#✨│ai-help
1 messages · Page 50 of 1
yes
so you can test
dont touch pretrain
yeap
index u can do before and after
its the index file
for accents
train is the pth
you could try use youtube or help rvc for any other questions
also #1159513888199540817
have a good day 
@half covewhat tutorial should i watch
Ayo? @brittle wing level 6 !!! 
are there different tutorials for each thing
what are you trying to do
to use the rvc
inference or training
idk how to use it
i want to use existing models
i dont wanna create them
screenshot the folder of your model
my model or the rvc
model
unzip it
ok
should say extract all
what 😭
yes
ok so
you have index and pth file right?
sometimes you will have index sometimes it wont have
yes
ok
go on your rvc folder
then the assets folder
then weights
then put your pth
tell me when you have
okay
you done it?
move the pth file only into your weights folder
in the weights folder is there other pth files?
theres two weights folders
should look somthing like this
so its good?
put your pth in
it looks like yours has more but ig those are your ones
i copied it there
yes there is
you can change your index name
it might take like a min
does that screen matter
it should have opened a cmd
it looks complicated
you arnt going to use the cmd
yeah okay
also the website opened
btw ur using a old version of the webui
yes thats where you use it
space in filename too
i can see the file name
😭
0/10
should be a dropdown
how*

put the pth model you did
i did
try shift right clicking it then
click it
yeah i did
then paste it in Enter the path of the audio file to be processed
is it in the up of where i uploaded index
its a different language so im figuring out
yes
okay
those are for fine tuning
my audio has no music btw
whats that
Ayo? @half cove level 114 !!! 
now click convert
it says error
This whole time they could have just googled "how to make AI cover in rvc"
whats does it say on the cmd
😭
i gotta go eat rn 
I think that's called terminal
theres two convert buttons
Enjoy
one of them said something
traceback and stuff it said something like that
this sucks
i just wanna talk like a character damn
lol
what do i do
i will click the convert again
nah it said error again
dang
what file type was it
mp3
Ayo? @brittle wing level 7 !!! 
so i gotta move it?
now?
okay
so i gotta delete this one
and download another one?
okay
what is the difference
is mine slower?
damn
i cutted and pasted and i think its gonna take forever
i will just delete it and redownload it
nah its done
i didnt have to delete it
lol
okay i will try same thing again
poopmaster
i moved the folder
didnt work again
what do you use
maybe i gotta use it too
maybe mine doesnt work
mine is ''original''
i thought itd be more basic and stuff
idk
the folder is in my desktop now
so it is fully extracted
i will just delete it
and do everything over
i will extract it again from the compressed file
i wanna make audio for a video
do they work same way or do i gotta learn how ilaria works
il*ria
do i gotta download something
i thought it was a website
ilaria
because it looked like that
idk how is it supposed to work
''chose the model''
and doesnt let me do anything
how will i chose
it says ai covers
i wont do that but is it still the same way?
it says link
link of what
the model
google drive and huggingface
yea
as a folder?
what is a model?
them all combined?
what should i upload, a rar file?
should i upload the index or something
or the whole folder which has index and stuff
hey
what am i gonna upload to google drive
-huggingface follow the first two steps
English guide
Link: https://rentry.org/fdg_guide_new
Credits: FDG
Spanish Guide
Link: https://rentry.org/AI_Hub_guide_esp
Credits: Julia (ailen2091)
Hindi guide
Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes
Turkish guide
Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes
ok
paste on the download link
but the ilaria vc says ''hugging face link''
i didnt
do it
are you using the duplicated space
does it matter
i dont know if im using it
i clicked duplicate space again
its showing me texts on a notepad thingy
screenshot it
nevermind i guess i clicked something else
im duplicating it
it says ''building''
wait for it to build
i thought i duplicated space before, what if i did it a second time now?
then one should be on your huggingface profile
it will prob say this just click reset then wait
can i refresh the page before that
if u want
yep
where should i see the model now
''choose the model''
yes
it doesnt give me anything to choose
click refresh before u do it btw
yea
so i dont need to use rvc? i will just use that il a ria?
so i dont need to use the rvc?
i dont need to download it?
i dont need to have it on my pc?
if your using ila rvc you only need to download the infered audio
it uses the cloud
so like another computer does it for you
why would i need to download rvc then if il a ri a does the same thing
hey
what do i click
Can someone help me, im using the colab rvc and it keeps on disconnecting every 5 mins "due to executing code that is disallowed in our free of charge tier"
ila rvc has no training
i uploaded audio and pth
also it uses cpu so it can be slow
so it will change the audio that i put in
Ayo? @brittle wing level 8 !!! 
right?
yea
what are pitches
what are they
is there a pitch for every number or something
like 1 2 3 4
or just the ones that it says
makes the pitch higher and lower
so like
higher number the higher the pitch
like for example think of a piano
a higher pitch note will sound higher
thats for tts
text to speech
okay
tbh i never found a use on that because it doesnt match with the instrumental
i mean if your converting voices that are very different then u prob should
like female to male
as females usually have higher pitch
damn its still processing
vice versa
might take a while
yea i agree but does it fit with the instrumentals? its so weird when I try to do it
fine for me
yess
Ayo? @cinder bobcat level 1 !!! 
woah then that description ilaria is so confusing for no reason
im not sure where to ask this but are all the voice models safe to download or should i be cautious
il*ria
what rvc are you using? im still on mangio rvc local
just mess around with it untill you get the pitch you like
what does local mean
i keep hearing it on places
on some places
youtube vids and stuff
converting on my pc not on google colab
what does local mean
local = your pc
wdym by google colab
oh i just joined the server, maybe its outdated
damn
ai
is such a weird thing
im listening to the audio now
it changed it
damn
but it sounds too robotic
well ig its not ilarias fault
its just the model
right?
the fork wont change the output
ops sorry
then im asking you, whats the best rn, Mangio or RVC Mainline for local use?
Rn im using mangio but i wonder if the others are different?
im not sure where to ask this but are all the voice models safe to download or should i be cautious
i mean if your using a old version you might not have rmvpe
i have it actually
rmvpe ftw 🗣️
is training on rmvpe also worth?
you put the .pth file in Weights folder and index in Log folder
yes
Ayo? @fossil edge level 1 !!! 
you install from weights
ok
do i gotta make the google drive link thingy again whenever i go back to site?
it doesnt remember the model that i uploaded
what weights folder
@half covedo i gotta copy and paste the google drive link again everytime i go into the site?
because it forgot it
and theres some models that i didnt upload
idk where did they come from
Did you try googling "how to use RVC tutorial" before?
There are plenty of them on YouTube too
yea
wit
wait
no
u asking if u can reuse the google drive link?
@brittle wing
i dont see the model on ''chose the model''
and instead i see models that i didnt upload
the models you did before wont save
theyre so complicated theyre all about animes and songs and covers
and ofc these super smart people got this really smart idea of using tts ai or voice changing ai on their videos
Unless you use your own duplicated space. They are saved until you reboot the space
wow how incredible and unique is that
and the way they still move their lips if theres a facecam
these guys are just weirdos
they record themselves talking and then they make it an ai voice
How old are you by the way?
for what
it doesnt matter
You don't seem very grown
fr?
im older than that
Yeah! That's the cool thing about HF
you hate me for asking rvc questions on help-rvc
I'm not even a mod 😭
When did I say that I hate you?
I barely even know who you are
its obvious that you dont want me in the server
Idc if you are here. Idk what makes you think that I hate you
well idc if you hate me or not

you are insignificant
That's 
well whatever
ai is a super weird thing
i already knew about ai but seeing it done feels weir
d
weird in a cool way
I apologize for the interruption, but I was about to train a 41 minute and 56 second dataset with the default settings and I tried to train this at 300 epochs but it took damn more than 1 hour that I just ran out of GPU backend.
These are the files that I put inside my zip file:
I recommend not using MP3 in the first place. Here's why https://rentry.org/losslessguide
Were you using Colab?
Yeah, I used the RVC v2 Disconnected notebook.
colab has a gpu timeout limit
then you cant use for 24h
or 12h
leme check
i think 12
help i cant figure out how to make models! i was told that the mainline version could but i can't figure it out, how?
Ayo? @boreal snow level 2 !!! 
When I try to run the Easy GUI Colab, this error now appears: “Ngrok requires a verified account” but it worked yesterday
why no cuda? and why out of date? and i was told this version would train models!
did i just wast my time downloading 9.5 GIGS??
im done
isn't that just the index of your gpu? the computer usually counts with zero as the first number, so that would mean that it detects your 4060 as the primary gpu
or idk, I just downloaded RVC and it shows me this as well
is it outdated?
i dont see a way to train models ether
you mean in the voicechanger?
Ayo? @naive sparrow level 1 !!! 
that is not supposed to be used to train models
can't really blame him as I came here for stupid question as well
oof
I don't see the weights folder where the models should be located, is there a way for me to inport my own model?
i dont think ill be training models any time soon
not at all, RVC is a pack of two things at once, there should be a second .bat file
That's nice and all, but do you know the answer? English is not my primary language and I would be really glad if somebody just told me what to do in order to import it instead of reading through it all over again
What is the recommended amount of Epochs for a model?
depends on the length of your dataset
What if I do a 10 minutes long?
ty
it's pinned here in this channel, so that's good to know if you ever want to look it up. For 10 minutes about 300 epochs should be about right
but as I said it depends on the length and also quality of the dataset, there is a tool called tensorboard that shows you a graph of how well the training goes, but I can't really tell you how to get it working as I have to find it out myself as well
Ty! Imma apply it to my WILLOW model then
every time a try to start the client demo it makes some sort of internet connection and somehow causes steam to think there it no connection. i have bad internet but i think the version i have is a virus i can still use discord though... i tried to run it when the internet was not connected and it just asced if it could acsess wirless but it still worked (i dont have wirless)
help
client demo? and how does the "internet connection" look like?
Ayo? @naive sparrow level 2 !!! 
you should start the web UI by opening the go-web.bat
i have a start_http and a start_https
that shouldn't be the issue
Ayo? @boreal snow level 3 !!! 
Im sorry but where is the client demo located?
RVC Guides (How to Make AI Cover)
Translation by country
aaaaah
well in that case you are in the wrong help channel bro
this is for RVC and not okada
ooohhhhhhh
okada voice changer, it's kinda like the realtime part of RVC but better
i keep getting audo block passed
and the voice is super duper choppy cant understand it one bit
is it free
whenever i click the gui thing it loads and then stops woerking after a bit
I'm trying to make an ai vocal cover but it doesn't sound like the character I'm using, is there any way I could fix this?
guys I'm trying to open mmvc but it's stuck here (wait web server...0 http://127.0.0.1:18888/)
it's just mlp characters
I've only used okada so far. What program should I use for .wav files?
Is there any difference between different rvc softwares, I’m using rcv-gui
-local
Suggestions for @glad ravine
- Applio-RVC-Fork, by IA Hispano GitHub
- Mangio-RVC-Fork, by Mangio621 Huggingface
- RVC Studio, by SayanoAI Huggingface
- AICoverGen, by SociallyIneptWeeb GitHub
- Replay, by Replay Team Website
- Original RVC, by the RVC-Project team Huggingface
You can find more info on the #1159513888199540817 channel. If you can't find your answer, feel free to ask for help in #✨│ai-help. Credits to Faze Masta and Antasma for compiling these links.
mangio or original rvc can do it
Yes, mainly in feature set and GUI, but the tech is mostly the same
I need help
Ayo? @snow hazel level 7 !!! 
oof i think you screwed up while converting the vocals to a diff file format or sample rate
could you try using the original file, perhaps?
Wdym?
.mp3.mp3?
Ayo? @finite galleon level 23 !!! 
apparently you seemed to have converted the file somewhere and that screwed up the sample rate
so do you have the original file.
?*
I have the one that I haven't split the instrumental and vocals
Want me to redo that process?
Did you even convert the isolated file
Like the Hz?
Yeah maybe that could help?
Yeah
I did that and it was at 44100 so idk.
did you try rising or lowering the pitch?
I'd just keep it where it is tbh
Yes
not really
I can redo the process again and see if that fixes it.
Alright, do that and lmk if anything happens
It looks like it succeded but I don't know where it went
-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-HF, by r3gf Huggingface Spaces
- AICoverGen, by r3gf Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- Advanced RVC Inference, by neuclya Huggingface Spaces
is there any reason why rvc just stops responding after I press start audio conversion?
Ayo? @brittle wing level 1 !!! 
I'm using mangio btw. should of said it.
I am also having a similar issue. When I begin the conversion it says something like google collab stopping my process and wanting me to pay for a subscription.
I fixed mine just by going to the command prompt that opened up and pressing enter on it
You can look up covers or live performances because sometimes the original vocals doesn't cut it
Phew a lot of ppl in the help channel
Is it in the weights or logs folder?
Hello! I'm using Mangio v23.7.0 and go-web works fine. When I go into \Fixes and open Launch_Tensorboard.bat a command promp opens for a second then crashes and nothing else happens. Also, there's no other launch_tensorboard file anywhere
Alright so, your model has two files. A .pth and a .index
The .index named "added" is the one you should pick
@viscid moss they finished training. Which .pth is the right one?
That's really weird. Did you try reinstalling Mangio?
Before you downloaded it, did you whitelist Mangio and the folder from the antivirus?
Try doing that
I disabled my antivirus completely but I'll try that
Ayo? @wispy hull level 1 !!! 
what do i do with it? i thought i needed a zip file to use the model
Let's explain this
how do i use an ai voice?
A .zip file is a type of folder that compresses the files you put in, making it easier for transmission. It is like putting stuff in a box, it makes it easier to move it from one place to the other. So think of a .zip file as a type of folder, not a model itself
Did this but it still opens and closes instantly
An RVC voice model is made out of two files. The index and pth one (the index is not always necessary but to make it simple let's assume it is)
Ur pth's files is on "weights" folder
And sorry, take your time to reply I know you're helping someone else
Choose the latest one and this ADDED_XXXXXXXX.index file
And I still need to help a bunch of other ppl 😭 thank you for understanding
Zip that 2 files and that's it!
ok, take your time, i didn't unpack the zip yet lmao
Ayo? @misty quest level 1 !!! 
theres a pth in my weights, do i zip that with the added index?
You don't have any other antivirus other than Windows Defender, right?
No problem. I managed to grab a screenshot of what was on the console before it closed let me know if you'd like that
Did you download Python too?
I have AVG but I whitelisted the folder and turned it off
Yeah DM it pls
would it be the top one?
yep
both files into a zip
i am understanding that
-pth files are important
-i need python (?)
@viscid moss any idea of what the problem is here?
Why is Tensorboard unstable at times
You don't know what python is?
yes i know what is python, the question means that is actually required
Yup apparently
Or else it will start giving you errors
thank you 🙏🏻🙏🏻
Idk how to launch tensorboard on Mangio 😦
so....
cmd > python > ?
Seems like you run out of usage time. That's the thing about Colab. I think it has a daily 6 hour limit. Were you doing inference or training models?
Because if it's the first one, you can do it somewhere else with no limits
https://rentry.org/ilarvc_inf_guide here ^-^
~ Guide made by Julia ~
~ Improved by Alex and Squishy ~
Note: It works on mobile too! If your device is slow and can't run RVC locally, don't worry. This software doesn't use your device's resources. All you need is a stable internet connection.
READ THE GUIDE TO THE END. It will be worth it!
~T...
Probably u re using a banned colab notebook!
That could certainly be it, I was gaming with some buddies and had it in the background. Didn't know it was time based.
What would that be?
U can use Ilaria RVC like @molten pecan said or the new versions of colab notebooks.
-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-HF, by r3gf Huggingface Spaces
- AICoverGen, by r3gf Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- Advanced RVC Inference, by neuclya Huggingface Spaces
@proper shale any idea of what the problem is here?
This guide shows how to separate the vocals and the instrumental of songs using UVR locally
Link: https://docs.aihub.wtf/vocal-cloning-guide/vocal-and-music-separation/uvr
Link: https://docs.google.com/document/d/1M3rEj5RPSE60f17RdM9MaIKKkS2bLrj-dwwuCUOdRaY/edit?usp=drivesdk
got some issues with applio now
when i try to train my model i get this near straight away
RVC Guides (How to Make AI Cover)
Translation by country
Damn thanks for this
My model's graph is really fucked, I am currently following the guide (https://rentry.org/RVC_making-models)
Is it a requirement to use iZotope RX and/or Adobe Audition for a more precise model?
Or is merely using UVR and Audacity sufficient enough for a good model?
RVC Guides (How to Make AI Cover)
Translation by country
ehmm ilaria rvc is on hugginface
dont know whatcu mean tbh
So I just use Hugging Face?
Oh this wasn't my problem
Someone dmed their problem so I just asked here for the solution
But thanks grvyscale
They already fixed it
omg ty
Hi Im looking for a rule of thumb guide for how many epochs/Batch size to use for how large a dataset. For instance i have 3 mins of audio so, generally speaking how should I set the parameters accordingly?
All-In-One Guide on how to make a good model
This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models
Credits: LUSBERT 
Automated Overtraining Detection (AOD)
Will be available soon in #1159513888199540817
Credits: grvyscale
u can try 16/8
batch size
and epochs, go for 500 and check for overtraining
guys
what should the pitch be for guys
in ilaria RVC
i know girls are 12
but what about guys
which one is for training online (on a free colab without getting banned)
RVC Disconnected
alr thanks
you're welcome!
ty
girls to boys is -12 and boys to girls is 12 i believe
what if its deep
Just keep at 0
ok ty
np ;)
by the way there's a guide for it in #1159513888199540817, strongly recommend reading it cuz rvc disconnected can be a bit confusing in the beginning
Hello AI HUB Community ! I search how to train a voice model now... Help me, i am french !
is the rmvp still the same
or 61
still dont sound
like the original voice model
hi there
anyone know how to change the gpu option? its not showing my gpu on my rvc apps
#1159513888199540817 can help
do you have an AMD GPU?
nvidia
are you using W-Okada?
yes
hmm how to check it, its japanese lang. i dont understand T.T
ooooor maybe, you don't have CUDA toolkit installed
just make sure that you're running the onnxgpu-cuda version
ahh, gonna try to download cuda 1st
Ayo? @verbal ridge level 1 !!! 
already download it, but she option is not showing 😦
- 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-HF, by r3gf Huggingface Spaces
- AICoverGen, by r3gf Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- Advanced RVC Inference, by neuclya Huggingface Spaces
y
uhhhhhhhhhh
Ayo? @distant wagon level 5 !!! 
Do I need to press that cross?
yeah just click on that and convert
ok
Is there any next step?
now just download ur ai converted thing
Yes but... I can't understand qwq
It's not that complicated
for?
idk, but it’s written on the steps
what?
ohh
D:?
I smell tea
So... I can do without capcut?
oo ok
Do I need to archive at Ilaria RVC?
what do you mean by this
I'm afraid that if I leave that page, the things I made will disappear
you downloaded the audio right?
yes
then the audio will be saved on your device
go on where u downloaded it then u can play the audio
I mean if I close my page all the stuff I made will be gone
like the models?
uh...yes(?
@proper shale what should the pitch be
i know its 0
but i mean
whats the rmvcp
it doesnt sound like dainsleif
i think it does save
models do save, audios dont
yea thats what i said
what about the rmvp
???
it's ok i guess
do i need to change anything in the advanced setting
prolly not
I just changed it, do I reset?
just test out diff pitch values
nah no need
think about how the audio you're gonna convert sounds pitch-wise and think about ur model's character
then ur gonna pitch up or down to make it sound more like the character
What's the link to this site???
secret
Ayo? @quartz frost level 4 !!! 
If it save, would it be here?
I am getting this error "UnicodeDecodeError: 'utf-8' codec can't decode byte 0xfc in position 32: invalid start byte"
oh ok
Ayo? @distant wagon level 6 !!! 
-colab
Suggestions for @grand mango
- 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-HF, by r3gf Huggingface Spaces
- AICoverGen, by r3gf Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- Advanced RVC Inference, by neuclya Huggingface Spaces
ty
I need help with this.
It shouldn't be able to be used for voice, right?
is there anything for non local training
-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-HF, by r3gf Huggingface Spaces
- AICoverGen, by r3gf Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- Advanced RVC Inference, by neuclya Huggingface Spaces
@proper shale Can you help me?
What's up?
I am getting this error while training a model
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xfc in position 32: invalid start byte
What is your model's name and does it have any special characters?
hmm
model name is just "aaa"
Ayo? @lavish meteor level 2 !!! 
ive been trying but never fixed.
is their a guide for RVC disconnected?
Ayo? @misty cypress level 1 !!! 
Yes, it's in #1159513888199540817.
Hmmm, what's the name of the folder the application is in?
C:\Users\amaan\OneDrive\Masaüstü\RVC1006Nvidia
maybe the ü's were the problem
Masaüstü means desktop
also please move it out of one drive
to where
maybe ur downloads folder, wherever u want
let me try
which RVC are you using?
Mangio Fork is what i am using
It is not stuck, you have to wait.
oof i think you had to only use 2 cpu threads
NVIDIA GeForce RTX 3070 Ti this is the gpu i have
should use 5?
...no?
just 2, anything else glitches out, at least for mangio lol
what for RVC WebUI?
Mangio-RVC-Fork
2-4 threads are typically fine, and rmvpe-gpu is way faster yet no difference (which the latest original RVC and applio have)
Go for 500 and check the graphs for all models
go for 6 or 8
...no
this isn't how it works
you can't estimate how many epochs you'll need
oh
it's... kinda impossible
sadly
so just go for 500, dont think itll be overtraining by then, but who knows
u should have a TensorBoard launcher .bat file or something
where
anyone in the chat?
LOL
deez nuts
/removeskin @half cove
RVC Guides (How to Make AI Cover)
Translation by country
in rvc training do higher batches speed up training time? I have no idea what batches are
it trains faster per epoch but uses more VRAM
I have 8 giga of vram, how many batches should I put?
between 6-8
ty
It's the amount of data that's processed at once. The ideal amount depends on your GPU's VRAM. For example, for a GPU with 8GB of VRAM, you would use a batch size of 8. This option influences the speed, not the quality.
If you get the infamous RuntimeError: CUDA error: out of memory error, calculate your VRAM and substract 2, and select that batch size. For example, 6GB VRAM - 2 equals 4. 4 would be the ideal batch size.
ty
edit prevention, if it's older than 5 mins, the bot deletes it
Avoids issues even though it's inconvenient
I thought OOM error barely exists on recent versions of RVC and GPU driver, only performance bottleneck
Hi, I'm trying to use the RVC GUI (https://github.com/Tiger14n/RVC-GUI), I've launched the GUI, imported a .zip with the weights and the index, but the UI is outputting "Please select a model and input audio file", even though I normally had them already selected. The cmd doesn't output any error. Can anybody help me ?
I'd recommend using another local fork since that one is outdated
-local
- Applio-RVC-Fork, by IA Hispano GitHub
- Mangio-RVC-Fork, by Mangio621 Huggingface
- RVC Studio, by SayanoAI Huggingface
- AICoverGen, by SociallyIneptWeeb GitHub
- Replay, by Replay Team Website
- Original RVC, by the RVC-Project team GitHub
You can find more info on the "❔┋guides" channel. If you can't find your answer, feel free to ask for help in "help-rvc". Credits to Faze Masta and Antasma for compiling these links.
Thanks, and what would you recommend ?
Mangio and Original RVC would probably be best
Thanks I'l try that
is there a way to use rvc without downloading anything?
-colab
Suggestions for @wooden ruin
- 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-HF, by r3gf Huggingface Spaces
- AICoverGen, by r3gf Huggingface Spaces
- RVC v2 Huggingface version, by Clebersla Huggingface Spaces
- Advanced RVC Inference, by neuclya Huggingface Spaces
is there any of them u would recommend
ok ill try ty
Check the guide for it in #1159513888199540817 too
@honest pond
Hello, I would like to request for some help with the Wordband bot D:
I asked him to cover a video, he told me "You're in the queue! Your cover will be ready soon. Number of people in front of you: 0 (Status: 🟢)" but he took a little bit too much so I asked him again a couple of times more
And now I can't ask for more covers because the bot tells me "You already have too many generations in the queue! Please wait for some of your generations to finish before adding more."
Ö
It's been more than a day and nothing has happened
Also, I'm really sorry for asking him too many times to cover the same video
I'd just recommend making covers with other stuff
is it ok that my .index file size is about 160mb?
I'm using applio with 32 batch size and ~20mins dataset
and whats the difference between added and trained index?
May anyone answer?
oh my god that's a huge batch size, you'd only need 1-20.
Anyway, yes it's ok
ty
Added is the one that can be used in inference
If you want better models, yes.
it seems to me that in the add-on the already processed dataset is added and the trained one is not. that's why it weighs so little but it's just a suspicion.
https://rentry.org/ilarvc_inf_guide my guide ^-^
~ Guide made by Julia ~
~ Improved by Alex and Squishy ~
Note: It works on mobile too! If your device is slow and can't run RVC locally, don't worry. This software doesn't use your device's resources. All you need is a stable internet connection.
READ THE GUIDE TO THE END. It will be worth it!
~T...
its been 1:40 hours my guy
Ayo? @wooden ruin level 1 !!! 
but ty
Hey I was busy with other stuff
i mean its fine
but ive already read that guide by the time that passed
lol
What settings should i use
:3
No matter what model i try
the voice that comes out
sounds like shit

that f0 is perfect
maybe the problem is the mic itself. hear yourself talking through it without the app
idk about okada so ask on #🔍│help-w-okada
RVC is the name of the voice cloning AI software
Ohhh
there are different versions (forks) of RVC, like Mangio or Applio. but the main one we call it "mainline RVC"
mainline RVC includes a realtime voice changer
and yeah, okada is built using RVC
so, the diff between the two.
I heard okada is more convenient than realtime (prob because of a user-friendly UI)
realtime RVC is more accurate than okada but is more GPU-intensive

