#✨│ai-help
1 messages · Page 29 of 1
Uh no sé acerca de w-okada
Intenta preguntar en #🔍│help-w-okada
Pero vas a tener que preguntar en inglés
Should they have the same file?
I mean like should I place the unzipped file in both of them?
Nope
Where it says "drag your .pth file here" you put the pth one
And you put the index one in the other one
Oh now it's donw
Yep so now I should press download?
You don't have to
Just go back to the main tab
And hit refresh
The model should be there
@lofty zephyr did it work?
The page is lagging 😭
guys how do i open the RVC soft ware i havent used it in ages
-rvc
RVC Guides (How to Make AI Cover)
Suggestions for @long ingot
Translation by country
thank you
Today is my birthday

Is it normal now it took already 600 s
It depends on the length of the audio
And the algorithm
How long is it?
I still didn't place the audio
I just pressed refresh
This is the loading of refresh
Hm I see
This is really weird
Try reloading the pagw
And hit refresh again
Don't upload the files again, just hit refresh after you reload
You need to duplicate the space
Mangio Crepe takes longer that’s why
When I have the link from the kitai, am i supposed to make an account to download the file there?
i already have the software setup and i just wanna test some models
I use the Applio from Hispano fork
the What's the Problem
Ahhh I'm tired 😭
I'll check later 
Ayo? @lofty zephyr level 5 !!! 
im still a bit confused
Ayo? @vapid depot level 1 !!! 
ive got the kitai link and idk what to do with it
Applio RVC Fork Locally
Problem: after i update it, now it have this module not found ( ffmpeg ) problem
any solution/help?
How?
Because it uses a different method for pitch extraction
How much time approximately/.
Depends on song length
No way this is normal
😭
Why is the refreshing taking alot
I still didn't press convert
Did you duplicate the space?
Yup you did
Try it in another web browser
Like Firefox or Brave
Alr
Is there a gradio colab link with a download button?
I have found 777 results for the search ariana...
Owner: B&A Productions
Uploaded: <t:1693758518:R>
This model can be found on the Kits.AI platform, visit their page for more information.
how do i continue training my model in apollio, after i stopped it yestersay
It's been a long time since I opened the app, then I decided to open it and a message appeared for me today saying: your notebook could go to the trash or something like that and it said that it can be deleted permanently and said that I should contact the owner or Back to colab I decided to keep using it anyway, what could this mean?
Which colab
applio.
Applio colab doesn’t work anymore
-colab
- AICoverGen-NoWebUI, by Ardha Google Colab
- RVC Disconnected, Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- 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
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.
Which colab rvc works now? that is online and not local?
If you want to use RVC not locally, either use the huggingface spaces or colab
-colab
- AICoverGen-NoWebUI, by Ardha Google Colab
- RVC Disconnected, Kit Lemonfoot Google Colab
- easyGUI, by rejects Google Colab
- 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
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.
mine gives an error. I'm trying to do voice with a zip file and it won't work.
I can only do the ones that have a web interface.
Why?
I can't understand it even when I watch a video or something, or I always make mistakes. or never understand.
web ui I think is easier.
Makes sense
-overtrain
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
this here and to create voice model. I just want to use the template.
I do voice model with RV2 disconnected. I know how to model. I just want to use voice model with web ui. because locally I can't because it gives me an error to install or something like that.
Use this one
You only have to unzip it and open go-web and that's all
Don't close the cmd command window thing until you are done using it, otherwise it won't work
Good luck!
Is this here to make a voice model?
how do i continue training my model in apollio, after i stopped it yestersay
set the same model name, exactly same settings except more epochs as you want, then just click train model, skipping the preprocessing and feature extraction.
I don't want to create a voice model, I just want to make a vocal song with a ready-made model but I can't do it through the app and it doesn't work with the error none. Anyone who can help me would be grateful.
Did you duplicate the space?
If your using HF Spaces
And that colab doesn’t work anymore
what's the best voice changer to use for conversations live etc?
Nope
the other colabs is to create AI models. I don't want to create a model, I just want to use a model that has already been created.
Voicemod.
i mean like ai
W-Okada
think i'm using it but i can't tell, is there a way to confirm?
Ayo? @teal quartz level 1 !!! 
Send a ss
i sent it
Anybody have colab of speaker diarization ?
RVC Guides (How to Make AI Cover)
Translation by country
tf is this?
ohh ok, ill try right now
Traceback (most recent call last):
File "/workspace/Applio-RVC-Fork/infer-web.py", line 3095, in <module>
app = GradioSetup()
File "/workspace/Applio-RVC-Fork/infer-web.py", line 1908, in GradioSetup
record_button = gr.Audio(
File "/usr/local/lib/python3.10/dist-packages/gradio/component_meta.py", line 152, in wrapper
return fn(self, **kwargs)
TypeError: Audio.init() got an unexpected keyword argument 'source'
2023-11-19 06:32:49 | WARNING | lib.tools.loader_themes | Error Loading: Base.set() got an unexpected keyword argument 'chatbot_code_background_color'
Traceback (most recent call last):
File "/workspace/Applio-RVC-Fork/infer-web.py", line 3095, in <module>
app = GradioSetup()
File "/workspace/Applio-RVC-Fork/infer-web.py", line 1908, in GradioSetup
record_button = gr.Audio(
File "/usr/local/lib/python3.10/dist-packages/gradio/component_meta.py", line 152, in wrapper
return fn(self, **kwargs)
TypeError: Audio.init() got an unexpected keyword argument 'source'
oh i already did that
thanks anyway tho
it worked too
How to run?
I press more info but it shows this
No python
Ask someone else to help it’s too late for me
what happened to the big amount of threads with everyone sharing models
Ayo? @plush pilot level 1 !!! 
i cant find any models i need now lol
/search command in #🔍│find-models
(IF U DIDNT PLEASE CHECK PINNED MESSAGES OR #✨│announcements)
What is the name of the model you are looking for?
Minecraft villager
I have it but the client doesn't show it
🤔
I am telling you about the client
Not the voice
I'm reinstalling thev voice
Yes?
Ok
Ayo? @lone oxide level 2 !!! 
How do I check
I clicked a Folder shortcut and it opened And I install the voice But it's not showing the option
Press refresh
Ok
Youtube two weeks ago
This link
Ok
Ok so I will download a other client
Yeah
If I have more problems, I will tell you.
For now i'm downloading the new client
Ok thanks
-svc
Yeah I still use so-vits-SVC
RVC Guides (How to Make AI Cover)
Translation by country
Hello there! I have a question. How do I covert a .pth model to a .pt model while not knowing the version of pyTorch used to make the original .pth file?
What causes a model to sound like literal white noise/artifacts? I’m training a singing model.
Should I just let it finish training? Because I stop it around 30 epoch cause I’m just curious what it sounds like
how to setting in 1060 ?
how the fuck did you get the gif
you're talking to a bot
no shit
am shit
does anyone know how to get mainline rvc to use dml with amd gpu I have RVC1006AMD_Intel and mainline rvc but its not using dml or my amd gpu its keeps using cpu
guys where should i download clean and last version a guy suggested copy paste update and my everything ducked up
You can't train on that
inference should be accelerated with GPU though
yeah i know it just says this when i use it for inference
i just want to use it because i don't want to use bad applio
Is the 4050 good for local training
I think it has the power but idk about the VRAM
Would 16GB be enough?
Ah
does anyone know why mainline rvc has this error on some audio and not others using RVC1006AMD_Intel version with amd gpu
Try using the beta one
From what I saw the AMD one is unstable
where can i get the beta version
but why that one its 4 months old and at that point applio is better but the quality of the RVC1006AMD_Intel version sounds better then applio
Really? Imo they sound the same quality wise
I don't know where to get a more up-to-date one
Are there any other resources for voice models rather than weights.gg?
Use /search on #🔍│find-models
But use the ModelAI bot, not the Applio bot
You also have applio.org/models
Hello, I have issue with overall pitch change. When I change pitch on this function, I have error: "Invalid file_type. Must be one of []". I don't know what to do. Thank you for help
Do u have also another one?
also by https://huggingface.co/models but some may not be voice models but other ais tho
Mmm I don't think I'll find that model I want there
or use the weights.gg bot in #🔍│find-models also
what model are u searching for?
also maybe its js not made done yet
I got an error RMVPE
Riize Anton
Oh
i Got an error stored locally
https://huggingface.co/meaisae/RIIZEMODELS/resolve/main/anton-modelrvc2_rmvpe.zip?download=true i found this, you can try it
Oh it's just that I've been facing a problem in HFsince yesterday Ive been using a specific model unzipping it's file and then uploading the pth and index but still and when I press refresh the loading takes like forever Idk y
Ar
Question about RVC: I have a trained voice model (pth) and vocals (wav) that should be replaced. They are found and recognized by RVC. When I click on convert (pm or harvest), it takes some time, reports success but I got an all silent audio file. Does any1 know, what might be the problem here?
I run it on CPU, an i7, 16 GB RAM, load max. 30%. Reinstalled RVC, but still silence.
Hi everyone. I need help with weights.gg
Where should i put downloaded model to make it work? I'm clueless, because normally i use https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing, and it needs only link to model to work, which doesn't work with weights
you use ilaria rvc?
No I previously used it and uploaded the model's link but it kept saying there's an error
id suggest you to use https://discord.com/channels/1159260121998827560/1167844377046040769 which is more updated,no time limit but a bit slower, btw if u really still wanna use colab for making covers, for both colab and ilaria rvc when theres a download link on weights.gg that starts with like models.weights.gg, follow what is said in the following pic
so this error is on ilaria rvc right? did u duplicate the space ?
The link problem was in ilaria rvc but this one is in Aesthetic_RVC_Inference_HF
ohhh
id suggest you to duplicate the space (ilaria rvc or aesthetic one)
They already did
Yeah I did
(I have installed version)
Can you try it on another device?
Can anyone help me I just want to make a song then I already have a voice model now how do I do it with the song?
On a better phone or computer
-rvc
RVC Guides (How to Make AI Cover)
Suggestions for @spiral osprey
Translation by country
Can someone help out with deeper knowledge for RVC?
Good luck!
Sure, what's on your mind?
:3
I'll see thanks!
wheres the voice changer channel
for help Realtime Voice Changer
#🔍│help-w-okada
I get an empty audio at the end of the RVC process when dubbing a voice.
It could be that you didn't properly select your model
What software are you using?
Can anyone help me?
Pinokio installed RVC. I have a trained voice model (pth) and vocals (wav) that should be replaced. They are found and recognized by RVC. When I click on convert (pm or harvest), it takes some time, reports success but I got an all silent audio file.
Ayo? @sudden pilot level 1 !!! 
Depends on what issue you're having...
@molten pecan Can you help me?
Never heard of Pinokio
Does your model have an index file?
I got an error on RVC Intel version
Haven't heard of that one...
I assume it's on older version since you don't use RMVPE
I don't know much about models itself but you need a .index file for inference
Does anyone know what this error means?: 'Invalid file_type. Must be one of []'. Its shows when I use overall pitch function. I have installed last version.
Ayo? @dreamy prairie level 1 !!! 
From what I saw the Intel and AMD ones are unstable
It's just the platform from which you launch RVC. It installs it all in a sandbox. RVC finds all the files, but gives me an empty audio at the end.
Try using that one
hi I tried to download a model from weights. gg but It keeps canceling the download is there any way I can download the model
We can't help if you don't tell or show the issue...
Guys I need help, I'm receiving no sound at all
Have you uploaded the .index file of the model?
Or is it detecting one?
Try downloading it from another web browser
There is no idex file in all the voice models. It detects none.
@nocturne solar
@wispy burrow
So, I got a hardware error on RVC For AMD and intel version
Ayo? @outer isle level 6 !!! 
hi, is my first time with this program, and i want to know how to put a new voice
What program
a friend already told me how to select the voice, but not how to set up
voice changer client demo
Please, help me, if it possible.
What's the issue
this
When does this error show up? What were you trying to do?
What file type are you using?
I am sorry, what do you mean?
I'm trying to infer an voice to another voice that I infer and this error pop up
What's the audio's filetype
Are you using the Mac version of RVC?
I use mp3, and youtube. I am not sure is it what you ask
Which one , hugging face ?
it says I dont have the authority to download it when I tried to download a model in weights gg
Ayo? @shadow mulch level 1 !!! 
Are you logged in?
Did you log in to weights.gg?
Can I get a sc?
Did you download the model?
Try using a .wav file
Yes, I use an Ariana Grande model
ok, give me a sec
Try googling that error
Make sure the Model name has no spaces.
There should be a fix for that
perhaps the download speed is too slow and shuts it off. There isnt a clear method to fix this problem just assumptions
An intenal iissue prvented the driverfrom carry out the specified operation
Ariana Grande.pth could cause an issue.
So rename it to ArianaGrande.pth or Ariana_Grande.pth if that's the case.
The driver's state is probaly sus
And the Applicatin will not continue
what does that mean
I got an error
I got same error
dude`s If I doesnt make a index for model and have deleted the preprocessed dataset can I make index for that model? Or I need to make model again?
index training only needs feature extraction done
If you still have files made during 'Feature Extraction' in logs you should still be able to train Feature Index.
damn makes model again... thx for help
It's just 'process Data', 'Feature Extraction' and 'Train Feature Index'.
Shouldn't take long.
Hi. I tried to load model onto Llaria RVC, but everytime it says ERROR - The download failed. Check if the link is valid.
What do i do wrong? I tried link from huggingface, same
Downloaded from weights.gg, loaded model to my google drive and then put link to llaria, still same error
Ayo? @brittle wing level 1 !!! 
Got an Error again
I think you were a bit too fast.
I see no "all feature done" or something like that.
but colab says code is done
it could be something in your dataset was wrong, did u make it with capcut?
Hello im on macos and i can run the gui but there arent any visuals i can click and interact without seeing elements
I am not using an Ariana model, it's the app itself
Did you remove silence?
Sure!
Maybe the path to your audio file is wrong.
pls help me.
That does not look like a huggingface link....
i cant upload any images to here why?
and how do I turn it into a huggingface link?
You need to be level 1.
do you know anyone that can help me for macos related issues
Ayo? @cold tendon level 1 !!! 
i can interact with the gui but cant see
Audio path is right
mmm kinda weird then, what if u try to put it in audacity and export is as a wav here?
pls help-me
I had already entered the name but it does not recognize the drive link.
can you share the download link? i js wanna see if its private
thanks it's working now.
Ayo? @spiral osprey level 4 !!! 
RVC-Gui is deprecated
im glad i helped
how do you copy a model?
what can i use on macos
or atleast move them to a different slot?
Does it end in .wav or .mp3?
Retrieval-Based-Voice-Conversion-WebUI
thank you
wdym copy it? you trying to use a voice model?
no like
It's on mp3
k
They don't seem to have releases for mac though
let me tell you the story
Oh nvm there's a run.sh script
It's on Mp3
I used the merge lab in the voice ai software or whatever you want to call it
and I like it
and I want to make anotehr one
but
it always uses the same slot
It's on mp3
so
how do i
like
move the model I just made in merge lab
to a different slot
so I can make a new one
if it makes sense
u mean the program voice.ai for real time talk? im sorry i cant help with that, we use different programs for real time ai (free ones ofc) and i dont really know mucb for real time ai tbh
Mind uploading a full screenshot of RVC?
With everything filled in?
you can look some ai real time in #1159513888199540817
Realtime voice changer client is what im using
Ayo? @worldly gale level 1 !!! 
im sorry but i cant help you with voice.ai
aw dang
W-Okada?
yes
Well try asking in #🔍│help-w-okada
thx
That’s not voice.ai 💀
he said voice.ai before wym
“Real Time Voice Changer Client” is Okada
What they probably meant was they trained the model on voice.ai?
I mean a full screenshot of RVC, the webUI.
atleast from what i understoof he was talking abt voice.ai
how to tranform drive to huggingface?
if the model is yours(you made it) , https://huggingface.co/new you can upload it here on hugging face
how to upload?
first,did you make that model?
i make it
okay then js make an account on hugging face https://huggingface.co/new and upload it here
but he literally said abt voice ai before 😭 i got confused
and next?
after you made the model, give the model a name, go into files and click on the + and upload the .zip of your model files (a .zip of the .index and .pth from your model training)
okey, thhx
yw
After uploading my zip file what should I do below?
I have same exact error. I didn't make model tho, i just downloaded it and wanted to use
I can't figure out what do i do
https://rentry.org/fdg_guide_new
This explain how to add model to your huggingface account.
Step 2.
man said gui turned u
Ayo? @restive yacht level 1 !!! 
pls helpme
Try removing everything after .zip
i'ts don't working without .zip
Remove the "".
Copy the download link by right clicking on the download button next to the .zip in hugging face
i've done this.
The link should end with .zip.
So remove everything after it.
Did u duplicate Ilaria rvc
what?
Did u duplicate the Ilaria rvc space ?
nope I'm just closing the page and opening it again to see if it comes back, that is, it's like an f5.
you should use this link to duplicate the space so mabe it will fix as u wont have queue https://huggingface.co/spaces/TheStinger/Ilaria_RVC?duplicate=true
I created it but I didn't wait for it to load and I accidentally closed it. There's a problem now I can't duplicate the space anymore.
icl i think mangio is easier than ilaria
pls send the link
the course i sent you should have it linked in the description of one of the videos
infinite queue.
yeah pt 1
send in private.
RVC Guides (How to Make AI Cover)
Translation by country
How do you convert a weights.gg model to hugging face
put weights and index in a file then compress
infinite queue
js go on your profile and click again on the ilaria rvc space
its not converting them but if the download link is smt like models.weights.gg go look in pinned msgs
i don't understand.
And I cannot download the model
Ayo? @tawdry linden level 1 !!! 
It does not work
Please help me with python
I managed to download model from weights, but i can't make it work on Ilaria RVC. I get error after error and pinned guide didn't really helped me out, so yeah, same problems
i mean this msg #✨│ai-help message
I saw it. I still get that ERROR - The download failed. Check if the link is valid error
Ayo? @red badger level 2 !!! 
um guys in hina ai google collab why i can't add the weight model ai i was about to add with ai model but it dosen't work everytime appear the error
🇺🇸 English guide
Link: https://rentry.org/fdg_guide_new
Credits: FDG
🇮🇳 Hindi guide
Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes
🇹🇷 Turkish guide
Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes
heres how to upload
how do i fix this bruh
Weights links can't download
I loaded exact same model to google drive and linked it from there, same error
make sure everyone w link can see
Not sure about gdrive
it is, everyone can
if u try https://models.weights.gg/clm72mahk07epcctcd8subhdx.zip it could work maybe?
Scout TF2 (Team Fortress 2) | Here is the Scout 600 Epoch RVC 2 Model, I hope you enjoy it! https://huggingface.co/Sr-Nightmare/ScoutTF2/resolve/main/scoutTF2.zip | RVC Model, free download. RVCv2, SO-VITS-SVC, RVMPE, and more. Official AIHub Discord. | Fictional Character, RVC v2, English
That's exactly what i try to use
I just loaded it i guess finally (atleast ilaria says so) but now this happened
i downloaded zip and put it to google drive. it worked out few times later (i did exact same thing previously and it didn't worked for me, i got errors)
i still get them tho. sad life
like everythin seems fine?
try a different pitch extractor (in advanced options). id recommend dio
Did you duplicate the space?
Nope you didn’t
ooh wait they didn't
Top right there’s 3 dots
Click it
“Duplicate this space”
Set it to public and duplicate
You only have to do this once, wait for it to say running
Then try using it then
Yep
See where it says “building”
When it says running in green click the X middle right
It did, nice
I'll try to use voice again now
Make sure it’s a HuggingFace.co link for the voice
I can use only huggingface models?
I removed the " Apostrophe " and it's gives me the same eror
How can i use models from weights tho?
I got that error while loading again 😦
Ok
Weights models
Used to link to HuggingFace
But they started to host them on their own website
Download the model and upload it as your own
i need help the sound of this ai elvis voice is bit off everytime i make one it sounds like this
I cannot download any weights.gg model
@wispy burrow
What’s the page?
i wanna make it sound good
To the model?
Are you using RVMPE?
Yes
I'll dm you the model
Okaaay i'll try that
Is there any guide about how to load model to huggingface?
@wispy burrow @molten pecan
-upload
Suggestion for @red badger
🇺🇸 English guide
Link: https://rentry.org/fdg_guide_new
Credits: FDG
🇮🇳 Hindi guide
Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes
🇹🇷 Turkish guide
Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes
Go to step 2
@violet heron the model yes i used it with rmvpe but it sounds better but still worse
but it was trainned with voicemy.ai
Ignore step 3
Alan are you using the beta one?
Try a different elvis model
it doesn't give me the index file, is this normal?, it just gives me the pth file I'm using:
RVC v2 Disconnected, can you recommend a good internal colab or promt for train model? Thanks
the one i have is better but i wanna make this 77 elvis model sound good
is there way
No, I'm using the original RVC intel version
I Should use the beta version
Yeah
I mean that's what I would do if I was you
umm
and No, I'm using the original RVC intel Version
it doesn't give me the index file, is this normal?, it just gives me the pth file I'm using:
RVC v2 Disconnected, can you recommend a good internal colab or promt for train model? Thanks
It's not duplicated
?
How to duplicate it then?
Ayo? @red badger level 3 !!! 
On the three dots in the upper right corner
what kind of pitch does kits ai trains there models with
What doesn't give you the index file?
No clue
Try asking on #📑│making-models
yes, don't give it to me, maybe you're wrong to use it? I'll send screen shots of what I do, wait
Sure
doesn't it mean i duplicated it already?
it means i should do that everytime i want to use voice?
the entry works but only with the pth file
see
Go to your profile
Click on the circle on the upper right corner
And click on your username
Your duplicated spaces are there
-infer
You don't have to duplicate them each time you want to use it, don't worry
-svc
Yeah I still use so-vits-SVC
I don't see it
Send sc
???
There it is
Click on it
The pink rectangle
That's the duplicated space
Use that space instead. That one should work
Still same error
I got it now, tho
Change blob to resolve
Blob?
text in the link
how can i make an zi;p file url?
-upload Step 2 only
Suggestion for @sinful tree
🇺🇸 English guide
Link: https://rentry.org/fdg_guide_new
Credits: FDG
🇮🇳 Hindi guide
Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes
🇹🇷 Turkish guide
Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes
Name it Test
Old what?
Old model, i mean
That one named scout
Will delete automatically after 2 days
what does Pretrained weights not downloaded. Disabling training tab. mean
Try asking in #📑│making-models
ok
Ayo? @distant turtle level 3 !!! 
Does anyone know how to open rvc ? 😭
Ayo? @vestal crescent level 1 !!! 
Are you using Applio or Gradio?
On the file folder ?
Aight so i'm back with questions
I have infinite converting there
I tried restart few times with no result
how long is the file
05:02
ooh that'll take 500-700 secs
oooooooh i see i restarted around 400-500
Yeah no that takes me far more than that 😭
Oh I never use dio
run the training variables step
use it
it's faster than rvmpe
OHH thanks
np
YESS it works now
Yeah but I prefer quality to be honest
yo @proper shale
what frequency should i set
in training
cause it's said 25 or 50
set whichever one u want
Does it matter?
its up for you to decide
and total_epochs?
not much, plus if u restart training later u can change
oh okay
set to 500
yh
keep the page running
so Im gonna do it tomorrow i guess
yes
How long?
based on your experience what is your approximate time for training a model?
hmm
does the batch size matter?
yes
it is 30 megabytes
well for me it's already over 1000 sec
@julia Can you give me the link to RVC Beta?
so like what are the most used batch sizes?
I have it set 8 on default
can it take more than 1 hour in my case?
@molten pecan Can you give me the link to RVC Beta?
2 mins 50 sec audio 8 batch size 25 frequency
yeah
wym by save it
like the google fields
id suggest training for 10 mins
hmm
So I can like stop the trainning in any moment?
and then coming back tomorrow wh u have more time
oh so I can resume it???
yeah sorta
just dont stop it on saving epochs
and saving epochs take long?
Ayo? @boreal valve level 3 !!! 
and its gonna be alright
i can set it on like 1 hour today
yeah
damn its more advanced than i thought
and do you possibly know what does the frequency do?
like 25 and 50
is it like 50 = better quality or smth?
ohh
and ive got 500 epochs
so i was telling u to be careful when it hits one of those epochs when it saves
because ur training files can screw up
after ur done with training for today u click on export model and it'll be on google drive
what should I do to make the training the best quality ?
okay and then when i come back tomorrow what should i do?
have a good dataset
run the resuming steps and set the same settings and all
np ;)
but like wait
So now I train for hour, I save model then I open the page tomorrow
guys i have a probleme someone can help me ?
And like, the google page will be fresh ye?
it'll be new yeah
so like
What is it?
how may I load those unfinished models
what's up?
same training variables
and uh
same settings in the train step
thats about it
im new here im trying to download the voice changer
i already dowload it but the probleme is that instantly crash
or feature extract again
check pins in #🔍│help-w-okada
yeah
anyone have a guide on this: Mangio-RVC-Fork HF, how do I insert the audio file for train model?
so the second option, not the first one?
wait what
like
you said dont preprocess
oh wiat
ohhh okay
nevermind
I do both those steps ye
yeah
preprocess was before
just the resuming steps
yee thank you so much
np
anyone have a guide on this: Mangio-RVC-Fork HF, how do I insert the audio file for train model?
ı need a voice models but ı dont know how can ı get voice models
you cant train models on huggingface
Website weights.gg
The largest collection of free RVC Voice AI Models to download. RVCv2, SO-VITS-SVC, RVMPE, and more. Official AIHub Discord.
is applio
well I actually want a model of myself haha
so like I can record as many audio as I want
but like im not sure what kind of audio should I record in order for the AI to make the best model
tell me another functional site to find models, it's fine not a site but an internal prompt or program, with the site: RVC v2 Disconnected, it doesn't give me the index file, it only gives me the pth (working)
because you didn't train the index
id recommend just reading stuff
do a bit of expressions if u want to
maybe a bit of singing
In what sense?
...you didn't train the index. As in... you didn't run the step to train the index.
yeah like basically I just put audio of me singing
yeah, but you need to make sure to trim silences and noise gate in audacity
I need to trim silences?
yeah
itll screw up training
u can use the trim silence effect
if ur using a recent version its under the special category in effects
These are the steps I took!
theres literally a step named "Index training" right below export model
I see! Thanks
and -35db threshold
np!
Через какую программу вы делаете AI кавера?/
What program do you use to make AI covers?
yh
Спасибо бро
/
Thank you bro
ok thanks, I hadn't noticed, the settings I set are fine, I took a 3 minute audio but I'm making a cleaner 10 minute one, are the settings I used in the photo ok?
No problem!
Без проблем!
Yeah
train epoch 68 means training
but when it says saving then i should not stop ye?
when its saving d/g files yeah
if I use a longer audio file like 10 minutes (very clean audio), is it better than 3 minutes (very clean audio)?, total-epochs, should I increase it from 300 to 550 for 10 minutes of video?
longer clean audio is better
keep it at 500, but you dont need to use epoch 500 necessarily
Thanks so much for the help!
np
in llaria rvc i download model with google drive refresh the model but theres nothing there so should i do for the huggieface link
hey guys can anyone help me?
yes
what's up?
so when I try to download a model or make a cover its just stuck
and the cmd isn't showing anything
nothing
man I'd recommend trying the normal mangio gui
easygui is apparently broken locally
okay I'll try
oof
the feature index file is the .pth?
so how do i google drive model turn to huggieface model link? im curious
is just the index
oh okay so I don't need the .pth?
alright
because i need to know how to add model google drive model to huggiface model it's saids new model
🇺🇸 English guide
Link: https://rentry.org/fdg_guide_new
Credits: FDG
🇮🇳 Hindi guide
Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes
🇹🇷 Turkish guide
Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes
guys is there a tutorial to the local mangio-rvc-v23.7.0?
thank you!
np
i need some help !!
Ayo? @graceful owl level 1 !!! 
pls guys
ok i see if i can try
im tryna lunch the ivoice generator but it instantly closed
and this message appears
dunno if someone can help me
Ew opera
haha leave me alone
is it rlly a problem?
ill see
Might not fix the one issue you are having, but it allows it to work
it doesnt work
Does it auto close ?
If it doesn't auto close, you should be able to use it at http://localhost:18888
in any browser
no i closed it myself
Keep it alive, discard the log and try via any other browser
For some reason the stripped down browser they ship sometimes just refuses to work
Ah fuck
Some people fixed it by putting it on the C: drive
Not sure why it does that, but it just does
gn
what do you mean ?
bro it worked !!!
Ayo? @graceful owl level 2 !!! 
I think its broken I just clicked on refresh voice list index path and audio files
and this is not the easy gui
rvc creepe v2 in step 2 red it didn't work on drive model
Ayo? @fiery shadow level 3 !!! 
What is it then?
Colab?
If so send the colab link
its just go web
Ok
Is it Mangio?
And did you try to install EasyGUI?
yes I installed it and tried it
it got stuck as well
That's the problem then.
Installing EasyGUI breaks Mangio.
ooh
is there a way to uninstall easygui?
Best to delete the entire RVC folder and extract the zip again.
Back up your models if you have any.
okay, thank you so much, I'll do that. I don't have any models yet, but I'll make some.
a model creation takes like and hour and half?
That depends.
What GPU you have, how long the Dataset and the Batch Size you use all affect how long training will take.
My last Model took 14 hours...
hey i'm using a RX6600 so need to use onx model how can i convert them?
Alright, thanks for the info.
:O
I have an Nividia GTX 1060 6GB...
So yeah...training is slow on my end...
Can't even go higher than batch size of 2 or I have OutOfMemoryError...
shoot that sounds sad friend :(
Ayo? @quaint dock level 4 !!! 
But apparently a low batch size is better for the quality-wise.
A higher value makes it faster faster to train but will be less accurate...
As it uses less of the Dataset per Epoch to train on...
But yeah...it hurts but at least I can train...
yea at least, and thank you for the tips!
Looks like W-Okada...
You might be better of asking in #🔍│help-w-okada
btw if I train a model its private, right?
What do you mean?
I meant like I train a model on local mangio
Okay, thank you
are there any reuploads of community models what were on kits.ai?
or like a website of rvc models?
Kits never allowed to download the models so those need to be created again (but not with kits).
Any colabs that Google didn't fuck up?
alright, I think I'll do that
thats why I downloaded local, cause It kept stopping randomly
I don't have space
Blame Stable Diffusion lol
Don’t use kits.ai
my antivirus keeps blocking it idk why , but thanks
okay
Local is always the best way to go
Also
If you want models try using /search with @shy zealot in #🔍│find-models
I don't have the space