Voice Changer

• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[plush bane]

/guides realtime

[proud inlet]

Thank you

[azure marsh]

🎙️ Realtime Voice Changer (Local)

💻 (Local) Realtime Voice Changer

• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[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, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• 🆕 UVR5 UI, by Eddy Google Colab

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

[reef lark]

Hey, I'm trying to learn more about TTS and voice cloning with open source models. Having a very hard time. Can anybody recommend resources? I'm fairly technical/familiar with programming but have no background knowledge in ML

[violet heron]

Do you just mess around with voice cloning?

[left rampart]

help me...

[rare goblet]

Ayo? @left rampart level 1 !!!

[left rampart]

help please

[violet heron]

Go to that url

[left rampart]

?

okey, thank you

[left rampart]

help

@violet heron

[violet heron]

What do you need help with

[left rampart]

when I go to the url it says that there is no such site

@violet heron

[violet heron]

https://github.com/deiteris/voice-changer try this fork

[left rampart]

is there a new version out?

[left rampart]

I downloaded it, what should I do next?, just put it in the folder with the voice changer?

[rare goblet]

Ayo? @left rampart level 2 !!!

[left rampart]

@violet heron

[violet heron]

No you can delete the old one

[left rampart]

It seems that the folder contains not the full version, there are only some files from the voice changer

[violet heron]

It’s a fork of it

It’s more optimized

In this version you open the exe

[left rampart]

Got it, I'll try it now, thanks

Do I need to unpack everything from the archive?

where is the exe?

@violet heron

[violet heron]

What did you download?

[rare goblet]

Ayo? @violet heron level 106 !!!

[left rampart]

I can’t find where to download it, I downloaded the link above

@violet heron

[violet heron]

That’s not what you were supposed to click

Read the read me

It’s there for a reason

[left rampart]

please help me find where to download that file or a download link right away

[violet heron]

What’s your GPU?

[left rampart]

GeForce RTX 2060 SUPER

[violet heron]

https://github.com/deiteris/voice-changer#how-to-use

Read this

It should tell you how to download it

Since the file comes in 2 parts

[left rampart]

I found another link, you can see if this is it or not

@violet heron

https://github.com/deiteris/voice-changer/releases

[shrewd thistle]

how do i make a model for rvc

[violet heron]

Download the 001 and 002 files

The read me tells you how to unzip them since it’s a bit different compared to normal zip files

[violet heron]

-docs

[azure marsh]

Suggestions for @shrewd thistle

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

[left rampart]

These ones?

@violet heron

[violet heron]

You get the files that end in 001 and 002

Once again, read the readme, it should tell you this

Along with how to unzip it since it’s a 001 and 002 file

[left rampart]

There are different versions, can I download them for Windows?

these 2?

@violet heron

[violet heron]

Yes you download both like I said

[hard gust]

What's the difference between using Hubert-Large for training and using it only when infering an audio?

[celest cypress]

Hey I'm new here, but recently tried out w-okata voice changer and the original rvc gui, whereas the rvc gui runs a little faster than w-okata but a little laggy/choppy. Is this fork you linked any better? I mean I have a new RTX 4070 (but how can that be too slow)? Maybe my settings are cracked but I dont have any idea how to sound fluently like a japanese girl who speaks english without any chopping

[tame mica]

Kaggle Mainline file exp doesnt load for some reason? :p

nevermind

[mystic tangle]

yo @tame mica

[tame mica]

hello

[wispy lodge]

Depends on the settings. It's definitely better than original w-okada, but it should provide performance similar or a bit better performance compared to RVC Realtime

[north seal]

Hello everyone

[rare goblet]

Ayo? @north seal level 1 !!!

[twin pagoda]

i updated my drivers it didn't crash this time but i got this error

[rare goblet]

Ayo? @twin pagoda level 2 !!!

[wispy lodge]

Try starting preprocess with less number of cores like 4-8. I think I had similar issue

[meager quarry]

what do I need to continue training an unfinished model on Disconnected colab?
I've found no answer to this in the guide for it

[twin pagoda]

that fixed my problem thanks

[meager quarry]

Ok, I've kinda figured it out by now the Colab can't open the rvcLogs.ZIP file containing the preproccessed dataset
can I run feature extraction and preprocessing again as an alternative to this?

[lone berry]

What are the options for real time? Just mainline and w-okada? Mainline has much better performance from what I see (EXTRA)

[wispy lodge]

Modified w-okada version is also an option
https://github.com/deiteris/voice-changer

[vagrant marsh]

ElevenLabs_2024-07-24T12_01_04_Laura_pre_s50_sb75_se0_b_m2.wav->Traceback (most recent call last):
  File "/kaggle/tmp/training/infer/modules/vc/modules.py", line 185, in vc_single
    audio_opt = self.pipeline.pipeline(
  File "/kaggle/tmp/training/infer/modules/vc/pipeline.py", line 410, in pipeline
    self.vc(
  File "/kaggle/tmp/training/infer/modules/vc/pipeline.py", line 221, in vc
    if protect < 0.5 and pitch is not None and pitchf is not None:
TypeError: '<' not supported between instances of 'dict' and 'float'

What does this error means?

Where does the audio is saved?

Cauthless didn't explain how we can test the audio on the guide

on the batch inference

[gray crow]

sup guys, when training 44khz model what sample rate should i pick in rvc disconnected? 48 or 40?

[vagrant marsh]

Nvm, I found the result here

[brittle wing]

Heya, is there any I can know the specified number of the step I pointed in tensorboard.
Since it only shows 14.2k, but I want to know what's on the number "14,2**"

[lone berry]

Ahh cool I’ve seen this. You know if this has the same performance issue as W-Okada? Because you can max Extra on RVC GUI with very little latency but okada’s problematic

[wispy lodge]

No, there's no such issue. On the videos in readme you can see that it's set to 2.7s (which is equivalent to 131k in the original version), and none of them has CPU usage issues

[west abyss]

is there a way to record the AI voice changer, as in importing audio into it so thatb it could sound smooth than being choppy

[vagrant marsh]

Does this option control the instrumental?

[dense gazelle]

hey im trying to use this link: https://drive.google.com/file/u/0/d/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7/edit
and it doesn't work, says that this file has been blocked off of google drive

[low shard]

where did you get that link

its not valid

are you trying to do inference (use ai model, make ai cover)

[dense gazelle]

I am trying to make an ai cover

[rare goblet]

Ayo? @dense gazelle level 1 !!!

[low shard]

Use Ilaria RVC Zero which is the fastest, free easiest way to do rvc on cloud

[dense gazelle]

I really can't undertstand it

I already have the voice model

[low shard]

what of the guide you don't understand ?

[dense gazelle]

I just need to turn my song into the voice model that I have

[dense gazelle]

just how to take the voice file that I have and the song that i have and genrate it

[low shard]

its explained in the guide how to use it

you just need to read it

[hearty iron]

yo im tryna use voice.ai

and i get some error message

#33339

[low shard]

we dont offer support to voice.ai

we use wokada

which is free

[hearty iron]

oh shi

[rare goblet]

Ayo? @hearty iron level 1 !!!

[hearty iron]

ill get that then

[low shard]

-rt

[azure marsh]

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

[low shard]

you got a good pc?

[hearty iron]

yes

[low shard]

else you can use the google colab

[hearty iron]

3060

gpu

[low shard]

alright

thats good

[hearty iron]

16gb tam

ram

[low shard]

use the first guide link, wokada

[hearty iron]

are there any

yt vids

for this

i dont get it at all

[low shard]

nope, all yt vids are outdated

you just need to read the guide

[hearty iron]

can i get this

or do i needa get the 3gb file

[proven hill]

whats this?

[heady gorge]

how am i unable to upload an audio file as an example for the voice model i'm requesting in my post?

[proven hill]

why do you need that

[hearty iron]

idk im trying to follow a guide

[hearty iron]

ø

[proven hill]

vb cable?

[hearty iron]

uh idk first time i been using ai

im js tryna get a real time voice changer

ai

yk how?

[drowsy pumice]

guys i dont know why but when im using ai voice in discord it always comes out really crispy and glitchy what can i do to fix that

[rare goblet]

Ayo? @drowsy pumice level 1 !!!

[drowsy pumice]

its working good inside the app

[azure marsh]

🎙️ Realtime Voice Changer (Local)

💻 (Local) Realtime Voice Changer

• Guide for W-Okada's RealTimeVoiceChangerClient
• How to use real time conversion with RVC (voice changer)
• RVC TTS

[azure marsh]

Suggestion for @azure marsh

Documentación

📚 AI HUB Docs

https://docs.aihub.wtf/v/spain

❔ Perguntas Frecuentes

https://docs.aihub.wtf/v/spain/faq

🔗 Servidor IA Hispano

https://discord.gg/iahispano

Créditos: IA Hispano

✨ Otras Guías

• Guía de Uso del Colab AICoverGen

Link: Google Docs
Créditos: Eddy (Spanish Helper)

• Cómo Hacer un Cover de IA Usando Ilaria RVC

Link: Rentry
Créditos: Julia (ailen2091)

Suggestion for @azure marsh

🍏 Tutorial de Applio en español

Link: YouTube
Autor: IA Hispano

Servidor IA Hispano

https://discord.gg/iahispano

[final cairn]

hii can we train models with sd.next software?

[proven hill]

what

[brittle wing]

When using RVC V2 Disconnected, although I finished training model with 500 Epochs, is there a way to use the model’s previous epoch?

Like download the 200 Epoch version instead of the latest one.

[knotty moth]

https://tenor.com/view/deaf-cat-deaf-aagghh-aagh-cross-eye-cat-gif-4205790747075254931

[proven hill]

if you checked the setting yes iirc

[knotty moth]

click folder icon in left pane then find it in mangio-RVC/weights

[brittle wing]

@proven hill @knotty moth Thank you so much!
And now I probably knew that I have to edit the save frequency to train the model again from scratch or backup just to match the epoch I want to get lol.

I might be stupid or not, but I thought in “Export Model from Notebook to Drive”, you can download the previous epoch you want by inputting step and epoch count, since I got curious on how do I know that it has the specified epoch I want.

[thin stump]

I don't think it's meant for inferencing because another person besides you tried to upload 8 hours of audio. They said the maximum is 3 hours long and that RVC creates raw audio files

Idk what they've done prior to use all their ngrok html requests

[vagrant marsh]

Hmm

Well

I already updated my model

I tested with the .flac format lol

[thin stump]

-hf

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• 🆕 UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by Ilaria Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[thin stump]

Use ilaria RVC zero

[vagrant marsh]

Next time I'll use Ilaria

[rare goblet]

Ayo? @vagrant marsh level 10 !!!

[vagrant marsh]

But does it download a model to test?

[thin stump]

You upload your model .pth and added index and viola

[vagrant marsh]

Ah

Okay

Thx for the recommendation

[proven hill]

:3

[final cairn]

i mean ZLUDA sry

[narrow nova]

hey guys
I'm trying to use this colab and it seems to be out, could anyone check if it's like this on my machine or on yours it's also out, thank you very much
https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS

[narrow nova]

it stopped working today, after a code update made yesterday

[feral bough]

Hello! I’m new to AI and I want to do a Pjsk AI cover! Any recommendations

?

[quiet ginkgo]

Man what the fuck

[lofty heart]

-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

[quiet ginkgo]

Yeah I tried most possible solutions I know I have no Idea how to fix It

[worldly oasis]

I need help

Please use pip<24.1 if you need to use this version.
WARNING: Ignoring version 2.0.2 of omegaconf since it has invalid metadata:
Requested omegaconf<2.1 from https://files.pythonhosted.org/packages/72/fe/f8d162aa059fb4f327fd75144dd69aa7e8acbb6d8d37013e4638c8490e0b/omegaconf-2.0.2-py3-none-any.whl (from fairseq==0.12.2->-r requirements.txt (line 2)) has invalid metadata: .* suffix can only be used with == or != operators
PyYAML (>=5.1.*)
~~^
Please use pip<24.

[timid valve]

should i do pitch extraction using mangio crepe or rmvpe?

[worldly oasis]

How i can use mine collab it isnt working it's gimme errors all the time please help

Traceback (most recent call last):
File "/content/AICoverGen-NO-UI-en/src/main.py", line 11, in <module>
import gradio as gr
ModuleNotFoundError: No module named 'gradio'
1

Please someone assist me

i used that link https://colab.research.google.com/drive/1u1brjK8IZt647UsbZuGYfW29oFM2I4tk?usp=sharing&authuser=2#scrollTo=ToLpkj2rkU1M

[tranquil sinew]

!pip install pip==23.1

[gloomy jungle]

dyg know which male ai model is good for singing

without a really deep voice

[worthy lantern]

have someone a good webiste for public url?

[worldly oasis]

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
chex 0.1.86 requires numpy>=1.24.1, but you have numpy 1.23.5 which is incompatible.
pandas-stubs 2.0.3.230814 requires numpy>=1.25.0; python_version >= "3.9", but you have numpy 1.23.5 which is incompatible.
torchtext 0.18.0 requires torch>=2.3.0, but you have torch 2.0.1+cu118 which is incompatible.
torchvision 0.18.1+cu121 requires torch==2.3.1, but you have torch 2.0.1+cu118 which is incompatible

[leaden crag]

I'm having the same issue too. Can't fix it for some reason 🤷‍♂️

[worthy lantern]

why im never getting a public url

[tranquil sinew]

Well, about that, I do not know how to help you. I saved my notebooks from that problem by downgrading the pip version.🤷‍♂️

[low shard]

@viscid moss aicovergen is broken again

[quiet ginkgo]

I Ignored that and I'm training without it Is It fine

[rare goblet]

Ayo? @quiet ginkgo level 4 !!!

[quiet ginkgo]

Nvm It doesn't even work at all

[violet heron]

If your trying to make ai covers use ilaira RVC instead

[quiet ginkgo]

Oh I'm training on CPU

[brittle wing]

AICoverGen WebUI aint working
when i run the "Run WebUI" i get this and i got no idea how to fix it
is there a solution

[quiet ginkgo]

People told me train on Disconnected cuz everything else will kill It

[violet heron]

seems like multiple people are having issues with colab rn

[quiet ginkgo]

Oh

[brittle wing]

any similar alternative

[violet heron]

Kaggle

[quiet ginkgo]

Yeah It's that cuz I'm having that as well rn

[violet heron]

-kaggle

[azure marsh]

Kaggle Notebooks

• 🆕 UVR5 NO UI by Eddy
• How to use RVC Mainline Kaggle by Cauthess
• 🍏 Applio RVC by IA Hispano
• ✨ RVC Mainline by Hina

More commands: -colabs, -hf, -local, -rvc, -help

[violet heron]

Dunno if they would still work

[viscid moss]

hmmmm

[quiet ginkgo]

Does It work well on CPU training

[viscid moss]

also Web UI

[brittle wing]

any alternative with a UI

[violet heron]

Kaggles and Colabs use GPUs there are hosted on the cloud so your specs don’t matter

[violet heron]

For covers?

Or training

[brittle wing]

covers

[violet heron]

Use Ilaria RVC zero on hf

[proven hill]

Ilaria RVC Zero

[violet heron]

-spaces

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• 🆕 UVR5 UI, by Eddy and Ilaria Huggingface Spaces
• Ilaria RVC Zero, by Ilaria Huggingface Spaces
• RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[viscid moss]

a

[low shard]

adding that !pip install pip=23.1 worked even if there's some errors

[quiet ginkgo]

Doesn't GUI work fine

[low shard]

it works fine but still shows those idk why