#✨│ai-help
1 messages · Page 74 of 1
its clearly understandable
install nvidia one lol
let me find guide
Written by Julia Images by Poopmaster Last update: Jan 29, 2024
please someone help me i’ve tried mangio and ilaria and neither are working for me despite being successful in downloading the voice model to each system and putting the pth file into weights for mangio, i just get an error on mangio with no explanation and on ilaria the model doesn’t come up in the drop down no matter how many times i refresh
and i tried a different computer in case the operating system had anything to do with it
and i just … feel defeated :///
if I'm using rin_e3 with an hour long dataset what epoch and batch size should I put?
@half cove you know?
with rin_e3 i dont really know
you dont need a high step to epoch so for batch size just use as much as your gpu can handle
cuz long dataset
havent tested long datasets with it
well it's pretrain is made for long datasets
hence why not a lot of ppl use it I think
ive heard it messes up with longer datasets
Its says pth file but mines are mp3
but other people say its good
pth is your voice model
wav is the dataset
no 😭
okie
"How to do Inference"
go down
your trying to train a model right?
go to the train model bit
pls help me too

use rvmpe also what isnt working?
F0 det only rmvpe_onnx .... i want to change to other
Ayo? @dark sparrow level 1 !!! 
voicechanger right?
is not succeed
are you using the directml one
direct ml can only do those
how did you put the file?
this didnt worked
i use cuda. my gpu rtx 3050
i mean the aduios
dont use directml one
use what??
didnt worked
let me get link
ok broo.. i will waiting u
use rmvpe gpu and i dont really know whats wrong tbh
ask in #1192011222023950368
im too tired lol
does anyone know how to use ai voices, to make custom tts voices
i tried Ilaria-RVC
it did not work
did not work in what way?
using google collab to run Ilaria-RVC
thank u brooooooooo
the tts, i could only use their voices, not my custom one
i hope your day always good
you use the tts then it saves in audio then you use your voice model and click convert
You use it to make a TTS sample that you can then convert with RVC.
There is no direct TTS for RVC yet.
ok
Ayo? @brittle wing level 3 !!! 
is there any other way
There is only this way
oh it does work, thank you so much, i hope you have a great day
U can gather TTS from other places though.
You're welcome :)
How can i make an AI sing or say quotes using an already existing model?
I used to know but i havent touched AI in a year
You need an audio sample then just convert using a fork
Written by Julia Eddy Last update: Jan 30, 2024
Should help
aight thx @proper shale
yw :)
hey does anyone have a link to an ai voice changer article
i mean... theres this https://rentry.co/VoiceChangerGuide
This guide is back to receiving regular updates, however do not ping me for anything on it. The guide itself speaks for me. - Blanc
Other Links
Antasma's Local Error Fixes
Antasma's Colab guide
Sushi's useful Links - You need to follow this if you are on an AMD or INTEL ARC graphics card
Frequent...
thanks
what is the difference between rvc 1 and rvc 2?
@half cove i m futher
what
did you click...
yea
train index
if you clicked it then it should be in logs
i did train model and this
it should have said to train index in a prior step
what did it say in cmd when you trained index
thats with the model and a beat i think its okay but can be better
its goodd
thank you
this is the output when i tap train featureindex @half cove
it look like your done
Ayo? @worldly finch level 5 !!! 
what does it say in logs lol
(22361, 768),573
training
adding
成功构建索引,added_IVF573_Flat_nprobe_1_beyza_v2.index
go to your folder and look for logs there should be a .pth file thats the trained model
in your logs folder it should have the added index
dont use trained index
there is no index
send screenshot
ait wait
where was u looking lol
the batch size right that option turn yes
no?
i found..
viper on some next level drugs
hahahahhaha
Ayo? @cobalt trout level 2 !!! 
wym i m losed
the right option after batch size you have turned off
i using on
its better
you dont need that one tbh
its for retraining at any save checkpoint
id click this tho
so every checkpoint it makes a model
so you can test what one is best
i need now help from here
i know i m close to end
i have the index file what do i need to rn
get the pth file from weights
has he trained yet
i mean to use it you can just leave it there
ya umm i dont have weight file
then train model now
oh
where can i download the app ?
what app
the ai voice changer
⠀
Download for Nvidia GPUs 
Version 17b cuda
Download for AMD GPUs 
Version 17b directml
Download for Intel GPUs 
Version 17b directml
Download for Mac 
Version 17b Mac
⠀
weight appeared what now?
^^
srry
thanks 😉
should be (modelname).pth
if training has finished u can test
leave it there and it will detect
just go inference page
refresh models
should appear in dropdown
yea
what does cmd say
use a lower batch size
@worldly finch what GPU do you have
shall do again train index?
no
and train model?
no need
only this, but first
_ _
which one specifically, just wanna know cuz of vram
nvidia geforce gtx 1650
ig you'll be stuck with 2 batch size
result
what do i need to do
still not my name in the weight pth file
??
@half cove
is anything using ur gpu?
it should be able to use batch size 2
also why r u doing 20k epochs 😭
epochs?
id recommend using colab for u tbh
the guide had 20
amount of times it goes over dataset
oh wait u used 20
i mixed it up lol
dw bout that
ur gpu didnt have enough memory
i read it
or no wait its loading
Ayo? @worldly finch level 6 !!! 
i dont wanna change from options
hm?
its stopping due to being out of vram
how to check ahahah srry
like if you have any games open or anything
id just recommend colab
will my results the same
what do you mean by that?
- 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
- Ilaria RVC, by thestingerx Hugginface Spaces
- RVC-HFv2, by r3gm Huggingface Spaces
- AICoverGen, by r3gm Huggingface Spaces
- Advanced RVC Inference, by r3gm Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- GPT-SoVITS (demo), by litagin Huggingface Spaces
output will be the same
alr
ye i know couldnt find
HELP i tried training kasane teto and got this
is applio rvc broken
even when i set pitch to -12 the result is identical
RVC Guides (How to Make AI Cover)
Translation by country
Why does MVSEP have a queue?
Vocal separation requires computational resources
also on mac it finally showed up on privacy to open anyways but it doesn't open.
It just bounces at the bottom for a bit than goes away
You could just try the a colab
I've never tried uvr on mac so I can't really help with troubleshooting that
is there any way to get the overtrain detector back on applio rvc
why did they remove it on the new update
it was literally so useful
I think they are working on adding it again or somthing similar
I have no idea how to do dis
How many epoch can I do for a 10 minutes long dataset?
is there any kind of video guide i cant find anywhere
Is it ok if I go for 350? I just want to make sure I'm not overtraining
how come in thre odaka voice changer my gpu only shows cpu and nothing else?
Written by Julia Eddy Last update: Jan 30, 2024
Check the epochs page
Hey the app just stopped picking up me talking all together any fixes?
I'm trying to use rvc locally and I was trying to get tensorboard but I can't seem to find the rod file anywhere on github. Do you not need it anymore Im very confused?
help please
Written by Julia Eddy Last update: Jan 30, 2024
Try reading the guide from there
I did but their link for the rod file shows a 404 error
Use the file hosted in this site
yw!
Anyone know a fix where the app isn't picking up anything from a mic?
I just finished training my model but I don't see the index training, will it work with only the pth file?
Nevermind 😅
I just saw it lol
guys my gpu wont show up when i download the odaka thing i have a amd ryzen but it just shows cpu
which method is the easiet to create AI covers of a hole folder of audio files?
Ayo? @brittle wing level 1 !!! 
@tough ice I have had the worst experience with cartoon copyright issues
@tough ice I would say for fair use you could remix the song and throw in a voice, then there would be no copyright
I use to have an app on my old mac that I downloaded here that let me batch convert files to AI covers, does anyone know the name of it?
Do you think this can be able to be approved for being posted?
If u hear some cracks at the end is because I didn't remove the hidden vocals lol
Doesn't sound very good
I think the model was poorly trained
You used TensorBoard for training, right?
I did 350 epochs ಥ‿ಥ
Yes, and it wasn't overtraining
Oh don't tell me you used an arbitrary number of epochs 😭
Did you?
How does that work? I'm sorry, I'm a newbie on all of this
Don't worry, we all start somewhere
Written by Julia Eddy Last update: Jan 30, 2024
Check the Tensorboard page
im not gonna lie, im lost trying to find what i want in the guides and help sections. is there still a collaboration thing on google that i can just talk into or input some voice clips and change them into any model for fun?
https://aihubdocs.github.io/en/ using this site is much more convenient
Written by Julia Eddy Last update: Jan 30, 2024
Sorry about the mess regarding guides, we are revamping all of that

You'll find everything you need here
Tysm! Btw I already saw that guide and made every single step so, I also cleaned my dataset so I don't know ಥ‿ಥ
thank you
yw!
You read it and still don't know how to use it?
I do, but I don't know why the model came out with a low quality, I think I messed up with the epochs
Ayo? @young halo level 6 !!! 
And i was making sure it wasn't overtraining
10 minutes
For how long did you let it run training?
You mean the amount of Epochs?
Just time
About two hours, something like that
Yes! I've always used it for my trainings
Were you using the original pretrain?
What do you mean by that? Sorry
Seems like we might have found the problem here
https://aihubdocs.github.io/en/ read the rvc disconnected guide
Written by Julia Eddy Last update: Jan 30, 2024
Use the original pretrain this time. You'll read about it later in the page
Basically they introduced a new "feature" to RVC-D, a new pretrain
Which helps you train models for a shorter amount of time
It's still in development so it's more buggy
That's why I recommend to newbies using the stable pretrain instead
Oooo, tbh, I didn't notice that feature and I probably left it by default
I guess I'll try to train it again
And make sure to be using that feature
Yup. Good luck!
Thank you so much! I hope it works better this time!
Your guides are really helpful btw
Hehe glad to know! It's in early development so pls lmk any suggestions you may have 🙏
Sure!
Hope the model goes well now
There's isn't really a recommended amount
Neither a way to know beforehand, as it depends on the length, quality & size of the dataset
Read the epochs page and you'll understand more
Sure! And thank you again
Forgot to say, I recommend inputting 2000 epochs but not because that's the correct amount, but for making sure you train for long enough to detect overtraining
It's an arbitrarily large number
That way you won't have to re-train
In case you run out of epochs
I got it! And then I'll just stop the training before it gets overtrained, right?
You can stop the training at anytime, your model will be fine if you reach the point past overtraining
You stop the training when it's very evident it's overtraining, not before it overtrains (since you can't predict the future oangpwmtlq)
You stop it and pick the checkpoint that was closest to before the OT point
Any questions?
If I stop the training, will the pth and index files save on my Google drive?
By default they get saved on the Colab space, but you can move them to your GD storage anytime
At the end of the RVC-D I explained where they are located at
But once you end the runtime of the Colab space, all the files are gone
Alright! Thank u so much for answering all of my questions 😁
np! Ping me if you have more
can anyone help the tensorvenv bat file works but when I put in the url it shows a blank screen
Ayo? @formal bear level 1 !!! 
?
hey, so… im having a really hard time trying to download the music after converting it… maybe because im using an ios device… but someone know how to download it?
@molten pecan can you help me?
Ayo? @dense pagoda level 1 !!! 
what is chunk?
Ayo? @rugged kindle level 1 !!! 
Weird. Try it in another browser
Converting it where exactly?
hey guys what is RVC
Written by Julia Eddy Last update: Jan 30, 2024
Here's our documentation website
;3
thanks!
yw1
tried it on chrome still a blank screen
Try enabling third-party cookies
Any easy way to convert a batch of files to AI cover?
any website that can allow you to make a free ai model to say whatever you want it to
the one i was using became ultra greedy and now you can't do any of that without paying money
- 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
- Ilaria RVC, by thestingerx Hugginface Spaces
- RVC-HFv2, by r3gm Huggingface Spaces
- AICoverGen, by r3gm Huggingface Spaces
- Advanced RVC Inference, by r3gm Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- GPT-SoVITS (demo), by litagin Huggingface Spaces
RVC Disconnected, by Kit Lemonfoot Google Colab
easyGUI, by rejects Google Colab
for training
RVC Guides (How to Make AI Cover)
Translation by country
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
as they say in japan "arigto gozaimas"
For the real time voice changer everytime i use ngrok verison in google colab it never works so i wass told to run it localy but i dont see the run locan tab??
run local means you have to download the voice changer and run it with your hardware
you need a gpu to run it locally
so i would need a good gpu im guessing
yes
rip
Ayo? @oak echo level 1 !!! 
which colab are you using btw
is it possible to download through kit.ai or is rvc the only good one? Also can anyone show me any other ways to get good voices not trash oens
you mean shits.ai?
well
you cant download from kits, just use rvc
look in #1175430844685484042 for the tag "Real Time" since most of those models are the best sounding ones
i just followed this guys vid and it said to click google collab and it all went smoothly till i tried to upload a voice and it kept crashing
give the link of the colab
use
is there a tut to use it??
thx
ok thanks
same as the other one
the only different is the model upload, you can ignore it
its the same one but this one actually works
ok cheers appreciated
how can i make a voice to voice ai run locally? i've already tried mangio but i keep getting AttributeError: 'NoneType' object has no attribute 'tobytes' as an error
how did you add the model? with the upload button right?
how do i make no voice delay
is there an rvc voicechanger that is actually good and not delayed
Ayo? @silver notch level 1 !!! 
no
unless you have a nasa computer, there will always be a delay
by not delayed i mean i speak and i hear it 5 mins later
lowest you can do for delay is around half a sec
thats really good the one i was using rn had a really long one
are you running local?
idk what that is lmao
did you install it on your pc or are you using google for it?
yes:)
Can someone help? It writes "waiting generate pipeline" "Pipeline is not initialized"
Ayo? @winter bough level 1 !!! 
-colab
- AICoverGen-WebUI, modded by Hina Google Colab
- AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
- AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Ilaria RVC, by thestingerx Hugginface Spaces
- RVC-HFv2, by r3gm Huggingface Spaces
- AICoverGen, by r3gm Huggingface Spaces
- Advanced RVC Inference, by r3gm Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- GPT-SoVITS (demo), by litagin Huggingface Spaces
can you scroll up before it starts spamming the same stuff, and re-screenshot ?
Great !
Try deleting the folders named pretrain and model_dir under the voice-changer folder
and then relaunch. It should redownload the needed files. Let me know if it doesn't work, I have a different way to download if it re-fails
We don't use voice.ai much here (if at all). Not sure anyone will be able to help with it 
can anyone tell me how to connect the W Okada collab file to my google drive so it installs there? In the web version of the guide i've used it mentions connecting to G Drive but I can't find out how to do it - https://www.toolify.ai/ai-news/transform-your-voice-in-realtime-no-graphics-card-needed-76706
that's most likely outdated
the google colab code used is now flagged and won't run
I can get it to run it's just frustrating when I get disconnected as I have to install all my files again
try this one instead
I'm not sure if it has gdrive support, but at least it won't disconnect
ah okay thanks, is it as easy to use as the other one? on that one you could just click a voice and it would change and easily add other voices as you go. I'm not great at coding
Ayo? @unkempt roost level 1 !!! 
guys should i need to split my 30 minutes dataset in to 10 seconds of chunk ? Or single file is good ?
It's mostly the same, so yeah you shouldn't be lost
You can add voices via the graphical interface, it shouldn't be a big deal. The "OPTIONAL" boxes can be avoided for simplicity
It doesn't matter, keep it the way they are
Okay , thanks. I want to improve the quality of my dataset, vocals etc. Should i use Audacity or is there a better tool for the process ?
think i've sorted it thanks mate
I wouldn't really know, I don't touch model training
Ok
i just want to query what do the chunk and extra buttons do?
Can someone guide me
How do I make the training process faster??
I want to make voice ai with my own voice but i need to train it first
Any helpers maybe?
yesnt, its using normal gdrive so probably wont work anymore for drive links
yes for doing it online
I m now on guide page but i still cant go futher
do you have the recording of your voice?
how long is it?
15 sec
longer
rlly..
how long does it need to be
i tought a bit of voice is enough
@glad zealot
is there someone who can help me
guys, does anyone know... how to avoid getting copyrighted when uploading videos
I've changed the instrument, it's still there🗿
@helpers
RVC Guides (How to Make AI Cover)
Translation by country
alr tried with guide
It's always gonna be flagged if it's copyrighted. There is no way of getting away with that.
Hello, how to create an AI song? Can you help me😭
it says no enough space
but
i m using 0.1/4gb
anyone who can helppppp
@stoic cloud pika?
VB cable isnt working anymore anyone knows any other Virtual audio cable
@stoic cloud
Ayo? @worldly finch level 7 !!! 
No, it won't.
Nope
You'd need someone who is willing to comb through terribly written code and optimize the heck out of it
-colab
- AICoverGen-WebUI, modded by Hina Google Colab
- AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
- AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Ilaria RVC, by thestingerx Hugginface Spaces
- RVC-HFv2, by r3gm Huggingface Spaces
- AICoverGen, by r3gm Huggingface Spaces
- Advanced RVC Inference, by r3gm Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- GPT-SoVITS (demo), by litagin Huggingface Spaces
Neat
the link
its just a fancy way of saying "colab page"
literally just follow what the guide says and you should be ok
It's a clickable link
back in MY day
Yeah just click the blue text
Don't start
When I open the rcv application this appears
oh yes I will start
did u move the bat out of the folder or something
also, make sure folders have no spaces
[Voice Changer] Web Server(https) Launch Exception, [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from pretrain/rmvpe.onnx failed:Protobuf parsing failed.
this is my error code, any ideas?
I transferred the file to disk (D:)
delete the folders called pretrain and model_dir and relaunch
should redownload everything it needs on next launch
thanks!
Now I extracted the winrar file to the folder and downloaded it to the desktop but it's still the same.
On my mac I'm trying to open UVR but it just bonces at the bottom but won't open.
I went on privacy and pressed open anyways and it doesn't open.
its started to work, GPU however will not show up, any solutions?
Ayo? @earnest pumice level 1 !!! 
Mac is running 10.15.7 btw
What's your GPU ?
is there an easy way to find out? not the biggest PC guy
definelty an AMD gpu though
Oh if it's AMD, you might have gotten the wrong version
how do this?????
Does the filename match ?
is there a specific AMD version?
Ayo? @sand iris level 66 !!! 
congrats btw
nope, I guess I download this?
help
Yup, that would be the correct version then
Back to square one I guess, gonna go eat and then do that, thanks fired!
There's a full guide there if you need : https://rentry.co/VoiceChangerGuide
But feel free to ask if you need anything
This guide is back to receiving regular updates, however do not ping me for anything on it. The guide itself speaks for me. - Blanc
Other Links
Antasma's Local Error Fixes
Antasma's Colab guide
Sushi's useful Links - You need to follow this if you are on an AMD or INTEL ARC graphics card
Frequent...
What do I do about my problem?
Oh boy, RVC-Gui is kinda outdated by now
I'd help but I don't have much time rn, and I don't have a mac to test on
It's probably because UVR doesn't run on that version? Hm
Did your Mac stop being supported after Big Sur
Managing to run UVR through the terminal could help getting useful logs, but I don't know how everything goes on mac regarding executables
.dmg seems like a bastardized format
I have no idea. I'm normally never on this mac because it's my mothers.
Ayo? @snow hazel level 9 !!! 
UVR might not support it cause it is a 12 year old computer though.
mine??
oh kk np
It's kind of weird because it shows on the app on the screen.
someone who can help?
u mean this 2?
No, before the one that errored
i didnt run the cell befor
I just wanna share a model in mobile now i cant upload audio sample
can you tell me how to download this application without errors and details ?
Suggestion for @craggy brook
Full AI Voice Model Training Guide (Local)
Link: YouTube
Credits: Christopher Villanueva
Model training with Mainline RVC
Link: Rentry
credits: Raven (ravencutie21)
AICoverGen Colab Guide
Link: Google Docs
Credits: Eddy (Spanish Helper)
Create a model with RVC disconnected (colab)
Link: Google Docs
Credits: Angetyde
How To Make an AI Cover With Ilaria RVC
Link: Rentry
Credits: 👽 Julia (ailen2091)
...I do have a 2012 MacBook Pro too
Haven't tried out UVR tho
Most likely wouldn't run
I honestly don't like apple stuff because windows has more freedom
At least you're getting great optimization
Also, technically, you can do most things Windows can anyways
You can get really comfortable with the settings and environment they ( Apple ) give you.
But you will be serious hard to try to get out from their zone
yeah but... 4090
if you really wanna splurge, get 4090 and linux
dont use apple for ml
ive heard that getting like 48GB VRAM is crazy tho
u can do a lot of ml
Google uses linux.. and so does everyone who does ml seriously
at that point get like a tesla card
the Apple arm chips truly are something
or quadro
i think they're crazy
Yeah
What about it ?
because, they're good
but weird
https://www.youtube.com/watch?v=RHz5--LjnhE this is same ?
like m1/m2 ultra, theyre just 2 chips glued together 💀
The bandwidth available is just beyond insane for a CPU, and even for the GPU it's quire beefy. Sure it doesn't match GPUs that offer the same VRAM pools, but consider the prices
but it works
not only for graphical uses, but it also serve as RAM to other channel as well like CPU, storage
It just works™
Which... to be fair, make the RAM and storage not upgradable.
Can someone send me a DM and explain in detail how I can unlock this annoying thing?
Ayo? @craggy brook level 3 !!! 
that's the main issue i have
The storage being soldered is pure bullshit
but... ig that's just how it is
for the memory, why not, when you consider the architecture. It's unfortunate but it's a side effect
For me it is just something like EU tried to forced every manufacturer to produce removable battery.
Congratulations, you have acquired a brick
👍
please
in my 2012 i can replace the ram, ssd, even take out the optical drive for a HDD
what is it?
4
I guess you guy will really like Framework's laptop
I've seen this before. I don't think you've installed it correctly but I'm not 1000% sure
@proper shale ?
now I download
Did you click "add to path" when installing it?
no
( actually did RVC support Python 3.12 )
Yeah you should have
(idk it should? but who knows)
Nope. Only 3.10 and lower
(nevermind_
Unless you go out of your way to fix one of it's its dependencies
it will be an issue as which you need to manually identify Python's path to the start.bat
not only for this RVC thing
(im sso sorry I have to)
its*
oop
yeah that's what I am thinking
thx
you can try to call Python in cmd
if it is showing error, then it is not able to access to the public channel
it show like this when you enable it to get into PATH
certainly you need to downgrade your Python to 3.10 and below first
and an opportunity for you to enable the Python to PATH during installation
My voice changer works, now how do I connect it to an output?
(Like discord or something similar)
Set output as voice cable input or something, and then set discord's voice input to voice cable autput
Okay, thanks!
but at last line its stucked at 27%
cause youre using the same session
it finishes so fast that it doesnt have time to say it is completed
but it has
so there is no problem?
Nope
Yujin and Gaeul Ive please 🙏🙏🙏🙏
What do you mean?
Do 200-250
idk who picked the defaults
Alright thats fine
so or do u mean choose between
I dnot think you could do total epoch:200-250 lol
Ayo? @peak urchin level 5 !!! 
ok just do 225
@peak urchin VERIFIED CHAT LESSGOO
you can try 300 and saving the 200 for testing uses
hey i prefer green
my cmd says waiting to join pipeline
what do i do?
and it says pipeline is not initialised right after
[Voice Changer] Waiting generate pipeline...
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
me too
I fixed it by using a vpn (dns but i don't think is different for this). If you connected at least once with the extracted folder, then you have to extract to a fresh folder again, then turn on vpn and run start_http.bat, if it started downloading 52 MB files and 1 ~372MB file or something like that then you are failed probably. if it started downloading many ~372MB files then you are doing great. many many other files download after that. at least that was my case. this is photo of successful connect (not a bad connect.)
Thanks bro, i did it. It works.
founded
for now just train the model
and then use the cell that sends it to your google drive
use the cell
or the codeblock
alr did
so i got the weight file
but idk what i need to do rn
....
@stoic cloud
/......
@stoic cloud
what do i need to dooo
Sorry I'm busy and have a bit of a life but
Now use synthesis notebook
It is in your drive yeah
use your drive to download it
Ayo? @worldly finch level 8 !!! 
У кого есть голос Urbanichka?
did you save model to drive?
have you done this?
yea
it should save to drive
Hello guys. l have this message when l set more the 2 dre lighe in my projet: "Multiple dri lights are competing to be the singole one used for forward sharing." anyone kown how to fix that?
what did it say?
no
what does hop length mean?
if you just click it without selecting anything it will save the latest one
so answer is yes?
poopy
poopy
julia
poopmastaa
help
yesterday i trained a model on mainline rvc
and today i just realized the epoches arent enough
can i retrain it ?
yea
just put in the exact same settings and click train
do i need to replace d222.pth and g222.pth ?
it should use the latest pth to start the training
replace with what?
Load pre-trained base model G path:
its like what the model uses as a base kinda
what the model trains itself using
Beyza.pth is the one you use for inference
yea
do i need to make a zip file for pth and index?
and than in rvc gui import odel
so is it basically mean D233.pth and G233.pth of the model are no use for us ?
no those are training files
yeah
how about the one from model
idk for rvc gui but on rvc webui put pth in assets then weights and index in logs
both are training files lol
just one is used for pretrains
hey guys, i need some help
i downloaded a model from hugging face cuz i couldn't find here and when i try to select one of the pth files (g or d) i got this error "No eligible .pth files found in ./models/sarah"
here's the model i downloaded https://huggingface.co/QuickWick/Music-AI-Voices/tree/main/Sarah Bonito (Kero Kero Bonito) 9k
Ayo? @hushed comet level 1 !!! 
If i give you the D233.pth and G233.pth of the model and the dataset and same settings , will you be able to train with last epoches where i left off ?
oh,,
unfortunately it's the only one with her voice on the entire internet..
oh damn
if you gave me preprocessed stuff + g/d + settings, yes
does it always take so damn long to extract the Odkada Voice Changer?
you should #1159289738314919936
OR
you could just make your own
ig that depends on your CPU and file size
i mean i have a 12900k and its going pretty slow the file size is 4gb
Eh, it probably is normal
i see, i ended up asking there but I would like to know if there is any way to convert to rvc since i'm a bit of a novice when it comes to training models myself.. but thanks anyway ^^
i also keep seeing people say the voice model they downloaded contained a virus is this even real or is it just cap
Don't believe there is, sadly
It could most likely be cap, but it's a possibility because of the file type
But again, consider the fact that most model makers don't know how RVC works entirely, so I highly doubt it could be widespread
how to do in rvc webui
is there a guide
do i need to drag mine models pth to rvc weights file?
i dont have weight file in rvc
gotcha
what do they mean
i need to choose which sound i want cover? with my selected model
poopmasta
copy the path of your audio
file path
yea but which audio? the audio of my voice or the audio i want to cover with my voice?
the audio you want to convert
hey guys i have a problem, i can't find "cable input" in my output audio, is it normal?
do i also need to select the index
yea
use dark mode
press the moon symbol
for a better experience
aight ty i tought why is mine white
what does cmd say
cant find moon
it's in the docs website 😭
here
oh u mean doc guide
@half cove
use a colab to use it instead
upload it to huggingface first
look at the guides 😭
read the bit about doing inference on colab then
i said it here
where is that
Originally written images by FDG Re-written by Julia Last update: Jan 20, 2024
yes this was what i needed
