#✨│ai-help
1 messages · Page 98 of 1
rvc v2 is for inference and clone voices, is not a voice changer
Yes but there are many versions there
can i see the name of the zip you downloaded?
You need to download only one, however the version you downloaded is wrong, i will give you the one correct for amd
-colab
- Applio, by IA Hispano Google 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
- Ilaria RVC, by thestingerx Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Applio, by IA Hispano Huggingface Spaces
- 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
Ayo? @tropic trail level 1 !!! 
Download this for amd
Which colab is da best one?
hello, is there any rvc colab where I could train a model without/with the least limitations?
thankyou very much
im downloading
you are the best
You welcome 😄
love you

whats a good way to make a dataset for airy voices
-colab
- Applio, by IA Hispano Google 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
- Ilaria RVC, by thestingerx Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Applio, by IA Hispano Huggingface Spaces
- 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
how do I invite the astra labs bot to my server?
Anyone still having problems getting models to download on llaria I left it yesterday because they wouldn't work and came back today and none of the links I try work.
I'm getting ERROR - The download failed. Check if the link is valid. I was thinking it was me but I'm copying the link with no spaces like I always do.
try again with another model
Unfortunately I tried every model zip link posted for that character but none worked. so I went and tried a completely new character that I've had work in the past and still no luck. same error. I might just try clearing my cache in google or resetting my router or something because I'm not sure what else it could be.
how do i fix # File exists, proceed with creation of creds and client
use a newer colab
-colab
- Applio, by IA Hispano Google 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
- Ilaria RVC, by thestingerx Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Applio, by IA Hispano Huggingface Spaces
- 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
thx
Using EasyGUI
why does it happen?
it wasn't doing that before, has anyone faced this issue before?
finished installing, but there's no "paper.ipynb"
Even when I refersh it there is sile no "paper.ipynb"
@odd shale
You are using the updated colab right? Check if your dataset is zipped and all the audios are wav format
I used .flac files before
It's better compatibility if you use .wav maybe
I got a support from one of your colleagues and he recommended .flac so it won't lose any quality
that was like maybe a week ago after my post was deleted due to a artifacts that my voice model had
Ayo? @jaunty shale level 7 !!! 
it worked before just as fine, now it doesnt
I'll try anyway, but I doubt it will work.
You could try wav, its kinda weird it does that, if it still does that, try to click runtime, disconnect runtime, and restart all over
yep. still happening
Delete and disconnect the rubtime and try to run everything again
doing that rn
nothing.
but i had this
before
I feel like someone had this issue before
Its the easygui colab not the "easy gui with rvc improved downloader" right?
Yea 
yeah i have the same issue then
in the feature extraction
also yes its the new one
We just seen it's a bug of easygui colab
We will look into this
can I use another alternative for now?
Sure, for training you can use rvcdisconnected google colab https://aihubdocs.github.io/en/rvc/cloud/rvc-disconnected/
Last update: Feb 28, 2024
This is a deep guide for it
And it's also what i personally use
is it better than easygui?
How do i open RVC V2
Ayo? @smoky crag level 1 !!! 
Honestly i use it myself, it's no ui, but good, i dont use essygui personally tbh
a unit of measuring the training cycles of an AI model.
He is just saying the training data of the models, so saying how many epochs he trained it on
k
Guys, please tell me.
what is the best program to use for the best AI voice reproduction?
I just noticed that there is a difference in what program you use for your RVC model.
RVC Guides (How to Make AI Cover)
Translation by country
у меня проблема я слышу эхо и почему-то голос не распознает в вокале что мне делать?
I have a problem, I hear an echo and for some reason the voice does not recognize the vocals. what should I do?
this seems quite similar but different setup, It's not that difficult
Ye, i think that's what most users use tbh
yeah not only that, but its quite fast too?
well my dataset is not that long so-
It uses the same free gpu of google colab that you used for other colabs, i dont know if you found easygui slightly slower with the same dataset and epoch?
It's mostly uncomparable, that depends on the dataset
sometimes I cannot tell
only thing that it's kinda tedious is the set up
just a little bit, but it's not that big of a deal
it also writes the epochs to the board quite fast.
Yea i know, i asked u if u already tried training the same model on easy gui
well I couldn't because of the error
thanks man, I'll let know if it done any result
because sometimes there can be moments where like, it will save each epoch
Well, there isn't a difference of quality if you train it locally or online, just speed (that could depend), maybe you meant the pretrains, we have some test pretrains that could help making train faster but its just a test, check out our docs https://aihubdocs.github.io/en
Last update: Mar 1, 2024
You're welcome
and it will do something like this
as an example
which is what's not supposed to do
What do you guys do when geting voices from anime?
For dataset
Oh thats the save frequency, i think you left it to 10 epochs, so its saving a type of the model every 10 epochs, the save frequency is used in case theres overtraining, you can stop the training, see the steps before the overtraining in the tensorboard and see the model which steps are the most close to that point
Clean the dataset to get only cleaned vocals
can someone help me with making a travis scott song w ai
https://youtu.be/p2yIN_8ZbxQ?si=j9v0whEVUtVyPGfI this song with ai
I know but I asking how do people get voices from anime to get for therr datset
Hello Everyone! Super new to the RVC. I have a question, i did train my first model but it's size 1.6 gb how can i convert it to the 53 mb model?
Quick question.
The lowest point in the graph means that it's overtraining from what I believe, that means I have to find the checkpoint closest to the step
so instead of 400epochs like I set it up, it's gonna be 210 instead, What if I use the normal one? (that has 400epochs normally)
in applio whats the best settings for the songs that has background noises or voices
Does anyone know why my models in RVC GUI sound very robotic? The issue isn't with my track because I've tested it on many others, not just mine. The models are also of good quality. I've installed all the necessary packages for Python, harvest/crepe the same. I'm running out of ideas.
i have rvc gui 23.05.25 with rvc v2 support. are you talking about this version?
Ayo? @echo zealot level 1 !!! 
Ideally you want mel, kl, and d to still be going down while g is constant. But if g doesn't go down then stop
how to remove the echo on the voice?
Deecho-dereverb if there's a lot of it
im back, basically overtraining is the lowest point before it goes up again, as you seen it was around 2600 steps, so you need to use a model pth before that steps
so let's say a 210epoch with 2530steps for example?
Last update: Feb 29, 2024
okay, last question, would it be more precise to put smoothing to 0 and determine which one is the lowest?
it would be better if you put it to the max
manual said to 0.987
yea thats what i meant
oh ok nvm I see a different low point now
yea
already finished
this one then
if thats the lowest point before it goes up again, then use a model with a bit like lower steps than that then
yea
okay, what if I use originally 400 epoch like I set it up? Would there be any difference?
overtraining could cause the model to lose quality
that's fair
Anyone have any clue how I run Applio I can't find the RVC-GUI.bat file
run-applio.bat
Are you saying the batch file?
yes
Now what It just shows running on local url
Ayo? @pearl tulip level 1 !!! 
has it opened?
Yes
should open a tab in your browser
It did is that it?
yes
Thank you
if you want to know how to use it you can read this applio guide
if you have any questions or problems you can ask here again
JSONDecodeError Traceback (most recent call last)
<ipython-input-10-16f981cb2116> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:
5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
Ayo? @shell gulch level 1 !!! 
Im trying to use rvc v2 colab but i keep getting this on the download model stage
your using an old colab
same
i cant pass step 2
-colab
Suggestions for @frail vortex
- Applio, by IA Hispano Google 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
- Ilaria RVC, by thestingerx Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Applio, by IA Hispano Huggingface Spaces
- 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
i recommmend ila rvc for inference and disconnected for training
your using the old kalomaze easygui colab
that doesnt work anymore
@frail vortex @shell gulch
ilaria rvc?
Ayo? @frail vortex level 1 !!! 
its just a fork of easygui
ila rvc
ty ill try it
Question how do I fix the connection error for applio?
you will need to upload to huggingface first
elaborate
any model or that i want
if its on weights.gg you have to download it and reupload it cuz it needs auth to download
reupload to where
what does cmd say
Whats the button for it again?
huggingface like here
oh i found a model at hugging face
if your on local it should open a cmd after you click run-applio.bat
yes
👍
did you try blend 2 models that didnt exist?
ty bro i made it :D
Yes
np
thats why then
ohhhhhhhhhhh
xd
so basically i use kits ai but my download time expired
and i want to make a fanmade melanie martinez song but i dont have it back yet
so whats a good thing i could use instead of kits.ai
google colab
right so how do i use it?
do you want to make covers or models?
covers
just ai covers
because im gonna be making ai covers and yeah
i recommend using ila rvc we have a guide on it here
its online
mhm but can i get a link to the downloads section?
like where to download models on ila rvc or where to get the model links from?
where to download models
you need to go on the download tab on ila rvc (shows in guide) the models are in #1175430844685484042
kk tysm
Hi. I can't find my model on Tensor Board. I put the flag file in the logs where it's supposed to go. Even whenever I refresh it, it doesn't show up.
what are you tryna generate
A character's voice?
i thought you meant tensor ai mb
anyways i dont use that so idk
So how do I select the model if it doesn't show up?
Yeah I meant that.
What do you use?
i thought you was talking bout the ai art
tensor
for vocals i use ilaria or kits.ai
Can someone help me with how do I covert like these files into one file?
you can make a folder and put them all in it
or I can just put them in a audio joiner
How do I use ilaria? And what channel should I use them in?
hai who needs help?
https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/ here's the usage guide. It's through a website
=3
hi i need help
btw is not required for RVC to work with a single audio file. you can use that folder as the dataset
but with the w-okada thing
idk how to use real-time stuff
gotta ask in #🔍│help-w-okada
I need help with Tensor Board.
How To Troubleshoot 
- Don't simply mention your issue, like "
my rvc is not working". - Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
- The more context, the better.
- Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
- It's okay if you're frustrated, but don't take it into this server.
- Don't DM without prior consent.
- Don't ask for every little instruction. Put your own effort & test things by yourself.
- Don't ask to ask.
- Check if your answer is a Google search away/on our guides website.
Can someone help me? When I turn on the AI, only my voice comes out, the voice I put in does not come out, can someone help me please?
if it's about voice changer/okada, ask in #🔍│help-w-okada
I’m guessing there’s not quite a difference of quality if I just use multiple rather than a single right
yeah. theoretically there shouldn't be
I can't find my model on Tensor Board. I put the flac file in the logs where it's supposed to go. Even whenever I refresh it, it doesn't show up.
flac files in the logs? you mean the folder or the output box?
Although I’m pretty comfortable, we’re just using one audio file I don’t know is it because of the training speed if it makes a difference or not
The folder.
you are not supposed to put an audio file in the logs folder
idk where you got that from
https://aihubdocs.github.io/en/ read our guides here
Last update: Mar 1, 2024
if there's any issue, lmk and give context pls
why the voice changer don't load?
while using rvc i keep getting an error message that just says "none"
RVC Guides (How to Make AI Cover)
Translation by country
how do i download mangio rvc, this is really confusing 😭
I'm currently trying the applio and llaria HF spaces since the colab's broke. Both are stuck on processing. Is it normal to take a really long time compared to the colab, or are they broken? I've only uploaded a short clip to convert. Thanks in advance 🙂
Okay llaria worked it just took a while. Is that the best option right now? I was spoiled with how fast the colab was lol
scroll down where it says resume training
and you do have to enter the step count 2333333 apparently
Das if u clicked save only latest cpkt
Oh my bad, yeah you have to stop it before getting kicked out
why it takes so long to fix rvc iilaria on google collab
How do i even open RVC V2?
@proper shale Hey! Sorry to ping you but can you help me with colab?
What's the problem with it?
Send the console error
which colab notebook?
Can you send the error you got with it?
on colab
@meager brook download it from here
Should I use the 22k over 8k steps?
Im still confused guys, how do i open RVC V2
locally or on colab and for what purpose
can you give me some details. how you are trying to open it
mel, kl, d rised up but then went down further again for context
But which goddarn file does that
send me picture of what you are trying to do
Which file in Mangio-RVC-v23.7.0 opens the thing
go-web.bat
what did it say ?
python crap and then it opened a website
I think i got it now
local host ?
idk
you say web .. localhost:
do you now have a purple interface
I don't get what any of this means, im just trying to make a simple ai cover
Ayo? @smoky crag level 2 !!! 
Everything iooks purple, yes
On the website atleast
here you can make your cover
ok
but don't close the system32 thing
That closes the website aswell?
no. but the interface will show errors
alr
do you want to know how to make a cover?
why isn't harvest working on RVC-GUI?
Im only doing this so i can make the petition guy say "cheese"
from postal
colab or local host ?
local host
But sure
do you have voice model of that character
Yes
zip file ?
is it that my audio is like 7 minutes long?
So you weren't ment to extract it?
mangio-rvc or applio or mainli
extract it. you will get 2 files. index and pth
uhhhhhhhh "Tiger14n/RVC-GUI"
Ayo? @vast tiger level 1 !!! 
and not in a zip
copy. index to logs folder in mangio-rvc
this is the one ive used for a while
I don't know about this stuff
no . just put it in logs
and put .pth in weights
This stuff has gotten complex since i last downloaded lol
why you have downloaded it ? you can use applio
now?
go to the purple interface
and click on refresh index path.... something like this
k, and i selected postal dude for both pth and index buttons
now copy the path of your audio [which you want to convert] and paste it
help
k, i will get the sound effect i want
my stuff is echoing after it crashed
how do i do that?
Cause im using a .mp4 i downloaded
-local
- 🍏 Applio, 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
- GPT-SoVITS, by RVC-Boss GitHub
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.
convert it by using VLC
This?
yes
k
have you converted mp4 to mp3
eh no
Ayo? @smoky crag level 3 !!! 
after this. select your pitch extension algorithm [rvmp, crepe, mangio-rvc]
https://www.youtube.com/watch?v=T-PfTM_JdpE Well considering im using this small audio clip, i might choose the highest quailty one
if it is doing so let it do.. czz I have never tried to convert mp4.
and of course its not embled
It fails to do so
so yeah
Not a option
convert it into mp3
Done already
there is an ungodly echo pls help, its also picking up me typing
And why does it go through the process bar when im converting
oh it failed again
Thats why, i think i forgot something to do
why ? have you closed system32 thing
send me a picture of that system32 thing
what about the "Auto detect audio path and select from the dropdown:" thing
For some reason i blurred out the pc's name and replaced it with "user"
idk why but i just felt like it
send me a picture of that purple interface
I only blurred out the name because i have not changed it
ok
no matter
a full picture
full screen? just screen down
I will take more screenshots
can u give me a rvc colab github repo?
Ayo? @hexed hull level 1 !!! 
I don't get it
(Also yeah my user is called Family PC but it's only me using it now so i need to change it)
no... put your audio file's path
Think i get it now
The "Auto detect audio path and select from the dropdown:" is still not letting me select the audio file i put in audios
refresh?
is this path is correct?
Well what is it ment to iook like
oh wait i have to change audio.wav to audio.mp3 or something
what is the name of your audio which you want to convert audio.wav or audio.mp3
put mp3 instead 0f .wav and hit convert
done already
converting?
Yes
And then how will i be able to listen to my ai?
i still have somegirl selected so do i refresh
my Ai ? the output you have generated?
yes
Because i cant find the audio i put in while using auto detect
ah
Ayo? @smoky crag level 4 !!! 
somegirl
so change the path
this path ..
Ayo? @latent kettle level 22 !!! 
i turn audio.mp3 to the name of my audio?
no
Well im using the default path
And i accidently pressed convert already so its gonna fail
-colab
- Applio, by IA Hispano Google 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
- Ilaria RVC, by thestingerx Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Applio, by IA Hispano Huggingface Spaces
- 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
note: copy the path of your audio which you want to convert. for example C:/Users/...../[name of your audio]
press shift and right click on your audio and click on [copy as path]
and paste it in path and remove "
Why do i keep clicking convert everytime you tell info i need
Well now i cant convert when i removed both ""
It cant convert
😵
I don't know what are you doing
embled fail
This is what it iooks like
Im just gonna uninstall because this is too complicated considering im just trying to get postal dude to say this
Um... I'm using the latest Applio... and I get a white screen when trying to see the Tensorboard.
What's going on?
I know I'm on the latest version... so is there something I need to do?
@smoky crag Tryreplay.io. It's ran by the Weights.gg team for AI covers
Wdym there's a version history on their website. It's almost impossible to break
Just wondering if anyone knows how to set it up properly to use discord? I tried using virtual audio cable to hook it to discord but when i do that it goes all stuttery but its fine when i just loop it to my headphones
Do you want me to send u the link? It won’t let me upload any links to the model part
This is the one I use
update it seems to be the whole audio on my pc? i dont know if has soemthing to do with voice meeter
but people on the receving end here it fine
Ayo? @umbral moss level 1 !!! 
RVC Guides (How to Make AI Cover)
Translation by country
do i do 1k steps or 2k steps
actually
look
which point do i do
is it overtraining
if i do
3.4k steps
Wait
hey someone can help me?
my colab always reconnect
i'm just started training and become reconnect
u mean it says for innactivity? Its normal, click the page once every 5 or 10 mins
what colab are u using btw
rvcDisconnected
yea its normal then
does it say anything about innactivity?
or about gpu?
like what does the popup/warning say
no
i doesn't get anything
except the notification about connecting with google drive and captcha
and now connecting again
you need to connect to drive
yes is already connect
Ayo? @pliant sail level 3 !!! 
i dont think so
click runtime,disconnect and delete runtime, then re run everything again, also look at this https://rentry.org/colab_workarounds
try reconnect
and after that, 5 mins later when start training, become connecting
i'm already reconnect
and be sure u always connect to drive
You want to compare values so 33 > 34.20 or whatever. You can keep training if d, kl, mel keeps dipping.
okay thanks, i will chat u if my colab was problem again
ok i downloaded the lowest one
k, also kl and mel can keep dipping infinitely. But if g stops being constant and rises forever it's done
hmm alright
ok i tried the model on a song
a LOT of artifacts on the very high notes
but other than that its fine
if i remember
mangio-crepe and crepe is for like stuff with sudden pitch changes right
Question,can you use RVC on a discord VC?
for using models in realtime, you can use wokada, #🔍│help-w-okada
Thanks
yw
anyone active here?
hi. i got a problem with rvc v2 step 2 download voice model by using google colab
i was using the voice model in this discord and paste the link to make it download the sound but it show up with the code
JSONDecodeError Traceback (most recent call last)
<ipython-input-11-5bbe4d5947f3> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:
5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
what colab are you using
send link
Ayo? @uncut cedar level 1 !!! 
Hi uh do i have to pay for a ai Character to use?
@edgy bloom so i followed the guide downloaded the RCV n get a voice record and stuff but i kinda confuse on the .PTH one
do u mean the one in tensionboard?
tensor*
whatever piece of text you are stuck on @echo dagger
Ayo? @echo dagger level 2 !!! 
here
this part is related to adding voice models that were already made
is that what you wanna do
you said you wanted to make a model, right
can someone help me please
skip over the inference section for now @echo dagger and go back to it once you are done reading the training section and actually training the model
im getting the JSONDECODE error
in step 2
JSONDecodeError Traceback (most recent call last)
<ipython-input-6-7406109395cf> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:
5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
which colab is this
so start from here
yes
google collab maybe
that collab is really fucking old i think
what should i do then
Ayo? @proven lynx level 1 !!! 
Hello do I need a laptop to make a ai cover?
Ayo? @brittle wing level 1 !!! 
its easiest on something like a desktop or laptop but people SOMEHOW do it with phones
beats me lol
ok im trying them
Ok thank you sorry I'm very new here so I have no clue what to do
can u teach me how to make ai voice covers
im literally blank
@edgy bloom ok im back again by dataset i just put the voice record into the path file right n how do i know it actually received when i click process
tbh im not even obligated to help people here anymore i just got really bored, i dont mind pointing in the right direction but i dont got the time to do any step by step stuff
Ayo? @edgy bloom level 36 !!! 
go look at our guides and see what works
link pls
Who can help us ?
are the voices here safe?
the audio goes into a folder inside of the datasets folder, you path to that so like "dataset/voice" whatever you named the folder, without the quotes
and then hit process and it should show a bunch of shit in the box next to it
where is the guide?
@edgy bloom @proven lynx Can I ask if its safe to download voice folders
the voices are safe
idk im new here
because when i click on the link the discord default says it can be mallicious
garry do you have a good pc, the guides for the online ones just seem to be down
so i have a doubt
i have gtx 1650 and amd ryzen 5 4600h
discord does this based on how often the link has previously been clicked on, and if its a zip file it will always nag you about potential danger
will that work?
it should work for using the voices but not making your own
so the voices arent viruses ryt?
Ayo? @brittle wing level 1 !!! 
they are perfectly safe
thanks
???
trying
@edgy bloom hi yes im back again so im at this step but i can only see the default one
have you actually began training yet
it usually takes an hour or two to finish
just depends on your audio
mine is kinda short so it kinda end fast or i do something wrong
try taking those quotes outta the dataset path and then hit the buttons again
this one?
guess is working since this dont instant output
yeah, thats not going to work
let me find an online guide
your own pc wont be able to make models
sad weak laptop moment
Ayo? @echo dagger level 3 !!! 
what batch size is best for 5-10 minutes of dataset
@edgy bloom how i know it over train
ok so
the loss category
look at loss/g/mel and loss/g/total and when those start to consistently go up
they may spike but when its like, not going down anymore
thats generally the stopping point
it stopped itself
it shouldnt do that
RVC Guides (How to Make AI Cover)
Translation by country
did it go all the way to 200 and finish
so i just turn 200 to 400
check total/g
where can i see it
scroll to the bottom and press this
now scroll to the bottom and change the 1 to 3
page 3
im just following the guide so
10min
yup no iea
pain
ig just test the model
I've only been using it for a few minutes and it's already disconnected, what happend with this colab?
It's inactivity
you went afk
You need to click once every 5-10 mins
heem
Dont go afk
I just saved the model and continue training
but only for 12 mins
Click sometimes randomly on the page
damn
huftt
@low shard what you think of this?
bro just went offline
@glad zealot guess i stop this now
erm
@glad zealot so where is this
@edgy bloom
Where I can train my own model?
https://aihubdocs.github.io/en/essentials/how-to-make-voice-models/ You should read this first
In the context of RVC, the dataset is an audio file containing the voice the model will replicate. It can be either speaking or singing.
https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing
Is this colab outdated? I keep getting an error popping up when I try to download a model
Its kinda weird the way it goes up and down idk tbh
Im literally out rn
at least u touched grass
Ayo? @echo dagger level 4 !!! 
Yes it is
Okay thanks
it worked but the word is not clear on certain word
maybe i need a longer data
of speaking almost every english word
It's on the webui
anyone that train model suggest me how to make a clearer speaking model?
Last update: Feb 29, 2024
It explain you how to clean a dataset
for make a hq model
i already cleaned the data using that
i just wonder how u guys take the data and how many time is suggested cause i used a 10min one
10 min is also good
do i have to include every word possible or nah
Depends from the voice we want to take, for example for singers with original songs and we clean with uvr the voice
No don't worry
15-20 is the best but 10 is also ok
so only the 3-5min song then loop it?
or multiple song
No different songs
And for the dataset format for singers i recommend to use .flac
Never mp3
or m4a
wav and flac are the best
i used wav
Is ok also wav
guess i needed more example
i will just make a new one next time
im dead making one already
you guys know how to cover full high note song like this (idk what it call)?
https://www.youtube.com/watch?v=FvOpPeKSf_4
some part cant cover
Ayo? @pale geyser level 1 !!! 
-local
- 🍏 Applio, 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
- GPT-SoVITS, by RVC-Boss GitHub
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.
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
-realtime
This interaction has expired, use the command
/guides realtimeif you wish to see it again.
Ayo? @maiden raft level 1 !!! 
-rvc
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)
the easy colab is not working
which one are you using?
im guessing the old one
easyGUI
and whats not working?
when i tried to train he voice, it says error on the gradio live
Ayo? @low cliff level 2 !!! 
show what error 
Does anyone have the psych2go model ?
-audio
- Perfecting Audio Isolation on Low-End Rigs: A Practical Guide, by Litsa The Dancer and Faze Masta Google Docs
- Gathering and Isolating Audio, by SCRFilms Google Docs
- Creating Datasets for RVC using iZotope RX, by LUSBERT
Rentry - Vocal Mixing Tutorial, by Roomie YouTube
Audio Separation/Isolation
is there a guide for the best infer settings?
something weird i have noticed with RVC - when OBS is opened in the background it slows my PC down extremely
like i cannot have OBS running at the same time as RVC
how long does training usually takes
wait click on the third option below loss g total
...yeah
Ok thanks but why the percentual number on the epochs is still to 0%?
I mean right hete
Here
?
It is Norma?
Normal*
Yesterday I did another training and the percentage of epochs increased, but not now
yeah
Why?
it just tells you how many % it has trained of the epoch at that moment
sometimes it actually has done more than 0% so yeah
Yeah ;)
Thank you so much 🙏
You're welcome ;)
I take it there's a limit as to how well the VC can convert my voice?
I have to speak quite clearly for the changer to sound normal
Yes
Problem is, I can barely speak straight as it is thanks to dyslexia
My lips can literally get out of sync with my thoughts causing a word salad
I understand a better GPU won't help?
Oddio stavo trainando raga
E si è disconnesso
Ora che faccio?
😱😱
Helpp😭
Devo rifarlo da capo? 😭
guys i can use AICOVERGEN WEBUI in a low end pc pls help
Hello, Ive tried the ai-voicechanger, but the app is not claiming any voice-audio from my pc.
Where
your using ila rvc correct?
on the gradio page
it should have 2 tabs
one for inference one for downloading models
It has the play button thing on the left
screenshot
So I clicked it on the first part to download it and then i dont know what to do for the gradio thing
Im not on my pc at the moment but ill get on my pc soon
did you read the guide?
I watched this video on how to make a ai cover
But the google colab was the old one u said
so you run both cells
on the second one it will give you a gradio link
you click on that then it opens the ui
What does the link look like
at the top of the ui it will have 2 tabs
Oh ok
And then after that u import the voice model and the cover song thing ur doing?
yep
this should tell you everything https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/
Ty
@half cove can I please get help too?
with?
This
screenshot your settings
Which channel ? :D
Ayo? @remote pecan level 1 !!! 
now you should be able to also voicechanger channel is #🔍│help-w-okada
EasyGUI Colab is not working, i tried to train it. the error keep popping up when i tried to train the voice.
-rvc
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)
RVC Guides (How to Make AI Cover)
Translation by country
What does JSON Decode error mean
So I trained a model with Applio, Am I SOL If I want to use that with AI Cover gen etc.? How would I go about turning that into a model/zipping it up properly to upload?
i thought this was general chat wtf my wifi sorry
are you using the applio colab or locally?
could you be more specific? like what do u use and maybe show a screenshot
Sorry its already fixed
Ayo? @violet badger level 1 !!! 
well glad to hear
send the colab link, to check you are using the updated one, and also show a ss of the error
-colab
- Applio, by IA Hispano Google 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
- Ilaria RVC, by thestingerx Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Applio, by IA Hispano Huggingface Spaces
- 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
hey guys i got a question. The voice changer isnt working when i click on start. The guy in the video said something about i need python?
can somebody tell me why the dio, harvest, crepe, crepe_full,rvmpe options in the "f.0 det" are n/a? how can i enable them?
do ya have any other details?
what's your specs? how's the cmd logs looking?
you're on the AMD version
oh, i have nvidia
from where i can install thenvidia version?
icant upload pictures unfortunately.
Ayo? @coarse geode level 1 !!! 
also bro please change your username 💀
welp, now you can
thanks, mate
I guess i need another programm since the guy in the video doesnt linked them
k, that's the amd version
aight
so i have to download the intel one and it works?
wait what is your gpu exactly
you're welcome btw :)
nvdia geforce 2070 super
yeah then use the one i linked for the other person
Kk thx ill try it
are these settings right?
dont use crepe use rmvpe instead your output has nothing and chunk and extra are too high
Im trying to upload a new voice but its not working
!howtoask
How To Troubleshoot 
- Don't simply mention your issue, like "
my rvc is not working". - Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
- The more context, the better.
- Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
- It's okay if you're frustrated, but don't take it into this server.
- Don't DM without prior consent.
- Don't ask for every little instruction. Put your own effort & test things by yourself.
- Don't ask to ask.
- Check if your answer is a Google search away/on our guides website.
what are you using specifically
if you're using the realtime voice changer - just post that on#🔍│help-w-okada
Hey i need help, so i did everything right, all the models from the voicechanger work, but not the custom ones that i downloaded here, they bring me to 1000ms, what can i do?
im on amd btw
Ayo? @hearty ether level 1 !!! 
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
Frequently Asked Questions
Jump To The Important Sections
Other Links
Download NVIDIA
Download AMD/INTEL GPUs
Download Mac
Online Hosted Alt...
basically, convert to onnx
Hey, its been a long while since I've used RVC and I've noticed it is taking a lot of space on my hard drive. I did train a model and I got about 30 gigs from a 500 epoch voice, but its on logs, not assets or other places.
Are those files from the "save every X epochs" to avoid crashes leading to info lost and whatnot? As in, if I delete them, will the voice still be available?
Its been a while and I'd rather just ask than search info online
Or is it saved in weights alone?
Im getting this error when running Step 2:
JSONDecodeError Traceback (most recent call last)
<ipython-input-14-c72d8c0ea2c9> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:
5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
Those are probably the D/G files. You may have checked one of the options that saves D/G files separately every saving cycle
Pretty sure, I just checked and there is a .pth in files from the voice I made, so I'm guessing if I keep that, the logs are safe to be deleted?
I mean, if you wanna keep training keep only the most recent D/G files. Else, delete them and you'll be fine
And yeah the usable pth is in weights
Its at epoch 500 so I don't think I'll do much more than that, I'm happy with the results
Yeah then you can delete the d/g files
Thanks! Its mostly 'cus I was cleaning my drives and I got spooped to see like 50 gigs in the RVC folder
anyone know how to help with this?
might be using an outdated colab
https://aihubdocs.github.io/en/ has the recent stuff
Last update: Mar 2, 2024
and tutorials on how to use em
which cloud service should i use?
Ilara RVC?
if you just wanna use models, ye
im following the tutorial but it turns up an error
Ayo? @indigo reef level 1 !!! 
its ok
-colab
- Applio, by IA Hispano Google 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
- Ilaria RVC, by thestingerx Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Applio, by IA Hispano Huggingface Spaces
- 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
-local
- 🍏 Applio, 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
- GPT-SoVITS, by RVC-Boss GitHub
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.
I need help
I ran out if GPU I was about to reach 300 epochs but everything's interrupted
Do I have to start all over?
I RAN OUT OF GPU
WHAT I SHOULD START ALL.OVER WHAT A WASTE OF TIME
you should have stopped after an hour or two
and exported model
that way you prevent stuff like this
I can't export model while it's training also it's a request not my own model not my problem
Ayo? @brittle wing level 3 !!! 
Also why tf the moment I trained my own model it didn't crash at 300 epochs what's wrong with the colab WTF
Wouldn't it resume sibce it has a saving frequency
I will try and resume training with new account the rest idk.
Probably it'll work
It used to work in the old training colab crepe + persistent storage
you can simply stop training and export
two cells can't be running at the same time unfortunetaly
are there any advanced dataset cleaning guides? (I have RX)
there's one by lusbert: https://rentry.org/RVC-dataset-RX
In this guide i will be explaining how to use a Paid Software to clean audio for training models.
iZotope RX is known to be the software for denoising audio and the one used by "every" good model maker.
In this guide i'll be recommending mostly free VSTs/Plugins and some paid plugins wh...
🎙️ Realtime Voice Changer (Local)
Description: This one has more options and is somewhat more convenient to use than Real-time RVC.
Link: Rentry
Credits: ravencutie21
Description: More accurate than W-Okada's, but requires a decent GPU.
Link: Google Docs
Credits: teas.hi
Description: Text to speech with RVC
Link: Rentry
The Deconstruct step boosts vocal volume but it can cause volume peaking in some parts so you have to manually lower the gain if the auburn noise gate and kilohearts doesn't work 
help pls
Change ur runtime.
Follow the first steps of this video: #1208578250943504465 message
I wish I could use rvc on my phone cuz my laptop broke
U can use Colab os HF Spaces
thank u so much ❤️
-colab
!rvc
- Applio, by IA Hispano Google 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
- Ilaria RVC, by thestingerx Google Colab
- RVC Disconnected, by Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- Applio, by IA Hispano Huggingface Spaces
- 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
U can found the guides here: https://aihubdocs.github.io/en
Last update: Mar 2, 2024
Which one do I click
U want to convert some audio or train?
Convert audio
I’m sorry it’s been a bit since I made or used models
dw u can use Ilaria RVC or Applio
But since they open in another tab, on mobile you can be disconnected. So Hugging Face is better but it is quite slow
Does that one remove vocals from audio for you like covergen on pc?
Ayo? @dense marsh level 3 !!! 
Ok
What is the best vocal removing site?
UVR5 (App or colab) or MVSEP
Because I usually use vocalremover.io but it only lets u use it like 3 times a day
I see
Well...
UVR5 It's pretty good, MVSEP too but have queue
How long does the queue usually take
It depends tbh.
Alright so I got the gradio thing and followed the steps for that and It still is giving me errors when I convert
what does the error say
lemme get a screenshot
Ayo? @mortal ibex level 3 !!! 
what does the error show
