#✨│ai-help
1 messages · Page 97 of 1
Only this in there
And inside the commons folder there is nothing
Ah got it
Was in Appdata/Local/Packages/Python/Local/Scripts
Or something like that
Hello, is it just happening to me or is the CoverGen installation not working? A few hours ago it was working fine.
hugging face is down
will likely be back tomorrow
(or in a few hours if we are lucky)
the one night i want to use my colab gpu credits lmao
Failed to execute script 'MMVCServerSIO' due to unhandled exception!
im getting this message
The client turns white when i execute it
Is it possible to increse epchos nuber in applio(there is no infer-web.py in folder)
Past 1000
oh really?
is that why i have a blank screen?
when i try to launch it
hehehe hazbin hotel RP server =3
with okada allowed
that's cute
yep
dunno about the okada colab but yeah, most likely
The official one doesnt allow it
that sucks
for some reason
I mean its an SFW server. odd
Should be NSFW since its a show for adults only
Ayo? @lucid leaf level 2 !!! 
No wonder they banned roleplaying
[VCClient] wait web server...150 http://127.0.0.1:xxxx loop here
noob question: how do i download a model on the local rvc easy gui? i understood how to do it in colab but now im lost lol
^ i have this question too
i remember there was a how to video of it but i no longer have it
darn
Ayo? @quartz jetty level 1 !!! 
anyone out there with such vid 🥺
https://aihubdocs.github.io/en
Check the website 
Last update: Feb 28, 2024
guys rvc is back up!
Ayo? @storm holly level 4 !!! 
the instructions say 2 and a half hours but is it normal if it's working at a rate of minutes?
does it usually take forever for a song to convert if your model doesn't have an index?
how do I make my mic only pick up the voice not my voice
RVC Guides (How to Make AI Cover)
Translation by country
Ayo? @spare wadi level 1 !!! 
- JSONDecodeError Traceback (most recent call last)
<ipython-input-9-0ef3af68c63f> 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 does the error mean
Ayo? @uneven cedar level 2 !!! 
Really dumb but how do I delete model slot configs?
Ayo? @narrow igloo level 1 !!! 
show your cmd/console
i click and nothing happens
I need some help please, does it matter if you just talk while making a dataset ?
its for either singing or rap
got a XLR studio mic not sure if that will make any difference
yeah
-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
that is wokada, ask help for that in #🔍│help-w-okada
hey guys i need help when i upload a voice model this happens to me
it tells me could not initalize voice model
JSONDecodeError Traceback (most recent call last)
<ipython-input-3-d207da5c4fef> 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 is it
you're on an outdated 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
Check out Ilaria RVC
Ilaria RVC, by thestingerx Hugginface Spaces this?
Ayo? @slow swallow level 1 !!! 
yeah
Did you train it yourself ?
I'm not sure how you upload downloaded voice models, but if it is one found online, you can just put the link to it
more info over at https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/
Last update: Feb 25, 2024
When I make AI Covers and upload a ZIP, how do I know what pitch setting is the best? by trying? Used the Billie Eilish Model on a Bruno Mars Song and set it on 12. And I tell you it's having its problems
ask for that in #🔍│help-w-okada
tyty
usually the model makers leave a note on which should be the best pitch, but yes by testing yourself, as the input depends it too
Hi. I'm trying to use Google Colab to get models, but I keep getting stuck on the Running in Public URL. For some reason, it looks like I'm getting errors and I don't see that link. I've tried copying the google drive link, the discord link, and the huggingface link and I keep getting the same results. What am I doing wrong?
RVC Guides (How to Make AI Cover)
Translation by country
Your input will influence the pitch too
So just do it by trying
And see if its too low or too high until its perfect
What colab are you using
/content/voice-changer/server
Traceback (most recent call last):
File "/content/voice-changer/server/MMVCServerSIO.py", line 10, in <module>
from downloader.SampleDownloader import downloadInitialSamples
File "/content/voice-changer/server/downloader/SampleDownloader.py", line 12, in <module>
from voice_changer.RVC.RVCModelSlotGenerator import RVCModelSlotGenerator
File "/content/voice-changer/server/voice_changer/RVC/RVCModelSlotGenerator.py", line 5, in <module>
import onnxruntime
ModuleNotFoundError: No module named 'onnxruntime'
WARNING:pyngrok.process.ngrok:t=2024-02-29T12:17:55+0000 lvl=warn msg="Stopping forwarder" name=http-8000-f50785c0-8eea-436c-8bd5-d8f2f5771bbb acceptErr="failed to accept connection: Listener closed"
The runtime ends with the above contents
Ayo? @severe ocean level 1 !!! 
Outdated colab
There is an up to date one in the pinned messages over at #🔍│help-w-okada
CAN SOMEONE HELP ME PLSSSSSSSSSSSS?
I need help. I read the tutorial, but it was really confusing and I'm trying to follow the attached video, but i get stuck on Running on Public URL. For some reason, it isn't showing up. On the download model step, it looks like it had some errors, but I'm not sure if that's supposed to happen. Can someone help please?
!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.
im having trouble with iiaria rvc im coppying the link and naming the model but every model i try come uo with ERROR - The download failed. Check if the link is valid.
that video is old
and it uses outdated colab
most yt videos have outdated ways
with okada .-.
Ayo? @orchid plume level 1 !!! 
i tried online mode but it didnt worked or idk how to use
you can either use the new easy gui which has inference, training, but ui banned for free users https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw#scrollTo=Sb5fzhzEXK8X , or either check our docs https://aihubdocs.github.io/en that have ui inference like ilaria rvc
Last update: Feb 28, 2024
be specific, and the help channel for wokada is #🔍│help-w-okada or #1192011222023950368 , not here
ok thank you. I thought i might be on hugging faces side
If I wanna download multiple models, do I have to install it everytime? Or is it just a one time thing?
Ayo? @brittle wing level 1 !!! 
Oh and I"m using Ilaria (in case that makes a difference)
The hugging face space or colab?
I don't know. Sorry, I'm brand new to ai so much of this is a bit confusing but I'm following this guide
But I'm actually having issues with ilaria too. I installed it, and it seemed to install propeerly since it said done cloning repo, but when I tried starting it, it didin't have the gradio.live link. this is what shows up
Why isn't it working?
Could it be that my computer specs don't support it? I'm using a i5 10th gen cpu and a 1660 ti gpu (I think)
Do you have any ideas how I can get it to work?
currently seeing if it works on my side
Which it should, but we had someone else with that error before
What does the crossed out text mean?
Did it work on your end?
So no, but I think some of the python org's servers are under maintenance
or had issues
Maybe try this, and relaunch
It took it a few tries for it to successfully work on my end this time
Do I have to reinstall it after clicking that?
Ayo? @brittle wing level 2 !!! 
yeah
idea is to make it start from the start again
I FINALLY got it to work lol. Thanks for your help and patience!
Who knows? 🤷♀️
Not to be a pain, but any estimate for when the models will start work to again. I've tried a few and they still don't work.
all I've seen is sometimes it will work some times it wont. So I kept trying different models and so far none of them have worked. I completely closed out of the runtime and disconnected, restarted and refreshed but no luck with any hugging face link.
Why is the voice lagging when using Mangio RVC?
Sometimes it is lagging
Sometimes not
Like stuttering
think i found the problem
And what settings should I use? On Response threshold, Index Rate, Sample length, harvest, Fade Lenght, Extra inference time. I don't even know what these settings do
people say i need to check the arrays, what does that mean?
What's the benefit of using forks like Applio instead of RVC Webui?
Some extra thingies like integrated TTS
instead of having to get tts audio and manually convert it
- also multi-algorithm pitch extraction, which is kinda nice
hi, can you tell me how to fix the arrays in the applio colab?
Thx, I see.
This is related to training, correct?
for both training and to make covers
Failed to train index: need at least one array to concatenate
why this?
as far as i know the index shit is a GPU problem, as all my audio files are .wav
how much time worth of audio files are you feeding it ?
error seems odd
50 seconds
oh wait that doesn't seem to be a GPU colab runtime
expected to see "T4" somewhere there
so the problem is that i only used a minute's worth of audio
because the character i used only has a minutes' worth of dialogue
what's the problem here
You want to connect to a colab runtime which has a GPU
Instead of a CPU-only one
Unless I am mistaken and there is a GPU on this
still don't know how to fix the index crap
try with a GPU runtime
it seems to expect to use a GPU, which it cannot find
even though it shouldn't need one
how do i activate the GPU runtime
You sure ? It doesn't look like this for me
these are my settings
weird
I really don't know, I haven't used applio myself recently
but the error you are having is really weird
might actually download applio to see if the problem persists
Voice keeps cutting out liek a voice chat on bad internet?
i cant really understand the time limit before a disconnection in RVC V2 with free colab
it's impossibile to go over 400 epoche since it random disconnected
and if runtime disconnected it's impossibile insert the next day a preprocessed model
can someone help me with setting up sm?
Sure 😄
how do i start making the ai voice models?
I send you a guide 😄
Last update: Feb 10, 2024
um
Una domanda nella fase training quando inserisco set training variables devo far partire il tasto play?
Nelle. Guide non è specificato
Scusate che significa questa cosa?
Durante il training
Spunta tensor board con questo messaggio è un errore?
Sto usando rvc v2 disconnected
Ragazzi
Ma è normale che la guida inglese dice di mettere 2000 epochs
?
Non sono troppi?
RVC Guides (How to Make AI Cover)
Translation by country
Come si aggiorna?
https://aihubdocs.github.io/en/rvc-resources/epochs-overtraining--tensorboard/#usage-guide qua c'è una guida più dettagliata sulla tensorboard se vuoi
Last update: Feb 10, 2024
E riguardo gli epochs? 2000 non sono troppi?
Se non aggiorno tensor board continua comunque il suo training?
depends on your dataset
I have 5 file audio waw of 5 minutes
Total
1 minute for each other
How many epochs?
300-350 are enough..
Thank you
@shut willow is getting this error . help him to fix this
im a boy
is it OKay ?
could giving my .pth any random .index file work?
@sand iris
np
space in folder name
I think yes
really? thatsall?
I don't think it can cause this error
its the (2) that messed up the whole thing bruh
It will cause issues later, so it must be fixed either way
is anyone have trouble running the cell that opens gradio?
what does it do ?
on which notebook
ilaria
okay i'll try
working or not?
breezip
not really
that might also be an issue
as far as I'm aware, it only unzips properly with 7zip
try to unzip with winrar
is it onmicrosoft store
try winrar
nope 
is it on microsoft edge
check your colab
is runtime is connected or not
Sorry to interrupt, this problem probably came up a lot but i couldnt find any help in previous chats: basically it would repeat this over and over except with Beatrice
[Voice Changer] Pipeline is not initialized. [Voice Changer] Waiting generate pipeline... [Voice Changer] Pipeline is not initialized. [Voice Changer] Waiting generate pipeline... [Voice Changer] Pipeline is not initialized. [Voice Changer] Waiting generate pipeline... [Voice Changer] Pipeline is not initialized. [Voice Changer] Waiting generate pipeline... [Voice Changer] Pipeline is not initialized. [Voice Changer] Waiting generate pipeline... [Voice Changer] Pipeline is not initialized. [Voice Changer] Waiting generate pipeline...
@sand iris thanks it worked
what do you mean?
are you using Applio on colab?
yea
check runtime
I vividly remember that I was getting Same error. but I couldn't find any solution
this thing?
Ayo? @storm holly level 5 !!! 
yes... your runtime is disconnected
really? It isnt working with any custom models either, its weird
I works for others.. but not for me and you .. idk why 😭
so you've never been able to use the app properly?
wait
you can ask in #🔍│help-w-okada
WAIT
okay ty
Ayo? @solid root level 1 !!! 
I THINK ITWONT WORK BECAUSE I NEED NVIDIA
LOOK AT THE NAME
RVC1006Nvidia
BRO SO UNFAIRR
Ayo? @latent kettle level 21 !!! 
?
Ayo? @shut willow level 5 !!! 
it work .. but idk why it is not working for you
wait i dont need nvidia?
800+ seconds
you can try this.. if you think it is not working due to Nvidia.. I use this one
kkk
time to waste 5 more gb
your runtime is disconnected. connect it again and run everything again
waste ?
bro ive already wasted almost 118 gb
BTW.. extract it with winrar ..
he said 7z works best
I said winrar.. czz I use it .. yes 7z is good .. but I never used 7z.. I prefer winrar
what was you doing 🥵
why do you keep doing ..
bro is not likeee that im a dev
a familyy friendly one
what ?
you keep saying ".."
noted: connect before installing anything in colab
... 😆 hobby....
itmakes you sound sad or suspicious
if you are inactive for a long time.. colab will disconnect you
also i have imagee perms yiiiippeeee
seriously?
yes
like in a
i lost my grandma... sitttttuation
or a "no i didnt do anythiiiiiiiiiiing illegal..."
my keyboards so glitchy
I don't think bro. it's just like a person is thinking when he is speaking
noppe
okay. I'll try to avoid this
have you downloaded RVC ?
? what ?
he is getting this error
trying to fix it. else nothing
Found the error. There's a space in the path file
remove this space. [rename the folder
I think it's already done
sorry it wouldnt let me type so i plaayed deltarune and all 3 characters died but i need to eat also i aalready tried that nothing worked
everyone is saying.. it is due to space
just rename your folder in windows explorer
and launch it again
I got Ilaria to make and download a model correctly, but I forgot how I did it and the guide on github is a little confusing. I'm having 2 issues. 1. On the colab page, it's saying there's no gpu found, which was new, and 2. On the gradio .live page, when I put my audio in the box on that page, am I supposed to use the acapella track, or the backing track? I can't remember
i fixed it, i just redownloaded it on my main hard drive, i was trying to use it on my external hard drive
not on colab?
colab?
I think you are using it on colab
whats that? 
you don't know about Google colab
i dont i never used it before
either way idk if u had it on ur main drive or not but it works that way unless i did something different
do you have a powerful gpu?
- change your runtime to GPU.. 2ND IDK
How do i do that? I just started using ilaria last night so I'm brand new to all this
on colab
Sorry, this i probably a dumb statement, but I pretty much need a step by step guide on how to do that. I tried using the guide on github but it didn't help.
I tried installing it, and thats when i ran into the issue with the gpu
if you are using colab ... just click on change runtime type and run both cells again
locally or on Google colab ?
i think colab
the interface is like this
Wait...I think it's working. The conversion process was just EXTREMELY slow lol
this notebook?
is it working now?
I think so. It finally converted and had a download link available
But it just took like 20 min or so to convert lol
did you got like an 'cannot connect to gpu backend' popup?
Yeah but I think it's working now. I thought it was stuck on the converting process, but It just took a very long time to convert
Ragazzi aiutatemi ho appena finito di trainare un modello ho il file index e poi una cartella con tutte queste cose dove trovo il file . Pth?
che colab hai usato ? manda il link
uhhh not rlly
Rvc 2 disconnected
Ayo? @ripe flicker level 5 !!! 
ohh, well that means you finished your daily free gpu on colab which is around 4 hours for high task like this, so yea now its running on cpu, which is wayyyyyyyy slowerrr than gpu
Ti serve il link della colab?
Ho sbagliato qualcosa? Il training è stato completato ma dove si trova quel file?
si se puoi
sembra molto strano
ti dovrebbe uscire una cosa simile a questa col nome che gli hai dato
Purtroppo mi spuntano quelle cartelle 😭
Il file index
Ce l'ho
Ma non trovo il pth
Oooooh. Ok, that makes sense. I was confused cuz it was originally a LOT faster when it was using my gup. So is it safe to say that the allotment refreshes every day?
Aiutatemi please sono 2 ore che provo a fare il training
Tutto terminato ma mi da sti file
Ho seguito passo per passo la guida inglese
Spero di non dover rifare il processo da capo😢
check your drives folder or send a screenshot of it please
After model blending where do I find the Index
both files should be there if it didnt abort
did you train on colab or locally?
Yeahs probably
its either in the root of the drive or in your logs folder
I'm not training I'm blending models...
My own models
Together
With applio
Ayo? @brittle wing level 1 !!! 
Here is a screenshot I finish trained my model. And I don't find pth.
You don't even know why I'm talking about
Sei italiano
thats the logs folder. it says rvclogs. i meant a screenshot of the drive
Quale colab stai usando
its the same process, check your logs and weights folder
It's not and the blend already generated it an index and .pth file you don't even know what I'm asking about
Ma usa Easy guo
Gui*
Metti il dataset in una cartella in drive
I audio file non fare zip archivio
Scusa la mia ignoranza ci sono delle guide per easy Guo?
Gui*
Si ma devi trovare quella in italiano io visto quella in inglese
bruh i know what you mean
im telling you
dont you mean merging two existing models
dont use that one
Last update: Feb 28, 2024
use this instead
that one is outdated
What's wrong with the one that I used before?
?!?
Outdated?
i think you used the correct one, wasnt it rvc disconnected? i just meant the guide sweet rose sent is very outdated
please dont link that... this is the correct guide https://aihubdocs.github.io/en/rvc/cloud/rvc-disconnected/
Last update: Feb 28, 2024
that one doesnt even work anymore
Oh yes it is, but I don't know what I'm gonna do after finish that training, because I don't find that weight folders
Maybe I do something wrong
Ohhh so that's the script
I use this guide to train before
can you please send a screenshot of your drive? i keep asking that
your google drive
It's the one I used for my newest model I thought it was broken
i dont think it will work on your phone.
Someone wants me to train their model and said they used RVC disconnected and told me it crashes bruh they lied to me
I train my models on my tablet
This is what I have
They just want to get away
You mean this?
docs.aihub.wtf are VERY OUTDATED. never use them
L index e qui ma dov'è il .pth file del modello
using colab?
mi sa che non avevi finito di trainare non ce il pth
yes i mean that
So that's why I couldn't train this dude s model through easy GUI hah
looks like you didnt finish training
Nick puoi aiutarmi? A risolvere il problema?
non hai finito a trainare
Exactly
dovresti riprovare
Hello there! Any way to use a custom .pth in order to make it sing?
Cosa mi manca?
i kept saying this but they said idk what im talking about
I was asking about MODEL BLENDING WITH APPLIO.nit about training
Ayo? @brittle wing level 2 !!! 
the free colab gpu should recharge around 12-24 hours
Respectful miss sir
I'm confuse guys
I want to blend my voice models
we are talking about two different things rn, please dont mix it up
Not to train
i know what you mean
Yeah
but you just mentioned training too, i was answering that
e mi sa che si è disconnessa mentre trainava
i downloaded the vc thing but its not opening?
Ayo? @noble cipher level 1 !!! 
like i said, it wont work on your phone im pretty sure
Because someone asked me to train a model for them.
In realtà ha vinto i 300 epoche solo non ho capito se poi ho saltato qualche passaggio
And I was wondering why I failed training with easy GUI
Si è disconnesso dopo aver finito i 300 epochs
I know why...now
Allora inizia di nuovo mi dispiace
ehh mi sa che ce stato qualche problema, dovrestri riprovare di sicuro è un problema di disconnessione
Ho già provato prima e mi dava gli stessi file
Non è che sbaglio qualcosa io?
Ho provato varie volte ma manca sempre il. Pth
Ayo? @ripe flicker level 6 !!! 
per evitare che si disconnetti, devi fare un click ogni tanto, secondo me è un problema che lo lasci afk cosi si disconnette
Si facevo così ma purtroppo niente 😢
Una cosa la barra arancione
In basso
Durante il training va toccata?
durante il training lunica cose che devi vedere e se overtraina la tensorboard
Ecco intendevo questo cosa devo fare con la tensorboard? Perché in passato ero riuscito a trainare un modello ma ancora non c'era la tensorboard
Poi nella fase di pre training salto le fasi di load reprocessed dataset files e import model
Nella guida diceva così..
devi solo vederla per vedere se non overtraina come in questo caso
È corretto?
A me non spunta il grafico mentre traino
ah si quelle sono per continuare il training
Posso tralasciare?
dovresti fare l'icona che aggiorna il grafico mentre traina
Tralasciare?
si
Mi dice questo dove c'è la
Tensorboard
Come faccio a mostrare il grafico?
Ho il timore che facendolo una terza volta si ripresenti lo stesso problema
Oltretutto ho raggiunto il limite di utilizzo gpu in colab non posso fare più nulla mi sa 😭
Purtroppo la mia gpu non regge senza colab
seleziona scalars e aggiornalo
Riprova domani l altro giorno oppure fai un nuovo account Google e ricomincia da lì
Cosa non lo sapevo
Last update: Feb 10, 2024
Semmai dovesse andare oltre il grafico, posso fare qualcosa?
where can i download pytorch
ok. And is it possible when using my cpu (or gpu, as well, but especially cpu) that they might be overtaxed and overheat/get ruined? Or would I be ok?
Ayo? @brittle wing level 3 !!! 
you mean if the colab cpu would make the model quality worse ? nope just that it would be wayyy slowerrr
No, I mean my actual cpu/gpu
like that's in my pc
Sto riprovando incrocio le dita
stai usando un altro acc?
ohh, how much ram does your gpu have and what processor?
Si è l'unico modo
I think it's 16 gb and I've got a i5 9th or 10th gen and a 1660 ti (i think)
Una cosa il secondo punto.. Set training variables va fatto partire col play? Oppure no?
Perché nelle guide non lo specifica
yea you can use that for inferencing dw
intendi il pulsante che fa eseguire le celle di codice?
ok thanks
Si dove c'è experiment name
Il rate e tutto il resto
Ok allora i passaggi penso siano corretto
Corretti fin qui
Vado avanti
Nel caso in cui si dovesse disconnettere prima della fine del training devo ricominciare da capo?
😢
si a caso
Ho il timore che anche stavolta vada a vuoto
È il terzo tentativo
Che faccio
Invece save_extra_files_to drive va spuntato?
whats wrong? i am trying to install pytorch
Su scalars non da nulla perché?
Non devi devi eseguire delle celle
Comunque sto notando una differenza rispetto il precedente training
Cioè prima la percentuale accanto ad epoch rimaneva a 0%
Ora ogni epoch ha una percentuale
Cosa significa?
fai refresh
Nel senso che per recuperarlo devo eseguire le altre due celle? Che si tralasciano?
è solo la percentuale che sta finendo l'epoch, non preuoccuparti
Refresh qui intendi?
Segui la guida
della tensorboard si
in https://huggingface.co/spaces/TheStinger/Ilaria_RVC i can also import the models from google drive right
Ah ne sono grata 😄
Come faccio a capire se si interrompe?
Nel senso spunta un errore? Qualcosa?
Perché nei precedenti training non è spuntato alcun errore
O messaggio di disconnessione
Infatti andava avanti fino alla fine del training per questo non capisco il perchè non mi ha dato i file
@sand iris could u help me please
Se la ram da questa linea vuol dire che sta per disconnettere?
with?
Ragazziii Ho appena capito cosa sbagliavo prima... Il file . Pth si trova nella cartella mangio rvc fork Ecco perché non lo trovavo... Io guardavo nella my drive.. Dove ci stava solo il file index
😱😱😱
i am trying to get pytorch but it doesnt work? and when itry to open vc by clicking "start http" it doesnt work
you don't need python or pytorch
oh why it doesnt work then?
no, dovrebbe stare pure nel drive quando finisce di trainare
Ah si? Ma eventualmente posso prenderlo dalla cartella mangio rvc fork? Se non c'è li?
Manda uno screenshot
Anyone ?
Ayo? @ripe flicker level 7 !!! 
Quelli sono i salvataggi automatici in base al save frequency che hai messo, non ha ancora finito di trainare
Quando finisce appare nella cartella di drive
Si
Ma perché nel precedente training le percentuali erano ferme a 0%? Sul numero di epoche?
Cioè non andavano avanti allo 0
Mentre ora si
Forse stavi con la cpu quindi non riusciva a trainare credo
Ahh ecco allora potrebbe essere questo il problema
Tra l'altro adesso sta impiegando il doppio del tempo di prima
Già solamente a trainare
Allora si poteva essere quello il problema
Comunque grazie mi sei veramente d'aiuto purtroppo sono un po' newbie
trqq
metti smoothness al massimo
oh 
Maybe try renaming the long folder name to like VoiceChanger
instead of MMVCServerSIO_win_onnx[...].15
the folder path might be too long 
yeah try renaming it, but with no spaces
btw you didnt really do a good job of censoring
ill just delete the image since you clearly dont want it to be known
thx and uhm it still says the same
what gpu do you have?
So I've trained a model using several audio clips, totaling at one and a half hour of audio now and stored it as 40 bit voice model. Is it possible to convert this pth to 32 or do I have to do it all again?
40 bit? 32bit? do you mean the models sample rate?
yes, the sample rate
and no, once its trained on 32k for example you would have to train it again
but dont just use random sample rates
did you check your dataset in a spectrogram beforehand? do you know what sample rate to use?
I didnt check anything, I figured i'd do it at 40 based on the guides or recomendations i read/found
i guess you already trained your model already so it wont matter now
but
could you put your dataset into Spek, or this? https://huggingface.co/spaces/TheStinger/Ilaria_Audio_Analyzer
see this for more https://aihubdocs.github.io/en/rvc-resources/audio-formats--sample-rate/#section-1
Last update: Feb 10, 2024
uhm AMD?
Ayo? @noble cipher level 2 !!! 
looks like you downloaded a older version
where did you get it from?
and follow this guide https://rentry.co/VoiceChangerGuide
okayy
It shows 48
Ayo? @royal vapor level 1 !!! 
can you send a screenshot?
Thing is, I have 2 other models at 32 that I would like to merge it with. Not only to create a customized model that I partly can modify but also to get a more well rounded model that can handle a wider spectre... Unless I'm mistaken on how the merge works that is
Finito il training è il momento della verità 😢














this audio is really really lq and tbh i wouldnt even considered using it in that state, your best bet would have been training it at 32k too if you really need to use that. also did you use multiple audio clips? you said its over 1h total, how are those files looking like? they should not be varying
where did you get the audio from?
dont look at the samples per second. the graph is the important part, look at where it peaks
They are all of the same qual, its a 3 hour long clip that I split into clips between 10-30 minutes and trained/upgraded
Hello, I'm new to this topic. I've just downloaded Mangio RVC v23.7.0 from github.com (the infer train.bat version), and when I install the Mangio RVC file, I don't get a new zip file with the same name in which I can proceed with further steps. Simply put, nothing changes after the installation. I can't find any new file throughout my computer. What could be wrong?
ig it works now?
the issue was that they got a older version lol, but ive seen that error a couple of times now. weird
if you havent installed vb cable already, please do that. otherwise it wont work
and follow the rest of the guide to get the best settings 👍
its all in the guide, please read it carefully in full
okay
What was the source of the audio? And wdym by upgraded? If you look at the screenshot you send you can see that its actually very lq so im a bit worried. also you dont need a big dataset like that. more than 40min is really not needed and more than 2h would not even change anything, completely useless. if you really want to use a big dataset, make sure its hq and that it is 20-30min (imo)
The source is a private recording that sounds pretty good, but seems like the qual is bad after all. the endresult so far is decent and usable imo, but its far from several of the models ive found here
Hello, does anyone have a link to an RVC that allows you to create AI covers? I've made a model but the page where I made my AI Covers is down so I can't post the model without a sample at least
Also, the RVC I have only allows you to train vocals but not do AI covers.
Ayo? @brittle wing level 3 !!! 
Does merging several models make the endresult potentially better?
if the datasets bad, nothing can save it
what does that mean?
okay
so like...where do i put a voice model on this site?
download tab
after go back on inference and click refresh
it will appear on the dropdowns
So by peaks, you mean the light lines that dont go higher than 17-18?
Ayo? @royal vapor level 2 !!! 
I removed what I could, but there are clicking sounds etc that I didnt manage to get rid of
Figured they wouldn't harm the end result too much
would you be able to send a small part of the dataset to see what its like
also it does harm the dataset
if its not clean, dont include it
processing it too much makes the quality worse too so be careful
uh what shall i do?
my applio was bugging so someone recommended me to install mangio, it wont let me train and keeps giving me this error, I have the infer-train version
So you're somehow having a model version mismatch
...Not sure how that would happen, but it's attempting to load a v2 file, but expecting a v1 model
if you did extract it and didnt run it as administrator it might have actually installed it and the error doesnt mean anything. reboot to make sure and check again if you have vb cable installed
dont type random lyrics here 💀
btw where can i see if i got it already
your sound/default audio devices settings window
or you can just restart the voicechanger again and see if it shows up
ok, Im trying to use ov2 and at the start I set the version to v2, but at the bottom where you can ad your own pretrains, it defaulted to v1, I manually changed it to make it v2 but im not sure how to amke it compatible
nevermind, i fixed it
Hi, its been months since ive made any ai covers since they banned rvc on google colab, how do u guys do it now? Ive heard sum about downloading a program but idk
do you wanna do it locally or online
Do we have to pay now in order to make ai covers or what?
no
Online preferably
- 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
check our docs! i suggest you ilaria rvc colab
Last update: Feb 29, 2024
should be even easier
download is in the gui instead of a colab cell
yeah
Alright, ill check it out but if theres a video tutorial it would be even better
Ayo? @granite falcon level 1 !!! 
soon 
so what should i do with this?
use newer guide 
they are using huggingface moment
aight so what do i use?
if your using the huggingface space click on this and duplicate it (i recommend colab)
..oh that's nick's guide
yeah, which app or site do i use in that guide cause I'm confused as hell
Ayo? @upbeat ice level 5 !!! 
hi
hi
the best quality one is the edge tts one
...a bit
so what should i do if my tensorboard is kinda not there
maybe try using a diff browser
alright
Ayo? @uneven quail level 2 !!! 
then i think your dataset screwed up
my dataset is like 1 minute btw
(i also set the sample rate to 32000 due to the thing saying the audio was 22050 Hz)
yo i need help
C'è l'ho fatta era proprio come ti dicevo il file pth andava preso dalla cartella mangio rvc spunta lì alla fine, il modello funziona perfettamente grazie per l'aiuto 😭
then any one that has the "mixing with rvc" next to it
in realta sta nella cartella drive ma vabe, almeno ti ha funzionato
@molten pecan are u gonna revamp the tts guide or should i
Purtroppo non c'è li ma l'importante che sono riuscito a farlo funzionare 🤣
so i need help idk how it work??? so i can change my voice
check our docs! https://aihubdocs.github.io/en here its explained how to make ai cover
Last update: Feb 29, 2024
say, is there a required length at minimum for a dataset
not really
its more of a recommendation
not really but its suggested to have atleast 2-5 mins for a decent model
5+ mins works best, 10-30 mins is perfect
oh
wel lguess i should expand my dataset
i dont have much data but since its a CDI model i can probably just use sentence mixed stuff
yeah I'm trying something called applio and my god it takes a long time just to generate a small thing...
huggingface spaces take a long time...
Yea because the hugging face dpace uses cpu that is veryy sloww
any alternatives or just those huggingspace stuff?
colab
mhm, do they work with phones?
it keeps saying this despite me copying the path
RVC Guides (How to Make AI Cover)
Translation by country
Hello, I'm new to this topic. I've just downloaded Mangio RVC v23.7.0 from github.com (the infer train.bat version), and when I install the Mangio RVC file, I don't get a new zip file with the same name in which I can proceed with further steps. Simply put, nothing changes after the installation. I can't find any new file throughout my computer. What could be wrong?
welp at this point i might just give up on finding an elevenlabs alternative...
i got the rvc file but i dont see the start apollo???
Ayo? @brittle wing level 1 !!! 
Yes
ok
i got this
doria is not the real name of mine btw
pc name
shi
use 7zip
ok
unzip rvc using 7zip
ok
There's a difference??
give it a min
winrar and normal one fucks up
Kewk
Salve visto che sei italiano volevo chiederti aiuti per una cosa del training
Dopo la disconnessione del colab che avviene totalmente a caso senza un periodo preciso assoluto, il giorno dopo non fa caricare il save nel reprocessed perché dice no such file or directory
Quindi stai avendo un problema a continuare il training?
Chiedo per un amico, comunque si è un loop dove ogni giorno non riesce a sospendere il training manualmente e il giorno dopo non fa manco caricare quel save ma deve ricominciare da capo
Basta che esegui le celle delle dipendenze, usi le stesse esatte training variables, e poi skippi fino alle 2 celle per continuare il training qua
Se vuoi sta un tutorial qua https://aihubdocs.github.io/en/rvc/cloud/rvc-disconnected/#resuming-
Last update: Feb 28, 2024
Abbiamo seguito il tutorial il fatto è che ci fotte la disconnessione che arriva a random, l'altro giorno avevamo il rar non apribile perché danneggiato da non so cosa, comunque proveremo grazie ti aggiornerò
Rar? Non comprimere mai un modello o dataset con rar
Ah bruh
Per la disconnessione basta che fai qualche click ogni 5 o 10 min
Ma una volta iniziato il training dopo quante ore avviene?
Col doppio monitor non si dovrebbe avere 12 ore?
Doppio monitor? Non centra nulla, avviene tutto sul cloud online in google colab
Dipende, non si puo dire precisamente
L'altro giorno a 50 minuti
Dipende dal dataset e dagli epochs
hey is done
Su windows basta che fai click destro e comprimi in zip
now run go.bat or whatever it is
hmm
go-web
something is weard
Ayo? @brittle wing level 2 !!! 
what does it say
where did you unzip it to
yeah
let me do it again give me s ec
download
im extracting it again
this is why i use winrar
with rvc winrar doesnt extract properly
eh
?
did you download it from github?
look
that the page
?
so
screenshot your file rq
cuz why tf is it looking in appdata
you unzipped it before right?
yeah but it dosent show the file
i did the open achive
dont open achive
to see that would work
it needs to be unzipped
alr ima do it
look where im estracting to
alr is done
Alr is working
i see it
now click go-web.bat
ok
tell me if it works
i there smt i need to do to make it work
I keep hearing my original voice really clearly "under" the new voice from the model I used. Any tips on which settings might help minimize that?
like, I hear Lois singing the song I want, but I also very clearly hear myself singing as part of the resulting mp3
something about google colab?
I want to make a voice model. But Idk how. Plus, Idk how to make a pth file with the vocals.
Ayo? @unreal sand level 2 !!! 
70% of it is spending time compiling and creating a clean dataset otherwise you hear noises in your model https://aihubdocs.github.io/en/vocal-isolation--datasets/datasets/#datasets
then you can use rvc disconnected colab or applio to create your model
Anyone working on api for rvc
how do i even open the thing-
you can adjust pitch and also you can change your voice model. I think the model you are using is not properly trained
which thing ?
What ??
Api for rvc
why do you need an API for rvc
So we don't need to depend on collab and we can implement it anywhere
you need a local version of RVC ?
No not local but a version where we can just directly call api from any website and it does job without interface
which you can use on your computer?
Ayo? @bold ermine level 3 !!! 
Not on pc but even at mobile we will impleme t api on cloud
I think no. czz There is no need to do this .
RVC V2
d/total, kl, mel are still going down while G is rising. Do I keep going?
jajaj
Ayo? @edgy parrot level 1 !!! 
amigo me pasa lo mismo
it only happens when I shift the pitch, very odd
you can hear both the original and the modified voice here
THIS WORKS WITH AMD ?
I'm not sure, I'm pretty much a noob at this
RVC Guides (How to Make AI Cover)
Translation by country
RVC or real time voice change ?
Yeah. What is your GPU ?
-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.
how do we use these voice models
UVR is using CPU instead of GPU. How to fix that ? I checked the gpu conversion option. I have rtx 4050 laptop
check our docs https://aihubdocs.github.io/en
Last update: Feb 29, 2024
-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
RVC Guides (How to Make AI Cover)
Translation by country
Ayo? @edgy parrot level 2 !!! 
Which gpu you have and which version did you download?
I have rx6600xt
Amd
And which version of okada you downloaded?
i can see the name of the zip?
Of course 😄
I downloaded the following version
is voice changer client demo rvc2?
Which?

