Oh yeahh that thing
#✨│ai-help
finite galleon
Year old me was hyped about that until i got news its discontinued since mangio focused on his studies
lavish lintelBOT
finite galleon
When the hell did i get a raboot
low shard
When the hell did i get a raboot
yup
i don’t remember that
i think u want a realtime voice changer for calls, tell me ur pc gpu in #🔍│help-w-okada , don’t use yt tuts
low shard
damn
hallow thistle
craggy brook
https://www.youtube.com/watch?v=QT0oUARKw68&t=1s In this video I will use Meggy Spletzer as the sound. How can I edit and adjust the sound as beautiful, clear and realistic? Can you give me advice on how to adjust it?
jovial dock
tame finch
hi im training a model with appllio (local) and i get this error "UnboundLocalError: local variable 'reference' referenced before assignment
An error occurred extracting the index: need at least one array to concatenate
If you are running this code in a virtual environment, make sure you have enough GPU available to generate the Index file.
"
pine rose
Where can I see all the voices
craggy brook
https://www.youtube.com/watch?v=QT0oUARKw68&t=1s In this video I will use Meggy ...
how is it done
crystal bridge
Any issue with rvc and discord ? Even with vb cable input it doesn't get any sound out
pearl cove
best software for denoising
AI HUB Docs
tame finch
hi im training a model with appllio (local) and i get this error "UnboundLocalEr...
can anyone help me figure out why im getting this
gleaming badge
Hi. I have a trained model with a voice and i am currently processing the audios with rvc gui. However i would like to process the videos from the command line or from python. Is there any other tool i can use ?
low shard
hi im training a model with appllio (local) and i get this error "UnboundLocalEr...
what's ur pc gpu
tame finch
3080 @low shard
low shard
Where can I see all the voices
You can search rvc ai voice models at:
- #1175430844685484042
- In #🔍│find-models , Do /find with @earnest musk
- https://weights.gg/ (login required)
- https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
- https://voice-models.com/
- https://thevoicemodels.com/ (for Turkish Models, login required with discord and level 2 on their server)
if there isnt one, you can:
- #1159289738314919936
- #1191429836321849435
- make it yourself with our docs guides https://docs.ai-hub.wtf/essentials/how-to-make-voice-models/
earnest muskBOT
You can search rvc ai voice models at:
- <#1175430844685484042>
- In <#11635920...
:wave: @low shard, How can I help?
Available Commands:
• @weights find <query> or /find <query> - Search for RVC Voice Models
• /create - Create an AI Cover
• /image - Generate an Image
low shard
Any issue with rvc and discord ? Even with vb cable input it doesn't get any sou...
that's not RVC, It's wokada which uses RVC for realtime, use #🔍│help-w-okada and explain what tut link u followed, what's ur pc gpu too
low shard
How do I resume training w Applio no UI???!?!
not sure if @nocturne mural maybe has any guide on it
low shard
Hi. I have a trained model with a voice and i am currently processing the audios...
RVC GUI is ol
the only cli u could try is https://github.com/blaisewf/rvc-cli
what's ur pc gpu
don't follow old yt tuts
low shard
3080 <@911742715019001897>
mm, at which specific step do u get that issue
gleaming badge
what's ur pc gpu
I got a 1660 super
tame finch
perhaps i did a earlier step wrong
gleaming badge
but somwtimes i have to process audios on my laptop with no gpu
low shard
I got a 1660 super
can be used for training and inference, not much suggested for training tho even if it's possible
low shard
but somwtimes i have to process audios on my laptop with no gpu
u can just use cloud (remote good pc)
gleaming badge
the only cli u could try is https://github.com/blaisewf/rvc-cli
Thanks man, gonna try it later
low shard
Train (make) RVC Models on cloud:
- Prepare the Dataset
- Setup RVC:
Choose a cloud way to use RVC,
- Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
- Applio (ui)
- Mainline (UI)
- RVCDISCONNECTED (no ui)
- Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
- Mainline (UI)
- Applio by Vidal (UI, no guide as of right now)
- Applio by Shirou (UI, no guide as of right now)
- Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly):
- Be sure to know about the tensorboard
Google Colab = Easier but risk of getting disconnected
Kaggle = Harder but way more gpu time
If you are looking for the easiest way and for free, is using https://weights.gg which ofc uses RVC
RVC Inference (use models) on pre-recorded audio on Cloud
You can use either:
- Weights.gg: Easiest Possible Ever Automatic
- Ilaria RVC Zero: Fastest free on cloud
- Applio (ui)
brittle wing
not sure if <@989772388508000306> maybe has any guide on it
Uh
tame finch
can you give me a guide to the appllio local?
maybe ill try following one
and see if the issues persists
low shard
mm, did you do feature extraction
low shard
can you give me a guide to the appllio local?
sure, our docs https://docs.aihub.gg/rvc/local/applio/
gleaming badge
# Train (make) RVC Models on cloud:
1. [Prepare the Dataset](<https://docs.ai-hu...
Thanks again but i think i am gonna stick with processing the audios locally
low shard
i pressed "extract features" haha so i think so? let me start over
check also the output of that, be sure it's not 'no-feature-todo'
low shard
Thanks again but i think i am gonna stick with processing the audios locally
alright, I'm just telling you that it will be slower locally because of your hardware
but your choice at the end of the day
gleaming badge
alright, I'm just telling you that it will be slower locally because of your har...
i know but it doesn't matter alot
because after processing i need to send a message to a kafka broker
And i dont know if i can do that with the cloud option
tame finch
check also the output of that, be sure it's not 'no-feature-todo'
just did it again, output says model extracted sucessfully
low shard
because after processing i need to send a message to a kafka broker
kafka broker ?
gleaming badge
kafka broker ?
its for microservices
low shard
did u uncheck audio cutting in preprocess?
low shard
its for microservices
yeah I dunno about it
low shard
i wanted it to cut the space for me
could u show me ur preprocssing settings
low shard
It seems like u didn't make a folder inside the datasts folder
low shard
or is just the zip of acapellas enough
it should be folder inside of the datasets folder with ur audios in it, not a zip
tame finch
ohh so only the cloud uses a zip?
low shard
like assets/datasets/acas, and inside the acas folder, put the audio files
tame finch
ok got it
trying that way
ok seemed to have worked now.. only took 18 seconds is that normal?
ah wait the gui said "training successful".. but the command prompt says
ah wait when preprocessing, i see this error
nvm i just had to update the path to include the folder loll
wanton frost
"Preprocess completed in 0.00 seconds on 00:00:00 seconds of audio."
BRO THE AUDIOS ARE LITERALLY THERE???
i'll explain my problem in a minute hold on
thorn carbon
guys approximately how long of voicedata do i need to make a good sounding voice model?
@acoustic anchor did you find the newest local RVC training thing? Cuz i need it too 😭
wanton frost
i'll explain my problem in a minute hold on
above that reply, you'll see that there's an explanation to use applio... as the old rvcDisconnected way to train models is finally gone.
in that face-to-face tutorial between nick and kc, nick says to kc that he should use his own audio folders in his device instead of zipping them or using google drive as it works easier that way
so what i've done was to create a folderin my phone called "assets". then a "datasets" folder inside it. and then finally a "acas" folder to put my wav lead vocal audios in it
then onto the training part, i started applio. and then entered the public link. to where i should be adding my audio files, i wrote in "assets/datasets/acas" and when i presses the "preprocess files" button. it said "model sueda(the name of the singer i wanna train) preprocessed"
but... the colab says this instead...
what do i do? however i try, i can't seem to start the "epoch 1, epoch 2, epoch 3" count. i know how to use tensorboard too do it's not about tensorboard
wanton frost
above that reply, you'll see that there's an explanation to use applio... as the...
(the TB is just below the screenshot. i just scrolled up bc i thought it would be easier to read out the writings and see what there was wrong)
thorn carbon
I've been using Mangio RVC but i wonder if there is a better RVC voice cloner cause Mangio didn't seem to work very well for me.
wanton frost
I've been using Mangio RVC but i wonder if there is a better RVC voice cloner ca...
the wiser people around here recommend applio. and my questions above were about it. i started learning about applio yesterday so idk if ur asking me but that's all i could add lol. we still gotta wait for them to come tho
thorn carbon
alright, thank you, i'll give that a check
brittle wing
the wiser people around here recommend applio. and my questions above were about...
I wanna use Applio no UI for training, how do I exactly resume the training for a model also my data should be audio files in a folder, not a zip file?
What should the folders be named. Also correct path?
This is where I place my pretrain links I guess
wanton frost
I wanna use Applio no UI for training, how do I exactly resume the training for...
im using the UI one sadly so idrk about your question sorry💔. might rn even switch to no UI to try it out. maybe i can successfully train on that so i could get back to you before the wise ppl do
(i am a noob on that too stop asking me im begging😭😭😭)
simple ore
what do i do? however i try, i can't seem to start the "epoch 1, epoch 2, epoch ...
slice the files
brittle wing
im using the UI one sadly so idrk about your question sorry💔. might rn even swi...
Uh I'm sorry
brittle wing
slice the files
Hi, how do I resume training on Applio no UI?
simple ore
Hi, how do I resume training on Applio no UI?
as long as the files are there, you just start training with the same parameters and it resumes
brittle wing
as long as the files are there, you just start training with the same parameters...
There's a resuming cell.
simple ore
no
brittle wing
simple ore
okay, so it copies the files from backup back to logs
brittle wing
okay, so it copies the files from backup back to logs
What are the resuming steps
simple ore
and it sets the training values
brittle wing
So I don't mess up
brittle wing
then you just go back to train, I guess?
Yes but what cells should I avoid when resuming
brittle wing
Is there any guide
brittle wing
no preprocess/no extract features
Uh that's the same as RVC disconnected
The resuming cell loads the preprocessed files
Someone just sent some only fans link and immediately deleted it 💀
-guides
thorn carbon
guys how do i change the language of applio
thorn carbon
oh
thorn carbon
okay i did it thanks
thorn carbon
in settings
i genuinely couldn't find it at first i'm sleepy 😭
terse nymph
Can I run Ilaria RVC locally?
I can only find ways to run it on huggingface. I want to use my GPU
hallow thistle
I can only find ways to run it on huggingface. I want to use my GPU
Ilaria RVC from Hugging Face? The author of this program hasn't made it available to run locally outside the Hugging Face space unfortunately. But you can try another fully version of RVC program that can use GPU instead.
terse nymph
Ilaria RVC from Hugging Face? The author of this program hasn't made it availabl...
If it provides similar results I'm open to the others.
hallow thistle
-rvc
hallow thistle
Applio
gleaming badge
the only cli u could try is https://github.com/blaisewf/rvc-cli
hey man, i just set this up and it worked perfectly. Thanks !
hallow thistle
That's great.
placid stump
Hi anyone know why I cant add this voice? Pressed upload and nothing happen #1214406843590377482 message
@rigid mango sorry for tagging
rigid mango
Because it’s not an RVC model
It’s made for a completely different architecture
Use this model if you want to do stuff with Zeta
hallow thistle
"RVC" program can only use RVC voice models, as the name says.
dreamy barn
/ reate
fleet trail
wew
simple ore
Will kmeans make my index better?
it is only when you have 200,000 slices
brittle wing
it is only when you have 200,000 slices
How many mins of dataset in simple words
https://docs.applio.org/applio/getting-started/other-alternatives
I found how resuming is done maybe...
Should I run preprocessing, feature extraction
brittle wing
not for resume
Scroll down there's a resuming section
hallow thistle
-colab
Why you keep provoking the Automaze bot after you ask anything? This is like not the first time you've done this. 
simple ore
yes, it copies data from backup, that includes sliced audios and extracted features
so no preprocess is needed
brittle wing
yes, it copies data from backup, that includes sliced audios and extracted featu...
Same as RVC disconnected
brittle wing
Why you keep provoking the Automaze bot after you ask anything? This is like not...
And what's the problem.Its my keyboard, my choice and stuff
Well since I like to find info by myself too
hallow thistle
And what's the problem.Its my keyboard, my choice and stuff
hallow thistle
Are you even an admin or staff?
No.
hallow thistle
brittle wing
Then why do you keep making remarks towards me?
brittle wing
<:kittybruh:1167393213661839433>
?
hallow thistle
Oh, so you called me a kid just so I could cry faster? Nah, I asked you in the most calmed way, and you completely lost it, taken it way too seriously.
brittle wing
Oh, so you called me a kid just so I could cry faster? Nah, I asked you in the m...
Why you keep making remarks
Why you have beef w me?
I just wanted to train w Applio this time.
hallow thistle
Because I keep seeing you doing this anytime you asked anything, for like the past days. And I wanna know why you still doing this even to this time, this day.
brittle wing
Because I keep seeing you doing this anytime you asked anything, for like the pa...
It's a habit 🖤
hallow thistle
It's a habit 🖤
brittle wing
yes, it copies data from backup, that includes sliced audios and extracted featu...
So
Should I run this first and the training ones after
And rerun this for restarting resume point?
Wait maybe I understood
brittle wing
<:shrug:1159570387953258616>
I apologize if I called you a kid.
brittle wing
1) resume-training to load data back and set the values
And then go back to where you put the epochs in
Mmm
Does this colab really save the weights in my drive
hallow thistle
I apologize if I called you a kid.
Well, next time just don't do this again.
brittle wing
Well, next time just don't do this again.
Yes it was rude
brittle wing
I said that kids are a nice thing
hallow thistle
what was that
hallow thistle
will my gpu explode
What is your GPU? If it's an old GPU, with super hard tasks going on, it will sure explode anytime.
brittle wing
I read somewhere you can't post models of Hitler for ex
Can I post models of any singer I want?
barren vault
the difference between rvc and w-okada
hallow thistle
I read somewhere you can't post models of Hitler for ex
There are voice models of him available. Not sure why they say they wanna ban this though.
brittle wing
There are voice models of him available. Not sure why they say they wanna ban th...
What the-
hallow thistle
the difference between rvc and w-okada
W-Okada is the "realtime" voice changer. RVC is the audio changer, or audio conversion.
barren vault
W-Okada is the "realtime" voice changer. RVC is the audio changer, or audio conv...
will w okada colab explodes my gpu
hallow thistle
will w okada colab explodes my gpu
Uh. Google Colab uses GPU from its own. It doesn't use your GPU to process that.
barren vault
Uh. Google Colab uses GPU from its own. It doesn't use your GPU to process that.
So it wont explode
hallow thistle
It won't explode. Unless you do ray tracing 4K ultra HD on your GPU, it will sure explode.
hallow thistle
W-Okada, RVC, Stable Diffusion and any related Python program can work around with GPU. With proper settings, you won't make your GPU to explode.
barren vault
-colab
hallow thistle
With proper cooling system for your GPU, you won't have to worry about if it gonna overheat.
hallow thistle
The links where you provoked Automaze to send are links of Google Colab RVC notebooks.
barren vault
What is the use of every single one
hallow thistle
I'm not a developer, unfortunately, so I can't tell the detail any of these link one by one. But first three links have an option to train a voice model.
silent turtle
I trained a voice with Applio using a dataset of someone only speaking, and it doesn't reach certain notes. If I use covers I made with that model and train a new one but only with the singing voice, could it achieve more tones?
hallow thistle
An audio with monotone voice can only do monotone just that.
Trying to get "monotone" voice model to reach at that high note can result in the most goofy way.
silent turtle
Trying to get "monotone" voice model to reach at that high note can result in th...
I just had a doubt about whether doing this would improve the models quality. Thank you.
wanton frost
damn thank god i switched to try "applio no UI". it worked for me and was almost like the "rvcDisconnected v2 by KitLemonFoot" that i was used to. if anyone wants some tips about training a model in a mobile device, with "Applio no UI" now MAYYYYYBE i could finally be able to help.
idk if ur a staff seeing this, you could note me down somewhere and @ me when needed
low shard
nvm i just had to update the path to include the folder loll
Is this fixed now?
low shard
guys approximately how long of voicedata do i need to make a good sounding voice...
Well I'd say minimum 5 or 10 mins for a decent one, max 1 hour
tribal cosmos
is there a guide for making good data sets?
opal cobalt
hello somone now how can i make that work with amd gpu(rx6750xt) pls ?https://github.com/Eddycrack864/UVR5-UI
low shard
hello somone now how can i make that work with amd gpu(rx6750xt) pls ?https://gi...
I don't think @viscid moss's project has AMD support, maybe you can try to patch it yourself with Zluda or wait for Eddy's response
opal cobalt
I don't think <@274566299349155851>'s project has AMD support, maybe you can try...
ok thx
knotty moth
hello somone now how can i make that work with amd gpu(rx6750xt) pls ?https://gi...
use the original UVR and it has AMD gpu support
opal cobalt
use the original UVR and it has AMD gpu support
yes but this one doesn't support the mdx23c models
knotty moth
yes but this one doesn't support the mdx23c models
do you refer to the UVR UI? I dont recall the original one is so
you can try find the models in Settings -> Download Center
also, make sure to update UVR to the latest patch that enables downloading roformer models
opal cobalt
but that doesn't include mdx23c model no ?
knotty moth
but that doesn't include mdx23c model no ?
just try do what I said instead of just assuming that
opal cobalt
just try do what I said instead of just assuming that
yes this is what i am doing rn but i i am jsut searching a way to have those model in local
knotty moth
yes this is what i am doing rn but i i am jsut searching a way to have those mod...
because again it is not preincluded in the installation, but through the "DLC"
opal cobalt
because again it is not preincluded in the installation, but through the "DLC"
ok sorry ( i don't understand really well english)
i download the beta and i didn't have reformer in the download center did i need to do more step pls ?
and another click on the link 🥲
viscid moss
hello somone now how can i make that work with amd gpu(rx6750xt) pls ?https://gi...
Nope, doesn't have AMD support yet.
viscid moss
i download the beta and i didn't have reformer in the download center did i need...
If u have the beta version, go to the download center and download the model u want
opal cobalt
If u have the beta version, go to the download center and download the model u w...
ok thx you 🙂
and another time sorry @knotty moth
simple ore
Nope, doesn't have AMD support yet.
i bet all it needs is a zluda patch like Applio
knotty moth
i bet it could perform better than directml backend in the original UVR
viscid moss
i bet all it needs is a zluda patch like Applio
I'll check Applio's code to try adding that AMD support but in a while I'm busy with irl stuff atm
simple ore
it is literally - patch torch libraries, add import zluda into main script, then start it with zluda\zluda -- python script.py
torch 2.3.1+cu118
can be lower
viscid moss
torch 2.3.1+cu118
Wait, torch version must be 2.3.1 or lower?
I'm using the latest
2.5.1 iirc
simple ore
lower is fine, 2.5.1 is not compatible
viscid moss
Damn
simple ore
not much benefit there anyway
viscid moss
Well i can downgrade it
knotty moth
Wait, torch version must be 2.3.1 or lower?
2.4 and above are mostly not compatible
and so are some comfyui nodes I ever had
simple ore
normally you just want to use separate virtual environments
so 2.3.1 for one project, 2.5.1 for other
viscid moss
Ye that's why i actually do, but using the latest
knotty moth
well I mostly have some colabs and kaggle notebooks
viscid moss
I'll try in like... 1 week. I'm working on computer vision models atm (YOLOv11)
thorn carbon
Well I'd say minimum 5 or 10 mins for a decent one, max 1 hour
alright thank you
vale raptor
how can i use rvc locally
low shard
how can i use rvc locally
what's ur pc gpu and what are u looking for
vale raptor
what's ur pc gpu and what are u looking for
2060 super, just changing voice
low shard
2060 super, just changing voice
realtime voice changer for calls? Then it's wokada which uses RVC for realtime, not just RVC
I will give u the link in #🔍│help-w-okada
vale raptor
im sure you mean the non-realtime one
thanks
low shard
im sure you mean the non-realtime one
actually both lol #🔍│help-w-okada message
tame finch
Is this fixed now?
yes it worked thank you
low shard
yes it worked thank you
yw
safe moat
tried running the gui thing and this turns up
simple ore
why are you in system32 my dude
safe moat
as admin btw.. it doesnt seem to do anything if im not running it as administrator
simple ore
why are you running it as admin
safe moat
bc it doesnt seem to do anything otherwise
simple ore
go to the actual folder with the application
unless you're brain dead and unzipped it into windows/system32
safe moat
ill just let the bat run while i mkae dinner incase it decides to do something
safe moat
unless you're brain dead and unzipped it into windows/system32
be nice
i did not do that..
simple ore
so then why you're trying to run it from system32
safe moat
it does that when i run it as admin
knotty moth
-gui
simple ore
dont run it as admin
knotty moth
tried running the gui thing and this turns up
btw yea should just run the bat file as is
safe moat
bro i just followed the guide sent by a helper 😭
safe moat
oke how do i get the new one
simple ore
voice changer - https://rentry.co/forkvoicechangerguide#download-nvidia
safe moat
i will try that in due time thank u
knotty moth
oh wait, it is the realtime-gui
but the regular go-web.bat should run fine
simple ore
he did not say what he wanted
voice changer or rvc
the command line he used was for the voice changer
hallow thistle
C:\WINDOWS\system32 isn't where the RVC folder located.
simple ore
unless you're dealing with a computer genius
hallow thistle
The actual RVC folder path could be something like E:\RVC or C:\Users\your username\Downloads\RVC, depends on where you extracted.
pastel oak
voice changer or rvc
ame wants voice changer
knotty moth
as admin btw.. it doesnt seem to do anything if im not running it as administrat...
for me it's always the rvc's current directory, not the system32 as its "working directory"
lavish lintelBOT
safe moat
ive no idea why it does that when i open it as admin
i feel like a primate when i have to work with programs i do not know
safe moat
voice changer - https://rentry.co/forkvoicechangerguide#download-nvidia
okay.. ive tried this one. i still get the exact same error i was getting on w-okada where it just errors when i put in my microphone
mint anvil
safe moat
other microphones work fine.. it just hates one in particular
mint anvil
what am i supposed to do here?
low shard
what google colab link are u using and what u want to do
mint anvil
i want to process files for creating a model and im using easygui
queen fjord
AICoverGen WebUI
Why can't I use it?
Can you help us?
error
"'NoneType' object has no attribute 'setdefault'"
Up like this.
simple ore
use the collab that works
queen fjord
Teach us a little.
wary stone
what's your go-to voice remover?
low shard
i want to process files for creating a model and im using easygui
easygui is broken
check #📰│dev-updates
use smt else
low shard
AICoverGen WebUI
send link of what u using
shrewd bolt
how to do ai models
shrewd bolt
https://docs.aihub.gg/essentials/how-to-make-voice-models/
thx
pastel oak
if you have an nvidia RTX gpu, on step 2 do locally -> applio
if you dont then use colab/cloud options
shrewd bolt
if you have an nvidia RTX gpu, on step 2 do locally -> applio
if you dont then ...
i have thx 1650
gtx
pastel oak
i have thx 1650
Could work but its below the recommendation
Someone else should guide you on that
shrewd bolt
how much epochs i need set for 1 minute dataset?
low shard
https://colab.research.google.com/github/SociallyIneptWeeb/AICoverGen/blob/main/...
That is old
-colab
queen fjord
@low shard Can you recommend it? I'm a newbie.
queen fjord
Exception: [Errno 2] No such file or directory
low shard
<@911742715019001897> Can you recommend it? I'm a newbie.
You can use Hina mod ai cover gen which is a fixed version of what u want to use
Also, don't use yt tuts
low shard
Exception: [Errno 2] No such file or directory
By using what
low shard
I can access the user page.
What user page
pastel oak
im not sure if you should be sharing that link
low shard
AICoverGen WebUI (Modded by Hina)
Alright, what did you do step by step
low shard
Will you teach me?
btw, do you want to use AI Cover Gen which has no guide but it's perk is that it will automatically separate vocals and instrumentals, or would you rather use another colab that has a guide like Applio?
it's kinda ur choice on what to use
the easiest thing ever is https://weights.gg for example, that's cloud but not a google colab, still uses RVC and automatically separates vocals & instrumentals
golden sphinx
someone help kaggle not working rn
simple ore
I working with CoverGen_NO_UI_v2_en but see this error
using colab with torch 2.6, that's what
lofty shuttle
Do you have the same errors?
queen fjord
@lofty shuttleI am the same as you
simple ore
you need to use colab that has depencencies fixed
queen fjord
Where to find it?
simple ore
silent stratus
how can i open tensorboard in lightning ai applio
queen fjord
I really don't know where to start.
silent stratus
I really don't know where to start.
what are u looking for?
lofty shuttle
it's good link and working colab?
https://colab.research.google.com/drive/1u1brjK8IZt647UsbZuGYfW29oFM2I4tk?usp=sharing
lofty shuttle
is it int he list?
from list of colabs
silent stratus
it's good link and working colab?
https://colab.research.google.com/drive/1u1brj...
https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb try this or weights.gg
way easier and up 2 date
lofty shuttle
https://colab.research.google.com/github/iahispano/applio/blob/master/assets/App...
it for music voice cover?
silent stratus
it for music voice cover?
search the voice u want and click create
then put ur song in
lofty shuttle
https://www.weights.gg/
google colab will no longer work?
silent stratus
google colab will no longer work?
weights is just way easier
plus i dont think it stems the audio for u
so weights is just better
plus u dont gotta import models u just search
lofty shuttle
weights is just way easier
it free for use or have limits?
silent stratus
it free for use or have limits?
free
lofty shuttle
just tried it, but the result is noticeably worse than in google colab . the original voice is almost unchanged
will colab work again in the future?
latent rain
Running with the system Python.
Traceback (most recent call last):
File "C:\Users\admin\Downloads\RVC-GUI-main (1)\rvcgui.py", line 4, in <module>
import soundfile as sf
ModuleNotFoundError: No module named 'soundfile'
help
brittle wing
Do I have to download & store the pretrains in my drive somewhere???
hallow thistle
Running with the system Python.
Traceback (most recent call last):
File "C:\Us...
-gui
hallow thistle
Using RVC GUI in 2025? That's crazy. This program is from 2023, and no more such an update to be updated.
hallow thistle
If you wanna use RVC for basic AI cover, I'd suggest to go for Applio instead.
-rvc
hallow thistle
latent rain
i cant send screenshots
hallow thistle
You know what you gonna do next, right?
hallow thistle
You must talk a little bit until your name turns light blue, so now you will be able to send an image here.
latent rain
ok
hallow thistle
If the program files still in a zip, just extract them to a folder.
hallow thistle
If you see this the first time, it's pretty normal. However, if you got a slow CPU, it can make Applio to launch much slower.
hallow thistle
This behavior is normal, which means Applio has launched successfully.
You're welcome. 
final violet
so
when I try to run the rvc models, it lags and goes reeeaaalllyy slow
any ideas why?
hallow thistle
If you're looking for W-Okada, this channel isn't where you asking for W-Okada. Go to #🔍│help-w-okada instead.
hallow thistle
W-Okada is the real-time voice changing program.
I don't think an "RVC" program can make one system to lag while processing.
silent stratus
yo thanks for using my model lmao, try gewtting the newest applio build or infer in a cloud notebook
hallow thistle
yo thanks for using my model lmao, try gewtting the newest applio build or infer...
brittle wing
Is it possible for it to overtrain already at 30e?the dataset is about 30 minutes, I also used ov2super.
Ah no it's going even more down at 70k, my mistake
copper elbow
What pre trained base model do you guys recommend? I see that there are some up on hugging face.
(still very new at all this)
I feel like I can get better results using things that aren't default models that come with WebRVC
(Yes I know the golden "Models will never be better than it's training data rule" my sets are clean, just wondering what would you guys recommend)
knotty moth
What pre trained base model do you guys recommend? I see that there are some up ...
if not sure, use the default one
knotty moth
(Yes I know the golden "Models will never be better than it's training data rule...
may be true for novices, but you gotta wonder why deepseek R1 can be so good despite trained using some data generated from GPT or another LLM
simple ore
seriously, none of you have ever been in the real 'overtrained' zone
copper elbow
may be true for novices, but you gotta wonder why deepseek R1 can be so good des...
Still very new to this stuff I know little to nothing, thankyou for your input
red kayak
@misty forge heyy
misty forge
Hey.
red kayak
So you need to go to the very last step
And unselect pretrained
Refinegan doesn't have any pretrains
So you need to train one on your own
Hence why it's showing the error
misty forge
But… huh?
misty forge
This…
This is a pretrained refinegan model, isn’t it?
This is the one I’m trying to use.
red kayak
This is the one I’m trying to use.
Dog brain moment...
Well the pretrains says it's 44k
crude flame
Codenames fork doesnt work with the newer refinegan pretrains because he hasnt updated the generator to work with it
red kayak
Not 44.1k
red kayak
Codenames fork doesnt work with the newer refinegan pretrains because he hasnt u...
There you go
misty forge
Oh.
lavish lintelBOT
misty forge
So… what should I do then?
crude flame
use applio
crude flame
just download the zip and run the install bat file
lavish lintelBOT
misty forge
You mean the latest version? When I used the latest version, I didn’t see any settings for refine…
crude flame
You mean the latest version? When I used the latest version, I didn’t see any se...
thats the precompiled version
misty forge
Oh…
meager comet
how badly does local training burn a gpu
low shard
I working with CoverGen_NO_UI_v2_en but see this error
send the link of the colab
low shard
how badly does local training burn a gpu
you should be fine as long as you got good ventilation, it shoudn’t burn just for a single training
meager comet
you should be fine as long as you got good ventilation, it shoudn’t burn just fo...
longevity wise
knotty moth
how badly does local training burn a gpu
perhaps if you are using a laptop gpu that easily overheats, unless you do temp limit and/or simply use a desktop one with decent cooling
lofty shuttle
send the link of the colab
low shard
<@274566299349155851> https://colab.research.google.com/drive/1qezaBvCn4eSG1U2L4...
seems broken rn, u gotta wait for a fix
low shard
https://colab.research.google.com/drive/1u1brjK8IZt647UsbZuGYfW29oFM2I4tk?usp=sh...
its broken rn, use smt else or wait for a fix
@viscid moss that colab is broken rn
lofty shuttle
<@274566299349155851> that colab is broken rn
what is "smt"?
pastel oak
what is "smt"?
smt = something
viscid moss
<@274566299349155851> that colab is broken rn
ye, but not RVC related this time, Colab is finally blocking yt-dlp at all
The only way to use that notebook atm, it's using an audio from Drive or manually uploaded to Colab
viscid moss
wait whaaat
misty forge
Applio should work
??????
I'm using applio, I downloaded the zip file from the github.
analog obsidian
Use a refinegan pretrain
red kayak
lyery u filthy gay help me
misty forge
Yeah, I am.
viscid moss
This error was fixed!
viscid moss
I working with CoverGen_NO_UI_v2_en but see this error
Fixed!
analog obsidian
show me the trianing gui
sure 
misty forge
show me the trianing gui
red kayak
ah owo
misty forge
Anything off?
I was on codename's gui, and it didn't work. Now, I'm on applio's recent branch, and it gives me the same error.
What could possibly be the issue? Is the pretrained model too old?
red kayak
What could possibly be the issue? Is the pretrained model too old?
noooo
idk but they named it 44k
not 44.1k
maybe it was trained with a sample rate of 44k not 44.1k on accident idk
but
try
copuing the g path
and going over here and see if you can check the mdoel info
red kayak
rip the silly info extractor no work as well as it did in the past
analog obsidian
which pretrain are u using
misty forge
which pretrain are u using
analog obsidian
https://canary.discord.com/channels/1159260121998827560/1319621721950785586
In applio
Latest main branch
Not codename fork
red kayak
Not codename fork
they using silly applio ow
analog obsidian
they using silly applio ow
Maybe it’s the pretrain bc klm was trained using the old generator
while noobies pretrain was trained using the new one
red kayak
woof idk
misty forge
Use this https://discord.com/channels/1159260121998827560/1327283027776245810
I think it works.
Yep, it's working.
Thanks.
analog obsidian
Yea
It was that
seoul has to train klm using the new generator to make it work in the latest update
misty forge
Got it.
median tree
i was using mainline kaggle to train, i just restart the session and the pth files in /training/assets/weights are gone, does that mean i failed?
simple ore
they should be in your google drive backup, no?
median tree
they should be in your google drive backup, no?
i did'nt use google drive in the whole process
raw forum
How do I put the modifier as voicemeeter audio output?
or is there some other app that does this?
median tree
they should be in your google drive backup, no?
i have download e100_s7900.pth though, how do i operate the inference ui in this situation?
karmic oliveBOT
topaz zinc
Is the current recommended pretrain https://discord.com/channels/1159260121998827560/1327283027776245810 ?
simple ore
Is the current recommended pretrain https://discord.com/channels/115926012199882...
it is an experimental pretrain for the new vocoder
topaz zinc
it is an experimental pretrain for the new vocoder
is it good enough yet? (for speaking models)
simple ore
it is for you to try a finetuning and provide a feedback
simple ore
it depends
topaz zinc
What does it depend on?
simple ore
you train a model on top of it, you test it, you say "it sucks", then I go train it a bit more
or you say "it is good" and I dont
right now i'm running 32k training
so 44k is on hold
topaz zinc
you train a model on top of it, you test it, you say "it sucks", then I go train...
surely you've tried it yourself? maybe I'll try it I'm not sure I just wanted to know if it's worth trying
simple ore
I've tried many experimental things, I can not provide a feedback for every single one.
topaz zinc
ok thanks I guess
opal cobalt
hello i want to separate 2 voice from an interview someone know how can i do that please ?
simple ore
show the voicemeeter config / voice changer setting
raw forum
I can't send it to this channel, I'll send it to you dm
simple ore
you need to figure out how audio routing works
like what A1, A2 buses are
voice changers input should be an actual microphone
voice changer's output should be a virtual cable
voicemeeter's physical input 1 should be virtual cable
simple ore
@steel forge cleanup
dusk arrow
hello guys can anyone help me
i tried to start the client and it says wait web server and its loading loading and its not starting
onnxruntime.capi.onnxruntime_pybind11_state.InvalidProtobuf: [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from pretrain/rmvpe.onnx failed:Protobuf parsing failed.
it says this anytime i try loading any voice models
brittle wing
Is Ultimate Vocal Remover v5 still the recommended way to separate vocals locally? The Github page of it is not really active now
fathom ore
So I’ve been using RVC Cover maker by Shirou and it’s stopped working. Does anyone have a working colab cover maker link? Thanks
low shard
So I’ve been using RVC Cover maker by Shirou and it’s stopped working. Does anyo...
@warm shell u gotta update ur colab too, check #💬│engineer-chat and #📰│dev-updates
low shard
So I’ve been using RVC Cover maker by Shirou and it’s stopped working. Does anyo...
btw, check #📰│dev-updates for fixed colabs
fathom ore
thank youu
warm shell
<@1032433063713575003> u gotta update ur colab too, check <#1159515072704557127>...
I'd even forgotten I'd done that
fathom ore
@warm shell please fix it it’s the best one 😭
warm shell
<@1032433063713575003> please fix it it’s the best one 😭
I think it's pretty bad, to be honest... But I'll see if I can do something
I've never fixed all the problems, and it's already out of date
low shard
I've never fixed all the problems, and it's already out of date
Btw if u don't want to take care of it anymore u should archive it then
So people know that they shouldn't rely on it anymore and would be removed from our links
Or if u want to take care of it, I pinged u in another chat
You're the creator of that fork afterall
visual wind
best colab to train? it wont let me on disconnected rvc
low shard
best colab to train? it wont let me on disconnected rvc
Train (make) RVC Models on cloud:
- Prepare the Dataset
- Setup RVC:
Choose a cloud way to use RVC,
- Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
- Applio (ui)
- Mainline (UI)
- RVCDISCONNECTED (no ui)
- Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
- Mainline (UI)
- Applio by Vidal (UI, no guide as of right now)
- Applio by Shirou (UI, no guide as of right now)
- Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly):
- Be sure to know about the tensorboard
Google Colab = Easier but risk of getting disconnected
Kaggle = Harder but way more gpu time
If you are looking for the easiest way and for free, is using https://weights.gg which ofc uses RVC
RVC Inference (use models) on pre-recorded audio on Cloud
You can use either:
- Weights.gg: Easiest Possible Ever Automatic
- Ilaria RVC Zero: Fastest free on cloud
- Applio (ui)
either applio colab or applio kaggle
be sure to check if ur pc is good enough for local first tho
silent stratus
how do i get tensorboard up on lightning ai applio
ivory surge
hey everyone! im havin a problem in the aicovergenwebui by hina. when i try to generate the audio it never generates and on colab i get a message in red and yellow telling me theres a problem with CUDA or something like that. hod do i fix that?
serene crown
rvcdisconnected and i got this error eventho i already ran feature extraction along with preprocessing, how do i do this?
simple ore
do i need to turn on pitch guidance while training a model?
always yes
serene crown
rvcdisconnected and i got this error eventho i already ran feature extraction al...
How about this problem? 😭 i kept trying again and again it just gave me the same error output
rain urchin
Hello, once im in here on the browser on the program this doesnt recognize any microphone i use please help me
vague sage
Hello, once im in here on the browser on the program this doesnt recognize any m...
youre using the program?
rain urchin
youre using the program?
im using the google colab vertion
vague sage
im using the google colab vertion
oh, idk how it works, i use and recommend the program version
lavish sleet
Help me
raw forum
Cadê um brasileiro pra ajudar 😭😭😭
low shard
rvcdisconnected and i got this error eventho i already ran feature extraction al...
Rvc disconnected has issues again
thin edge
Rvc disconnected has issues again
so that is normal ?
I've repeated it several times and the error I think is my fault...
low shard
so that is normal ?
You have to use another thing
Or wait for a fix, but tbh rvc disconnected runs on old code, it's slower than before
Check #📰│dev-updates
low shard
Cadê um brasileiro pra ajudar 😭😭😭
We are English only
patent trellis
raw forum
Sorri bro
low shard
Hello, once im in here on the browser on the program this doesnt recognize any m...
Wrong channel, use #🔍│help-w-okada
raw forum
<@155030383648440320> It's giving me audio from my PC
Help
thin edge
You have to use another thing
I want to try to model voice in weight, but I don't know where to put the voice file.
low shard
I want to try to model voice in weight, but I don't know where to put the voice ...
You have to go to train, then import voice model
low shard
<@155030383648440320> It's giving me audio from my PC
Wdym
low shard
<@1032433063713575003> please fix it it’s the best one 😭
Unfortunately it's archived and the creator won't maintain it nor fix it, you have to use smt else
thin edge
You have to go to train, then import voice model
aight
thorn widget
is hina modified realtime still work, I've tried several times but the resolution is running fast and doesn't make any sound at all
low shard
is hina modified realtime still work, I've tried several times but the resolutio...
Elaborate more your issue with a screenshot in #🔍│help-w-okada , this is the wrong channel
thorn widget
aight thanks
low shard
I didn't understand
I don't know neither what u are doing, want to do, nor know what's ur PC GPU or tut link ur following
raw forum
rtx 3070, the audio is duplicating along with my AI voice on voicemeeter
raw forum
rtx 3070, the audio is duplicating along with my AI voice on voicemeeter
B1 it's playing audio from my PC
raw forum
B1 it's playing audio from my PC
raw forum
B1 it's playing audio from my PC
is sending the audio of the PC to the microphone
low shard
Oh that's Wokada not rvc, let's talk in #🔍│help-w-okada
raw forum
ok
serene crown
Or wait for a fix, but tbh rvc disconnected runs on old code, it's slower than b...
😪😪😪
Other colabs ive checked so far dont have magio crepe
Anyone know if rvcdisconnected is being fixed?
Or not abandoned at least
low shard
Other colabs ive checked so far dont have magio crepe
They do, Applio has mangio crepe, it's just renamed as normal crepe replacing it
low shard
Or not abandoned at least
Idk if it will be fixed, 3rd time broken this week, and the creator is kind of busy irl
But yeah probably
I'd not suggest it much tho, to fix it he had to use the mangio rvc fork, which is old and slower, bc Alex mangios rvc tweaks got deleted
Applio and mainline would be more updated and faster
serene crown
They do, Applio has mangio crepe, it's just renamed as normal crepe replacing it
I see
I'll try training on Appilo later
low shard
I see
I'll try training on Appilo later
Train (make) RVC Models on cloud:
- Prepare the Dataset
- Setup RVC:
Choose a cloud way to use RVC,
- Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
- Applio (ui)
- Mainline (UI)
- RVCDISCONNECTED (no ui)
- Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
- Mainline (UI)
- Applio by Vidal (UI, no guide as of right now)
- Applio by Shirou (UI, no guide as of right now)
- Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly):
- Be sure to know about the tensorboard
Google Colab = Easier but risk of getting disconnected
Kaggle = Harder but way more gpu time
If you are looking for the easiest way and for free, is using https://weights.gg which ofc uses RVC
RVC Inference (use models) on pre-recorded audio on Cloud
You can use either:
- Weights.gg: Easiest Possible Ever Automatic
- Ilaria RVC Zero: Fastest free on cloud
- Applio (ui)
This can help
Well rvc disconnected isn't working, it's a copy pasta but you get it
serene crown
I think im too used with rvcdisconnected lol its kind of a shock for me to use a different code
serene crown
# Train (make) RVC Models on cloud:
1. [Prepare the Dataset](<https://docs.ai-hu...
Thanks
low shard
I think im too used with rvcdisconnected lol its kind of a shock for me to use a...
well, either wait to use a slower colab based on an abandoned fork or use smt else
cosmic epoch
can somone please explain to me what the hell this error means: ['extract_feature_print.py', 'device', '1', '0', '0', '/content/Mangio-RVC-Fork/logs/The Biskit Twins', 'v2']
/content/Mangio-RVC-Fork/logs/The Biskit Twins
load model(s) from hubert_base.pt
Traceback (most recent call last):
File "/content/Mangio-RVC-Fork/extract_feature_print.py", line 73, in <module>
models, saved_cfg, task = checkpoint_utils.load_model_ensemble_and_task(
File "/usr/local/lib/python3.10/dist-packages/fairseq/checkpoint_utils.py", line 425, in load_model_ensemble_and_task
state = load_checkpoint_to_cpu(filename, arg_overrides)
File "/usr/local/lib/python3.10/dist-packages/fairseq/checkpoint_utils.py", line 315, in load_checkpoint_to_cpu
state = torch.load(f, map_location=torch.device("cpu"))
File "/usr/local/lib/python3.10/dist-packages/torch/serialization.py", line 1470, in load
raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint.
(1) In PyTorch 2.6, we changed the default value of the weights_only argument in torch.load from False to True. Re-running torch.load with weights_only set to False will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
(2) Alternatively, to load with weights_only=True please check the recommended steps in the following error message.
WeightsUnpickler error: Unsupported global: GLOBAL fairseq.data.dictionary.Dictionary was not an allowed global by default. Please use torch.serialization.add_safe_globals([Dictionary]) or the torch.serialization.safe_globals([Dictionary]) context manager to allowlist this global if you trust this class/function.
simple ore
can somone please explain to me what the hell this error means: ['extract_featur...
you need to downgrade torch in that colab to 2.3.1
or use a colab that has it fixed
cosmic epoch
you need to downgrade torch in that colab to 2.3.1
how?
simple ore
in install requirement cell
there should be torch installation, probably without a pinned version
cosmic epoch
in install requirement cell
the "install requirement cell", is it the first or the second cell?
cosmic epoch
in install requirement cell
i'm working with this colab colab by the way https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS#scrollTo=Q7JhsK5oYByc
simple ore
there needs to be !pip install torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 --upgrade --index-url https://download.pytorch.org/whl/cu121
as I see that colab has torch 2.5.1+cu121
so it is not surprising that old code fails to load weights
fresh kayak
@warm shell Hello, I was a user of the AI Maker Colab you created, and I’ve been using it with great satisfaction.
Today, however, I’m unable to access it. Do you have plans to fix the issue?
The results from your Colab are far better than other AI Colabs, and I really prefer using yours. Please, I hope you can make it accessible again.
Please understand that I used GPT-4 for the translation, so there might be some awkward phrasing.
Thank you!
knotty moth
<@1032433063713575003> Hello, I was a user of the AI Maker Colab you created, an...
what kind of error is it? perhaps torch related like mentioned just above?
small depot
guys can u use this vc changer while playing games cuz when i play games ik it uses my gpu which vc also does but i wanna know if i can use it without massive delay
cosmic epoch
there needs to be ``!pip install torch==2.3.1 torchvision==0.18.1 torchaudio==2....
and where do i put the "!pip install torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 --upgrade --index-url https://download.pytorch.org/whl/cu121"
brittle wing
Is the overtraining detector truly effective?
simple ore
and where do i put the "!pip install torch==2.3.1 torchvision==0.18.1 torchaudio...
in the cell that installs requirements
simple ore
Is the overtraining detector truly effective?
no
cosmic epoch
in the cell that installs requirements
you mean in the code of the first cell?
low shard
i'm working with this colab colab by the way https://colab.research.google.com/d...
rvcdisconnected has issues, check #📰│dev-updates , also I wouldn't suggest it since alex's mangio rvc tweaks got delted, it uses mangio rvc fork which is old and slower
brittle wing
@simple ore
low shard
@grave tundra #🧬│ai-chat message
grave tundra
im just looking for rvc lastest version
but i only find 2023 version
is it lastest?
low shard
alright, so your PC is good enough to set it up locally (run on ur pc gpu), meaning that it will take some more time to setup but won't have time limits like cloud (remote good pc used for people with a bad pc)
low shard
im just looking for rvc lastest version
the latest can be found in our docs https://docs.aihub.gg
you can use applio or mainline
low shard
but i only find 2023 version
i feel like you watched a youtube video where it told you to use mangio rvc, it's old
grave tundra
yes
low shard
mainline is the original rvc which gets updated, applio is a fork (modified version) which has some more tweaks
low shard
yes
don't follow yt tuts for this, they are old
please read our docs for setting up either mainline or applio (your choice, honestly i suggest applio since it has a better user interface)
grave tundra
okay thank yuu
low shard
okay thank yuu
yw, let me know how it goes
low shard
i will write here if i had sum problems
alr, yw
cosmic epoch
Okay, can someone show me what to do to solve this?
simple ore
use a colab that works #📰│dev-updates
low shard
Okay, can someone show me what to do to solve this?
rvc disconnected has issues, use a colab that works #📰│dev-updates
cosmic epoch
rvc disconnected has issues, use a colab that works <#1159380240271953940>
can you please give me a link to a colab that works?
low shard
can you please give me a link to a colab that works?
they are all linked in #📰│dev-updates
I'd honestly suggest either Applio or Mainline
high halo
hey how to enter rvc page?
low shard
hey how to enter rvc page?
please tell me:
- ur pc gpu
- what u want to do
cosmic epoch
is there any colab that WORKS where i can train models?
specifically one where i can train
simple ore
applio noui
cosmic epoch
applio noui
can you give me a link to applio noui?
simple ore
wise crater
sadly im having this error on the clone repository
Cloning into 'Mangio-RVC-Tweaks'...
fatal: could not read Username for 'https://github.com': No such device or address
FileNotFoundError Traceback (most recent call last)
<ipython-input-2-66c1853df4d6> in <cell line: 0>()
15 get_ipython().system('git clone -b pr-optimization --single-branch https://github.com/alexlnkp/Mangio-RVC-Tweaks.git')
16 #Rename to keep backwards compatibility with old variants of Disconnected
---> 17 os.rename("/content/Mangio-RVC-Tweaks", "/content/Mangio-RVC-Fork")
18 get_ipython().system('git clone https://github.com/maxrmorrison/torchcrepe.git')
19 get_ipython().system('mv torchcrepe/torchcrepe Mangio-RVC-Fork/')
FileNotFoundError: [Errno 2] No such file or directory: '/content/Mangio-RVC-Tweaks' -> '/content/Mangio-RVC-Fork'
odd shale
sadly im having this error on the clone repository
Cloning into 'Mangio-RVC-Tw...
Mate, Mangio Fork is pretty outdated.
What's your GPU?
You can install Applio or Mainline instead.
wise crater
Mate, Mangio Fork is pretty outdated.
oh i'm using the google collab version of rvc disconnected
odd shale
oh i'm using the google collab version of rvc disconnected
Oh, i see.
but what's your gpu?
simple ore
sadly im having this error on the clone repository
Cloning into 'Mangio-RVC-Tw...
that repo got nuked/achived
wise crater
but what's your gpu?
its T4
low shard
Mate, Mangio Fork is pretty outdated.
Rvc disconnected used Alex mangios rvc tweaks before, but Alex deleted it, so now it uses the even older and slower mangio fork now
low shard
sadly im having this error on the clone repository
Cloning into 'Mangio-RVC-Tw...
It seems like you're using an old version of rvc disconnected that used that repo
You prob saved a version on your drive
The latest rvc disconnected uses the original mangio fork, but it's broken that too
So you have to adapt to another way to use rvc
Mind you tell me ur PC GPU
wise crater
It seems like you're using an old version of rvc disconnected that used that rep...
I'm using V0.29
wise crater
Mind you tell me ur PC GPU
sadly i only have RTX 3060 mobile
low shard
I'm using V0.29
the latest one is https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS, I feel like you're using an older one since in the code it says the original mangio fork
anyways, the latest one is broken too rn
low shard
sadly i only have RTX 3060 mobile
it can actually work locally on that, the only issue might be the low vrfam (6gb) which may limit the batch size you can put
wise crater
the latest one is https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b...
thank you so much, I'm gonna bookmark this
wise crater
anyways, the latest one is broken too rn
I see. welp I guess i have to wait until it gets fixed.
low shard
Train (make) RVC Models on cloud:
- Prepare the Dataset
- Setup RVC:
Choose a cloud way to use RVC,
- Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
- Applio (ui)
- Mainline (UI)
- RVCDISCONNECTED (no ui)
- Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
- Mainline (UI)
- Applio by Vidal (UI, no guide as of right now)
- Applio by Shirou (UI, no guide as of right now)
- Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly):
- Be sure to know about the tensorboard
Google Colab = Easier but risk of getting disconnected
Kaggle = Harder but way more gpu time
If you are looking for the easiest way and for free, is using https://weights.gg which ofc uses RVC
RVC Inference (use models) on pre-recorded audio on Cloud
You can use either:
- Weights.gg: Easiest Possible Ever Automatic
- Ilaria RVC Zero: Fastest free on cloud
- Applio (ui)
this info can help
low shard
I see. welp I guess i have to wait until it gets fixed.
Well, I'm still waiting for the creator response, it's the 3rd time it breaks this week, I would suggest you to use smt else, bc even when it will be fixed, it will use the mangio rvc fork which is slower and older than what it was using before
raw stirrup
is there any current mainline colab working dont fancy trraining locally atm
low shard
is there any current mainline colab working dont fancy trraining locally atm
mainline colab works
btw if u got a good pc gpu u should do it locally tbh
raw stirrup
mainline colab works
oh i tried to run the local gui thing last step but it wouldnt load#
raw stirrup
btw if u got a good pc gpu u should do it locally tbh
i usually do local but my pc having issues atm
low shard
oh i tried to run the local gui thing last step but it wouldnt load#
could you link me the colab u are using to check it
low shard
i usually do local but my pc having issues atm
May I ask which issues
raw stirrup
could you link me the colab u are using to check it
raw stirrup
May I ask which issues
keeps rebooting if i do heavy loaded things either my gpu dying or lack of paste
low shard
https://colab.research.google.com/github/hinabl/RVC-Online/blob/main/Mainline_Co...
I'm gonna check it out one sec
raw stirrup
I'm gonna check it out one sec
np
low shard
keeps rebooting if i do heavy loaded things either my gpu dying or lack of paste
and it worked all fine before?
I'm feeling like your GPU is dying, u should also check the temp while it runs
raw stirrup
and it worked all fine before?
hen i try to open ngrok it says this SecretNotFoundError: Secret Ngrok_Token does not exist.
low shard
hen i try to open ngrok it says this SecretNotFoundError: Secret Ngrok_Token doe...
did u do this?
low shard
temp is fine at 33 idle and 50 6- when running rvc local
weird
raw stirrup
weird
yea but this error ERR_NGROK_8012
Traffic was successfully tunneled to the ngrok agent, but the agent failed to establish a connection
low shard
yea but this error ERR_NGROK_8012
Traffic was successfully tunneled to the ngrok...
Are u sure u set the right Ngrok token
I tested myself and it works
raw stirrup
thats my token but doesnt work
i get all the options but when i click themn it fails to load
simple ore
thats my token but doesnt work
dont share you tokens my dude
they are secret for a reason
brittle wing
wanted to use hina rvc to make a mr. krabs meme and it keeps breaking, i wanna show the error code but i got no img perms lol
warm shell
<@1032433063713575003> Hello, I was a user of the AI Maker Colab you created, an...
I understand, I'm sorry, but I've decided that I'm just going to abandon the project... If you want the same or better results than my colab, just try out the audio separation models on mvsep, they're all available there, there are even some better ones. Again, sorry....
rocky maple
wanted to use hina rvc to make a mr. krabs meme and it keeps breaking, i wanna s...
same for me
i'm pretty sure hina's rvc doesn't work now either
low shard
thats my token but doesnt work
Don't share ur token here
low shard
wanted to use hina rvc to make a mr. krabs meme and it keeps breaking, i wanna s...
!give-media-perms 20m @brittle wing
low shard
Could u show a SS of what happens when u click them
lavish lintelBOT
brittle wing
try again i still got it
low shard
try again i still got it
!give-media-perms 30m @brittle wing
brittle wing
there
i tried it on an alt thinking maybe i just installed it wrong bc im still pretty new to this stuff, but its the same error consistently
low shard
Send which google colab link are u using
brittle wing
low shard
https://colab.research.google.com/github/ardha27/AI-Song-Cover-RVC/blob/main/Hin...
Someone re-uploaded an older version of her colab, it's not the official link..
@glad zealot ofc YouTubers reuploading your colab bruh
-colab
low shard
Here u can find the official link
@brittle wing @rocky maple both check this and be sure to be using the official link
And yes, all YouTube tuts are old
You shouldn't follow yt tuts for this
rocky maple
Here u can find the official link
it's the same link i use still
give me ~4 minutes i'll show the error
rocky maple
https://colab.research.google.com/github/ardha27/AI-Song-Cover-RVC/blob/main/Hin...
wait no it's just him using the yt link what
Timer: 00:05:03 2025-01-31 00:21:16.292671647 [E:onnxruntime:Default, provider_bridge_ort.cc:1836 TryGetProviderInfo_CUDA] /onnxruntime_src/onnxruntime/core/session/provider_bridge_ort.cc:1511 onnxruntime::Provider& onnxruntime::ProviderLibrary::Get() [ONNXRuntimeError] : 1 : FAIL : Failed to load library libonnxruntime_providers_cuda.so with error: libcudnn.so.8: cannot open shared object file: No such file or directory
2025-01-31 00:21:16.292707075 [W:onnxruntime:Default, onnxruntime_pybind_state.cc:921 CreateExecutionProviderInstance] Failed to create CUDAExecutionProvider. Please reference https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirementsto ensure all dependencies are met.
dependency error
-# i think
brittle wing
wait no it's just him using the yt link what
i just unfavorited the colab and went to the official one
all is good thats on me i didnt know it was a repo
rocky maple
ok maybe it's just a me issue then
brittle wing
yeah it works now
rocky maple
not sure why it's not installing the right dependancies if that's the case
glad zealot
<@444684887363026974> ofc YouTubers reuploading your colab bruh
ah its those russians again
knotty moth
Someone re-uploaded an older version of her colab, it's not the official link..
what is that link?
brittle wing
what is that link?
see this is the most confusing part
bc when i originally used it, it was like "if its broke contact hina" as if HINA made it and this was the og
it even had this weird anime ui and everything
meager glen
Hello, I just finished setting up my rvc on Okada, but when I test the output, I always hear my original voice. Is this a setting on Okada or on my cable output?
low shard
Hello, I just finished setting up my rvc on Okada, but when I test the output, I...
Elaborate more in #🔍│help-w-okada
low shard
not sure why it's not installing the right dependancies if that's the case
Do u still have the issue
low shard
ah its those russians again
Rip
brittle wing
Is batch size 15 okay for a 30 minutes dataset
raw stirrup
rvcdisconnected works its just the epochs go up very fast leading to a broken model anyoneunderstand why that is/
it was working fine like 4 days ago
low shard
rvcdisconnected works its just the epochs go up very fast leading to a broken m...
rvc disconnected is broken again, u should use smt else
robust cove
ik this might sound goofy but is there any fix for robotic sound when using games ik its putting pressure on gpu so what can i do to elviate that to avoid robotic sounds i have alls settings on low and its 40series gpu so it shouldnt really struggle
alpine lily
Can some one give me the download link? I have a GTX 970
And I need to convert audiofiles, not realtime
fervent rover
Umm, Help?
Is that some kind of glitch?
Because Look, even I though I already putting the in the google drive, the index training said it is not exist and it told me to run the feature extraction even though I already did it, like I said early it is a error or a glitch?
keen coyote
been using the guide cant find a chunk and extra that works for my rtx 4060 anyone got advice?
also should add here it cuts out and idk why
low shard
been using the guide cant find a chunk and extra that works for my rtx 4060 anyo...
use #🔍│help-w-okada and its in the guide
fervent rover
rvc disconnected is broken
Oh Okay, yeah go ahead and fix it
low shard
Oh Okay, yeah go ahead and fix it
I’m not the creator of that notebook, i already contacted him and he hasn’t replied yet
its the 3rd time is broken this week
plus he changed from alex’s mangio rvc tweaks to mangio fork, bc alex deleted his fork, so now rvc disconnected is based on even older and slower code
I suggest you to use smt else
fervent rover
What does smt stand for?
low shard
it always cuts out
its better u elaborate abt this in #🔍│help-w-okada
low shard
What does smt stand for?
something
fervent rover
Oh Okay, I better to just wait for it
low shard
Oh Okay, I better to just wait for it
It’s better you use something else since it’s going to be slower even if it will get fixed
I can suggest you alternatives
rvc disconnected creator is kinda busy in real life
fervent rover
I can suggest you alternatives
Oh Okay also what alternatives do you recommend?
low shard
Oh Okay also what alternatives do you recommend?
may you tell me your pc gpu?
fervent rover
Here
fervent rover
That’s your drivers not gpu
Okay, I don’t know alright, I don’t know of my GPU so yeah, that’s it all I had to tell you about it
Just tell me the other alternatives do you recommended for me?
low shard
Okay, I don’t know alright, I don’t know of my GPU so yeah, that’s it all I had ...
You can check your pc gpu via:
ctrl+shift+esc (task manager) -> Performance tab -> GPU
you can also send a screenshot
low shard
Just tell me the other alternatives do you recommended for me?
yeah im asking your pc gpu for that, in case its good enough for local
fervent rover
Here you go
low shard
Here you go
Yeah that gpu can do inference (use models) even if slow, but would be an issue for training (making) models
it’s better you use cloud (remote good pc)
fervent rover
Yeah
low shard
Train (make) RVC Models on cloud:
- Prepare the Dataset
- Setup RVC:
Choose a cloud way to use RVC,
- Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
- Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
- Mainline (UI)
- Applio by Vidal (UI, no guide as of right now)
- Applio by Shirou (UI, no guide as of right now)
- Lightning.ai (Kinda hard, needs login, no issue with web uis or anything, but only free 15 credits monthly):
- Be sure to know about the tensorboard
Google Colab = Easier but risk of getting disconnected
Kaggle = Harder but way more gpu time
If you are looking for the easiest way and for free, is using https://weights.gg which ofc uses RVC
RVC Inference (use models) on pre-recorded audio on Cloud
You can use either:
- Weights.gg: Easiest Possible Ever Automatic
- Ilaria RVC Zero: Fastest free on cloud
- Applio (ui)
fervent rover
# Train (make) RVC Models on cloud:
1. [Prepare the Dataset](<https://docs.ai-hu...
Yeah, I think I’m just going to wait for a few days until this google colab is finished so yeah
low shard
Yeah, I think I’m just going to wait for a few days until this google colab is f...
Alr, just know that even if it gets fixed it will be slower tho
fervent rover
Alr, just know that even if it gets fixed it will be slower tho
Okay
fervent rover
Alr, just know that even if it gets fixed it will be slower tho
What do you mean slower, you mean like training the voice model?
low shard
What do you mean slower, you mean like training the voice model?
yeah, it runs on an even older and more abandoned rvc fork than before
it used to run on alex’s mangio rvc tweaks, which was already old but was a better version of mangio fork, now alex deleted it since its too old, so rvc disconnected is based on mangio fork
it hasn’t been updated since 2023
this is why we suggesting users to use applio or mainline, they are more updated
fervent rover
yeah, it runs on an even older and more abandoned rvc fork than before
Oh Okay, Just Asking
unkempt vale
This notbook doesn't work anymore: https://colab.research.google.com/github/ardha27/AI-Song-Cover-RVC/blob/main/Hina_Mod_AICoverGen_colab.ipynb#scrollTo=NEglTq6Ya9d0
Do you know how it's possible to repair it?
low shard
This notbook doesn't work anymore: https://colab.research.google.com/github/ardh...
that is not hina mod aicover gen, a youtuber who reuploaded a version that now is old
check #✨│ai-help message
unkempt vale
It doesn't matter!
low shard
and use the official one
low shard
It doesn't matter!
it does
unkempt vale
Who stole it
low shard
its old
unkempt vale
It works perfectly
low shard
It works perfectly
youre saying it yourself it doesnt work
unkempt vale
I asked about another question
low shard
its old
its not made by that youtuber
its made by hina
check the convo i sent u
use the official link by hina
not by youtube. Youtube tuts are old
unkempt vale
It worked perfectly and sound was perfect with echo like it was produced by expensive microphone
2 weeks ago it broken
low shard
It worked perfectly and sound was perfect with echo like it was produced by expe...
its old, it wont be updated since its not even his, its by someone who reuploaded hinas colab
its not ours
use hina official one
-colab
low shard
its literally the same thing but updated, since the youtuber reuploaded a version that now is old
unkempt vale
I have seen this references. It's not competitve as that one(
low shard
I have seen this references. It's not competitve as that one(
bro its the same code, just that it was a version that now is old
that colab u sent wont work anymore
its a reupload of an old version
use the official one by hina, its listed above
youtube tuts are old, dont use them
unkempt vale
which one is made by Hina, there are some version from this order named by Hina
low shard
which one is made by Hina, there are some version from this order named by Hina
use the Hinas mod ai cover gen webui
knotty moth
which one is made by Hina, there are some version from this order named by Hina
this link looks strange
low shard
the colab you sent, is from a yotuber who just reuploaded that same colab some months ago, since its a reupload, it doesnt get updated,, the official one gets updated by hina
low shard
this link looks strange
where did u find that?
it seems like a discord lookup with @glad zealot discord id
knotty moth
where did u find that?
the ardha27's colab https://colab.research.google.com/github/ardha27/AI-Song-Cover-RVC/blob/main/Hina_Mod_AICoverGen_colab.ipynb
low shard
the ardha27's colab https://colab.research.google.com/github/ardha27/AI-Song-Cov...
lol he just reuploaded the same notebook
idk why people even do that
knotty moth
it seems like a discord lookup with <@444684887363026974> discord id
im not sure if that site is even legit one
unkempt vale
use the Hinas mod ai cover gen webui
Bro give me please direct link for this Hina's Mod!
low shard
Bro give me please direct link for this Hina's Mod!
just coick this, its an hyperkink
blue text that redirects u to the site
unkempt vale
It doesn't work anymore. I've just tried it by using Gradio and CloudFlare. It has run, but didn't changed voice for 2 minutes mp3.
Now I'm trying agian
low shard
It doesn't work anymore. I've just tried it by using Gradio and CloudFlare. It h...
its different from the one you tried
that one wasnt uploaded from hina
Same name but uploaded from different people, its different code
just use the direct link https://colab.research.google.com/github/hinabl/AICoverGen-Colab/blob/main/Hina_Mod_AICoverGen_colab.ipynb
unkempt vale
! Stop! I told about this link: https://colab.research.google.com/github/hinabl/AICoverGen-Colab/blob/main/Hina_Mod_AICoverGen_colab.ipynb
fervent rover
it used to run on alex’s mangio rvc tweaks, which was already old but was a bett...
Hey Man, What Does Mangio Fork Do to make it slower?
low shard
im not sure if that site is even legit one
it is i think
fervent rover
RVC V2 Disconnected
low shard
Hey Man, What Does Mangio Fork Do to make it slower?
its older code that isn’t maintained, it doesnt have optimizations that applio and mainline have
serene crown
anyone know the hardware requirements to train a voice model locally? i wanna see if i can do tht
fervent rover
its older code that isn’t maintained, it doesnt have optimizations that applio a...
Can’t you just update it than?
low shard
Can’t you just update it than?
im not the creator of it
You can just use applio or mainline
they are more uodated
@idle adder isn’t going to maintain the code anymore, iirc the github repo is even archived
There are other people who made other forks like applio which are more updated
fervent rover
its older code that isn’t maintained, it doesnt have optimizations that applio a...
I Mean can Kit Lemonfoot Just update the Mangio fork in RVC v2 Disconnected?
I mean like it hasn’t be updated since 2023 but now when technology is getting more advanced and updated of Ai, Can he just update it?
low shard
I Mean can Kit Lemonfoot Just update the Mangio fork in RVC v2 Disconnected?
He could fork it, but he’s busy too irl, this is for example a comment in the code of rvc disconnected
steel forge
please just use applio
unkempt vale
Error for official Hina's Mod: 2025-01-31 12:42:39.508530557 [W:onnxruntime:Default, onnxruntime_pybind_state.cc:921 CreateExecutionProviderInstance] Failed to create CUDAExecutionProvider. Please reference https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirementsto ensure all dependencies are met.
Timer: 00:06:31
steel forge
mangio is dead
low shard
just use applio fr
low shard
Error for official Hina's Mod: 2025-01-31 12:42:39.508530557 [W:onnxruntime:Defa...
are u using T4 gpu?
unkempt vale
yep
unkempt vale
when I start process to change voice for Mp3
This hang on the process for an endless time
low shard
when I start process to change voice for Mp3
I will check it myself right quick
glad zealot
This hang on the process for an endless time
it just takes way longer than normal cause for some reason its not using the gpu (again)
unkempt vale
please/ I'll wait
glad zealot
does it only happen with mp3 files or also happen with youtube links?
glad zealot
fixed
unkempt vale
That's it
glad zealot
Should I try again? Did it fix?
yup try it again
disconenct and delete runtime then run it again
it should be fixed
fervent rover
He could fork it, but he’s busy too irl, this is for example a comment in the co...
Oh Okay, well maybe if he is available and just to see if how he can to fix it, than maybe it will be faster than Alex’s but who knows, maybe he might be at work or maybe if he is sleeping or just going on vacation but yeah just asking, so yeah anyways goodnight and bye or and also make sure to update me if Kit had fixed the issue or did Alex decide to change his minds if he decide to bring his fork back into it but anyways I will be using the Applio Colab Tomorrow so I will kept you posting for updates of how was it going, anyway goodbye and goodnight
knotty moth
I mean like it hasn’t be updated since 2023 but now when technology is getting m...
mangio and its forks are discontinued, so it'd be the job for colab owners to maintain