✨│ai-help

[rare goblet]

Ayo? @next dove level 1 !!!

[next dove]

ytrewq

so whats wrong here?

[violet heron]

Can’t tell

Switch to a RVC model and turn down your in

[median knoll]

Is it just me or has creating AI become a LOT more difficult than 6 months ago?? I took a break and now I feel like I'm back at square one, the guides look like a foreign language to me

[molten pecan]

wdym by creating an AI?

You mean doing inference or training an RVC model?

[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

[thorn bobcat]

what is the difference between rmvpe and rmvpe_gpu of the pitch extraction algorithm?

[half cove]

second one uses gpu so its faster

[median knoll]

I would assume RVC, I'm not very good with computer stuff anyway but I feel like the whole layout with the websites and guides has been changed, I don't understand anything 😭

[rare goblet]

Ayo? @median knoll level 1 !!!

[molten pecan]

I know :( most guides are a little over the place and not very begginer-friendly

[thorn bobcat]

does this change the quality?

[molten pecan]

We all start somewhere. Don't worry

My tip is to just experiment. You'll learn through trial and error

That's how I learned

And now I have Helper role ^-^

[half cove]

dont think so

[median knoll]

Thank you! I had a tiny bit of experience with making AI covers and creating RVC models a while ago, but coming back into it I feel like I'm stuck at the beginning again haha

[thorn bobcat]

thanks

[median knoll]

Also idk if the automaze bot automatically sent the guides or if someone else did, but thank you, I can hopefully easily navigate the guide now haha

[glad zealot]

It automatically sends it

[molten pecan]

@median knoll yw! And good luck

@brittle wing use rmvpe. Crepe-tiny is kind of obsolete nowadays

[half cove]

i think they are using it for realtime

rmvpe is faster

use it

[brittle wing]

what does index do again?

[half cove]

accents

[molten pecan]

Accent

Usually models have two files. .pth file and .index file

The .index file is an additional file that makes the model sound better

[brittle wing]

so it automatically chooses the settings?

[molten pecan]

"it" what?

[brittle wing]

the index file

[molten pecan]

Uhh the index file just contains info about the accent and speech of the voice model

It's up to you to tweak the settings

[brittle wing]

thank you

[molten pecan]

yw!

Have fun

[median knoll]

Is there a set method for getting the best acapellas on UVR? or is it a matter of trial and error?

[brittle wing]

y’all

“connection timed out”

why

[patent pasture]

is it collab?

[brittle wing]

yes

on the gradio page

i was on it for like a minute wtf

[rare goblet]

Ayo? @brittle wing level 1 !!!

[patent pasture]

I guess it is because colab banned gradio module, no?

[brittle wing]

huuuuh

whattt

fr?

[patent pasture]

kinda you need to have premium to have acces

[brittle wing]

what do i do then

[patent pasture]

what notebook did you use?

[brittle wing]

wdym

i’m dumb with this stuff

been following a yt tutorial closely

https://youtu.be/gtjCmJ_6d0k?si=_HOG1iucgczxyMId

this

guy

yes

easy gui

lol

also i keep running out of runtime

[patent pasture]

really it is not actual method

[brittle wing]

then what is

[patent pasture]

here is one kaggle notebook, but will it clear for you

[patent pasture]

kaggle is like colab

[brittle wing]

link?

[patent pasture]

wait

[patent pasture]

https://www.kaggle.com/code/hinabl/multinotebook-ai-public

but in this method you need also register in ngrok

[proper shale]

or use stuff like this

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[proper shale]

a bunch of colabs

and Spaces

[patent pasture]

do they work, just I tested aicovergen and there GPU was so stange, waited 20 mins to make a cover

[proper shale]

probably had that uvr separation error again, hm

[brittle wing]

colab runtime is an issue tho

[proper shale]

all of these work tho, and plus runtime is 4 hours long

that's like

a long time

[proper shale]

well, most of them

[brittle wing]

fr? then why is the colab i’ve been using has like 3 mins runtime 😭😭

[proper shale]

@glad zealot has the UVR thing in AICG been solved?

[proper shale]

That one has been restricted probably

[sand iris]

The onnx lack of gpu accel ?

[proper shale]

Yeah

[vagrant oar]

i need help to cofig the voice changer

[brittle wing]

free registration?

[rare goblet]

Ayo? @brittle wing level 2 !!!

[sand iris]

the webui version should be fixed, but it takes roughly as much time to be fixed than a single slow inference

linux moment

[proper shale]

😭

[proper shale]

novideo 😭

[vagrant oar]

can you help me

[patent pasture]

Yes.... but you need also verificate using your phone number

[proper shale]

state the issue

[proper shale]

nah no need

[vagrant oar]

nao sei configurar

[rare goblet]

Ayo? @vagrant oar level 2 !!!

[proper shale]

just email confirmation should be fine

[patent pasture]

to use use gpu, without verify you can only use weak cpu

[proper shale]

mas oq exatamente tu n sabe?

[proper shale]

oh u meant Kaggle

ur right yeah

[patent pasture]

in ngrok - yes, sorry

[rare goblet]

Ayo? @patent pasture level 5 !!!

[vagrant oar]

tudo quando eu coloco a voz tem que regular uns negocio

[patent pasture]

wtf did I write?

[azure marsh]

🎙️ Realtime Voice Changer (Local)

Guide for W-Okada's RealTimeVoiceChangerClient

Description: This one has more options and is somewhat more convenient to use than Real-time RVC.
Link: Rentry
Credits: ravencutie21

How to use real time conversion with RVC (voice changer)

Description: More accurate than W-Okada's, but requires a decent GPU.
Link: Google Docs
Credits: teas.hi

RVC TTS

Description: Text to speech with RVC
Link: Rentry

[proper shale]

vê esse primeiro guia

vai ajudar bastante

[vagrant oar]

blz

[patent pasture]

Sorry, if you will use kaggle, there you need to register using email and verify using phone number. Then if you register in ngrok there you need only register using email

[vagrant oar]

oq é esse chunk

[patent pasture]

ngrok is to get token, which you need to paste in one place

[brittle wing]

ahh okay thanks

[proper shale]

basicamente, é o tanto que o voice changer vai processar de uma vez só (menos exige mais da gpu, mais aumenta o delay mas diminui o risco da voz ficar travada e tal)

[vagrant oar]

eu to usando um do google

pq me falaram que nao funciona com amd

[patent pasture]

do you want to make a cover?

right?

[proper shale]

depende da placa que for... Se for uma mais recente da AMD acho que funciona

hey, this might be copyrighted

[median knoll]

What do you mean?

[proper shale]

...You're sharing copyrighted music around

That can basically get our server banned again

It's alright for a dataset I suppose

[brittle wing]

yeah

[median knoll]

Ohh am I not allowed to share audio files that I'm using for a dataset?

[proper shale]

Yeah. It's all copyrighted

[median knoll]

Ahh my apologies I didn't know!

[rare goblet]

Ayo? @median knoll level 2 !!!

[brittle wing]

OMG I DID IT

but it sounds glitchy

what do i do 😭😭

[violet heron]

What program are you using

And what doesn’t work

What are you stuck on

[violet heron]

What file are you opening

[honest veldt]

Hey i was wondering if someone could tell me why does my model have such a bad quiality bc at this point i tired 3 diffrent google colabs and they all given be the same bad results, here I give the sample voice I used and what it produced (please dont mind the youtube video i used for test) :

[half cove]

what would using other colabs do?

[honest veldt]

give the same bad results

[rare goblet]

Ayo? @honest veldt level 2 !!!

[violet heron]

Use 40k sampling rate

[half cove]

no im asking why you would

[violet heron]

You open the bat file

[honest veldt]

to try get the good result and see if it was problem of this one

[rare goblet]

Ayo? @brittle wing level 2 !!!

[violet heron]

Which bat file are you opening

Outdated moment

-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 GitHub

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

[violet heron]

Download Original RVC or one of its forks like applio

[desert niche]

-huggingface

[azure marsh]

huggingface.co upload Guides

English guide

Link: https://rentry.org/fdg_guide_newer
Credits: FDG

Spanish Guide

Link: https://rentry.org/AI_Hub_guide_esp
Credits: Julia (ailen2091)

Hindi guide

Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes

Turkish guide

Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes

[ionic canopy]

I keep getting this error trying to infer a voice:

Traceback (most recent call last):
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\routes.py", line 321, in run_predict
output = await app.blocks.process_api(
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\blocks.py", line 1007, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\blocks.py", line 953, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\components.py", line 2076, in postprocess
processing_utils.audio_to_file(sample_rate, data, file.name)
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\processing_utils.py", line 206, in audio_to_file
data = convert_to_16_bit_wav(data)
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\processing_utils.py", line 219, in convert_to_16_bit_wav
if data.dtype in [np.float64, np.float32, np.float16]:
AttributeError: 'NoneType' object has no attribute 'dtype'

How do I fix it?

[violet heron]

Yeah!????

[tacit anvil]

what website should i use for rvc ai covers?

applio

[quasi umbra]

Can someone explain to me why my voice doesn't work on the app?

[ionic canopy]

I keep getting this error trying to infer a voice:

Traceback (most recent call last):
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\routes.py", line 321, in run_predict
output = await app.blocks.process_api(
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\blocks.py", line 1007, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\blocks.py", line 953, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\components.py", line 2076, in postprocess
processing_utils.audio_to_file(sample_rate, data, file.name)
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\processing_utils.py", line 206, in audio_to_file
data = convert_to_16_bit_wav(data)
File "E:\RVC0813Nvidia\runtime\lib\site-packages\gradio\processing_utils.py", line 219, in convert_to_16_bit_wav
if data.dtype in [np.float64, np.float32, np.float16]:
AttributeError: 'NoneType' object has no attribute 'dtype'

[proper shale]

something might've not been selected properly?

Oooh that's the Mac version

[violet heron]

Run the bat file if your on windows

.sh is for Linux and Mac

[proper shale]

Does the Original RVC seriously include .sh files even for Windows installs?

[ionic canopy]

Odd. Because I have the right pth and index, and I can train voices just fine...

[proper shale]

Maybe the audio path wasn't selected properly? idk tbh

[sand iris]

nope !

the windows release contains it too

if you extract with anything other than 7zip it bugs out like that IIRC

[proper shale]

oh.

[ionic canopy]

Here are the settings I have

I think it looks right, maybe I missed something...

[proper shale]

You should insert the path to the joker index manually, try that

[ionic canopy]

still the same error

[rare goblet]

Ayo? @ionic canopy level 3 !!!

[proper shale]

Oh yeah no worries

You wouldn't even run that file tbh

Yeah that run.sh file is for Mac/Linux users

[honest veldt]

annnnd it still sucks

do you know something else that might be cousing this ?

[violet heron]

How much epochs are you using?

[honest veldt]

250

[violet heron]

And how long is your dataset?

Try setting your batch size to 4

[honest veldt]

3:12

[honest veldt]

alright Im gonna try this, I hope this works but are there any other setting I should change in RVC v2 Disconnected that may be cousing this ?

[proper shale]

or just train more

[honest veldt]

like more epochs ?

i tried putting it on even 800 and it was the same resoults

[proper shale]

nah i listened to it, it's just the dataset probably

[honest veldt]

this is the dataset it is very clean I think

so any other ideas ?

[proper shale]

has a slight echo to it

[violet heron]

I’ve had plenty of models with echos and they’ve been fine

Just try batch size 4 at 300

[proper shale]

bro no

[violet heron]

Should sound good

[proper shale]

this is a dataset issue

[violet heron]

Oh

As if my models didn’t have echos and came out fine

[proper shale]

Adobe Enhance that

It might result in a better model

Just saying

[honest veldt]

alright i try these two thing you told tomorow and come back here if the issue still is there

[stiff goblet]

which sample rate should i use for training. this is one of my samples in my dataset

[proper shale]

32k probably

[stiff goblet]

is 48k sample rate important? it says audio is 48k

[proper shale]

yeah uh no, because it's limited at 15/16k

[proper shale]

see here how ir barely reaches 15k and is limited

so you should do 32k

[stiff goblet]

so we are multiplying by two to understand 16x2 = 32 khz right ?

[proper shale]

yh

[stiff goblet]

thx

[proper shale]

you're welcome :)

[rare goblet]

Ayo? @brittle wing level 3 !!!

[sand iris]

what did you use to extract it ?

Was it :

• WinRAR
• 7Zip
• Something built into windows

I think it doesn't unzip properly on winrar & builtin windows

try with 7zip and report back, the "unsupported file type" thing you had earlier caused multiple files not to unzip

[thin stump]

@vivid breach I wouldn't be able to help you much lol. If it were me, I would search up VA's that does roles for villainesses or comic dubs. It doesn't have to be bowsette related because that perfect voice doesn't exist

[vivid breach]

any little bit helps

[stoic rampart]

hi im having trouble with colab rvc

/content/Mangio-RVC-Fork/logs/NewHiyyih/3_feature768

Exception Traceback (most recent call last)
<ipython-input-12-9c42c5055659> in <cell line: 30>()
29 listdir_res = list(os.listdir(feature_dir))
30 if len(listdir_res) == 0:
---> 31 raise Exception("No features exist for this model yet. Did you run Feature Extraction?")
32
33 try:

Exception: No features exist for this model yet. Did you run Feature Extraction?

but feature extraction runs perfectly im not sure whats happening

[proper shale]

it probably says no-feature-todo

this could happen in 3 situations:

1. audio was zipped incorrectly (as a .rar)
2. audio wasn't in the right folder format (zip should have a folder with ur audios inside)
3. audio is with a wrong format (try re-export from Audacity)

[stoic rampart]

what format should the audio be in mine is mp3

[proper shale]

try a re export to wav

[stoic rampart]

ok

[stoic rampart]

tysm

[proper shale]

and make sure to follow the folder format thingy

[vast yoke]

same i went step by step and also the voices still dont work

[tepid blade]

[sand iris]

software thay supports lots of compression/decompression algorithms

pretty much an alternative to winrar tbh

it might be needed to extract RVC, if I remember right, I had the same issue

[half cove]

then install it

now unzip rvc

[stoic rampart]

it still errors out

[proper shale]

maybe your folder was set up in the wrong way? does it have any spaces

[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

[runic meteor]

Nope it doesn't

[runic meteor]

Is there an updated colab

[half cove]

same

shi forgot to tell u to do that lol

extract files

just click enter dont mess with anything

[proper shale]

It's not really the Colab's error

You might be doing something wrong ngl but idk what it is

[half cove]

bro is defo exploding 1000 bombs at once

[runic meteor]

Any other suggestions

[proper shale]

nah wait

Honestly I'm not sure

It probably is the zipping

[runic meteor]

Can I send you the dataset the way I have it zipped and you tell me if it's right

[rare goblet]

Ayo? @runic meteor level 1 !!!

Ayo? @brittle wing level 4 !!!

[mental basalt]

Are there extra steps to using a custom model on the AMD version?

[restive yacht]

just zip with model and index

[rugged snow]

soo how do I actually start the guı?

[pallid cradle]

Newbie here! I'm using RVC GUI and someone else's model, but it's very glitchy. Any suggestions on how to fix it?

[restive yacht]

run go web

[restive yacht]

when you say glitchy what do you mean?

[rugged snow]

Thank youu

[pallid cradle]

The voice is scratchy when it's not supposed to be.

[rare goblet]

Ayo? @pallid cradle level 1 !!!

[restive yacht]

hmm

could be the dataset the model was trained off

[half cove]

Old

[restive yacht]

does the audio you input have any background noise?

[half cove]

Rvc gui doesn't have rmvpe or anything and it's old

Unless you mean rvc webui

[pallid cradle]

Not sure, someone else made it.

[restive yacht]

mangio?

[pallid cradle]

Ok that could be it. Any recommendations on what I should use instead?

[restive yacht]

no i mean the audio you want to be converted

[half cove]

-links

[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 GitHub

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

[half cove]

Original rvc ig

[pallid cradle]

Oh! No there's no background noise

[pallid cradle]

Thanks!!

[runic meteor]

I can't get the index to train I need help with that

I need the link to the rvc colab please

[bronze hedge]

Hi there! I'm trying to use RVC to change one audio voice to another. can someone help me? ngl, im not on dicord often and am quite lost

[pallid cradle]

I'm probably being dumb but I can't get Original RVC to work. When I try to extract the files from the zip folder, a lot of files can't be extracted.

It keeps saying their unsupported file types

[brittle wing]

Someone can explain Why does this appear?

[robust hazel]

Is there a document for the best parts to run RVC?

[violet heron]

Very very outdated colab

[brittle wing]

And what is the newest colab?

[woven furnace]

can someone break everything down for me? there are things in the tutorials and it's hard because i don't get what different things are needed.

[lean burrow]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[violet heron]

-colab

[azure marsh]

Suggestions for @brittle wing

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[glad zealot]

pretty sure it works no problem

at least on the mod i have

[desert niche]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[rare goblet]

Ayo? @desert niche level 3 !!!

[coarse cedar]

-realtime

[azure marsh]

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

[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

[lofty island]

how to add model voice on google colab ? it blocked me when i upload the model voice

[thorny raven]

Why my hugging face keeps processing

Mine has run for 300+ secs

[rare goblet]

Ayo? @thorny raven level 1 !!!

[rugged ravine]

help me with this

[rare goblet]

Ayo? @rugged ravine level 1 !!!

[rugged ravine]

look the response time

[patent pasture]

is your gpu good?

[rugged ravine]

its a RX 580 8gb

it should´ve fine

[patent pasture]

ok, if you want, you can test colab version of wokada changer

https://colab.research.google.com/github/hinabl/voice-changer-colab/blob/master/Hina_Modified_Realtime_Voice_Changer_on_Colab.ipynb

[rugged ravine]

the issue its, only happens with the voice changer from CSGO, idk why

with another voices works normally

[patent pasture]

oh, then I dont know

maybe model is bad

[rugged ravine]

its happening with all custom voices

[patent pasture]

and do you load pth and index file?

[rugged ravine]

yeah

[patent pasture]

I really dont know then

[rugged ravine]

im using onnyxgpu version with a Radeon graphics card, it should work good right?

[patent pasture]

I think yes

check your castom voices in colab version

[wicked dome]

Which chunk/delay should I use with rtx 3060?

[rare goblet]

Ayo? @wicked dome level 1 !!!

[sand iris]

you need to press "export to onnx", save the resulting .onnx fle, and reimport the model (with the .onnx file this time) just like you did originally with the voice model

[brazen patrol]

i downloaded a few outputs without issue but im getting file not available on site, didnt i tick something wrong by accident?

[short mulch]

can we request recovery from old Ai Hub to be brought back

[rare goblet]

Ayo? @short mulch level 1 !!!

[woven furnace]

like what are all these terms? what?

so i need a program to make the voice sing and i need a file to put the voice that replicates a voice?

[brittle wing]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[woven furnace]

the steps on the tutorial are all like paragraphs and hard to read

[long scroll]

How do you make the ai sing the song? I already downloaded the voice I wanted in this discord and I have the mp3 song with the vocals already separated. This is my first time attempting this.

[rare goblet]

Ayo? @long scroll level 1 !!!

[copper pebble]

well since u separated the vocals i suggest to use Ilaria_RVC this is thr guide:
(youll find the link for that in it)
-https://discord.com/channels/1159260121998827560/1167844377046040769

[copper pebble]

hey u need help with smt?

[woven furnace]

yep

[rare goblet]

Ayo? @woven furnace level 3 !!!

[woven furnace]

nothing makes sense and it's like it's expected to be common knowledge or simple when it's not

[copper pebble]

what are we talking abt?

[woven furnace]

._.

[copper pebble]

training? inference?

[woven furnace]

this is a server where i can learn to make ai covers?

[woven furnace]

idk

that's what i'm saying the learning curve is crazy

[copper pebble]

well for sure

[woven furnace]

ok good 😭

[copper pebble]

inference stands for making covers or just conversions

training stands for making voice models that u can use

[woven furnace]

okay 😭

[copper pebble]

i remember of you tho what do you say if u dm me and i help you step by step?

[woven furnace]

yea thank you

[long scroll]

Thank you for answering. I am currently having it converted now. so just to double check do I put the vocals in the "drop files here" or and the "2. Choose the audio file" for the music?

[copper pebble]

RVC will choose the file by himself but if you want to be sure u can check too

[brittle wing]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[rare goblet]

Ayo? @brittle wing level 5 !!!

[long scroll]

it works thank yuo so much ive been stuck here for 7 hours!

10/10

[copper pebble]

Yw

[proper shale]

...rickie

i think they were being sarcastic

[brittle wing]

/notes

[long scroll]

me? No I'm not I was actually stuck researching for 7 hours.

and I wished I asked sooner.

[proper shale]

oh 💀

[long scroll]

We kewl bud haha happy new year!

[long scroll]

you too happy new year!

[proper shale]

glad Rickie was there to save the day :)

[long scroll]

My saving grace.

[proper shale]

Happy new year for you too! May 2024 be a prosperous year for all of us

[long scroll]

❤️

[copper pebble]

aah cmon i did nothing

[long scroll]

haha how humble.

[earnest jasper]

TypeError: Descriptors cannot not be created directly.
If this call came from a _pb2.py file, your generated code is out of date and must be regenerated with protoc >= 3.19.0.
If you cannot immediately regenerate your protos, some other possible workarounds are:

1. Downgrade the protobuf package to 3.20.x or lower.
2. Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python (but this will use pure-Python parsing and will be much slower).

More information: https://developers.google.com/protocol-buffers/docs/news/2022-05-06#python-updates
make: *** [Makefile:56: run-ui] Error 1

Error occured

[sand iris]

what are you running?

[patent pasture]

Where is overtarining started? 2000 or 2250?

[proven hill]

BRAND NEW FEMALE VOICE MODEL!!

Female e-girl sexy voice model for Okada (RVC V2) is now avaliable on my Ko-Fi!

Get it now here! https://ko-fi.com/s/734d5f276c

Or find other female models (including free ones) here: https://ko-fi.com/ilariaowo

20% discount if you DM me!

[mortal lion]

if i have a game open, with the voice changer running in background it pauses the voice changer and only does the stuff for it when i alt tab from the game

or it will be extremely delayed and buggy and will happen like a minute after i say stuff

[proven hill]

it will be delayed

depends on your gpu really

[mortal lion]

any way to fix or make the voice changer have higher prioritization?

[proud elbow]

set CPU affinity for audiodg.exe to CPU 2 only, check for more guide: https://rentry.co/VoiceChangerGuide

[mortal lion]

tyvm

[honest veldt]

annnd still nope

any other idea ?

[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

[honest veldt]

and this is the adobe enhanced its still sound bad but now the original voice of the character is worse

[proper shale]

it sounds better, the problem is the tone of the guy's voice really

[honest veldt]

well now you can bearly reconize the voice of the guys voice i sound like a bad voice of random men not even close to the samples

[proper shale]

again, you're using a voice that has a different pitch from the audio

[glad zealot]

what they tryin do?

[proper shale]

you gotta try something closer

and the quality sounds better

[honest veldt]

i think its more bc the enhancing of the audio with adobe enhancer kinda ruined the original voice , made it sound nicer but not like him

well it is sounding good but i kinda is not sounding like the original voice

[half cove]

please no instrumental cuz copyright

[honest veldt]

oops sorry

[half cove]

just del it rq

[honest veldt]

here no music

[half cove]

delete the one with music

[honest veldt]

i did

[half cove]

wat

[honest veldt]

i dont see it I deleted it

[half cove]

oh dw

it was my discord bugging

[honest veldt]

ah alr

[hidden drum]

if my dataset audio length is 6mins what epoch do i set it to

[proper shale]

go for 500 epochs

on every model

because you can check for overtraining

and if it is overtrained u can just download an earlier model

[quartz moon]

is there another website for making the ai thing? i used to use the google colab but its not working for me anymore.

[proper shale]

-colab

[azure marsh]

Suggestions for @quartz moon

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[proper shale]

u can use Huggingface spaces

and the colabs above r pretty nice too

[modern nest]

does anyone know hwo to remove artifacts from voiceless constonants n breathing sounds

[finite galleon]

I think increasing protect value (which means lowering the protect) helps

Also lowering the index rate since it can contain noise if your dataset is meh

[modern nest]

thanks

is it the search feature ratio slider

[finite galleon]

Protect is the protect voiceless consonants index rate is the feature ratio

[quartz moon]

how long does it normally take to convert it? it's been a few minutes and i dont remember it taking this long on the other one.

[proper shale]

if you're using a Huggingface Space it's because it's using the CPU

Colab used GPU so it was faster

[quartz moon]

ah i see. i was confused cuz it said it would only take 10.5s but it's been 400

how much longer should it take?

[proper shale]

depends on ur audio's length

4 mins took like 400-500 secs iirc for me

[quartz moon]

its about 3 minutes

[proper shale]

btw did you duplicate the space

[quartz moon]

no?

[proper shale]

oof hope it works then, some people don't even get their audios sometimes

next time duplicate the space

[quartz moon]

wdym? like just duplicate the link so its opened twice?

ok i just got it after like 700 seconds. wdym by duplicate it though?

[proper shale]

basically, replicating that space only for you

there's an ai cover guide for it on #1159513888199540817, shows you what you need to know

[proper shale]

(so you don't get queues)

[quartz moon]

i tried looking at the guide but i dont understand "duplicate" like open the link twice right?

[proper shale]

no, basically creates a new private space

see how there's a duplicate button on the top of the screen

[quartz moon]

no... umm where is that exactly?

i can dm you a screenshot of what i see

[proper shale]

oh you need a Huggingface account first

so make that

and then you should be able to duplicate the space

[molten raven]

[Voice Changer] use torch contentvec [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from pretrain/content_vec_500.onnx failed:Protobuf parsing failed.
[Voice Changer] exception! loading embedder PytorchStreamReader failed reading zip archive: failed finding central directory cuda:0

Anyone know how to fix? My voice and mic works in passthru, but it never changes

[sand iris]

delete the pretrain directory and relaunch

[distant wagon]

Can I ask how to adjust the pitch?

[proper shale]

there should be an option labeled "transpose", if ur using voice changer that is "tune"

[distant wagon]

i use Ilaria RVC

[proper shale]

yea theres a pitch option there

its either labeled pitch or transpode

transpose

change that number to positive to make it higher, negative to make it lower

[distant wagon]

this?

[proper shale]

nuh uh, its not in the advanced options

[distant wagon]

so is there any way to adjust the tone?

[proper shale]

yes, there's the pitch part

[distant wagon]

i mean

uh

[proper shale]

yeah thats high

u can decrease it with negative numbers

[distant wagon]

o k

let me try

[brittle current]

i have an error with some red text. what am i doing wrong?

[proper shale]

Can you give more info

[stray mauve]

can someone help me its not opening

[hidden drum]

But i use applio local

Oh nvm

Theres a tensorboard there

💀

[hidden drum]

ill try 500 if not that good than 300 or 250

[agile storm]

guys how do I continue training in easyGUI?

[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

[haughty tangle]

gradio.exceptions.Error: 'index -1 is out of bounds for axis 0 with size 0 how do I fix this error

[nimble vapor]

would it be possible to run okada on a linux server, and use that gpu to run the voice changer so i dont use resources on my main pc

cause i have a second pc setup running linux for like game servers, and as a media server as well

and if i could like run okada or rvc on it but use the mic from my main pc

only way i can think is having a interface using audio cables to go back and fourth but thatll be confusing

[sand iris]

yes

[sand iris]

the "client" mode for audio actually routes everything through the client app, that won't be needed

[nimble vapor]

wym

[sand iris]

the graphical interface is just a web page

[nimble vapor]

ahhh ok

[sand iris]

when audio mode is set to client, audio is handled via this web page

[nimble vapor]

OHH

okok

so that gui is just basically chromium

[sand iris]

yeah

[nimble vapor]

cause i have a entire linux server for game servers, plex, and i have my old gpu in it so it should relatively be able to do it fine

[sand iris]

i'll be back online tomorrow (not sure if i'll have much time), but I do have the instructions on how to get it running on a linux machine. from there, it's just a matter of connecting to that machine via web browser

[nimble vapor]

connecting to it via browser i can do myself ahah

[sand iris]

GPU just needs to be an AMD Radeon vega 56/64 or newer, or any GTX 10xx+ GPU

[nimble vapor]

ill either run a nginx reverse proxy or figure something out

lol

[sand iris]

it binds to 0.0.0.0:18888 by default, but should be tweakable

[nimble vapor]

yeah

either way a nginx reverse proxy or me going over the machines local ip should do

[viscid glacier]

my rvc gui doesnt wish to start for some reason, its just stuck at the cmd without the program starting

not quite sure what it could be

[vale pike]

-realtime

[azure marsh]

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

[wheat marten]

My RVC is asking for more memory anyone know how to fix that

[magic gale]

how do i use this ai voice thingy?

[wheat marten]

#1159513888199540817

[cyan spear]

idk whats going on this hasnt happened before.

[wheat marten]

same issuse lol

[cyan spear]

any idea how to fix??

or is this happening to everyone

[wheat marten]

nope

I have that and mine says it need more memory

so I got too issuse lol

[cyan spear]

this is really annoying

anyone else got the same problem?

[nimble vapor]

where can i get noise supression again

[rare umbra]

Does anyone know a solution to this?

[nimble vapor]

run in cmd pip install tensorboard

ez

[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

[wintry juniper]

trying to train a model but for some reason my tensorboard is empty any help?

[proper shale]

have u refreshed the thing

there's a refresh butto

button*

[wintry juniper]

no i mean the grad_norm_g and grad_norm_d are empty

[proper shale]

those really arent the ones u should be looking for

check scalars, search for g/total

[wintry juniper]

dont really see anything

only thing i see is loss/g/total and thats also empty

[proper shale]

have u refreshed the thing tho

and check cmd if it's even training

[wintry juniper]

yeah

its training it says

epoch 4 rn

[proper shale]

oh i think it should say at the nearest save checkpoint

[wintry juniper]

oh k that makes sense

ill see when it gets to 5 then

[proper shale]

alrighty, hope it works out :)

[wintry juniper]

kk thanks

[rare goblet]

Ayo? @wintry juniper level 1 !!!

[proper shale]

you're welcome. good luck on the model :)

[waxen lodge]

Guys, why when I click on the Rvc model training does it write to me "训练结束, 您可查看控制台训练日志或实验文件夹下的train.log"

[knotty nimbus]

is there a video guide to rvc? i have not made ai covers in a while

something in the web btw i do not have a good enough computer for the local version.

[glad zealot]

the only gui online that i know works is the kaggle version

https://www.kaggle.com/code/hinabl/multinotebook-ai-public

[molten pecan]

https://rentry.org/ilarvc_inf_guide

You are welcome ^-^

[knotty nimbus]

nvrm

[shy zealot]

**Method 1 🔎: ** The search accuracy is 100% 💕

NTTS

Upload the <t:1700186934>
Server uploaded: AI HUB
Channel uploaded: AI HUB Post
Profile Owner: modpotato

Author

mod

Discord User

modpotato

Internal ID

98464

Language

Unknown 🇺🇳

Type

RVC V2

Size

116 MB

Extractor

RVMPE

Epoch

48000

Steps

0

Powered by ModelAI.net, Made with Love ❤️

I'm sorry, I've tried everything 😭
It seems that I can't find any model with your search, please try another query.

[ionic adder]

1 result found

NTTS-AI-Voice-RVC

• Author: TheOriginalBox
• Downloads: 0
• Likes: 2
• Tags:

["license:apache-2.0","region:us"]```
- **Pipeline-Tag:** `undefined`
- [**URL**](https://huggingface.co/TheOriginalBox/NTTS-AI-Voice-RVC): https://huggingface.co/TheOriginalBox/NTTS-AI-Voice-RVC

[verbal oxide]

got error when using overall pitch change

[glad zealot]

hmm pretty sure i didnt change that

[verbal oxide]

idk if kaggle blocked it maybe

it automatically timed out

[glad zealot]

use headless

uhh the save all feature

[verbal oxide]

still same

[glad zealot]

when using covergen right?

[verbal oxide]

yes

[glad zealot]

lemme check

[languid plinth]

so Installed the program from hugface for the AI voice changer

I was told to click start_http after the download but that black screen does not pop up when i double click on it

any help would be appreciated

[glad zealot]

but using it without overall pitch change and normal pitch change works?

[verbal oxide]

it works but after second it was canceled by backend exit code:137

[glad zealot]

weird it worked fine for me

[young oasis]

How do I make a cover lol

[glad zealot]

no errors and stuff

lemme test the overall

[verbal oxide]

wait i screen record it

[glad zealot]

okek

[glad zealot]

either use mainline, mangio, applio, covergen

[young oasis]

Huh?

I haven’t been in this server since June idk what any of that is

[glad zealot]

ahh yu using overall crashed it

but using normal pitch change works

[verbal oxide]

yeah thats it

[glad zealot]

weird

[languid plinth]

will anyone be able to help me with my issue? 🥹

[rare goblet]

Ayo? @languid plinth level 1 !!!

[verbal oxide]

maybe kaggle blocked sox? since overall pitch change was using sox

[glad zealot]

no clue

lemme check it on colab

[verbal oxide]

in colab has no issue with overall pitch change but kaggle always getting timed out

[rare goblet]

Ayo? @verbal oxide level 7 !!!

[glad zealot]

ahh so its kaggle issue

since its using the same repo

[verbal oxide]

yeahh

[glad zealot]

ye no idea what can fix it

[languid plinth]

so no one can help me out?

[glad zealot]

#🔍│help-w-okada

dont really know how to do run it properly on local so cant help with that

[final thunder]

Does anyone know how to turn off the audio normalization in the preprocess.py script for Mainline RVC? The file path is "RVC1006Nvidia/infer/modules/train/preprocess.py" here is a direct link to the code, because I can't seem to be able to post any screenshots in here: https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/infer/modules/train/preprocess.py

[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

[brittle jay]

???

it didnt happen the first time i ran the cell

Im running the "Clone and Install" Cell

[sand iris]

don't mind it, it should work

Blame Nvidia for making stinky drivers smh

[brittle jay]

oh okay thank you

[latent kettle]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[worn prawn]

yall know any good models for a proudcer tag?

[vital mortar]

how do you download a model on weights.gg??? i cant find the download button anymore

[brittle wing]

triple dots, that says more actions

[vital mortar]

ok thank you so much. i was so confused

[silver rivet]

Is there a way to better train some ones AI to your own voice?

[pure scroll]

my producer tag

[agile storm]

guys is there a way to continue training in easyGUI?

[hallow tulip]

hello can I ask you for some help? I think i am in trouble with going on the website

i've downroaded certain rvcs, and everytime i try to press 'go-web', the web does not come out

what should i do to get in to the rvc web site?

[rare goblet]

Ayo? @hallow tulip level 1 !!!

[hallow tulip]

hello!

[hallow tulip]

can you please help..?

[glad zealot]

get more of their voice

[silver rivet]

So you cant train it specifically to change your voice

[drowsy solstice]

hey, can anyone tell me about the best method to train a RVC 2 model? I do have a GPU

[latent kettle]

Use Google colab. Four one RVC disconnect

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[glad zealot]

you really cant just train it to only work with your voice

thats why you have to set the tune on the voice changer to make it work with your voice

[blazing garnet]

How can I train a model without google colab or a nvidia cpu on rvc?

[glad zealot]

local

if you dont have gpu use colab/kaggle

if you dont want to train yourself make a requiest in #1159289738314919936

[blazing garnet]

I wanna train local but I don't have nvidia gpu

What is kaggle?

Colab doesn't work for me

It doesn't sound well

[spare wraith]

ehhh i dont understand how to use wordband bot

[blazing garnet]

Or it just don't work

[glad zealot]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• 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

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

[glad zealot]

use the rvc disconnected from that

the old one doesnt work anymore iirc

[blazing garnet]

K

I will try

[sand iris]

With an intel dedicated GPU

Or an AMD dedicated GPU, if you are running linux instead of windows

^ only ways to train locally without an Nvidia GPU

[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

[brave garnet]

Colab di Google

⠀

• Colab per usare modelli già esistenti e trainare

Link

• Colab di backup nel caso la precedente dia problemi

Link

⠀

Made with love by Ilaria 💖

[unkempt vale]

Can someone help me

[rare goblet]

Ayo? @unkempt vale level 1 !!!

[brittle wing]

um the voice changer won't work for me

[rare hull]

hello im using rvc and I want to model my friends voice that I obtained with a media player file. How am I supposed to convert the mp4 file to a voice model that I can use on rvc-gui?

[proper shale]

convert to an audio file, then do all the dataset treatment steps on audacity (trim silences, noise gate), and train via colab or locally

rvc-gui is outdated btw, I'd recommend going to a newer local fork

-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 GitHub

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

[proper shale]

mangio is a good option

[brittle wing]

hey the voice changer is laggy and choppy

any fix?

[proper shale]

do you have any info? specs? cmd output

anything

[brittle wing]

i got my specs

[proper shale]

just gpu really

[brittle wing]

GPU : RX 570 Series AMD
CPU : i7-4th Gen
Rams : 20

[proper shale]

do you know what version of W-Okada you're using?

[brittle wing]

what is w-okada?

[rare goblet]

Ayo? @brittle wing level 1 !!!

[brittle wing]

i'm using coda

[proper shale]

...the voice changer

[brittle wing]

latest

[proper shale]

oh ok that's the right one for your gpu

alright, you need to export the model you downloaded to onnx with the button on the GUI

then u reupload it

[brittle wing]

save to onnx?

[proper shale]

and it should work

[proper shale]

yeah

[brittle wing]

well i did that

[rare hull]

thank you

[brittle wing]

just now

[proper shale]

now reupload that file in another model slot

should work

[proper shale]

you're welcome :)

[brittle wing]

[Voice Changer] Web Server Launch Exception, [ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Exception during initialization: D:\a_work\2\s\onnxruntime\core\providers\dml\DmlExecutionProvider\src\ExecutionProvider.cpp(896)\onnxruntime_pybind11_state.pyd!00007FFB3003FE81: (caller: 00007FFB3003FC22) Exception(2) tid(465c) 887A0006 The GPU will not respond to more commands, most likely because of an invalid command passed by the calling application.

i get this when i try to open the voice changer

[proper shale]

well honestly idk that error

[brittle wing]

can i change the gpu from outside?

i changed to gpu4 and this happened

[sand iris]

what's the GPU ?

AH

[proper shale]

should've kept it on 1

nah wait

0

[brittle wing]

how can i change it from outside

[proper shale]

try restarting the thing

[sand iris]

stored_settings.json

or stored_config.json

[proper shale]

oh

[sand iris]

Not sure which one is the right name

[coarse crater]

help can someone please teach me how to solve this thing

[proper shale]

you didn't select the index file

select that from the dropdown

[brittle wing]

its faster now

but any way to make it faster?

[proper shale]

probably not, if you're already using rvmpe onnx

[brittle wing]

also what is passthru?

i can use crepe_tiny

[proper shale]

allows ur normal voice to pass through

[brittle wing]

like how?

pass through of normal voice so it can be converted instantly?

[proper shale]

it just stops the ai and only lets ur real mic pass through

[proper shale]

and no GPU can do that in the world

not even the 4090

[brittle wing]

what about index?

[proper shale]

index is basically all the mannerisms of a person (accent, how they say certain phonemes, etc) into a file that can modify the outputs

wouldn't recommend using it in real time stuff tho

[sand iris]

Something along the lines of how close to the source audio you want the conversion to be

[proper shale]

https://media.discordapp.net/attachments/1191364149083648050/1191364186639454279/image.png?ex=65a52b5c&is=6592b65c&hm=9548088725fbf7faddc6a7f12ca6185af0f9444d30f85ceba2cf6e7b8c624c7c&

This is your error right?

Alright so, it says down there that there is no free GPU VRAM

What were you trying to do

[rare goblet]

Ayo? @brittle wing level 1 !!!

[loud herald]

anyone know why the cable input and output are switched around?

[proper shale]

because that's how it supposed to be, the cable output should be the input on your other program

[blazing garnet]

If I leave the rvc disconnected running in background for like a hour with no movement will it disconnect or no?

[proper shale]

it will

[blazing garnet]

It will disconnect?

[proper shale]

check it every 5 mins, refresh TensorBoard to check the graphs

[proper shale]

it will

[loud herald]

I mean like in sound options for playback it is input

[rare goblet]

Ayo? @loud herald level 1 !!!

[loud herald]

but input is for microphone right?

[blazing garnet]

Bruh I will just start a autoclicker and leave it on then

[proper shale]

yeah

[loud herald]

its meant to be like that?

[proper shale]

yeah

you didn't do anything wrong, dw

[blazing garnet]

I will be very dissapointed if this doesn't work

[loud herald]

well when i try to test it with my friends nothing works and that little green ring when you speak doesnt appear

[proper shale]

The input output settings should be like this in W-Okada:

Input - Your mic
Output - CABLE Input

And in Discord:

Input - CABLE Output
Output: Your headphones

[loud herald]

i set input on discord to vcable aswell

[loud herald]

ok ill try rearranging the settings

[blazing garnet]

Can you upload rvc models to voicemod

I am contemplating to buy it or not

[blazing garnet]

Y'all watch ai tools search?

[proper shale]

no

voice mod just changes pitch

and adds some effects maybe

[blazing garnet]

Bruh so is there any way to run w okada locally?

[proper shale]

...yeah?

not if you don't have a gpu tho

[blazing garnet]

Is that a yes?

[proper shale]

if you have a good gpu yes

[blazing garnet]

Need an nvidia one?

[proper shale]

yeah

[blazing garnet]

BRUH

rvc stop being non nvidia racist

Jk

[proper shale]

amd would also work too, just gotta export models to onnx n stuff in the gui

[loud herald]

works now thanks 👍

[proper shale]

tell that to PyTorch

[blazing garnet]

K

[blazing garnet]

Who is that?

[proper shale]

you're welcome :)

[proper shale]

RVC uses PyTorch lol

[blazing garnet]

Oh...

[proper shale]

And PyTorch has horrible AMD/Intel support

[glad lance]

when i try to convert my voices it just said connection timed out how do i fix?

[proper shale]

😭

[proper shale]

what are you using

[glad lance]

what do you mean?

[proper shale]

like what fork, colab

that you're using rn

[blazing garnet]

Why not just install other python packs to make it compatable in this long history of python noone thought of this?

[glad lance]

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

[proper shale]

u can make DirectML work

[proper shale]

have you duplicated the space yet

[proper shale]

which is why there's the amd version

[glad lance]

no what is that

[proper shale]

oh yeah you should duplicate the space

it gives you a private link to it basically

no queued

queues*

and no errors lol

[glad lance]

and how i do that

[rare goblet]

Ayo? @glad lance level 1 !!!

[blazing garnet]

Ima just use temp mail in voice ai the people who are literally money mining so I can just use infinite models

[proper shale]

just need a huggingface account, there should be a duplicate button on the top

[blazing garnet]

They only give enough coins for one model

[proper shale]

voice.ai is worse tho

¯_(ツ)_/¯

you do you i guess

[blazing garnet]

Not if I scam them by making multiple accounts for multiple models

[glad lance]

thanks

[proper shale]

[proper shale]

you're welcome :)

[blazing garnet]

Bruh google colab is dying since so long I have only gotten to 26 epochs

[rare goblet]

Ayo? @blazing garnet level 4 !!!

[blazing garnet]

Been half an hour

[proper shale]

long datasetvlol

[blazing garnet]

If I copy it onto my drive will it work faster?

[proper shale]

uhhh I don't recommend training for over 2 hours so please stop it before 2 hours

[blazing garnet]

Why?

[proper shale]

...no? its a 40 minutes dataset it will be slow

[proper shale]

you could lose your model

if you dont run export model step

[blazing garnet]

It saves every 50 times how could I lose it?

[blazing garnet]

I ran mostly everything

Even the two for rusuming training

[proper shale]

still, if you dont run export model, it wont save to drive

[proper shale]

aaand you didn't need to do that

[blazing garnet]

REALLY?

BRUH

[proper shale]

yeah

it's to not strain the google drive/colab connection yk

[blazing garnet]

In some years people will just make free ai just removing all sound effects music background noises and train the model so fast

And I am gonna be like my work has been pointless

[proper shale]

theres already an ai to remove bg noise and music

and it is free

[blazing garnet]

Ik

But it doesn't do all

5 epochs in ten minutes holy moly

[loud herald]

@opal jolt btw is there any way to lower the buffer/delay?

[blazing garnet]

Wrong mj

Lol

[loud herald]

oh

😭

[loud herald]

@proper shale mb

😭

[proper shale]

what's the average res? you can check that in the gui

[blazing garnet]

How do I get into verified chat?

[proper shale]

Get level 5

[blazing garnet]

K

Hello poopmaster

[half cove]

shit i forgot i had silent typing off

cant even lurk in ai hub chat before getting caught anymore

[blazing garnet]

VENCORD USER SPOTTED

[half cove]

https://tenor.com/view/nuh-uh-beocord-no-lol-gif-24435520

[blazing garnet]

Bro just put yourself as offline

Wait u are offline

[half cove]

😭

[loud herald]

43 ms

[proper shale]

u can reduce chunks to 112 or even 96

[loud herald]

and that will reduce the delay?

[rare goblet]

Ayo? @loud herald level 2 !!!

[proper shale]

yeah

[loud herald]

ok thx

[blazing garnet]

EPOCH 50 REACHED IN 1 HOUR

[mortal vortex]

salut, c'est quoi les paramètre pour la meilleur qualité ?

[halcyon ledge]

who can help me

[proper shale]

just state your issue and I'll try my best

[halcyon ledge]

how do i upload a model

[rare goblet]

Ayo? @halcyon ledge level 1 !!!

[blazing garnet]

Just go to the weights folder and put in ur pth

Then go to logs and paste ur index file

[proper shale]

upload where

[latent kettle]

Where

[blazing garnet]

Good question I should've asked that

Did I solve his problem???

[halcyon ledge]

hmm

[blazing garnet]

Upload where?

[halcyon ledge]

on raltime voice changer

[blazing garnet]

W okada?

[halcyon ledge]

i have a issue

it says no mudle named "pyngrok"

[blazing garnet]

Ask @proper shale

[proper shale]

are you using the colab

[blazing garnet]

Rip bros net may be bad or he is looking at a tutorial

[halcyon ledge]

i fixed it

[proper shale]

oh

[halcyon ledge]

i need help uploading this one

Smii7y

how do i upload it

[proper shale]

u can check a W-Okada guide if u arent sure, just check the part where it talks bout models

[halcyon ledge]

do i click upload

on the model slot config

what voice change type do i pick

[proper shale]

yeah

[proper shale]

rvc

[halcyon ledge]

i let it upload

[proper shale]

alright

[halcyon ledge]

its on 99%

[rare goblet]

Ayo? @halcyon ledge level 2 !!!

[halcyon ledge]

stuck on 99%n

not working

moving

[halcyon ledge]

still stuck

[proper shale]

wait a lil

[blazing garnet]

It's been 2 hours do I stop it?

[agile storm]

guys, how to continue training in easy GUI?

[blazing garnet]

Simple

Don't use easygui

[proper shale]

💀

[blazing garnet]

Use rvc disconnected

[agile storm]

simple

its not working for me

[proper shale]

💀

[blazing garnet]

I tried easy gui for so long

[proper shale]

just insert same settings

[blazing garnet]

It sucks

[proper shale]

in the easy gui stuff

should be easy lol

[blazing garnet]

Do I stop it and how it's been 2 hours

[proper shale]

just stop the step, make sure it's not on a saving epoch tho

[agile storm]

do I need to preprocess?

[proper shale]

no

you've already done that

and no need to feature extract either

[blazing garnet]

K but it's been running for 2 hours and a half and nothing happened I am gonna stop anyways

[proper shale]

yeah u can stop most definitely

[halcyon ledge]

Frequnt Errors occur. Please check if the model of the framwork is being tsrgeted is loaded.

[blazing garnet]

I am gonna use the model if it sounds bad I only have 50 epochs I will give up

[proper shale]

okay sure

[halcyon ledge]

@proper shale

[proper shale]

is the thing working tho

[blazing garnet]

Can u send me a voice which doesn't sound like an ai to put in it?

[halcyon ledge]

idk

its on 99% still

[proper shale]

idk, you'll have to search for that manually tbh

[blazing garnet]

I searched and the voices I found sound very like an ai so it ruins the conversion too

[agile storm]

for some reason when I start training it just launches tensorboard

and does nothing after that

oh, I dont think I should show my ngrok token

well idc

[proper shale]

ehhh i honestly don't know how to resume on there

thought it would've been simple

[blazing garnet]

Who will literally take the effort to copy that

[glad zealot]

how long has it been when the tensorboard shows up

cause iirc it should only take less than 10 sec to start it fully

[agile storm]

yes, it takes like 10 seconds or less

to start it

[blazing garnet]

@proper shale IT WORKED SO WELL EVEN AT 50 EPOCHS

WORTH IT

[proper shale]

woo!

[blazing garnet]

look what i made 💀

[proper shale]

sounds ok tbh! if u train more it could sound even better

[solar jacinth]

mr . @glad zealot

everyone

I'm using hina 's muti kaggle notebook

I'm running mangio tweaks on it and

I drop a voice model.pth to data

[glad zealot]

you can ignore it

[solar jacinth]

how can i put the .pth file into weights folder

[blazing garnet]

Bruh just move it in

[glad zealot]

you have to upload it as a dataset and use commands to move it

[blazing garnet]

Go in mangio folder and find weights then put it IN

[glad zealot]

since you cant drag and drop

[blazing garnet]

What

I am confused

[glad zealot]

ye cabt drag and drop files to folders on kaggle

[solar jacinth]

i will try this one mr. hina

[solar jacinth]

yeah me too ToT

[brittle wing]

Yo guys, I could trained the models for the first time I installed applio, but now it's giving me this error when I tried to train a new model for the second time, anyone knows a fix?

File "C:\Users\User\Desktop\Apple\Apple\Applio-RVC-Fork\infer-web.py", line 1093, in click_train
& set([name.split(".")[0] for name in os.listdir(f0_dir)])

FileNotFoundError: [WinError 3] The system cannot find the specified path: 'C:\Users\User\Desktop\Apple\Apple\Applio-RVC-Fork/logs/deborah/2a_f0'

[blazing garnet]

u running on local?

[brittle wing]

Yep, I do

[rare goblet]

Ayo? @brittle wing level 1 !!!

[blazing garnet]

it is basically saying it cannot find the thing u out

so justt right click on the file and copy its path again

and remove the inverted commas

[blazing garnet]

[brittle wing]

I'll try, thank you!

[brittle wing]

[blazing garnet]

did it work?

[brittle wing]

I'll be trying to find the commas

[blazing garnet]

bruh i mean like the commas at the end and start if u have any

bro...

are you copying th file path or folder path?

[brittle wing]

The folder path

Sorry, this is my first time using applio lmao