wait
#✨│ai-help
finite galleon
what do i do if my model doesn't have a linl
Create one
What whats the command for the guide to uploading to HuggingFace again
-huggingface
azure marshBOT
finite galleon
There you go
pale spindle
ok thank you very much
sudden tartan
i downloaded rvc mangio, so where is the file that has the voice converter?
Im really confused 😦
hollow kettle
is there a working rvc for amd?
proper shale
i downloaded rvc mangio, so where is the file that has the voice converter?
just open go-web.bat
sudden tartan
just open go-web.bat
it worked, but where do you put the models?
proper shale
pth in weights, make a folder in logs with your model's name for the index
rare gobletBOT
Ayo? @sudden tartan level 3 !!! 
sand iris
is there a working rvc for amd?
For inference? Or making models ?
sudden tartan
nvm i figuered it out
hollow kettle
For inference? Or making models ?
for making models
rare gobletBOT
Ayo? @hollow kettle level 1 !!!
hollow kettle
Will it be possible to make models with Linux?
sudden tartan
@proper shale I’m getting a error when converting the voice
sand iris
Will it be possible to make models with Linux?
Yes, on any AMD GPUs newer than the rx 500 cards
...Or you can use google colab for training
proper shale
<@622856015444049937> I’m getting a error when converting the voice
something wasn't selected properly
sudden tartan
what wasnt selected?
hollow kettle
...Or you can use google colab for training
when I use rvc google colab I click train model and just endless loading
hollow kettle
when I use rvc google colab I click train model and just endless loading
How to fix it?
brittle wing
anyone here?
odd shale
<@622856015444049937> I’m getting a error when converting the voice
You did make sure you selected the .pth file?
sudden tartan
Yes
proper shale
Windows?
i just think it could be the path of the audio, maybe it's not selected properly
sudden tartan
how could you tell?
odd shale
i just think it could be the path of the audio, maybe it's not selected properly
the audio file must go on the audios folder of Mangio Fork root folder, right?
proper shale
how could you tell?
because that's just how most people get errors
hot lagoon
I need help finding the e girl modle
glad zealot
-egirl
glad zealot
I have an real E-Girl model tho <:nails:1159569314848972891>
i have me too :D
odd shale
i have me too :D
I bought one from Ilaria.
sudden tartan
ok i put the audio in the audio path but still getting a error
finite galleon
I bought one from Ilaria.
Sell it back to every other person that wants an e girl model
Easy profit fr
odd shale
Easy profit fr
That's what i'm planning to do.
There are a ton of randoms that are thirsty for an Egirl model.
sand iris
ok i put the audio in the audio path but still getting a error
what's the audio path
sudden tartan
what's the audio path
molten pecan
ok i put the audio in the audio path but still getting a error
Try removing the spaces of the audio file name
Instead of "bandicam 2023" rename it "bandicam2023"
snow wedge
which colab or site can be used for training?#
molten pecan
@snow wedge
snow wedge
thanks!
rare gobletBOT
Ayo? @snow wedge level 1 !!!
sudden tartan
Try removing the spaces of the audio file name
I still got an error.
molten pecan
I still got an error.
Did you try with a different model and audio?
Maybe one of them is bugged
sudden tartan
Im using the squidward model
Ill try a different audio
Tried a diffrent audio and still not working
Tf is even going on
molten pecan
Tried a diffrent audio and still not working
Tf is even going on
Did you try with a different model?
molten pecan
modern nest
where do i find the index file i just trained
sudden tartan
I literally did that
molten pecan
where do i find the index file i just trained
Did you use Colab?
violet heron
I literally did that
Locally or online method?
rigid tree
Any XXXtentacion models or juice or notti Osama
sudden tartan
Locally or online method?
Well its on chrome
rare gobletBOT
Ayo? @sudden tartan level 4 !!!
molten pecan
Any XXXtentacion models or juice or notti Osama
Did you check on weights.gg?
modern nest
Is it on the weights or logs folder?
the pth is in the weights
tho i havent checked logs
finite galleon
Well its on chrome
Yeah but what is the web ui
rigid tree
Did you check on weights.gg?
Yea but I downloaded models and it don’t sound like them
molten pecan
Yea but I downloaded models and it don’t sound like them
Did you use high quality sample?
Or just an mp3?
rigid tree
Did you use high quality sample?
I used a mp3
molten pecan
I used a mp3
Check dms
modern nest
I used a mp3
JAIL
violet heron
Well its on chrome
Did you open a .bat file
modern nest
USE FLACS
sudden tartan
Did you open a .bat file
Yes
violet heron
Did you close the terminal?
sudden tartan
Did you close the terminal?
Yes
storm dune
I need help :c
oblique rain
I need help :c
what's wrong
storm dune
When I start the voice changer, I get an error
oblique rain
When I start the voice changer, I get an error
for vc help, go here #🔍│help-w-okada
storm dune
okey, thanks
sudden tartan
Unless you want to stop using RVC
Do I have to delette rvc gui then
hollow kettle
rvc google colab works on all video cards?
rare gobletBOT
Ayo? @hollow kettle level 2 !!!
proper shale
rvc google colab works on all video cards?
yea cuz its a cloud machine
hollow kettle
yea cuz its a cloud machine
can you upload a working version of rvc to google colab please?
proper shale
-colab
violet heron
Do I have to delette rvc gui then
Close the tab and reopen the bat file
Do not close the terminal unless your done with RVC
hollow kettle
does easygui have tutorials?
molten pecan
does easygui have tutorials?
https://rentry.org/ilarvc_inf_guide here's my guide :b
hollow kettle
thx
sudden tartan
Close the tab and reopen the bat file
But I’ve done that
violet heron
But I’ve done that
Did you close the terminal
sudden tartan
Yes
violet heron
I still got a error
How did you get the wav file?
Did you just rename the mp3 or did you actually convert it
sudden tartan
I renamed the file so that could be the reason why
violet heron
I renamed the file so that could be the reason why
💀
That’s not how it works buddy
Convert the file
Not rename it
sudden tartan
I know im not new to the new rvc fork lol
sturdy ridge
Is the so-vits-svc 4.0 colab not compatible with RVC models? It's not working with mine that I downloaded from Huggingface & the configs seem to be pretty different
violet heron
Is the so-vits-svc 4.0 colab not compatible with RVC models? It's not working wi...
-svc
sturdy ridge
Lol haven't done anything related to this in about a year so Im a little out of the loop
rare gobletBOT
Ayo? @sturdy ridge level 1 !!!
molten pecan
Is the so-vits-svc 4.0 colab not compatible with RVC models? It's not working wi...
We don't offer support to svc since it's very outdated. Use RVC instead
sturdy ridge
Thanks!
sudden tartan
Wdym
sudden tartan
violet heron
Wait try dragging the contents of the infer_train folder to just the normal one
Replace if you need to
sudden tartan
Isnt the folder called train?
meager terrace
Hey, using Easy👨- 10/25
And I get this problem:
It stopped training for no apparent reason and when i press for it to restart training I get this
What could be the issue fam?
hoary estuary
hello, what is the best ai to create natural voice please ?
meager terrace
hello, what is the best ai to create natural voice please ?
-colab
hoary estuary
wich of them ?
violet heron
wich of them ?
RVC Disconnected for training a new voice
sturdy ridge
Does anyone know if my link in the wrong format or is there some other problem here?
violet heron
Does anyone know if my link in the wrong format or is there some other problem h...
Try a different model
meager terrace
Does anyone know if my link in the wrong format or is there some other problem h...
Maybe you already downloaded it
meager terrace
Hey, using Easy👨- 10/25
And I get this problem:
It stopped training for no appa...
@violet heron any ideas on this?
sturdy ridge
Try a different model
Got the same issue with a couple different models
sturdy ridge
Maybe you already downloaded it
It's not showing up in "Choose Your Model" so I don't think it did yet
meager terrace
It's not showing up in "Choose Your Model" so I don't think it did yet
Try to refresh/Make sure you duplicated the environment
violet heron
<@858923165379199007> any ideas on this?
Not really
Always had problems with easy👨
Try RVC Disconnected if your training a model
meager terrace
Try RVC Disconnected if your training a model
Sounded robotic + someone told me its broke somehow
So its not broken?
violet heron
Sounded robotic + someone told me its broke somehow
On RVC Disconnected?
meager terrace
On RVC Disconnected?
yup
meager terrace
Lower batch size to 4
You can't even chose batch size bro
sturdy ridge
Try to refresh/Make sure you duplicated the environment
Still nothing- also tried downloading it to Google Drive & using that link instead
Is this showing that it did in fact download?
meager terrace
Still nothing- also tried downloading it to Google Drive & using that link inste...
You duplicated it?
meager terrace
Still nothing- also tried downloading it to Google Drive & using that link inste...
Try with à different model
brittle wing
This thread needs to become more active
strong spindle
whats the most updated way to make a model because none have worked for me
violet heron
whats the most updated way to make a model because none have worked for me
RVC Disconnected
molten pecan
Is this showing that it did in fact download?
Yes
violet heron
-colab
strong spindle
RVC Disconnected
oh fr??
rare gobletBOT
Ayo? @strong spindle level 1 !!!
violet heron
any other way thats not a subscription
It’s free?
strong spindle
ik
but some arent
if theres another way to make a model thats not rvc bcs its down
violet heron
if theres another way to make a model thats not rvc bcs its down
RVC is not down?
There is no way for it to go “down”
strong spindle
you just said its disconected
brittle wing
hey, someone can teach me how I delete "model slots" on voices?
sudden tartan
@violet heron I've come back and used a wav file and its now saying " connection errored out "
meager terrace
Can't resume training. Easy Gui
Any thoughts?
meager terrace
Is this a tesla t4 runtime ?
I have no idea, I just put the basic info in the RVC and changed no setting
craggy brook
what is this
violet heron
what is this
You need to get python 3.9
languid viper
Hello everyone
How to download cover when it is generated with CoverGen?
What should I do with audio path?
slate pilot
hi guys
rare gobletBOT
Ayo? @languid viper level 1 !!!
slate pilot
i oppened go-applio
and it downloading files about 2 hours
pls help me
what should i need to do?
violet heron
Not done yet what it looks like
It should not take 2 hours
Unless your ISP limits you
slate pilot
i need just wait?
rare gobletBOT
Ayo? @slate pilot level 2 !!!
white wren
how do I get infer-web.py?
rare gobletBOT
Ayo? @white wren level 1 !!!
white wren
lethal crown
use RVC and because Mango might be the problem. because I used Mango and all the models I trained sounded the same. but RVC is working for me, I got mine from this video. good luck
violet heron
use RVC and because Mango might be the problem. because I used Mango and all the...
Mango
lethal crown
Mango
? why you just write Mango? lol I spelled Mango right. you may have miss read. anyway was saying Mango RVC 23 7.0 didn't work for me to train, so I switched to to main RVC. works great.
violet heron
? why you just write Mango? lol I spelled Mango right. you may have miss read. a...
It’s Mangio
Not mango
lethal crown
It’s Mangio
😂 so it is lol oh well I stopped using mango
rare gobletBOT
Ayo? @lethal crown level 3 !!!
violet heron
😂 so it is lol oh well I stopped using mango
YOU DID IT AGAIN
What’s next? Your going to call applio “apple”
lethal crown
YOU DID IT AGAIN
😂 IDC but for your peice of mind. Mangio
grave swallow
Hey, i use collab AICoverGenWebUI (first link)
But when I generate my audio, the original voice can still be heard behind the AI Voice, how comes?
violet heron
like old way...where we have to paste model link?
Locally or online
static garden
Locally or online
locally
azure marshBOT
Suggestions for @static garden
foggy cedar
-colab
sudden tartan
Ilaria RVC is alright so far, but what huggingface link do I use for the model to download it?
rare gobletBOT
Ayo? @sudden tartan level 5 !!!
sudden tartan
its saying that the models been downloaded to the infetence, but when I look in there its not there
tame mica
its saying that the models been downloaded to the infetence, but when I look in ...
refresh
sudden tartan
then refresh your web
I did that too
full imp
Please send the most noob friendly guide to train models for rvc, I use the huggingface interface to make the songs if that matters
brittle wing
ModuleNotFoundError: No module named 'ffmpeg' how do i fix this
rare gobletBOT
Ayo? @brittle wing level 1 !!!
cloud finch
is there a file limit on ilaria rvc?
tame mica
is there a file limit on ilaria rvc?
i don't think so..?
white wren
im not
violet heron
I did a 122mb file and
How long is it?
cloud finch
around 12 minutes
white wren
oop
molten pecan
Please send the most noob friendly guide to train models for rvc, I use the hugg...
Check the RVC Disconnected guide on #1159513888199540817
molten pecan
Thats normal
The amount of time depends on how long the audio is, the algorithm you are using, and if you duplicated the space
cloud finch
Use colab brotha
I thought you could only use colab if you paid or something
rare gobletBOT
Ayo? @cloud finch level 2 !!!
finite galleon
I thought you could only use colab if you paid or something
Thats for web uis
finite galleon
So what should i use then
Its a lil complicated as it uses the google colab forms
But the process is the same as web uis colabs (install, download model by link, and inference)
finite galleon
Yeah
cloud finch
bet
finite galleon
Thats the first thing you should do
violet heron
should i install harmonify?
You have to
finite galleon
And then do the other stuff in order
finite galleon
tf does it say optinal then
Mount drive is optional
cloud finch
ohhh
finite galleon
You can turn it off if you dont want it
violet heron
Installing it is not
cloud finch
what does it do?
finite galleon
what does it do?
Makes everything work
finite galleon
Ohh
It basically mounts your google drive into the colab
Some people have slow upload speeds to colab but not drive for some reasons 
cloud finch
do i have to install harmonify every time I do this?
violet heron
do i have to install harmonify every time I do this?
Only once per runtime
cloud finch
so everytime i open the tab i have to do it correct?
like if i close and reopen it
i have to reinstall it?
violet heron
i have to reinstall it?
Unless you resume your runtime or mount your drive, yeah
cloud finch
i mounted it to my drive
to upload audio would i first upload it to my google drive?
violet heron
to upload audio would i first upload it to my google drive?
Yes
violet heron
im getting that error
What did you put in
cloud finch
cloud finch
Yeah
rare gobletBOT
Ayo? @cloud finch level 3 !!!
cloud finch
I didn't know there was a way to obtain google drive file paths though
where do i get the file path?
finite galleon
I didn't know there was a way to obtain google drive file paths though
Press the colab file panel
Clcik on the folder named "drive" and then "MyDrive"
From there search your audio
Once done click on the file and there will be three dots on the right side of the file, click it and press the copy path button
Once that done you can paste the path on SOUND_PATH
finite galleon
there we go
If you wanna download the results do the same steps but open the harmonify folder and output folder press three dots on the audio and click download or simply turn on the DOWNLOAD option in the OUTPUT SETTINGS
violet yoke
I'm using RVC WebUI to train but I'm failing at the second step which is processing the data. I give it the path to the audio files and click process data and nothing is happening on the console and on the web ui it shows "queue: 1/1 | X seconds". X is now 10k which is almost 3 hours now...
What am I doing wrong?
cloud finch
If you wanna download the results do the same steps but open the harmonify folde...
Could i use colab for non singing things?
like people talking
or voice acting i mean ig
finite galleon
like people talking
Ye
cloud finch
using the same method?
cloud finch
bet
clear temple
rancid pine
Is there any chance EasyGUI will return?
pseudo steeple
Is there any chance EasyGUI will return?
Never lol, just learn the normal GUI
violet yoke
I'm using RVC WebUI to train but I'm failing at the second step which is process...
Anybody?
rare gobletBOT
Ayo? @violet yoke level 1 !!!
pseudo steeple
I'm using RVC WebUI to train but I'm failing at the second step which is process...
Is tensorboard up?
might just be updating on where the progress is
violet yoke
What is tensorboard?
brittle wing
how can I upload of locally trained RVC voice models to use in the colab for ai covers?
pseudo steeple
how can I upload of locally trained RVC voice models to use in the colab for ai ...
I believe you have to import the model into the folder where you place the models, then try to select it
violet yoke
I was following this guide https://docs.aihub.wtf/guide-to-create-a-model/model-training-rvc
I don't see tensorboard here
pseudo steeple
violet yoke
I didn't get to that part yet, I'm still here
pseudo steeple
I didn't get to that part yet, I'm still here
specify the folder's direct item
So like /content/dataset/Xiao
violet yoke
That's what it looks like
pseudo steeple
That's what it looks like
well
worst case scenario
just copy the folder location
from file explorer
violet yoke
That's what I did
And it says queue 1/1
And nothing is happening
On youtube people get the files instantly
pseudo steeple
perhaps the cpu extraction is too high?
Turn it down to 2
doesn't really matter tbh
violet yoke
I did that too
violet yoke
It doesn't even matter what I put in the path, it writes queue no matter what
light silo
I'm trying to use Applio to download an .pth and .index file but it's not working, I downloaded the folder and are trying to upload it, but no matter what I try, the index file will NOT show.
violet heron
I'm trying to use Applio to download an .pth and .index file but it's not workin...
Index file goes in logs
violet heron
\logs\name\added_probe (etc etc)
Try not putting it in a folder
violet heron
whats the clearest model for SPEECH
Model or f0?
violet heron
RVMPE
light silo
not +?
light silo
mk
rare gobletBOT
Ayo? @light silo level 5 !!!
light silo
also btw
how do you download zip from weights.gg onto applio
because it says download failed every time i try
pseudo steeple
What's an efficient way to make them go back to back?
Align tracks end to end doesn't work
since they're all in a track and not considered individually as tracks
molten pecan
because it says download failed every time i try
Are you logged in weights.gg?
light silo
Are you logged in weights.gg?
i mean using the applio downloader
molten pecan
i mean using the applio downloader
Are you using it locally?
light silo
yes
molten pecan
Or Colab or HF?
Did you try manually uploading the index and pth file into the folders?
light silo
Did you try manually uploading the index and pth file into the folders?
yes
but its not important anymore
pseudo steeple
Theory question about batch sizes.
Assumedly, a model with a larger dataset is more likely to be higher quality if it uses a lower batch size as if you use a higher batch size it would just use an example you have given it.
example:
a dataset with a 90 minute length (assume it's cut to perfection) will more likely be higher quality if set at batch size 4 compared to 8
finite galleon
Did you install all the requirements on requirements.tct file
slate pilot
fairseq doesnt downloading
burnt latch
Am I able to find voice models with rather searching for models via VAs?
junior vine
I have just watched a YouTube video about a tutorial for creating an ai voice
I have a project for school and I have basically rewritten the first verse of “The Man” by Taylor Swift. Now I want to change the lyrics for the first verse with her voice into what I have written, but I have not really an idea of how I’m going to do that
I really hope that makes sense
true lynx
Please I need help ! I uploaded a voice on Voice.ai but it's just stuck in the processing part and I dunno what to do >_<
rare gobletBOT
Ayo? @true lynx level 1 !!!
brittle wing
whats the link to google collab please anyone
sudden tartan
Im still getting an error when I use Ilaria RVC like wtf is going on?!#
every rvc i use theres always an error with the audio
like ffs
copper pebble
whats the link to google collab please anyone
inference or training
copper pebble
Im still getting an error when I use Ilaria RVC like wtf is going on?!#
did u duplicate the space??
slate pawn
Yo is there anyone that could help in fix an index file not working?
limber fractal
Yo is there anyone that could help in fix an index file not working?
im currently having the same issue
azure marshBOT
rotund zephyr
guys what does epoch mean?
exotic ember
What to do?
past sentinel
Does anyone know how to fix this?
thick otter
rvc web UI
Traceback (most recent call last):
File "E:\rvc\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\infer-web.py", line 292, in vc_single
audio = load_audio(input_audio_path0, 16000, DoFormant, Quefrency, Timbre)
File "E:\rvc\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\my_utils.py", line 103, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: ffmpeg error (see stderr output for detail)
Traceback (most recent call last):
File "E:\rvc\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\routes.py", line 437, in run_predict
output = await app.get_blocks().process_api(
File "E:\rvc\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1349, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "E:\rvc\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1283, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "E:\rvc\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 2586, in postprocess
file_path = self.audio_to_temp_file(
File "E:\rvc\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 360, in audio_to_temp_file
temp_dir = Path(dir) / self.hash_bytes(data.tobytes())
AttributeError: 'NoneType' object has no attribute 'tobytes'
loading rmvpe model
how to fix this
pliant turtle
when using the model there are strange noises coming from the back. like a humming sound. how can I fix it?
proper shale
try increasing s thresh and using sup2 if you're using W-Okada
proper shale
rvc web UI
```
Traceback (most recent call last):
File "E:\rvc\Mangio-RVC-v2...
something didn't seem to load properly
brittle wing
does anyone know why the download speed is so slow even though my wifi speed isnt that slow
glad zealot
the server where the download is coming from
even if you have the fastest internet in the world but if the file you are downloading are on a slow server/host its still gonna take forever
spice thorn
Hi there, can anyone help me with ai cover?
Colab not working and my pc is too low
I have model and vocals of song
rare gobletBOT
Ayo? @spice thorn level 2 !!!
glad zealot
Hi there, can anyone help me with ai cover?
Colab not working and my pc is too...
wdym by colab not working? how exactly is it not working?
spice thorn
wdym by colab not working? how exactly is it not working?
If you use the free version, google just turns it off
glad zealot
If you use the free version, google just turns it off
are you using the modified version or the original?
spice thorn
wym
glad zealot
this is the modified version
that shouldnt get turned off by google
unless you went afk for too long or reached 4 hours-ish
glad zealot
oke
copper pebble
do yall know an ez programm/site where i can get a waveform from an audio/video?
glad zealot
audacity
molten pecan
do yall know an ez programm/site where i can get a waveform from an audio/video?
Audacity is fine
copper pebble
wo i didnt know it could do that
copper pebble
Audacity is fine
oh i think u missunderstood
https://youtu.be/yHfayijc8ok?si=HcPGmV4uOQHAvyXb wanted to do. something like this
HAHAHA TIMBO I LOVE YOU
violet heron
https://youtu.be/yHfayijc8ok?si=HcPGmV4uOQHAvyXb wanted to do. something like th...
Adobe After Effects
finite galleon
Now how to do it on mobile 
glad zealot
https://youtu.be/yHfayijc8ok?si=HcPGmV4uOQHAvyXb wanted to do. something like th...
ahh no idea
ngl that would be nice for ai cover so its not just a still image
slow vortex
I've been looking for some singers
Like ahmet kaya or hayede
I found some voice models but they don't really work well
How can I make my own voice model ? And how can I make it better so their voice doesn't crack?
copper pebble
I've been looking for some singers
Like ahmet kaya or hayede
I found some voice ...
well there are many ways to make ur own model
- 1st: u can do it locally if ya got nice gpu and enough vram if y'aint that just go for colab
i would suggest the rvc disconnected one
molten pecan
https://youtu.be/yHfayijc8ok?si=HcPGmV4uOQHAvyXb wanted to do. something like th...
Ohhh that
copper pebble
while to get good dataset u just need to use a good audio extractor with the right models (UVR 5)
@slow vortex
molten pecan
That's with a video editing software, I think
copper pebble
That's with a video editing software, I think
yh Ae
molten pecan
What
copper pebble
will do that tomorrow
copper pebble
yh
copper pebble
Can you do me a favor?
yh
molten pecan
yh
https://rentry.org/ilarvc_inf_guide anything else you think I should add?
I need community feedback
Squishy will add images btw :b
copper pebble
lemme see
molten pecan
thats really good and easy to follow
Oh awesome!
Good to know ^-^
Anything else you think there should be?
copper pebble
Anything else you think there should be?
tbh its perfect to me
u really did a nice job
the images will complete it
violet yoke
How do I get Tensorboard? Is there a download anywhere?
molten pecan
How do I get Tensorboard? Is there a download anywhere?
Fuck I should make a guide about that too
molten pecan
How do I get Tensorboard? Is there a download anywhere?
@copper pebble help em
copper pebble
Fuck I should make a guide about that too
theres one
violet yoke
Well I was here
rare gobletBOT
Ayo? @violet yoke level 2 !!!
copper pebble
yh that one
violet yoke
But there is strikethrough on AOD and literally the next thing after getting python is running the install-requirements.bat which is where exactly?
slow vortex
* 1st: u can do it locally if ya got nice gpu and enough vram if y'aint that jus...
How can I do it locally? I got gpu
copper pebble
How can I do it locally? I got gpu
check here #1159513888199540817 all in one documentation
slow vortex
while to get good dataset u just need to use a good audio extractor with the rig...
Yea but they are both singers and both dead so most of the audio I can find of them is in old songs
slow vortex
check here <#1159513888199540817> all in one documentation
Thanks alot
violet yoke
So how do I get that Tensorboard?
rotund ridge
is there a way to add some emotions into the ai audio? I use TTS on Applio but, it sounds natural but there is a lack of emotions in the voice
violet heron
is there a way to add some emotions into the ai audio? I use TTS on Applio but, ...
You can convert your own voice
ornate ice
how to install and use rvc?
violet heron
how to install and use rvc?
By downloading it
Then extracting it and running the bat file
candid jetty
😦
wide ingot
On Paperspace, running Applio may result in a "No module named 'lxml.etree'" error. To fix, type this command in the terminal:
pip install --upgrade lxml
sand iris
On Paperspace, running Applio may result in a "No module named 'lxml.etree'" err...
@proven hill
Good call on that
thx for reporting
wide ingot
Yea, I had that issue when just doing a pip install requirements.txt. The fix from above made it disappear.
proven hill
Yea, I had that issue when just doing a pip install requirements.txt. The fix fr...
thank you so much for the heads up
sudden tartan
what does this mean
violet heron
what does this mean
What are you using???????(
tawdry scroll
kits ai LOL
sudden tartan
ik kits is a dogshit ai
violet heron
Everything is paywalled, only 15 minutes of conversion
violet heron
You can’t even download voice models you made
violet heron
but its the only one
Use “RVC Disconnected” or voicemy ai if you need something easy
sudden tartan
ok
jesus kit ai fell off hard
the community was the only good thing
idk why they took it down
sudden tartan
Kits Ai wants to make money and they're probably gonna lose it
violet heron
Kits Ai wants to make money and they're probably gonna lose it
As if they didn’t already
sudden tartan
voicemy.ai is taking a while to clone the voice but hopefully it works
violet heron
voicemy.ai is taking a while to clone the voice but hopefully it works
All methods do
sudden tartan
voicemy.ai is saying that the file is too short but I uploaded a 3 minute file
and now ilariarvc isnt working
im so pissed off
violet heron
voicemy.ai is saying that the file is too short but I uploaded a 3 minute file
What format?
sudden tartan
mp3
sudden tartan
yes
pulsar timber
For RVC, has anyone had this issue?
The first was because the file location was too long? I shortened it and a new error appeared
I'm on windows and downloaded the model _infer_train from https://github.com/Mangio621/Mangio-RVC-Fork/releases/tag/v23.7.0
proper shale
For RVC, has anyone had this issue?
Has a space on the name
"AI Tools"
no wait
"AI Voice"
If you remove that you should be good to go
pulsar timber
That's worked! Thank you 😄
violet heron
for nba youngboy
RVC
violet heron
do u need a laptop
What device are you on?
violet heron
Create a HuggingFace account THEN go to this link
https://huggingface.co/spaces/TheStinger/Ilaria_RVC?duplicate=true
void ivy
could i do it on my phone?
rare gobletBOT
Ayo? @void ivy level 2 !!!
void ivy
or laptop
fossil glen
doesn’t show more models
on settings check desktop/web mode
fossil glen
safari or chrome?
void ivy
safari
fossil glen
use chrome
void ivy
ok doing it rn
wide hornet
do you need a good mic for ths
pearl moon
fossil glen
not for AI covers but more so VCs
w-okada realtime
violet heron
Refresh
molten pecan
I uploaded the model as a zip
Where?
azure marshBOT
-realtime
This interaction has expired, use the command
/guides realtimeif you wish to see it again.
molten pecan
Yes
green olive
github?
green olive
yeah found it
molten pecan
> This interaction has expired, use the command `/guides realtime` if you wish t...
It's here
green olive
w-okada/voice-changer
crimson horizon
I'm trying to make a model. Why am I getting this error in step 2a?
start preprocess
['trainset_preprocess_pipeline_print.py', 'C:\\Users\\Lenovo\\Downloads\\Mangio-RVC-v23.7.0_INFER_TRAIN\\Mangio-RVC-v23.7.0\\datasets\\', '40000', '11', 'C:\\Users\\Lenovo\\Downloads\\Mangio-RVC-v23.7.0_INFER_TRAIN\\Mangio-RVC-v23.7.0/logs/KarleeAI', 'False']
C:\Users\Lenovo\Downloads\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\datasets\/KarleeAI.zip->Traceback (most recent call last):
File "C:\Users\Lenovo\Downloads\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\my_utils.py", line 96, in load_audio
ffmpeg.input(file, threads=0)
File "C:\Users\Lenovo\Downloads\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\ffmpeg\_run.py", line 325, in run
raise Error('ffmpeg', out, err)
ffmpeg._run.Error: ffmpeg error (see stderr output for detail)
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "C:\Users\Lenovo\Downloads\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\trainset_preprocess_pipeline_print.py", line 84, in pipeline
audio = load_audio(path, self.sr, DoFormant, Quefrency, Timbre)
File "C:\Users\Lenovo\Downloads\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\my_utils.py", line 103, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: ffmpeg error (see stderr output for detail)
end preprocess
thanks a lot julia
molten pecan
to huggingface
-huggingface it should say how in Step 2
azure marshBOT
Suggestions for @quaint dock
molten pecan
thanks a lot julia
yw! And from the river to the sea
quaint dock
-huggingface it should say how in Step 2
thanks!
green olive
yw! And from the river to the sea
haha ofc! and palestine will be free
rare gobletBOT
Ayo? @green olive level 1 !!!
hot lagoon
how to get yt-dlp-gui install?
viscid moss
rocky karma
Iget an error when trying to use the huggingface bot
Like I duplicated it and everything and it still gives me the error of nothing
limber fractal
any ideas? 2023-12-10 15:04:43 | ERROR | root | Exception in callback _ProactorBasePipeTransport._call_connection_lost(None) handle: <Handle _ProactorBasePipeTransport._call_connection_lost(None)> Traceback (most recent call last): File "asyncio\events.py", line 80, in _run File "asyncio\proactor_events.py", line 162, in _call_connection_lost ConnectionResetError: [WinError 10054] An existing connection was forcibly closed by the remote host
violet heron
Iget an error when trying to use the huggingface bot
Like I duplicated it and ev...
What was the link you used?
lilac sedge
Is the static noise issue a microphone issue or is it a setting I can change somewhere?
quaint dock
could anyone tell me the settings they use for covers?
like the feature extraction and stuff which are the best settings for those?
violet heron
like the feature extraction and stuff which are the best settings for those?
RVMPE
molten pecan
cause my ones sound robotic
https://rentry.org/ilarvc_inf_guide#the-voice-glitches-out-what-do-i-do read this part of my guide :b
If you want to know how to get the audio in lossless quality, lmk
quaint dock
If you want to know how to get the audio in lossless quality, lmk
sure, thank you!
molten pecan
sure, thank you!
yw! I highly recommend reading the whole thing too
quaint dock
yw! I highly recommend reading the whole thing too
Okay, thanks! I thought it was my model train fault that its really robotic lol
by that I mean like I didn't put enough audio
molten pecan
by that I mean like I didn't put enough audio
That would make a lot of sense. We recommend at least 30 mins of dataset
quaint dock
:O
quaint dock
And in lossless quality too
oof I can't even speak that much 😂
quaint dock
Happy early birthday, Rico
thank you, wait. Rico?
pulsar timber
Does anyone have tips on how to improve rhythm?
rare gobletBOT
Ayo? @pulsar timber level 2 !!!
strong belfry
How is it off? Both match with eachother
feral saffron
Can someone help me? I have colab pro but whenever I try to use the A100 gpu it doesn’t work and also it’s taking a very long time for feature extraction to load
strong belfry
Isnt the A100 overkill
feral saffron
I don’t think so
strong belfry
Do you only have the option to use a A100?
feral saffron
It takes a long time because I have a large dataset so I prefer to use the A100
I do have other options
strong belfry
How big is the dataset
feral saffron
It’s 38 songs
strong belfry
minutes
feral saffron
Each song ranges between 2-6 minutes
I don’t know exactly
But if I had to guess
2 hours
I have a theory I’d like to test which is why it’s so long
feral saffron
Oh really
All the other stuff went pretty quick which is why I thought maybe it’s acting up
strong belfry
Well anything less than 15m of dataset is pretty 'instant'
but ive had to wait a while with my 2hr datasets
id give it some time
not sure if colab you can see gpu cuda usage
also would think a100 be very overkill for rvc still and eat up all your credits
dont have colab so not sure how much it actually costs
feral saffron
It does but it’s speeds things up
rare gobletBOT
Ayo? @feral saffron level 1 !!!
feral saffron
not sure if colab you can see gpu cuda usage
I don’t even know what that is 😂
strong belfry
Would be at best 4x faster than a t4. Hope it doesnt cost more than that in credits
feral saffron
It does consume more credits for sure
But with a 2 hour dataset
With the settings I have and all that
It’d would take probably 10 hours to train with a t4
strong belfry
Yeah thats expected
feral saffron
And I just can’t afford that much time 😂
Yeah
Which is why I’m ok with A100
There is a V100 high ram
feral saffron
Im not sure
strong belfry
Could like batch 32
there should be a stats tab when you click the gpu
showing vram
feral saffron
I’ll have to check it out but im sure its high 😂
I usually use batch 15
But that’s for datasets under 2 hours
strong belfry
Well you are probably only using 16gb of vram
out of 40 or 80
could also instead load the dataset into vram
not sure if that is a option on colab notebooks'
faster training
feral saffron
Yeah
If it is I wouldn’t know I’m not very familiar with this type of thing unfortunately
pseudo steeple
Can someone help me? I have colab pro but whenever I try to use the A100 gpu it ...
i remember I have no emb perms
but jesus christ
strong belfry
dataset is usually read from disk cause vram is small, but you have so much spare vram you might be able to get away with it
pseudo steeple
a100??
feral saffron
dataset is usually read from disk cause vram is small, but you have so much spar...
Oh I see
pseudo steeple
go crazy with batch size but just be careful with your audio then
@strong belfry Question about batch sizes
Is it unironically better to use smaller batch sizes for large datasets?
strong belfry
Its a trade off between training speed tbh
Real world testing I saw very little difference between a optimally trained 2 batch vs 8
feral saffron
So what batch size would you recommend for a dataset that’s around 2 hours
strong belfry
But the 2 batch took a hella long time to get there
feral saffron
Oh wow
strong belfry
eh
pseudo steeple
So what batch size would you recommend for a dataset that’s around 2 hours
I'd figure max you could get tbh
If batch size qualities don't matter as much now
other than for time
strong belfry
16 or 8
strong belfry
16 if you want to go fast, also load dataset into vram if possible
I have no problem using 8 and leaving it for 10hrs cause i train locally and its much less expensive
a100 is like 9x more than a t4 for maybe 4x the performance
feral saffron
Oh I see
rare gobletBOT
Ayo? @feral saffron level 2 !!!
feral saffron
Alright thank you friends
feral saffron
Alright thank you
pseudo steeple
I can't wait for intel to be compatible with RVC....
strong belfry
yeah mainline is working on it
pseudo steeple
oh the blessings....
pseudo steeple
strong belfry
I wish I could use my amd card in my main rig too
pseudo steeple
I wish I could use my amd card in my main rig too
Another question
is there a way to efficiently rip voicelines from anime?
It took me 2 fucking hours to just compile a episode
So much easier using games...
strong belfry
I did see a script to use wisperx to identify, mark and snip a single audio file
no idea where its gone or who made it
pseudo steeple
it's joever...
strong belfry
If the anime has a game, might be possible to look there
pseudo steeple
nuh uh
strong belfry
which is it
strong belfry
ah
pulsar timber
How is it off? Both match with eachother
There's a noticable break in rhythm and flow imo for "FaraJason"
strong belfry
Anime is quite slow, but would take time to manually crop it.
Isolating vocals with like a weak seperation model
to avoid too much distortion
eh
fun times
I only bother with genshin and hsr for the time being cause I can automatically rip and prepare datasets with each version at this point
pseudo steeple
I only bother with genshin and hsr for the time being cause I can automatically ...
exactly...
rare gobletBOT
Ayo? @pseudo steeple level 5 !!!
pseudo steeple
you know
what would be a fun challenge
ripping files from stand up comedians or rakugo
strong belfry
I have a idle 3060 in my server i can put to work. i dont have infinite time to manually process characters
pseudo steeple
I have a idle 3060 in my server i can put to work. i dont have infinite time to ...
real...
proud elbow
I used vtubers' talking/singing streams rather than game playthrough & anime that have lot of sfxs that may not be able to separate
strong belfry
just talking streams work great
proud elbow
and sometimes acapella singing, luckily...
pulsar timber
I used vtubers' talking/singing streams rather than game playthrough & anime tha...
I can help isolate audio if you like
There's a plugin we use in the music industry
Here's an example where the voice is isolated at around 10 seconds
proud elbow
I can help isolate audio if you like
I'm not interested on some random paid plugins I don't know well
mossy ferry
i wanted to get back into ai but i've lost the google site that i originally used and am used to. I checked my google drive and i can only come up with the terms epoch and something around guardio
pulsar timber
I'm not interested on some random paid plugins I don't know well
If you send me the audio, I can help you
violet heron
i wanted to get back into ai but i've lost the google site that i originally use...
-colab
mossy ferry
thank you so much i found the one
proud elbow
If you send me the audio, I can help you
it's not like I'm asking for help, but anyway there's voice separation in Chinese capcut, some guide/tricks can be found here
https://docs.google.com/document/d/17fjNvJzj8ZGSer7c7OFe_CNfUKbAxEh_OBv94ZdRG5c/edit#heading=h.k3vca4e9ena8
(that can also separate some SFXs)
potent root
is there a way to buy gpu use on huggingface
molten pecan
(that can also separate some SFXs)
Yeah
violet heron
is there a way to buy gpu use on huggingface
potent root
which gpu is usable for training?
potent root
the T4s are usable
proven hill
which gpu is usable for training?
i hope you are joking
potent root
but idk if T4 small works
violet heron
which gpu is usable for training?
Did you used to be a model master????
potent root
also this seems very overpriced
proven hill
those are the prices
i mean, you can get a T4 small but i dont think you want to spend 432 bucks a month
potent root
those are the prices
it's billed PER MONTH?
proven hill
yes
molten pecan
That's crazy 😭
violet heron
those are the prices
Who’s paying $2,268 a month for ai cover 😭
potent root
it's either pay $2 a model on huggingface, pay $10 per month on colab, or local training (which i don't have the specs for)
you guys know any cheaper alternatives?
molten pecan
you guys know any cheaper alternatives?
Google colab :3
-colab use rvc disconnected
azure marshBOT
-colab use rvc disconnected
potent root
Google colab :3
i would use free colab for smaller scale models, but im gonna train an hour long dataset
manual saving sucks
finite galleon
manual saving sucks
Move the directory to your google drive and set the current directory to that main training directory
proven acorn
Hello, Just wondering
Is this still good to use?
https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=7vh6vphDwO0b
potent root
Move the directory to your google drive and set the current directory to that ma...
i mean transferring training data to a different acc
proven acorn
Also how long does a 2 minute reference audio take to render
potent root
the process i got used to is like, downloading all the required files and reuploading them on a different drive acc
finite galleon
i mean transferring training data to a different acc
Download it
proven acorn
Is there a tutorial on the EasyGUI by rejects
molten pecan
Is there a tutorial on the EasyGUI by rejects
https://rentry.org/ilarvc_inf_guide here's my guide on an updated version :b
proven acorn
The google collab one?
molten pecan
The google collab one?
This one uses Hugging Face
Is easier to use
And up-to-date
proven acorn
Is it bad that I'm still using this
https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=RrO4WDIYEPb8
rare gobletBOT
Ayo? @proven acorn level 1 !!!
molten pecan
Good luck
proven acorn
Though it takes longer to render the vocals
unlike the google collab
Does the local ones render the vocals faster or
not
violet heron
Does the local ones render the vocals faster or
Depends on your hardware but usually yes
violet heron
Which one do you recommend
-local
azure marshBOT
Suggestions for @proven acorn
violet heron
Original or Applio
violet heron
I’ve seen people having trouble with Mangio more lately
violet heron
Do they all function similarily?
They all make ai cover using guilt own audio files
Mangio and applio have Mangio-Crepe
Applio has TTS
proven acorn
Alright
Kinda concerned of the .bat stuff
is that gonna effect anything on my computer
violet heron
is that gonna effect anything on my computer
Nope
If your scared all of the stuff is open source so you can check the source code if you feel unsafe
proven acorn
i don't know how code works
is there an uninstaller thing when i download the program
violet heron
is there an uninstaller thing when i download the program
You just have to delete the folder
proven acorn
oh alright
rare gobletBOT
Ayo? @proven acorn level 2 !!!
proven acorn
thanks
Can I use "RVC0813Nvidia.7z" even if I use AMD and Intel or do I have to use "RVC0813AMD_Intel.7z"?
Apologies that the question is stupid
violet heron
Can I use "RVC0813Nvidia.7z" even if I use AMD and Intel or do I have to use "RV...
you have to use the one corresponding to your GPU
proven acorn
Alright
hushed cove
it says runtime\python.exe not recognized as an internal command or external in the go-web command prompt and when i press any key to continue nothing pops up afterwards
brittle wing
So... does anyone know how I could convert a .wav file I have to a voice changer? Not sure if this is the right channel to talk about this though.
violet heron
it says runtime\python.exe not recognized as an internal command or external in ...
Install python 3.8.9 and put it in runtime
hushed cove
Install python 3.8.9 and put it in runtime
alr
the latest version is the only one available, theres no way i can download an older version
rare gobletBOT
Ayo? @hushed cove level 1 !!!
hushed cove
Install python 3.8.9 and put it in runtime
alright i downloaded it. now what do i do? open the go-web again?
i downloaded python and im still getting the same message
fallow lava
I once succesfully made a cover using rvc & the second time I tired, it only gave me an error.
What was supposed to be an output was instead became an error box.
And the cmd prompt doesn't even show the error
fallow lava
I tried deleting & downloading it yet it still has the same problem
rare gobletBOT
Ayo? @fallow lava level 1 !!!
fallow lava
Did the audio file have any spaces in its name?
No & I tried testing it with the original file that worked & I still get the same error
Here
I even tried downloading the prerequisites afterwards such as the pip files & nothing has changed
Here's the evidence that it worked once when I used it
hushed cove
Here's the evidence that it worked once when I used it
it sounds so real WTF
sly oak
Hey everyone! I'm totally new to this. I was following Santatyde's tutorial on how to use RVC disconnected colab. It was going well but timed out a couple times while training, before I found how to prevent that. But then it told me I'd reached my usage limit.. people online say that they have some hidden algorithm that extends your cool down time each time you reach your limit. It only gave me about 2 hours today before the usage limit... Should I wait like a week to try use the colab again so I have enough 'time allowed' again to complete my model? I'm worried about them locking me out for a long time/permanently 😬 at this rate I'll never train a model
rare gobletBOT
Ayo? @sly oak level 1 !!!
mild hornet
Bruh the Google Collab stops after 5 mins of runtime.. why is it happening?
cloud finch
is there a way to stop voice cracks if my voice model goes to high?
cloud finch
i thought turning the semitone down would help but, it didn't
cedar tartan
Helo, can anyone help me find the public url? I've uploaded my audio zip in my dataset folder and ran the GUI, but I've been searching for 2 hours now and still haven't found the public url yet
woeful cave
hey ppl i have an inquiry
i trained an entire model but the dataset didnt go into the zip file and then after i deleted everything not knowing it was wrong
so i deleted all the save files
is there any way i can add the index file without retraining all 400 epochs...
ionic yew
Hey everyone! I'm totally new to this. I was following Santatyde's tutorial on h...
train at least enough amount of epochs it can handle and save the model before hitting to the 2 hour limit. I've been training rvc models with larger datasets for the past few days which that strategy works
for example if the dataset is large like 30 mins I train at least 50-100 epochs a day
also rvc disconnected hints a colab end prevention which if you need to train longer they said you can look it up
and to resume the training make sure the G and D pth files are also saved in Google drive and you then type in the "2333333" number for the amount of steps and then go put in the next amount of epochs you have
harsh scarabBOT
You're placed in Cover Generator Queue. Current position: 2
You're placed in Cover Generator Queue. Current position: 1
cerulean hedge
Our underwear
ornate slate
Would it be somehow possible to hook OpenAIs whisper model to the Inference/training to get cleaner pronounciation results by providing a script besides the audio? Just a thought I had
copper pebble
Would it be somehow possible to hook OpenAIs whisper model to the Inference/trai...
yh u can do a model of it
catch some HQ screen recording and extract the audio
brittle wing
Bals
mild wren
I was using the Aesthetic RVC and I downloaded a model onto it, fsr it keeps giving me Errors None
is it something wrong on my end or the model?
copper pebble
I was using the Aesthetic RVC and I downloaded a model onto it, fsr it keeps giv...
Hf right?
mild wren
ilaria RVC? Where can I find that
copper pebble
give me a sec
mild wren
Sure thing!
copper pebble
Sure thing!
https://discord.com/channels/1159260121998827560/1167844377046040769 heres a guide for that u will find the link in it
mild wren
https://discord.com/channels/1159260121998827560/1167844377046040769 heres a gui...
GAH! THANK YOU!
rare gobletBOT
Ayo? @mild wren level 1 !!!
copper pebble
just click on Ilaria RVC in blue and it will take u there
mild wren
Alright I'll give it a try, thanks a lot
copper pebble
yw
mild wren
HUH, now its saying the error - the link is invalid
i guess it is the problem of the model
YEAH NOW IM TRYING OTHER MODELS AND IT WORKS
gritty lily
is rvc and odaka same thing
or is rvc like the one that allows u to voice change while in vc
sand iris
or is rvc like the one that allows u to voice change while in vc
RVC does file conversion and model training (and also has a realtime thing which is... not that good)
okada's voice-changer does real time change
gritty lily
okada's voice-changer does real time change
can u help me set up odaka cause it closes instantly when i try launching it
sand iris
#🔍│help-w-okada pinned message