✨│ai-help

[brittle wing]

-colab

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[quaint dock]

sure, lemme start one again

[rare goblet]

Ayo? @quaint dock level 3 !!!

[brittle wing]

why doesn

t

it launch

[fervent tusk]

OK thanks

[proper shale]

Give us some useful information, like GPU, what you're using, what happens, etc

[brittle wing]

rtx 3060

and i load the cmd bat file

and do what it says for me to do

and it doesn

t

load up

like no program loads up after it tells me to press any button

[proper shale]

What fork are you using

[brittle wing]

real-time-gui.bat

[proper shale]

Oh, I'd recommend using W-Okada

[brittle wing]

It doesn't load for the covers either

and W-okada loads fine

i just don't want to sound shitty like its a boom mic

its not a boom mic

i have a good headset

i just idk why its acting up

it doesn't act up

its just the quality of the sound

people hear

it sounds for people like a boom mic or smth i think

[proper shale]

in/out gain settings might be baf

bad

[brittle wing]

are there good settings for it?

like what setting should i put it on

[fervent tusk]

If i am going to use RVC Disconnected to make a model whats the maximum dataset size i should upload? Currently I have 12 Wav files at 8 Gb uncompressed. Is this even an issue?

[brittle wing]

Oink oink

[soft stratus]

How i fix it

[winged condor]

Hello all I could really appreciate some help here as I've been really trying to follow the guides I've found here but I'm stuck. I've been trying to train a model and I'm not sure where I'm hitting a roadblock. Here's where I'm at:

• I successfully downloaded the Manga RVC fork and am running it locally on my mac
• I created a folder for the data i wanted to train the model on title "tony" with a 12 minute mp3 file in the folder.

In the screenshot it is saying:
-move model to cpu no-feature-todo

• Unfortunately, there is no compatible GPU available to support your training.

Am i doing something wrong or is training a model locally just not possible?

[soft stratus]

How i fix error

[fervent tusk]

Im having a similar problem using mangio rvc fork. From what im seeing, training a new model on mac locally is not possible and you have to use a google colab but im hoping this is wrong information and there is some way to do it.

[proper shale]

There isn't a GPU available so you can't train

[proper shale]

What GPU do you have?

[soft stratus]

I'm using GPU

[proper shale]

But it says there is none

So you can't train

[soft stratus]

But how i use GPU

[soft stratus]

But how i use GPU

[proper shale]

I don't know

What are you currently using to train

[soft stratus]

Easy GUI

[proper shale]

In Google Colab?

[fervent tusk]

Looks like were all having the same issue. These are my specs. Should i be able to trian locally?

[rare goblet]

Ayo? @fervent tusk level 2 !!!

[soft stratus]

Yes

[proper shale]

Yeah try RVC Disconnected

in #1159949278270193734

[soft stratus]

You mean colab

[proper shale]

The RVC Disconnected colab, yes

We call those "Notebooks"

[soft stratus]

But i tried but no way

[proper shale]

Then you're doing something wrong then

[soft stratus]

What is wrong i do

[proper shale]

That's the thing

I don't know

It's better to train on PC because phone is very limited

Those can be fixed if u have a decent PC, u can even run it on your own machine

[median monolith]

What is the best collab to make a model or cover on mobile?

[proper shale]

It's generally not recommended to make models on a phone, but it is possible by using RVC Disconnected. Check #1159513888199540817 and #1159949278270193734 for the guide and the link for the Colab, respectively.

[winged condor]

I'm not sure what GPU I have but my mac says that I have an M1 Chip and a "GPU" type

[proper shale]

I believe M1 can't train

[median monolith]

Well, im just trying to do it on a phone because my notebook is currently not available (its screen is all broken). But thanks for the lil help

[rare goblet]

Ayo? @median monolith level 1 !!!

[proper shale]

Or maybe you can if you use harvest as your pitch extractor

[proper shale]

Oh, alrighty, np ;)

[median monolith]

Btw. Does illaria rvc also works on mobile?

[proper shale]

Yeah

[median monolith]

Good

[winged condor]

Ok thanks if I can't train but can find a trained model I want to use would you recommend RVC Disconnected instead of running locally? Sounds like thats a better path

[rare goblet]

Ayo? @winged condor level 1 !!!

[proper shale]

For using a pre-existing model, your PC is fine

Thing is, these M1 Macs don't have support for some of those pitch extractors because of CUDA cores n stuff

RVC Disconnected is for training

[winged condor]

ah got it

[proper shale]

So if you wanna train models, thats a way

(I am also a Mac user haha, gonna upgrade to a M1 soon hopefully)

[winged condor]

I'm just trying to find the path of least resistance to cloning a specific characters voice over songs, I know of eleven labs and play.ht for text to speech and they work well for that but to get the inflection of someone's singing voice seems like RVC is the way to go I just need to find the right tool and someone who has cloined the voice im looking for

[proper shale]

I see, I see

For singing RVC is the best

[winged condor]

is there a scenario where I can use Google Colab and plug in the trained model if theres one already made for the voice im looking for?

[proper shale]

You can use local for inferences

Since the M1 can do that

• realtime too

[winged condor]

got it thanks I'll browse some more in the guides section but do you know the best resource for learning about using local? Sorry I'm not fully up to speed on this stuff

[proper shale]

It's okay, it's like that in the beginning. Anyways, the All In One Documentation in #1159513888199540817 is a pretty good one! #1159636677178953802 message

[winged condor]

perfect thank you so much for your patience with me! One last question - if I were to invest in a non Apple laptop, what's the most affordable one that could get the job done for what I'm looking to do?

[proper shale]

Since I'm in Brazil, I can't really give you good models because pricing in here is... horrible. But I can help you with the specs:

CPU won't matter much, but any i7 11th-13th gen (or Ryzen 7, 5000 or 7000 series) will get the job done.
GPU, an RTX 3060 would be great for AI training - it's the most important part of training
RAM, 16GBs is enough
SSD, I'd recommend 512 GBs at least, if u wanna do other stuff

[brittle wing]

Elliot Alderson voice ?

[buoyant crescent]

Dear All, beginner here so please excuse the silly question. I'm wondering what is the biggest factor that determines the quality of a voiceover cover song. Is it the quality of the model itself or are some postprocessing steps usually required for good results (which ones?). For example, this cover sounds amazing and I don't notice any artefacts at a glance:

https://www.youtube.com/watch?v=N0_NqE2cuvs

Is it that the model here is so good or are there some secret tricks to make a song sound convincing?

An RVC v2 model I trained on ~40min of good quality a-capella singing and 350 iterations sounds nowhere near as good and gives various distortions (e.g. occasional buzzing sounds). I'd appreciate some input here. Many thanks.

[proper shale]

A cover generally has a bit of postprocessing, but it's just mixing stuff normally. The thing about the models is that you need to make sure they are as clean as possible, and that the inferencing audio is also clean.

[buoyant crescent]

Thanks for the quick reply! What qualifies as clean? I used an album that contains only vocals, cello and piano and isolated the vocals with "MDX-UVR Vocal Model Voc_FT", then removed the reverb with "(de)Reverb HQ By FoxJoy", applied some noise filtering and gating in audacity and manually inspected the entire track. Is this good enough or is the fact that I don't have original studio vocals a prohibiting factor?

[proper shale]

A perfectly clean dataset would have to be: denoised manually, have no reverb or echo, and only the voice, so no background vocals allowed.

Not having studio vocals is not a prohibiting factor, especially for covers. It will still make quality worse than a studio dataset, though, because those filtering methods have instrumental bleeds on the vocals, even if they're not able to be heard.

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[buoyant crescent]

thanks a lot for your help!

[rare goblet]

Ayo? @buoyant crescent level 1 !!!

[proper shale]

No problem :)

[placid storm]

what we have to write on this @violet heron

can some one help me

pls

[errant shadow]

Hi

I was a newbie

How to train the model?

[vapid gust]

@acoustic thorn how to train the model

[long terrace]

Hi! I have a question

Now the best choose for make covers its Replay or still existing a good collab for avoid desconections? (I dont have HW for have it local)

[true geyser]

Hey, does anybody knows how to save my model on collab mid training so i can resume it tomorrow?

[proper shale]

Just stop the training, click on Export Model. Then just insert the same settings and use the resuming steps and the training step when you wanna return training.

This on RVC Disconnected

[long terrace]

If I close the collab its the same steps for re-train the model or is not possible?

[reef walrus]

Hi

Preciso un audio de Cristian Ronaldo pls

[wheat copper]

how the hell do i use rvc on paperspace

not mangio rvc

[wicked marsh]

Hello! This problem is probably answered already, but when I try to refresh on either the easier GUI or the regular go-web, it queues for a very long time and doesn't complete. What would be the best course of action to fix this?

I am using the mangio rvc on local

[true geyser]

thank you ❤️

[true geyser]

should I run code for index training or just simply export model to drive?

[rare goblet]

Ayo? @true geyser level 1 !!!

[twin orchid]

hi guy
i have a queation
how i continue training with my old model?

[robust vessel]

How to install gui for applio local? Please tell me 😭

I did not find it here https://docs.aihub.wtf/

[latent kettle]

I'm using webui on Google colab!

[nocturne mural]

https://docs.applio.org/

[open mural]

Idk how to use the 2 ai zip files for a ai cover 🥲

Oh wait theres some links

[rare goblet]

Ayo? @open mural level 1 !!!

[marble forge]

Field required [type=missing, input_value={'data': ['48k', True, 'v...on_hash': 'd0leqin86jf'}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.4/v/missing>
Traceback (most recent call last):
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 347, in process_events
client_awake = await self.gather_event_data(event)
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 220, in gather_event_data
data, client_awake = await self.get_message(event, timeout=receive_timeout)
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 456, in get_message
return PredictBody(**data), True
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\pydantic\main.py", line 164, in init
pydantic_self.pydantic_validator.validate_python(data, self_instance=pydantic_self)
pydantic_core._pydantic_core.ValidationError: 1 validation error for PredictBody
event_id
Field required [type=missing, input_value={'data': ['48k', True, 'v...on_hash': 'd0leqin86jf'}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.4/v/missing
2023-11-17 22:25:44 | ERROR | root | Task exception was never retrieved
future: <Task finished name='d0leqin86jf_15' coro=<Queue.process_events() done, defined at D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py:343> exception=1 validation error for PredictBody
event_id
Field required [type=missing, input_value={'data': ['mangio-crepe']...on_hash': 'd0leqin86jf'}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.4/v/missing>
Traceback (most recent call last):
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 347, in process_events
client_awake = await self.gather_event_data(event)
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 220, in gather_event_data
data, client_awake = await self.get_message(event, timeout=receive_timeout)
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\queueing.py", line 456, in get_message
return PredictBody(**data), True
File "D:\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\pydantic\main.py", line 164, in init
pydantic_self.pydantic_validator.validate_python(data, self_instance=pydantic_self)
pydantic_core._pydantic_core.ValidationError: 1 validation error for PredictBody
event_id
Field required [type=missing, input_value={'data': ['mangio-crepe']...on_hash': 'd0leqin86jf'}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.4/v/missing

[marble forge]

it just infinitly loads and outputs that into the console

[nocturne mural]

open cmd in mangio folder and put
runtime\python.exe -m pip install fastapi==0.88.0

runtime/python.exe ?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[quartz garnet]

Say i'm a bit new to this is there a way to use Kit.ai with Client?

like download models

[glad zealot]

No

[latent kettle]

What can I do to fix it ?

[glad zealot]

what is that? training? cover gen?

[latent kettle]

While training a model I have to face this issue.

[glad zealot]

which notebook is it? rvcDisconnected?

[latent kettle]

https://colab.research.google.com/github/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/Retrieval_based_Voice_Conversion_WebUI.ipynb

[glad zealot]

ahh yup that will die

[latent kettle]

What ?

[glad zealot]

that wont work

it gets terminated by colab

[latent kettle]

So what can I do?

[glad zealot]

you can use rvcDisconnected tho

-colab

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[glad zealot]

2nd link

[latent kettle]

Thanks for your support !!

[tribal cosmos]

yo guys how u download appolio on github
https://github.com/IAHispano/Applio-RVC-Fork

also whats this??

[tulip frigate]

guys i need some help

which model has the most "Umm, hmm, ah!, oh?" sounds?

cause none of the ones i found can register anything beyond words

i mean like pausing phrases

[copper pebble]

Ayo??

[tulip frigate]

nah bro

[rare goblet]

Ayo? @tulip frigate level 1 !!!

[tulip frigate]

im making a webshow

i need pausing words

[copper pebble]

did u record them?

[tulip frigate]

yeah

but my voice sounds too weird, like i want to convert a sigh into the characters voice too

but it sounds like hell when converted

u know any models that support this?

[copper pebble]

maybe u can try to use a tts

like 11labs

pretty good for dubbing

[tulip frigate]

im not using eng tho

[copper pebble]

there are a lot of them

[tulip frigate]

hmmm

which ones trained

alot

like aloooot

[copper pebble]

there are multilanguages

[copper pebble]

watchu searching for more specificaly

male/female

[tulip frigate]

both

[copper pebble]

mhh

[tulip frigate]

see

u did it right there

u used a pause

word

thing

i need to get that

[copper pebble]

i made a voice model of a radio speaker maybe that can help u

[tulip frigate]

well i was hoping for a person

o wait

u mean

like a host

on the radio

[tribal cosmos]

yo guys how u download appolio on github, i made dataset and i wanna train it
https://github.com/IAHispano/Applio-RVC-Fork
also whats this??

[copper pebble]

https://discord.com/channels/1159260121998827560/1171492691943624886 heres the voice

maybe this can fit u

[tulip frigate]

thanks bro

[copper pebble]

try to use 11labs for creating sounds

u can also just use its voice

without rvc models

[tulip frigate]

it dont have japanese tho

[copper pebble]

yh it has

[tulip frigate]

🤔 really?

[copper pebble]

yh

[tulip frigate]

dam alr ill try again, last time i went there it said only eng models

[copper pebble]

did some dub the other day

[tulip frigate]

i will check it out rn

[eternal thunder]

Hey there,

is there a guide somewhere that helps me to journey from zero to ai song creation and ultimately to TTS with your self-created ai model?
I looked at https://docs.aihub.wtf/vocal-cloning-guide/vocal-and-music-separation and #1159513888199540817 but to be honstet, i dont know where to start.

I have no issue with documentations, also got developement knowledge in different coding languages. i appreciate any advice ❤️

[tribal cosmos]

yo guys how u download appolio on github, i made dataset and i wanna train it
https://github.com/IAHispano/Applio-RVC-Fork
also whats this??

[fervent tusk]

HI, whats the best way to uninstall mangio rvc fork? I wanted to train with it locally but it turns out thats not possible on mac M1.

[solar jacinth]

Maybe Delete it

Check out Ilaria RVC

[fervent tusk]

that works with M1?

[solar jacinth]

It's not on local

It's on hugging face

Hold up , let me link that

[fervent tusk]

Thanks. Would you recommend that above rvc disconnected?

[solar jacinth]

https://discord.com/channels/1159260121998827560/1167844377046040769

[proper shale]

You must export the model or else you won't be able to retrain

[proper shale]

Train index and export model

Saves time

[proper shale]

If you're using RVC Disconnected, make sure the model you have was trained there. Then just insert the same settings and use the resuming steps and the training step when you wanna return training. Export model once you've reached your desired epochs

[proper shale]

Ilaria RVC is for inference only

Mangio is good already for inference

It's probably gonna take longer to infer in Ilaria's space than using locally

[copper pebble]

yo lil gojo how r u?

[fervent tusk]

Yeah and i need to make the model from scratch

[proper shale]

Yeah you need to use RVC Disconnected for making models

And then u can normally use the mangio fork for inference

[solar jacinth]

I become a KitKat 😭

[hexed pagoda]

i wanna make ai covers and i have rvc gui but when i start running it it doesnt work, how do i fix that

[solar jacinth]

Oh my god

[copper pebble]

did u try any hf spaces to do covers online??

[hexed pagoda]

whats that

[half cove]

[copper pebble]

ilaria rvc for example

[hexed pagoda]

liaria?

[rare goblet]

Ayo? @hexed pagoda level 1 !!!

[hexed pagoda]

ilaira

[half cove]

https://huggingface.co/spaces/TheStinger/Ilaria_RVC?duplicate=true

[proper shale]

NOOOOOO

[copper pebble]

im sorry

[proper shale]

I WAS WATCHING JJK

[hexed pagoda]

oooo

[solar jacinth]

This one

[hexed pagoda]

thank youu

[solar jacinth]

Fugaa !

[copper pebble]

left the chat

[proper shale]

demoted

[half cove]

newest episode was amazing

[copper pebble]

ma come fuga'??

[proper shale]

I CAN'T SEE IT

[solar jacinth]

That Fuga gave me chill

[proper shale]

I HAVE LIKE

[half cove]

https://cdn.discordapp.com/emojis/1143729437200887868.webp?size=48&name=whygod&quality=lossless

[copper pebble]

happened to me too

[proper shale]

OLD ASS EPISODES

i gotta pay for Crunchyroll

[half cove]

f

[solar jacinth]

I got cubed and after that I become a KitKat

[proper shale]

i thought yo ass died on episode 28 or something

[solar jacinth]

I'm still pirate

[proper shale]

piracy is bad...

[brazen gull]

Excuse me, i recently just tried to use the RVC. I got a double-voice problem, somewhat like an echo. Does anyone know the problem and how i can fix it?

[proper shale]

Have you checked the Echo option and increased S-Thresh? This helps most of the time

[brazen gull]

Yep, I did

[half cove]

[brazen gull]

I unticked Echo, Sup1 and Sup2

Thresh on 0.001 which is max

[hexed pagoda]

how you use that?

[proper shale]

I'd recommend ticking Sup2

And echo

[half cove]

#1159513888199540817

[brazen gull]

Still the same result

I've been trying for like 10 minutes and still can't get away from the echo problem

[rare goblet]

Ayo? @brazen gull level 1 !!!

[proper shale]

Are you using the monitor option

[brazen gull]

Oooh

That fixed it lmaoo

[proud elbow]

set monitor to your headphones, not your speakers

[brazen gull]

Alright alright

Thanks everyone XD

That was such a simple fix

[proper shale]

It's alright

We all have problems in the beginning

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[true geyser]

I'm using google collab and I want to resume training my model but after i've set everything in training tab i got an error
NotADirectoryError: [Errno 20] Not a directory: '/content/Mangio-RVC-Fork/logs/witunia/0_gt_wavs'
Any ideas?

[modern oxide]

what colab are you using?

disconnected?

[true geyser]

yes

[modern oxide]

when you were training did you press this section?

[true geyser]

yea, i did

[modern oxide]

and when you tried to resume it, did you load it back with this section?

[true geyser]

yes

[modern oxide]

hmm

sorry, no idea why it doesn't work

[true geyser]

aight, thanks for your time

[brittle wing]

what do i do here

[modern oxide]

type in the name of your zip file and launch it

[brittle wing]

how can you get free colab gpu? @modern oxide

@modern oxide

[modern oxide]

what?

google just gives you possibility to use their Tesla T4 gpu for everyday

they just give limited time

[brittle wing]

can you appear online even though your offline? @modern oxide

[modern oxide]

emm, yeah, it's a regular discord feature

[brittle wing]

oh

im so confused

its my first time making a ai model....

[modern oxide]

it's fine, you can read guides in #1159513888199540817 to learn about it

and if you'll have any questions after that you ask them here, and our helpers and mods will help you

[brittle wing]

@modern oxide how many epochs for a 22 minite model? is it too long?

[rare goblet]

Ayo? @brittle wing level 9 !!!

[modern oxide]

yeah, i'd say it's a bit long

10 minutes is maximum imo, but even 5 minutes is enough

[modern oxide]

usually people put like 1000 epochs and then look at the tensorboard to see when the model starts to overtrain

[brittle wing]

@modern oxide how many epochs for a 10 minite model

[modern oxide]

put 600 and then send me a screenshot of the tensorboard each 100 epochs

when the tensorboard will appear go to the scalars tab, here write down g/tot

it'll show the graph that you'll need to look at

[brittle wing]

how do i fix this @modern oxide

[stark sand]

Does anyone know how to fix discord mic cutting out

[brittle wing]

@modern oxide

[modern oxide]

do you use rvc v2 disconnected colab?

[brittle wing]

yes]

yes i do

[tribal cosmos]

i got apollio how do i train

[modern oxide]

tbh i don't use it that often, i use easy colab, because in my opinion it's better and easier to use

[tribal cosmos]

ohh ok

[modern oxide]

no it wasn't addressed to you

[brittle wing]

@modern oxide it just disconnects me after 200 epochs

[modern oxide]

easy colab?

[brittle wing]

yes

i remember using it....

[modern oxide]

probably because you were putting very big datasets, and it took too much time training it

[brittle wing]

ohh

overtraining

i see

[modern oxide]

if you put in 5-10 minutes it'll train them quickly

[modern oxide]

no, overtraining is when your g/total loss graph starts to rise instead of going down

[hot lagoon]

Was the best Ensemble mode to use in uvr ?

[brittle wing]

@modern oxide how do i fix this

@modern oxide

[modern oxide]

no idea, i don't use disconnected that much to know it's problems :c

[glad zealot]

dont mind me

-colab

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[glad zealot]

just me forgetting the link

[hot lagoon]

What shoud I do for Ensemble mode for vocals and instrumental

[half cove]

ill type this here before people ask

VOICE MODEL CHANNEL HAD A LITTLE PROBLEM IT WILL BE RESTORED SOON

Use this link to search models in the meantime:
https://www.weights.gg

[finite galleon]

rvc, sovits svc, and rvmpe?

[marble forge]

Let me try

[sick light]

How do i use models i find #🔍│find-models in colab?

[brittle wing]

can you help me with this dataset please @modern oxide

@modern oxide

[modern oxide]

yeah

i'm here

how can i help with the dataset?

[brittle wing]

can you make it into a model please?

[rare goblet]

Ayo? @brittle wing level 27 !!!

[brittle wing]

i just keep getting errors

[modern oxide]

emm, i'm afraid i can't do it for free

[brittle wing]

@wide parcel

[wide parcel]

i can when i get home

probably 6pm est

[eternal thunder]

hey guys,

why cant i find the train tab?
only have inference and download model.

i got Easy GUI v2 (rejekts) - adapted to Mangio-RVC-Fork 💻 [With extra features and fixes by kalomaze & alexlnkp] running locally.

[molten pecan]

@spare spear

The huggingface spaces

https://discord.com/channels/1159260121998827560/1167844377046040769

[spare spear]

ok so what app are you using to run the models?

and thank you very much for helping

[molten pecan]

I either use RVC locally or use huggingface

And you?

[spare spear]

I got MMVCSererSIO from huggingface

from a github repo

[molten pecan]

Interesting

Can you send a sc of how it looks like?

[spare spear]

isn't that what I should use?

[molten pecan]

Well you can use whatever it's more comfortable to you

[spare spear]

I am not confortable with any xd

I will try to use yours

[molten pecan]

Sure then

Ping me if you need any help!

[spare spear]

looks like this. I uploaded a model from this discord but no picture

and it sounds like a robot

[molten pecan]

Ohhhh you are using the real-time voice conversion

I thought you were converting audio files for song covers and stuff

I'm not familiar with w-okada

So try asking on #🔍│help-w-okada

[spare spear]

ahhhh

i thought you all used real time voice changers

because I want to prank my frinds on discord

friends*

[molten pecan]

We mostly use apps for converting audio files, not converting in real time with the microphone and stuff

Good luck with that!

[spare spear]

thank you

[rare goblet]

Ayo? @spare spear level 3 !!!

[eternal thunder]

hey guys, what are "normal" preprocessing times for training?

[rare goblet]

Ayo? @eternal thunder level 1 !!!

[eternal thunder]

@molten pecan

[molten pecan]

Idk about making models

Try asking on #📑│making-models

Good luck!

[fervent tusk]

Using rvc disconnected and i get a
NameError: name 'save_only_latest_ckpt' is not defined whie running the export cell. Any ideas how to fix?

[soft stratus]

I tried with your way but no way

[grand willow]

the models tape disappeared

[heady current]

Look at the announcements bro

[half cove]

Look at pinned messages

[wide parcel]

“a little problem”

[proper shale]

¯_(ツ)_/¯

[subtle glen]

hello

how can i download the app

the place i was downloading the app had problems

[proper shale]

which app

[subtle glen]

voice changer

[rare goblet]

Ayo? @subtle glen level 1 !!!

[proper shale]

#🔍│help-w-okada message

read this, it has download links

[subtle glen]

ok thank you

[soft stratus]

?

[soft stratus]

Answer me

[soft stratus]

What is this

[proper shale]

I don't really know how to fix ur issue

That's it

[soft stratus]

https://drive.google.com/file/d/1oc4WQeiEknnG2JJ_4I9D7rfp_kl1ysz6/view?usp=drivesdk

Here's my issue

[soft stratus]

Try with arabic and tell me

[soft stratus]

But can you train my voice model

[proper shale]

If you pay, sure

[soft stratus]

How much

[rare goblet]

Ayo? @soft stratus level 5 !!!

[soft stratus]

Is it free

[proper shale]

$10

But again, I'd recommend reading the guides n stuff and try it again

[soft stratus]

But after i trained my model

[soft stratus]

I tried the site which you're talking about... I failed

[brittle wing]

Can anyone help me clean up a recording?

[distant turtle]

I have a disconnected runtime problem, my computer is a little old, but it works to clone voices, the problem is that if I don't hurry and do all the steps to clone my voice, it gives me a disconnected runtime and I have to do it all over again, do you have a solution beyond this code?

// click connect button const clickConnect = function() { console.log("Working"); document.querySelector("colab-toolbar-button#connect").click(); } // interval to keep clicking the connect button every 60 seconds const keepClickingConnect = setInterval(clickConnect, 60000);

,it tells me once activated that it worked but then tells me runtime disconnected again, anyone?

[proper shale]

What Colab notebook are you using?

[distant turtle]

I'm using this site Easy GUI (for RVC v2, with crepe)

[proper shale]

Oh that explains ig then

it*

That Colab is restricted, even with this code it will not prevent disconnections.

[distant turtle]

is it paid? since it asks me to pay for colab pro?

this site also gives me the runtime disconnected but I haven't tried the code!

[rare goblet]

Ayo? @distant turtle level 1 !!!

[distant turtle]

RVC Crepe v2 + persistent storage / autosave

[heady current]

Hello friends, has anyone attempted to use the .pth file generated for TTS using Azure? I currently have a script in Python that uses Azure speech to read text but I'd like to change the voice it's using.

I saw people using the voice changer to kind of mimick that in real time but Im looking to use a full conversion inside the code like rvc does

[cyan osprey]

how long will it take to train 100 epochs on 6h dataset

also how to know if the model is overtraining or not

[lusty gulch]

Hi guys! What is the best webui for rvc vocal to vocal, tts, model training etc.

[weary fable]

So How do I use a speaker and how do I separate multiple speakers? Just new to this whole thing since voices are separative. including with separating multiple vocals in 1 mixed stem to separate speakers.

[cyan osprey]

karaoke models in uvr for 1 mixed one, probably wont work lots of the time

[midnight bison]

-colab

[cyan osprey]

for seperate just extract the vocals and just cut out the parts for each speaker

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[lusty gulch]

Hi guys! What is the best webui for rvc vocal to vocal, tts, model training etc.

Could use some help since their are so many github projects

[stuck ledge]

Which voice changer to use with these voice models?

[weary fable]

Actually what I am trying to do is separate multiple speakers.. thats what I am guessing. <_<

[rare goblet]

Ayo? @weary fable level 2 !!!

[stuck ledge]

Which voice changer to use with these voice models?

Which voice changer to use with these voice models?

Which voice changer to use with these voice models?

[astral jungle]

Hi

[soft stratus]

@proper shale i want to help me to making voice to another characters voice like SpongeBob, April O'neil and etc

[fallen oriole]

ModuleNotFoundError: No module named 'dotenv'

[wispy burrow]

Trying to install Applio?

[fallen oriole]

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/docs/en/README.en.md

[rare goblet]

Ayo? @fallen oriole level 2 !!!

[fallen oriole]

I wanted just update

then a guy suggested that

😄 and now broked

[wispy burrow]

Ah, normal RVC?

[fallen oriole]

yes

I was using RVC-beta0717

[soft stratus]

Did you not know how i create voice to voice like Shante etc

[distant turtle]

RVC Crepe v2 + persistent storage / autosave.
it gives me the runtime disconnected error, even though I used the codes, does anyone know why, my PC is a bit old

[wispy burrow]

Which one are you using or trying to use now?
I don't see that in the requirements...

[distant turtle]

?

[violet heron]

RVC

[violet heron]

RVC

[violet heron]

RVC

[violet heron]

That one doesn’t work anymore

-colab use one of these instead

[azure marsh]

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[distant turtle]

which one do I choose?, sorry but I'm not practical, I just have to train a model

[wispy burrow]

pip install python-dotenv
That should work...

[soft stratus]

Answer me how i convert our voice to female and male characters voice

[proud gyro]

Excuse me, but what exactly does crepe hop do?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[robust vessel]

I started training my model on Applio locally, is that ok what I see in the terminal?

[sand iris]

Keep in mind the runtime is what they are most likely using

[wide parcel]

install dotenv

[sand iris]

Looks normal to me ? I don't recall that log showing up that many times, but it's the usual deal

[wide parcel]

most uis say but last i checked is how quickly pitches can change

[wide parcel]

usually that clears out and it continues on with normal <epoch 1> (other info here)

if it’s just warnings it’s probably fine

probably

[robust vessel]

thanks! I'm used to it writing epochs in the console, but they don't exist here, so it bothered me

[wide parcel]

it generally does?

[drowsy grove]

i want to train my model but it wont let me in pth format what should i do

[wide parcel]

said logs usually clear after a moment

[wide parcel]

what ui

[proud gyro]

I see, so bigger values allow for faster pitch changes?

[wide parcel]

smaller is faster, not too small tho afaik

if you go too small it might pick up random noise or smth

[proud gyro]

alright thanks

[drowsy grove]

sorry for not understanding but whats an ui english its not my natiive language😭😭😭

[wide parcel]

what webui
mangio
applio
rvc

etc

[drowsy grove]

oh rvc

[rare goblet]

Ayo? @drowsy grove level 1 !!!

[wide parcel]

doing..?

[compact blaze]

how to install pytortch

[mellow flax]

type it on google like... 💀

[compact blaze]

did

brings me

to github

https://github.com/pytorch/pytorch

this one

[mellow flax]

https://www.youtube.com/watch?v=2RkK73h0BGY

should work

[teal shell]

how do I download the modles from https://app.kits.ai/ ?

[rare goblet]

Ayo? @teal shell level 1 !!!

[mellow flax]

-local

[azure marsh]

All Working Local Forks 🖥️

• Applio-RVC-Fork, 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 Huggingface

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[wispy burrow]

I checked several normal RVC versions but can't find it in any of the requirements files...

[distant turtle]

which one should I choose to train a model?

[soft stratus]

How i convert voice to character voice in rvc

[sand iris]

It should be pulled in as dependency

[soft stratus]

How i convert voice to character voice in rvc

Answer me

[compact blaze]

CUDA error: no kernel image is available for execution on the device CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with TORCH_USE_CUDA_DSAto enable device-side assertions. Traceback (most recent call last): File "voice_changer\VoiceChangerV2.py", line 265, in on_request File "voice_changer\RVC\RVCr2.py", line 248, in inference File "voice_changer\RVC\pipeline\Pipeline.py", line 161, in exec RuntimeError: CUDA error: no kernel image is available for execution on the device CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile withTORCH_USE_CUDA_DSA to enable device-side assertions.

[rare goblet]

Ayo? @compact blaze level 1 !!!

[compact blaze]

wth is this

[sand iris]

What's your GPU

My guess is that it's too old

[sand iris]

what ?

[soft stratus]

I want to convert our voice to Princess Peach and etc voices in RVC

[sand iris]

real time ?

Or convert files ?

[craggy light]

I need help, I wanna make a cover of Blitz singing Stay but I can’t find a good cover with the key I want. What do I do? And no I can’t sing it because I’m not a bass.

[compact blaze]

maybe the problem was i havent installed the

pytorch yet

but now i installed it

so it should not give me any errors

[wide parcel]

if i had to assume the input file isn’t valid or your pth/index isn’t valid

[wide parcel]

bro demands to be answered

[soft stratus]

Where

[distant turtle]

for this site: Easy👨- 10/25, can I use just one long audio?, how much do you recommend, is 5 minutes okay?

[wide parcel]

#1159513888199540817

[soft stratus]

Which is convert @wide parcel

[olive olive]

do the file need that we gonna train need to be mp3 or mp4 or it doesn't matter ?

[distant turtle]

for this site: Easy👨- 10/25, can I use just one long audio?, how much do you recommend, is 5 minutes okay?

[rare goblet]

Ayo? @distant turtle level 2 !!!

[sand iris]

What's your GPU

It most certainly isn't a pytorch error

You wouldn't get anything like that if it wasn't already running

[wide parcel]

inference

[wide parcel]

any audio
wav flac mp3

etc

in preprocessing they’ll just be converted into a wav tho

[distant turtle]

?

[olive olive]

Ok thx

[native agate]

yo what conveter do u use

[wide parcel]

?

[native agate]

cause mine isnt working for some reason

[wide parcel]

wdym converter

[native agate]

idk convert the voice

[wide parcel]

like for rvc?

which fork u mean?

[native agate]

i guess

i used the easy gui and it isnt working

[distant turtle]

for this site: Easy👨- 10/25, can I use just one long audio?, how much do you recommend, is 5 minutes okay??

[native agate]

any amount i guess

[compact blaze]

so if the GPU is old means theres no other way to make ethis voice changer work

[sand iris]

Yes unfortunately

[native agate]

can someone give me a conveter

converter

[robust vessel]

what do I do if my applio stopped training because of the timeout (1800000 ms)?

[rare goblet]

Ayo? @robust vessel level 2 !!!

[violet heron]

Click Ilaria RVC

[native agate]

no 3 dots

oh

then what

@violet heron

[violet heron]

3 dots -> settings -> scroll down and click “make public”

Go back to the page and try doing it again

[native agate]

beside the 3 dots?

says settings beside the 3 dots

yooo finally wowrked

[cloud violet]

how do i launch rvc i just downloaded it

[violet heron]

Extract all then open the BAT file

If the file you downloaded was just a bat file, open that and wait for it to install

[cloud violet]

thanks

[olive olive]

when i train my model why i don't have .pth file but only index

i train locally

[distant turtle]

with this site: RVC v2 Disconnected, I did everything until I got to training, does it stop automatically?, meaning does the command continue endlessly or does it stop once it's finished?

[soft stratus]

Is Ilaria can convert voice to spongebob voice and etc

Answer me

[midnight terrace]

https://youtu.be/Z2T1K-xN6QQ?si=jmvjOPBkxaJzHfCm

this is the song I wanna acapella

but im bad at uvr

[molten pecan]

First of all, is the song only available on YouTube?

[rare goblet]

Ayo? @molten pecan level 43 !!!

[molten pecan]

A

[rare goblet]

Vote Booster: 10% (5 hours remaining)

[midnight terrace]

dexndre is nugu but you can prob find it on other sites

I usually just convert it to an mp3

[molten pecan]

I swe

I'm asking because

When you upload something to YouTube, the platform compresses the quality of it

[midnight terrace]

oh yeah

[molten pecan]

Overall for the best results it's best to use an audio with a really great quality

If you want to use UVR or RVC

[molten pecan]

Try downloading the YouTube song but in wav format

mp3 is an audio format that compresses the quality in exchange of having a smaller size

Meanwhile wav doesn't compress the quality and doesn't lose any data about the song (lossless quality)

[midnight terrace]

unless theres a good soundcloud converter it would have to be mp3, atleast idk any other ones for yt

[molten pecan]

Oh you can download music from YouTube in any format

I use this website

https://y2down.cc/es/youtube-wav.html

Let me know when you are done downloading!

[midnight terrace]

is it a virus

[molten pecan]

No of course not 😭

I have been using it for years in a bunch of devices

[midnight terrace]

my dad is the police chief in florida he will arrest you if its a virus !!11!!11!11

OH FR

oops

caps

[magic badge]

Hi there, someone can give me applio link to make covers please

[rare goblet]

Ayo? @magic badge level 1 !!!

[magic badge]

:0

[molten pecan]

@magic badge -applio

[azure marsh]

All Working Local Forks 🖥️

• Applio-RVC-Fork, 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 Huggingface

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[molten pecan]

!!

[midnight terrace]

my internet is slow holdon

[rare goblet]

Ayo? @midnight terrace level 2 !!!

[molten pecan]

Yes

[midnight terrace]

it blue screened me

[magic badge]

My internet sometimes have 528 kb/s ;-;

[midnight terrace]

mine usually downloaded games at 20mb but today it was 235

wifi was cooking gorme

[magic badge]

F

[midnight terrace]

I did it

[molten pecan]

Okay now

How fast is your PC?

[midnight terrace]

uh

[molten pecan]

Is it a gaming PC or?

[midnight terrace]

its usable

ive made uvr things in a few seconds before if I use my gpu

its a 1070ti

[molten pecan]

Oh shit

You are lucky 😭

Alright then

[midnight terrace]

its kinda dying

got it in 2020

[molten pecan]

So you are trying to get the vocals from that song?

[midnight terrace]

yeah

[molten pecan]

Open UVR and select that audio file

[midnight terrace]

is loading

[molten pecan]

In Windows Size select 320

[midnight terrace]

is there a new uvr download, I prob got this downloaded in july

[molten pecan]

And for the algorithm use MDX23C

Uhm maybe. Try checking if there is any update

[distant turtle]

it's normal that on this site: RVC v2 Disconnected, it took me 2 minutes to pull the model, I didn't find the index file but only the pth, I don't know if on this site they don't give you the index file, and the result doesn't It's similar, I've set everything right I think

[midnight terrace]

oh it can update

[soft stratus]

@molten pecan Is Ilaria can convert voice to spongebob voice and etc

[molten pecan]

Yeah of course. Any model

[soft stratus]

But i tried with my model it failed

[molten pecan]

Are you logged in with your account in huggingface?

[soft stratus]

@molten pecan no

[molten pecan]

Otherwise you'll get errors

It's best if you log in first

[midnight terrace]

I got it open

[molten pecan]

Oh you downloaded it in m4a, not wav

😭

[midnight terrace]

I totally knew that

[molten pecan]

Ah

[midnight terrace]

I wasnt thinking holdon

ill redownload

[molten pecan]

Ping me when you are done

[midnight terrace]

this should be wav

wait

yes

@molten pecan

[molten pecan]

Hood

Good

Now open UVR5

[midnight terrace]

is open

[molten pecan]

Make the Window Size 320

Algorithm MDX23C

Got it?

[midnight terrace]

whats mdx under

[molten pecan]

Uhh the one that wasn't called VR

I don't remember the name

[midnight terrace]

is it with mdx net

[rare goblet]

Ayo? @midnight terrace level 3 !!!

[molten pecan]

Yes

If you can't find it you can download it in the settings

[midnight terrace]

its instvoc hq right

[molten pecan]

Uhm

Can I get a screenshot?

It should be called MDX23C

[midnight terrace]

thats the only 23c one

[molten pecan]

Oh yeah that one!

Okay you can now convert it

Let me know how it goes!

[midnight terrace]

[glossy oxide]

I can't find the channel "voice-models"

[rare goblet]

Ayo? @glossy oxide level 1 !!!

[molten pecan]

It's down for now

Use the website weights.gg

[wise swan]

how the level up

[rare goblet]

Ayo? @wise swan level 1 !!!

[wise swan]

levelll

level

please help mee

where is the voice models

plsss help me

[glossy oxide]

Do i need a softwere to create a model from a famous character or something?

[violet heron]

Use weights.gg or #🔍│find-models with @shy zealot, not @gusty kestrel

[violet heron]

You don’t need it if the character already exists as a model

[glossy oxide]

I meant i serched Meilin Lee's voice as model but it doesn't extists.

[violet heron]

Ok so you would have to train a new voice

How strong is your pc?

[soft stratus]

@molten pecan but my model c

[hot lagoon]

wich here is a good point for the modle?

[glossy oxide]

He has a RTX 3050 and 16 GB RAM. So i think It's strong enough

[violet heron]

Yep it is

[glossy oxide]

So what i have to do to create the voice that i wish?

[violet heron]

https://docs.aihub.wtf/guide-to-create-a-model/model-training-rvc

[glossy oxide]

Thank you

[pallid finch]

is there a more up-to-date web UI than the one linked here for locally hosting? https://docs.google.com/document/d/1KKKE7hoyGXMw-Lg0JWx16R8xz3OfxADjwEYJTqzDO1k/edit i know the colab one had an option to upload files directly from the web ui as well as a microphone option, would like to get those on my locally-hosted web UI if possible

[midnight terrace]

wait wtf

how did that one model isolate it so well

I thought u had to run like 5 or 10

naH this is so good wtf

[molten pecan]

What?

[molten pecan]

Hehe you probably used the wrong algorithm

Remember, when you convert the vocals on RVC make sure your audio sample doesn't have harmonies!!

[midnight terrace]

nah I swear I just did what everyone else did

[molten pecan]

That's odd then

[molten pecan]

-colab

[azure marsh]

Suggestions for @pallid finch

All working colabs / spaces ☁️

• AICoverGen-NoWebUI, by Ardha Google Colab
• RVC Disconnected, Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• RVC-HF, by r3gf Huggingface Spaces
• AICoverGen, by r3gf Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces
• Advanced RVC Inference, by neuclya Huggingface Spaces

Need some help? 🤔

You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.

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

[midnight terrace]

like on isolated it a bit, then the next cleaned it up, then one helped with bg noise, etc

[molten pecan]

I see

[marble reef]

who can help me please with this error

[violet heron]

Oh your using Okada.

Try a different Virtual Audio Cable driver

[marble reef]

ok

[normal grove]

Hey everyone! I seem to get an error when I press convert here, can someone explain me why that may happen?

Seems like I cant send any screenshots here, basically when I press convert in Ilaria RVC I just get an error

[rare goblet]

Ayo? @normal grove level 1 !!!

[normal grove]

Oh there you go

[proven hill]

did you restarted the space?

[normal grove]

What does that mean exactly ?

Sorry Im kinda new here

[violet heron]

Duplicate the space

Make sure it’s public because for some reason it doesn’t work when it’s private

[normal grove]

Ill try that in a sec

[violet heron]

And another thing make sure both audio name and model name has no spaces

[normal grove]

How long does it take for it to build after I duplicated the space ?

[violet heron]

A minute

[normal grove]

I followed the steps and for now theres no error but it seems like the processing time is taking a while, I havent got a result yet but Ive been waiting for about 5 minutes now

[hot lagoon]

wich here is a good point for a modle?

[hushed quartz]

set smoothing to 0.999 first

or whatever the max is

[midnight terrace]

how do I drag the audio in audacity

it wont let me move the track left or right at all

[lofty zephyr]

@violet heron

[violet heron]

You didn’t duplicate it

[lofty zephyr]

But when I try to duplicate it this happens

[magic badge]

and with me it doesn't go beyond processing

[nocturne mural]

runtime/python.exe pip uninstall ffmpeg_python runtime/python.exe pip uninstall ffmpeg-python runtime/python.exe pip install ffmpeg-python runtime/python.exe pip install ffmpy -q

also check if you have these files in the mangio folder

https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/ffprobe.exe
https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/ffmpeg.exe

[lofty zephyr]

What's the problem now?
I can't get it

[molten pecan]

Change the name

[lofty zephyr]

Should it contain uppercases?

[molten pecan]

Sure

If you want

[lofty zephyr]

I changed it but still

[rare goblet]

Ayo? @lofty zephyr level 3 !!!

[lofty zephyr]

The space is duplicated right

[molten pecan]

It says you already created this repo?

[lofty zephyr]

Yes

[molten pecan]

Go to your profile section

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

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

🇪🇸 Spain

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

🇫🇷 France

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

🇩🇪 Germany

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

🇯🇵 Japan

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

🇰🇷 Korea

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

🇵🇱 Poland

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

🇷🇺 Russia

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

🇮🇹 Italy

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

[molten pecan]

In huggingface

All of your duplicated spaces are there

[lofty zephyr]

To which section I should go in my profile (Sorry I'm a beginner and I've never used hf before)

[molten pecan]

Just click on the upper right corner

There should be a circle there

Click on your username

And you'll get them

[lofty zephyr]

Those?

[molten pecan]

Yes

Use whichever you like

Those are your duplicates spaces

Use them so you don't have to duplicate them each time you want to use them

[outer isle]

Problem with RVC, it's too slow for my old laptop

[molten pecan]

If you have a slow device you have two options. Either get a better one or use huggingface

[lofty zephyr]

And from here I'm supposed to change the visibility to public?

[outer isle]

I'm probably stick with So-VITS-SVC

[molten pecan]

You can set the space as private from the space itself

Go to the space and click on the three dots

There should be an option over there

[molten pecan]

Huggingface is much better

[molten pecan]

@half cove @violet heron 😭😭😭

Type the command y'all

[outer isle]

@molten pecan

[half cove]

avarage rvc beta male vs sigma svc user

[molten pecan]

[molten pecan]

Yes?

[violet heron]

-svc

[azure marsh]

Yeah I still use so-vits-SVC

https://tenor.com/view/crowd-avoiding-man-gif-25257570

[lofty zephyr]

This?

[molten pecan]

Yeah!

Good job loleeta ^-^

[outer isle]

@violet heron I'm probably stick with So-VITS-svc

[molten pecan]

Make an account first

[outer isle]

-svc

[azure marsh]

Yeah I still use so-vits-SVC

https://tenor.com/view/crowd-avoiding-man-gif-25257570

[molten pecan]

😭😭

Google it then

I'm sure uploading a model is pretty easy

[outer isle]

Kaggle it

[rare goblet]

Ayo? @soft kiln level 1 !!!

[lofty zephyr]

Whyyyy

It's still happening

[outer isle]

Which RVC Should be faster?

[molten pecan]

Have you downloaded a model with the same name?

[molten pecan]

RVC v2

[lofty zephyr]

Nope I've never downloaded a model in hf

[molten pecan]

Send me the link

I think I know the issue here

[lofty zephyr]

There are 2 links one I got from weights.gg and the other one was repaired by @violet heron

Which onw should I send

[molten pecan]

The one by timbovill

Brb

[lofty zephyr]

https://huggingface.co/TIMBOVILL/HuggingFaceReuploads/resolve/main/RIIZE Anton - Weights.gg Model.zip

[gloomy smelt]

Okay so, this is probably a dumb question but how do i put models on applio? like, not create them, just put already existing models on it. I'm asking because i'm used to mangio where the path file goes in one folder and the model itself in another folder, but in Applio it gets confusing for me to know where should i put the models

[rare goblet]

Ayo? @gloomy smelt level 3 !!!

[cyan osprey]

can you use rvc locally on android?

I can't use colab

[molten pecan]

Yeah!

Let me give you a link

[cyan osprey]

okay

[molten pecan]

https://discord.com/channels/1159260121998827560/1167844377046040769

Good luck!

Ping me if anything ^-^

[wispy burrow]

ModelName.pth in logs/weights.
added_IVF*.index in logs/ModelName.

[gloomy smelt]

Oh, i see now

[molten pecan]

@lofty zephyr https://huggingface.co/TIMBOVILL/HuggingFaceReuploads/resolve/main/RIIZE%2520Anton%2520-%2520Weights.gg%2520Model.zip

[gloomy smelt]

Thanks a lot

By the way

[molten pecan]

Try that one

[lofty zephyr]

Alr

[gloomy smelt]

So uhh i have like, a model that worked on RVC_GUI but like, it's only the model and it doesn't come with the path file

Does it still work

[lofty zephyr]

It's not working.....

Maybe the prblm is in smth else?

Not the link I mean

[molten pecan]

The problem is the link actually

[lofty zephyr]

Oh

[molten pecan]

When you don't trim the spaces huggingface types the spaces as %20

That's how you know he forgot

😭

Give me a sec

[lofty zephyr]

Okie

[wispy burrow]

The path file?

[molten pecan]

Well it worked for me so uhm 😭

[gloomy smelt]

Yeah

Hold on

[molten pecan]

Try naming the model "test"

[gloomy smelt]

Uhhh sorry, meant index

[molten pecan]

Without the quotes

[lofty zephyr]

Wait how... 😭

[lofty zephyr]

Ok

[wispy burrow]

Oh, it'll still work without the index file.

[lofty zephyr]

I changed the name to test still

[molten pecan]

That's really weris

Weird

[lofty zephyr]

What have u done exactly 😭😭😭😭😭😭

[gloomy smelt]

Oh okay

Because like

[molten pecan]

Opened my web browser and pasted the link

[gloomy smelt]

Sometimes i just use the model but it requires the index to make a cover?

i dunno

[molten pecan]

Okay let's try something else

One sec

[lofty zephyr]

Mhm alr

[molten pecan]

https://huggingface.co/spaces/r3gm/Aesthetic_RVC_Inference_HF

@lofty zephyr

Try using that one

Duplicate it too

In the Resources tab you paste the model

[wispy burrow]

The Index should be optional.
The pth file is the most important.

[lofty zephyr]

Wdym?

[rare goblet]

Ayo? @lofty zephyr level 4 !!!

[molten pecan]

Duplicate the space

[gloomy smelt]

i see

Let me try somethin

[lofty zephyr]

Isn't it already duplicated?

[molten pecan]

The one I just sent?

Nope

[gloomy smelt]

nvm

[molten pecan]

Hi Raid

[honest pond]

what error you have in the terminal?

[gloomy smelt]

Hmmm let me check

[honest pond]

hi

[honest pond]

try the second option under update

select again the model and try

[gloomy smelt]

Alright

[honest pond]

if is not this can you try exporting the audio like an mp3 if not idk

[gloomy smelt]

Well, got error

[gloomy smelt]

k

[honest pond]

is the same?

[gloomy smelt]

Let me checkkk

Seems like it, womp womp

[honest pond]

type, google something like audio to audio, and change the extension and sample rate, once it solved me like this

if not, it may be a model thing or something with the audio codec etc. if it is that, try with another one move it to see if it is a problem of the model or the audio itself.

[gloomy smelt]

I think its's just the model itself

[nocturne mural]

the file has spaces in the name

[gloomy smelt]

Well that's sad, but at least i can use it in mangio

[lofty zephyr]

Ok that what appears to me

[molten pecan]

Like I said, now you go to Resources and paste the link

Then come back and hit Refresh

[lofty zephyr]

So I can use this space whenever I want to do an ai voice cover?

[molten pecan]

Yeah!

Use the duplicated one

[cyan osprey]

@molten pecan why does the processing take very long?

[lofty zephyr]

And after that?

[molten pecan]

Select the model in Inference voice and Auto-detect index

Change the settings however you want and hit Convert

In mobile it might give you an error if you tap out the app, so I recommend staying there until it's done

[molten pecan]

The why depends a lot

[wispy burrow]

The mp3 file I take it?

[molten pecan]

Depends if your space is private, your audio length and the algorithm

You can make it faster by paying a few cents

[lofty zephyr]

Should I fill them?

[molten pecan]

Yeah of course

Tap the arrow to select them

[lofty zephyr]

I can't nothing appears

It just allows me to write lolz

[molten pecan]

Did you hit Refresh?

[lofty zephyr]

Yes

[molten pecan]

When you pasted the link in Resources

Did you hit Download?

[lofty zephyr]

When I try it tells me there's an error

[molten pecan]

I see

Let me try

[frigid bloom]

necesito ayuda

😭

[frigid bloom]

i need help

necesito ayuda

@viscid moss

[rare goblet]

Ayo? @frigid bloom level 1 !!!

[molten pecan]

Qué sucede?

[frigid bloom]

Cuando quiero modificar mi voz y pongo start suenan como susurros y se escucha lo que hablo después de mucho tiempo de hablarlo

no se si tengo que configurar algo

[molten pecan]

Well for some reason I can

I have an idea

[lofty zephyr]

HELP

[molten pecan]

Download the model

And unzip it

And then upload the pth and index file