no lol, it
#✨│ai-help
atomic zephyr
fading drum
yeah
atomic zephyr
uh yeah
atomic zephyr
not sure, i really dont want to OPEN my fucking pc case to try to make AI models
fading drum
not sure, i really dont want to OPEN my fucking pc case to try to make AI models
well if there isnt a seperate gpu, then you cant rvc, because intel gpus wont work, but something from amd and nvidia will
oh wait, its integrated graphics, so there ISNT another gpu nvm
atomic zephyr
i havent seen nvidia thing on my pc in the last two years of me having this
yeah theres no other gpu
so i guess i just cant do this bullshit 
fading drum
you could go to #1159289738314919936 or #1191429836321849435
or browse on weights.gg
fading drum
man, gotta request a fuckin model for myself
i mean, i could train it ifyw
atomic zephyr
i mean, i could train it ifyw
like, you make the model and i do the fuckin ai model thing? or you can do all that i'd need
bcs if so, that'd be greattt
fading drum
like, you make the model and i do the fuckin ai model thing? or you can do all t...
i can make the model yeah
fading drum
like, you make the model and i do the fuckin ai model thing? or you can do all t...
(the training and model is done together)
atomic zephyr
lets gooooo., thank you so much lemme dm you
fading drum
just dm me a onedrive link to the file
low shard
man, gotta request a fuckin model for myself
btw you can make models using google gpu
check inside the cloud folder inside the rvc folder of our docs https://docs.aihub.wtf
atomic zephyr
btw you can make models using google gpu
ourhg, its fine i dont wanna
low shard
oke
fading drum
turn down the batch gpu size
per every number is a gb of vram
so 10 is 10gb of vram
rustic locust
How do I know which model to download?
azure marshBOT
RVC Guides (How to Make AI Cover)
primal plank
When I try to train the model I get this error
训练结束, 您可查看控制台训练日志或实验文件夹下的train.log
and i cant open Tensorboard
rare gobletBOT
Ayo? @primal plank level 1 !!! 
rigid dirge
hey
im trying to get back into ai
after a while
is rvc still good for making ai voices
or is there a better option
wispy burrow
When I try to train the model I get this error
训练结束, 您可查看控制台训练日志或实验文件夹下的train.l...
I see no error...
It's saying it created the train.log file that you can take a look at after training.
Something like that...
wispy burrow
is rvc still good for making ai voices
We still use RVC.
wispy burrow
or is there a better option
But now there's alos SoVits-GPT to create TTS Models.
wispy burrow
is there a new way to use or download it
Nothing really changed in that regard.
Converting a voice from audio to a model's voice we still use RVC for that.
But now we have a very great guide.
https://docs.aihub.wtf/essentials/whats-rvc/
hidden snow
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
What im supose to do
wispy burrow
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pi...
Ask in #🔍│help-w-okada .
You should get a response/solution much faster there.
hidden snow
Thaks you
tender canyon
I am trying to train a voice model, and I keep getting error messages, alternating between this one:
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass
RuntimeError: GET was unable to find an engine to execute this computation
and:
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 10.00 MiB (GPU 0; 4.00 GiB total capacity; 2.44 GiB already allocated; 0 bytes free; 2.59 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
ive tried reducing the batch size but going down to even 1 still produces an error. what else can I do?
fleet pilot
I am trying to train a voice model, and I keep getting error messages, alternati...
what r u using
tender canyon
the mainline RVC webUI and tensorboard
following the instructions here https://docs.aihub.wtf/
brittle wing
Is there a website to train models (not Google Colab)?
untold frost
How do you test the voice and hear it back?
astral perch
What does a larger batch size do to your model?
marble forge
someone please help me bruh its been 2 days
willow shell
Does anyone know why Applio's playground and the Huggingface space take so long to proccess the inference
In playground it took like 800 seconds
tender canyon
how do we know if we are overtraining by monitoring tensorboard
rare gobletBOT
Ayo? @tender canyon level 1 !!!
tender canyon
will it just tell us?
crimson ruin
proud elbow
I am trying to train a voice model, and I keep getting error messages, alternati...
at least 6 gb VRAM needed
tender canyon
https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/#--monitoring
thx
wispy burrow
someone please help me bruh its been 2 days
Was there no error displayed in the cmd?
You couldn't the Index files in your logs folder?
primal plank
I see no error...
It's saying it created the train.log file that you can take a ...
does not continue
marble forge
Was there no error displayed in the cmd?
You couldn't the Index files in your lo...
no
lethal orbit
whats the best f0 det thing
proud elbow
you two, what's ur gpu?
fading drum
is rvc still good for making ai voices
it is good, takes a while for better quality, all voice cloners use a branch of it somehow
fading drum
How do I know which model to download?
what are you using it for?
main lichen
I said, Nvidia or AMD? cuz the settings depend on it
nvidia
brave garnetBOT
⠀
Settings for Nvidia GPUs 
F0 Det.: rmvpe (suggested for all series)
RTX 40-series: 80-96 chunk | +16384 extra
RTX 30-series: 96-112 chunk | +16384 extra
RTX 20-series: 112-128 chunk | +16384 extra
GTX 16-series: 128-192 chunk | +8192 extra
GTX 10-series: 128-192 chunk | +8192 extra
Advanced Settings
Protocol : Sio or Rest
Crossfade: 4096 start 0.2 end 0.8
Trancate: 300
Silencefront: Off
Protect: 0.5
RVC Quality: Low
⠀
Made with love by Ilaria 💖
main lichen
what
tidal cairn
I need a working vocal inference Google collab, can anyone please share me one?
main lichen
so i can do the max for extra
cuz it says +16384
i asked for f0 det
idk the full name but what do i do for that
proud elbow
i asked for f0 det
it clearly says rmvpe
main lichen
what where
proud elbow
this duh
reef frost
queen atlas
what program so i use if i got a gtx
untold frost
How do you access these other F0 Detectors?
untold frost
It says (N/A) next to most of them, and I can't click on it
main lichen
use rmvpe
untold frost
yo how to smoothen my voice changer
Is it cutting out all over?
rare gobletBOT
Ayo? @untold frost level 1 !!!
sturdy swan
rmvpe_onnx?
main lichen
sturdy swan
Is it cutting out all over?
yea sometimes
main lichen
nah
sturdy swan
its cutting hard af
sturdy swan
alright I did it
untold frost
Or the F0
untold frost
I can't 😭
main lichen
alright I did it
did you do the installation correctly?
sturdy swan
ye
main lichen
U chose the right gpu and stuff
main lichen
sturdy swan
LOL
main lichen
?
sturdy swan
yea
rare gobletBOT
Ayo? @sturdy swan level 1 !!!
sturdy swan
it feels just
a little bit delayed
it says it like 2-3 sec after
ima try 112
yea ima keep it at 112
where can I find models
primal plank
yo
why am i getting this error
ModuleNotFoundError: No module named 'gradio'
sturdy swan
Whats your gpu?
nvidia geforce rtx 3060
rare gobletBOT
Ayo? @main lichen level 6 !!!
untold frost
This is hella laggy
main lichen
set it on 16384
main lichen
wdym?
main lichen
Put it to 128 or 192 for it not to be laggy and put extra to 16384 or 32768
And make sure the model your using is well put
Also go to advanced settings on the bottom
Copy these settings
sturdy swan
ai done
main lichen
Here
main lichen
that are well put
Just listen to the samples to see if its good
sturdy swan
Alright
bronze maple
questions for experts : I have a dataSet of 4000 audio files, 3 seconds each
I launched a train for 500 epoche
in 20 minutes I am at the 4th epoch, step = 2200
It's normal ?
I think I have too many datasets, maybe there's no point in doing 500 epoche given the time it takes
I'm going to leave it until 100 epochs to see what happens
a little opinion?
sturdy swan
Just listen to the samples to see if its good
what are epochs
primal plank
I'm getting an error in applio
ModuleNotFoundError: No module named 'gradio'
rigid dirge
using rvc to make voices and this popped up lmfao can anyone hellp
apollio to be spefific
craggy dagger
Hey guys what am I doing wrong here for making a model?
bronze maple
cause its not progressing
patience*
wispy burrow
questions for experts : I have a dataSet of 4000 audio files, 3 seconds each
I ...
Don't tell me that Dataset is 3 hours long.
500 Epochs would be way too much.
bronze maple
Don't tell me that Dataset is 3 hours long.
500 Epochs would be way too much.
Yes, maybe I was being stupid
it's because I extracted the voice of a person from a video game and there were 4000 audio files of exceptional quality, so I put everything 😂
polar jackal
HELP ICANT STREAM ON ULIMATED
wispy burrow
using rvc to make voices and this popped up lmfao can anyone hellp
Any progress by now?
bronze maple
Look
untold frost
I've got these settings and it's cutting off as heck
Any recommendations?
And this vol section spikes up 5 seconds after I speak, and the audio cuts in with it, before cutting out again
proud elbow
I've got these settings and it's cutting off as heck
it's normal for cpu mode to get much delay, or better dont use it
untold frost
Haven't got another option there,
Though less so the delay, more so the absurd amounts of cutting out
sturdy swan
it's normal for cpu mode to get much delay, or better dont use it
yooo can u explain to me what epochs are
rare gobletBOT
Ayo? @sturdy swan level 2 !!!
queen atlas
my rvc doesnt launch
vast fractal
What rvc fork would you suggest
queen atlas
its telling me to open a ip adress port
bronze maple
Any progress by now?
I stopped the training, then I restarted it to have a backup of 5 epochs, do you think it's normal for the graph to go backwards like that?
proud elbow
cooling issue tbh
queen atlas
bro does nobody help here
finite galleon
I stopped the training, then I restarted it to have a backup of 5 epochs, do you...
Not normal
proud elbow
maybe not enough psu wattage, or try power limit it
finite galleon
When you restart you should restart from the last checkpoint that you have saved
bronze maple
Not normal
why does it do this? Applio v3 is crap then
rare gobletBOT
Ayo? @bronze maple level 5 !!!
vast fractal
maybe not enough psu wattage, or try power limit it
Like what is that?
finite galleon
why does it do this? Applio v3 is crap then
Did you change any of the training settings so that it doesnt save something?
proud elbow
Like what is that?
sorry not the power itself, but yea more likely temp issue
likely need to replace thermal paste or something
wispy burrow
Ask in #🔍│help-w-okada
untold frost
Ask in <#1159290161683767298>
What's the difference between the two channels
vast fractal
sorry not the power itself, but yea more likely temp issue
How do you suppose I fix the issue
bronze maple
Did you change any of the training settings so that it doesnt save something?
I had configured that it saves every 100 epoche, but it was very long, so I stopped the training why I modify the save every 5 epoche, so when restarting I just entered the name of the model folder in Applio /log
wispy burrow
What's the difference between the two channels
W-Okada one is for realtime conversion.
untold frost
I see
finite galleon
I had configured that it saves every 100 epoche, but it was very long, so I stop...
You should retrain after it saves so that it wouldnt restart the whole training process instead of retraining
bronze maple
You should retrain after it saves so that it wouldnt restart the whole training ...
on applio it's poorly done I don't know how to do that
finite galleon
If you want to stick to save per 5 epochs stick with it and dont change
finite galleon
on applio it's poorly done I don't know how to do that
It will tell you that its done saving on the cmd
Dont stop training until you see it has saved
bronze maple
Dont stop training until you see it has saved
Okay I understand better, so I have to start a new workout again because this one is dead then
finite galleon
Okay I understand better, so I have to start a new workout again because this on...
Yeah
Starting a new one would be ideal
proud elbow
Okay I understand better, so I have to start a new workout again because this on...
yeah if u dont want tensorboard looks screwed
bronze maple
yeah if u dont want tensorboard looks screwed
I understand better thanks guys
I have 2 hours of dataset, is that too much for you to think? what is the best Dataset and epoche ratio for a perfect voice?
I extracted the voice of a video game character, the voice is perfect so perfect dataset
primal plank
whyy
rare gobletBOT
Ayo? @primal plank level 2 !!!
wispy burrow
I have 2 hours of dataset, is that too much for you to think? what is the best D...
the recommendation was always 10-50 mintues for the Dataset.
proud elbow
I have 2 hours of dataset, is that too much for you to think? what is the best D...
quality, cleanness and best epoch to pick matter
proud elbow
I have 2 hours of dataset, is that too much for you to think? what is the best D...
also if u know how to modify the code, delete the minibatch k means part, or else just use less than 1 hour dataset
bronze maple
the recommendation was always 10-50 mintues for the Dataset.
and for epoches? because he shouldn't overtrain from what I understood but there is no way of knowing when he is overtraining
finite galleon
also if u know how to modify the code, delete the minibatch k means part, or els...
Is that for the index thing
proud elbow
Is that for the index thing
yup
bronze maple
also if u know how to modify the code, delete the minibatch k means part, or els...
so you will have to show me an example
proud elbow
so you will have to show me an example
cant do shit on discord mobile
bronze maple
cant do shit on discord mobile
I added you as friends, you will show me if you have time one of these days
proud elbow
I added you as friends, you will show me if you have time one of these days
nope, probably next week 💀
bronze maple
nope, probably next week 💀
thank you you are the best
in the cases when I stop training with Applio v3, the application restarts, but my GPU is still running at full speed like this it is still training without the program open, it's crap applio I don't understand I have to restart my PC
proud elbow
in the cases when I stop training with Applio v3, the application restarts, but ...
kill the python processes in task manager
rigid dirge
i have a question for applio when training a voice do you put the .flac file in a folder and if so when your putting the dataset path do you include the .flac file in the path or just the folder
meager quarry
Why do I get Runtime Error when I try training in RVC-Disconnected Colab?
I did all the steps right
something like "IndexError: list index out of range" and "RuntimeError: Error(s) in loading state_dict for SynthesizerTrnMs767NSFsid"
bronze maple
and for epoches? because he shouldn't overtrain from what I understood but there is no way of knowing when he is overtraining
meager quarry
Why do I get Runtime Error when I try training in RVC-Disconnected Colab?
Oh now, I understand
I was using OV2
so either I used some wack dataset or OV2 pretrained is broken on that colab rn
I didn't get the same error with the Italia pretrained model just now, will use that instead since I heard it's a good alternative to OV2
Do I have to be concerned if my model may sound like a too much like it's speaking in an Italian accent?
thin stump
@opal depot
thin stump
Oh now, I understand
yeah ov2 is broken so I was using ItaIla
thin stump
<@194069452113838080>
renegate is the improved version of audacity's noise gate and kilohearts supports it further
rugged stirrup
How is Ov2 broken?
thin stump
the f-1 in the code. It's the error you've seen
rugged stirrup
Oh
opal depot
<@194069452113838080>
That's so coo. And here i was doing it manually
rugged stirrup
I'm gonna try Ilaria RVC again
opal depot
So i then just package all the vocals together into a zip and train as normally?
thin stump
That's so coo. And here i was doing it manually
The smart people are our backbones in this server 🙏 . Yeah, it shouldn't include silences. My datasets were worse when I truncated
rugged stirrup
The smart people are our backbones in this server 🙏 . Yeah, it shouldn't includ...
Sorry to intrude but have you used Ilaria RVC before?
thin stump
Sorry to intrude but have you used Ilaria RVC before?
nope, I use replay
thin stump
Or is it just for inference
Right, there's no training. It's only a software for ai covers
rugged stirrup
Oh
fading drum
What rvc fork would you suggest
depends on what gpu you have, nvidia?
rare gobletBOT
Ayo? @fading drum level 5 !!!
opal depot
<@194069452113838080>
do I need to follow the setup for labels exact?
thin stump
do I need to follow the setup for labels exact?
It's also below in this guide. RVC will automatically splice audio that's not 76 seconds long so we're just doing it ourselves https://rentry.org/RVC-dataset-RX-updated
thin stump
it's ok, it's hard to explain because i'm trying to make sense of the settings
opal depot
It's also below in this guide. RVC will automatically splice audio that's not 76...
if I could I would go and hug you my man
thin stump
if I could I would go and hug you my man
yw, that's all you can do to improve datasets besides using RX10.
sonic agate
when i need to use ItaIla?
thin stump
when i need to use ItaIla?
for short datasets (3 minutes long) and italian audio. Use it if the Ov2 tensorboard isn't working
brittle wing
Hi guys
does anybody know why this error happens
wait
well
it says traceback something something
thin stump
Hi guys
You're using an outdated colab with the Jsondecode error. Check https://docs.aihub.wtf/ for the newer versions
brittle wing
So
brittle wing
You're using an outdated colab with the Jsondecode error. Check https://docs.aih...
Where would that be in the site?
thin stump
Are you making covers or voice models? The guide lays it out for you
cloud finch
I keep getting these files missing in system32 errors when i try to install APPLIO
brittle wing
Are you making covers or voice models? The guide lays it out for you
Covers
tidal cairn
Does anyone know what’s up with the ilaria RVC collab? Its not working
I need some help, this is really frustrating
I been trying to find a functioning collab all day and no one has provided me with a working one. 😞
rare gale
I usually did it from a tutorial on YouTube... there are few good guides on Russian YouTube. What similar collab can you recommend?
bronze maple
It's also below in this guide. RVC will automatically splice audio that's not 76...
wow, do you do all these settings for each audio file?
thin stump
wow, do you do all these settings for each audio file?
Yeah, it's the same idea as truncating silences. I only trust that Auburn renegate > kiloheart dynamics > then UVR denoise will completely get rid of the silences so that it's not included in audio labels https://rentry.org/RVC-dataset-RX-updated
tidal cairn
Wow, no help. As this community has grown it seems like people care less now in here to help others resolve their issues
bronze maple
Wow, no help. As this community has grown it seems like people care less now in ...
we all help each other don't worry
thin stump
Nah i'd win
tidal cairn
Well, can I get some assistance please? I would really appreciate it
rare gobletBOT
Ayo? @tidal cairn level 4 !!!
thin stump
Well, can I get some assistance please? I would really appreciate it
https://www.youtube.com/watch?v=wh4cLMLLkj0 @ her if it doesn't work
bronze maple
Yeah, it's the same idea as truncating silences. I only trust that Auburn renega...
you're hot brother, I use Dx Revive for the denoise and the deverb it works pretty well, but I should try with Izotope rx
fading drum
test
azure marshBOT
RVC Guides (How to Make AI Cover)
fading drum
Does anyone know what’s up with the ilaria RVC collab? Its not working
colabs sometimes break, even when in the middle of training, and take much longer to use than by just using your local gpu , ive found that its much easier to have it run on my local gpu too
tidal cairn
When I try to run on local it doesn’t let me, gives an error.
fading drum
When I try to run on local it doesn’t let me, gives an error.
what gpu do you have?
you can check in device manager
fading drum
The same one as you
download this and run go-web.bat then msg me
https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI
bronze maple
if you have an Nvidia graphics card that is fairly recent, download the RVC directly to your machine, it will go faster
faint moon
i have one voice audio, any one make ai voice
rare gobletBOT
Ayo? @faint moon level 1 !!!
tidal cairn
faint moon
i have one voice audio, anyone convert to pth file plz ???
tidal cairn
@fading drum
hasty acorn
is it possible to use the site without a mic or?
tidal cairn
I try to run go-web bat and it tells me (windows batch file) and it tells me path cannot be found
@fading drum @bronze maple
fading drum
Which one? <@578836590659371019>
go-web.bat
fading drum
It says path cannot be found
https://www.youtube.com/watch?v=hB7zFyP99CY follow this guys install tut
i didnt have to and it worked fine for me but i guess im lucky
elder wave
i cannot unzip the file
faint moon
https://www.youtube.com/watch?v=hB7zFyP99CY follow this guys install tut
i have one person voice audio. you convert PHT file ?
fading drum
i have one person voice audio. you convert PHT file ?
pht? or pth
and convert it to what?
faint moon
pht? or pth
sorry PTH
faint moon
ai voice chager
fading drum
so from a pth to a model zip?
faint moon
yes and sorry for the my english
proud elbow
i cannot unzip the file
which file?
low shard
I usually did it from a tutorial on YouTube... there are few good guides on Russ...
Nope, yt tuts are outdated asf, thats why i suggested you to see inside the cloud folder inside the rvc folder of our docs https://docs.aihub.wtf
faint moon
so from a pth to a model zip?
can i send voice audio file ?
fading drum
yes and sorry for the my english
im afraid having just the pth wont make it useable for a voice changer, i think the only way to use the pth is to train an index from the pth, and that pth is made during training with reference audio
TLDR having just a pth is not enough
||i think||
faint moon
i play gta v, i need actor vijay voice plz help
fading drum
if you have a model G and model D then that can be trained to an index
fading drum
i play gta v, i need actor vijay voice plz help
do you have these?
faint moon
if you have a model G and model D then that can be trained to an index
i need actor vijay voice, i dont know how to make model
proud elbow
like any of the files
7zip? or the file corrupted?
elder wave
wait can i ask which voice mod app you use?
rare gobletBOT
Ayo? @elder wave level 1 !!!
tender canyon
at least 6 gb VRAM needed
do you need 6 gb VRAM to inference with the webUI as well? I was able to successfully do it a couple times but all of a sudden started getting the error:
' if data.dtype in [np.float64, np.float32, np.float16]:
AttributeError: 'NoneType' object has no attribute 'dtype' '
with no change on my end of anything
elder wave
7zip? or the file corrupted?
for some reason i dont have the 7zip
proud elbow
for some reason i dont have the 7zip
winrar?
proud elbow
do you need 6 gb VRAM to inference with the webUI as well? I was able to success...
inference should be fine
tender canyon
inference should be fine
do you know what could be causing that error message, then?
proud elbow
the app im using only supports mp3
wdym mp3? isn't it model file?
proud elbow
do you know what could be causing that error message, then?
show me the whole gui as well
tender canyon
show me the whole gui as well
sorry if the pic is a bit scuffed have to use my phone cause my internet cut out
obsidian nymph
queen escarp
-colab
queen escarp
how to run gradio deploy from terminal
glad zealot
how to run gradio deploy from terminal
python infer-web.py or app.py depending on what you are using
brittle wing
What does it mean
rare gobletBOT
Ayo? @brittle wing level 7 !!!
loud summit
Making a custom pretrain takes so long?
I wanna try to make a custom pretrain for RVC, but idk how
brittle wing
I can't find the file?
how to open voice changer okada
please help me
Havr I overtrained?
brittle wing
I can't find the file?
@flint geyser
rare gobletBOT
Ayo? @brittle wing level 2 !!!
brittle wing
can you help pls
azure marshBOT
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
bronze maple
Nvidia
bronze maple
heh
nova trench
Hello! Is there any way to make the AI sound smoother? I'm using RVC and Gradio, but there are multiple parts in the song that just sound robotic no matter how much I play with the settings 😦
fading drum
Hello! Is there any way to make the AI sound smoother? I'm using RVC and Gradio,...
robotic is due to a model not being able to go as high or low as the pitch needs, or its because of the decrease in audio quality from getting only the vocals
nova trench
! I understand, thank you! I'll try to see if it's an issue with the model I'm using
fading drum
heres a free hq version of jack blacks fligugigu if you want to test
fading drum
https://podcast.adobe.com/enhance#
this doesnt fill in voice gaps, it just makes the voice sound "better"
like it was in a podcast, and sometimes it can fix the audio in some spots
fading drum
heres a free hq version of jack blacks fligugigu if you want to test
i use this as a benchmark to see if i need to train a model more
nova trench
heres a free hq version of jack blacks fligugigu if you want to test
Thank you so much for this! I'll try to give it a test run then
proper shale
like it was in a podcast, and sometimes it can fix the audio in some spots
Look, I love Podcast and all but...
...sometimes it screws up the person's voice. It's because they reconstruct the voice in the audio and sometimes it leads to making either weird audios or the person doesn't sound the same.
So I'd recommend avoiding it unless it's the only thing you can use to make it better
fading drum
Look, I love Podcast and all but...
...sometimes it screws up the person's voic...
"sometimes"
"some spots"
proper shale
"some spots"
You said that with a positive statement tho
fading drum
You said that with a positive statement tho <:nails:1159569314848972891>
better than leaving it robotic imo
proper shale
Well, either way, this doesn't mean that you can't use Podcast
But most of the time it's better to avoid it and prefer other tools ig
Removing bg audio for example, that can be done elsewhere
I'd use Podcast if nothing else works
Because hey, it works
fading drum
Hello! Is there any way to make the AI sound smoother? I'm using RVC and Gradio,...
oh yeah for removing backround audio use uvr5 with 6_HP-Karaoke-UVR Vocals only btw to get the lead vocals out
hard shadow
is there a guide on how to use a pretrain?
Like using custom pre-train model as a base to train your own model?
On the third box before you train or create index in Applio, there should be a checkbox that says "Custom Pretrained". Have a checkmark in it. There, you could upload the D/G files of the custom pretrained model. The program then will use that base model when training instead of the default one.
stiff elbow
I see thanks! Ill check it out
rare gobletBOT
Ayo? @stiff elbow level 1 !!!
nova trench
oh yeah for removing backround audio use uvr5 with 6_HP-Karaoke-UVR Vocals only ...
Thank you very much!! I'll keep this in mind!
rare gobletBOT
Ayo? @nova trench level 2 !!!
nova trench
Oh, uh, me again, sorry about that. I'm trying to train my old model now, but even though I've finished, the .PTH file is nowhere to be found, even if the index files are there. Did I do something wrong?
forest sluice
Havr I overtrained?
it looks like it
though you may want to wait for a bit to ensure
brittle wing
it looks like it
It started dropping after 320 so basically I saved the weight
forest sluice
Oh, uh, me again, sorry about that. I'm trying to train my old model now, but ev...
Check the weights folder in your RVC folder
fading drum
Oh, uh, me again, sorry about that. I'm trying to train my old model now, but ev...
the pth file is in the models folder (go to the main folder and search "models" then pick the model name)
nova trench
Check the `weights` folder in your RVC folder
It's still not there 😦
rare gobletBOT
Ayo? @fading drum level 6 !!!
fading drum
wait no nvm
fading drum
the pth file is in the models folder (go to the main folder and search "models" ...
im idot
fading drum
It's still not there 😦
when training is finished, it should save there, did the training get stopped before the final epoch?
nova trench
the pth file is in the models folder (go to the main folder and search "models" ...
Not there...
fading drum
Not there...
then... im afraid youll have to do the training again
nova trench
when training is finished, it should save there, did the training get stopped be...
Step 1: Processing data
runtime\python.exe trainset_preprocess_pipeline_print.py C:\Users\63929\Desktop\dataset\dataset_raw\Sinclair 48000 2 C:\RVC-beta0717/logs/sincle False
step2a:正在提取音高
runtime\python.exe extract_f0_print.py C:\RVC-beta0717/logs/sincle 2 harvest
Step 2b: Extracting features
runtime\python.exe extract_feature_print.py cpu 1 0 C:\RVC-beta0717/logs/sincle v2
Step 3a: Model training started
write filelist done
runtime\python.exe train_nsf_sim_cache_sid_load_pretrain.py -e sincle -sr 48k -f0 1 -bs 2 -te 1000 -se 50 -pg pretrained_v2/f0G48k.pth -pd pretrained_v2/f0D48k.pth -l 1 -c 0 -sw 0 -v v2
Training complete. You can check the training logs in the console or the 'train.log' file under the experiment folder.
(52833, 768),1354
training index
adding index
成功构建索引, added_IVF1354_Flat_nprobe_1_sincle_v2.index
All processes have been completed!
This was the message it finished with
fading drum
make sure to tick this thing
nova trench
make sure to tick this thing
Oh! I didn't have it ticked, haha
Thank you so much! I'll do it again, hopefully it'll work by then
fading drum
while im here, this thing is also very useful
like 10x faster useful, but it has to be under 8 minutes not 10
nova trench
like 10x faster useful, but it has to be under 8 minutes not 10
Got it! Will it work even if I don't have a GPU, though?
fading drum
Got it! Will it work even if I don't have a GPU, though?
ummmmmm not sure
it caches to your ram if you run out of vram
nova trench
Okay, I'll give it a spin when I can
fading drum
it makes you able to watch videos and play games (at slow speed) while training
just dont like... render a blender animation
nova trench
Haha got it! That's good then
Sorry for the trouble again, but somehow even if the training's finished, it still isn't there...
strong spoke
if I add to the trained model additional dataset material and continue training, would it train correctly?
fading drum
Sorry for the trouble again, but somehow even if the training's finished, it sti...
that, is puzzling
is 1 higher than 2?
proper shale
Got it! Will it work even if I don't have a GPU, though?
If you don't have a GPU, training will not work sadly
nova trench
is 1 higher than 2?
It isn't?
fading drum
If you don't have a GPU, training will not work sadly
yeahhh i was a lil skeptical about how fast it was training
proper shale
if I add to the trained model additional dataset material and continue training,...
I don't think so, it's best to make a new model ig
nova trench
Additionally, if I tried clicking on "Train model" on its own, it simply freezes and gives me this message
nova trench
If you don't have a GPU, training will not work sadly
Aw shucks, understood. Thank you so much for the help! And sorry for the trouble, too
fading drum
Aw shucks, understood. Thank you so much for the help! And sorry for the trouble...
its fine, its the point of this channel, you can request a model in here: https://discord.com/channels/1159260121998827560/1159289738314919936
or get a colab to work
nova trench
Alright, will do! I'll try Colab then, thank you so so much
rare gobletBOT
Ayo? @nova trench level 3 !!!
proper shale
Aw shucks, understood. Thank you so much for the help! And sorry for the trouble...
Ah no problem :)
proper shale
Alright, will do! I'll try Colab then, thank you so so much
Yeah Colab FTW
rare gale
Nope, yt tuts are outdated asf, thats why i suggested you to see inside the clou...
what's on your site? I don’t understand where to go and look.💔
tame mica
what's on your site? I don’t understand where to go and look.💔
seems like ur trying to inference. go https://docs.aihub.wtf/rvc/cloud/ilaria-rvc/
humble ember
hello really quickly where is like the links where we actually use the ai voice yk those collabs?
azure marshBOT
RVC Guides (How to Make AI Cover)
humble ember
thank u bot
low shard
what's on your site? I don’t understand where to go and look.💔
Thats the server site, it's a documentation, for knowing where to do ai covers online, you need to click the rvc folder on it, and then the cloud folder, it will make you show some guides of some collabs, please read the pros and cons and see which is the best for your needs
hard shadow
Yeah, sometimes it would not show up for me after reloading
strong spoke
brittle wing
yo how can i solve the sizzles when i talk
fading drum
yo how can i solve the sizzles when i talk
what program are you using, sizzles as in static?
if its running on a web server, try increasting the time that the server gets to process the audio, it should be called Chunk or smth
setting it higher will give the program more time to process and therefore a better sounding audio
(it should have a bunch of these numbers, fiddle around until you find one that works)
humble ember
why does it say Error and i get this message on the console: NoneType' object has no attribute 'dtype'
rare gobletBOT
Ayo? @humble ember level 1 !!!
fading drum
why does it say Error and i get this message on the console: NoneType' object ha...
colabs frequently break due to new colab code updates, perhaps wait and try again, or run it locally
stiff elbow
Yeah, sometimes it would not show up for me after reloading
yeah I think its just not updated anymore too bad still a good read though
bleak flame
Bro can you help me?
How to launch the RVC models on android/googlecollab
fading drum
Bro can you help me?
How to launch the RVC models on android/googlecollab
it should be easy as pressing the little "play" buttons at the side of the code
if it comes up with errors, read them carefully
bleak flame
it should be easy as pressing the little "play" buttons at the side of the code
Where do we get the CollaB research link for the launch model?
brave garnetBOT
⠀
Google Colabs 
⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.
AICoverGen-NoWebUI
Useful for making covers, doesn't inclued a UI, by Ardha, by Eddy, Hina and Gdr.
RVC Disconnected
To train new voice models, by Kit Lemonfoot.
EasyGUI
The OG interface, by Rejects.
⠀
Made with love by Ilaria 💖
fading drum
or any of these
bleak flame
Thanks bro👍
fading drum
id personally reccomend making it yourself, because free colab doesnt last long enough to sufficiently clone a good model
brave garnetBOT
Made with love by Ilaria 💖
vagrant bolt
it wont work on discord for me
proper shale
only amd and nvidia (gpus) supported
actually only NVIDIA stuff works
but yeah fair enough it'll work even w a CPU unless it's training
proper shale
it wont work on discord for me
what exactly
proper shale
okay, but what app are we talkin about here
vagrant bolt
RVC
proper shale
i believe W-Okada, then
so
check command prompt, see if there's any Tracebacks or any errors u can find
rustic ingot
Hi. Dumb question but what does RVC mean? Is it an app or...
vagrant bolt
i dont see any errors but i see Pipeline has been deleted
rare gobletBOT
Ayo? @vagrant bolt level 1 !!!
proper shale
i dont see any errors but i see ``Pipeline has been deleted``
hmmmm... seems like smth went wrong then
whats ur gpu?
vagrant bolt
GTX 1650
proper shale
Hi. Dumb question but what does RVC mean? Is it an app or...
Retrieval-based Voice Conversion
voice conversion w audio, basically.
rustic ingot
Thanks
proper shale
hm okay so
maybe try closing command prompt n shit, and delete the pretrain and model-dir folders inside MMVCServerSIO, then restart the program
see if anything goes wrong after that...
btw, have u read the W-Okada guide?
vagrant bolt
nope i havent
also another question
do you know why crepe is not selectable for me
it says N/A and i cant click it
pastel oak
do the gpu options for you say "gpu(dml)" and then cpu, gpu0, gpu1 etc.
vagrant bolt
i have gpu0 still cant
pastel oak
it was a yes or no question
vagrant bolt
oh mb
pastel oak
anyway you have the AMD version if you have a nvidia gpu then get the other
proper shale
anyway you have the AMD version if you have a nvidia gpu then get the other
ohhhh true
damn
brave garnetBOT
⠀
Download for Nvidia GPUs
Version 18a cuda
Download for AMD GPUs 
Version 18a directml
Download for Intel GPUs 
Version 18a directml
Download for Mac 
Version 17b Mac
⠀
Made with love by Ilaria 💖
proper shale
^^^ download the nvidia version
vagrant bolt
would that most likely also fix the discord problem
rigid dirge
hey using apollio and the model i was training is finally done, where does the pth file go to 😭
rare gobletBOT
Ayo? @rigid dirge level 4 !!!
rigid dirge
first time doing this lol
pastel oak
hey using apollio and the model i was training is finally done, where does the p...
dont have applio but it should be in a folder called weights, maybe in assets -> weights if it has that
rigid dirge
it does have assets but no weights il try and look through rq lol but i have no idea
proper shale
would that most likely also fix the discord problem
yeah
rigid dirge
i have a question for applio when training a voice do you put the .flac file in a folder and if so when your putting the dataset path do you include the .flac file in the path or just the folder
^^^
i asked this earlier does anyone know the answer btw 😭
proper shale
i have a question for applio when training a voice do you put the .flac file in ...
folder path
fading drum
actually only NVIDIA stuff works
i mean locally
so like, intel cpus wont work, and macs
🤮
no offense C:
rigid dirge
folder path
ight thanks i did that right lol thank god 😭
but to anyone who uses appolio does anyone wknow where the pth files go to
?
tall marten
hello
I am currently
https://www.youtube.com/watch?v=paVYNlZ5Xuk&list=RDpaVYNlZ5Xuk&start_radio=1
I'm trying to AI cover this song.
First, I downloaded the song and separated the vocals with UVR (I used Voc ft->karaoke2->no echo no reverb for the separation process)
Afterwards, the song was inferred using the Mejiro Mcqueen model downloaded in advance through AICoverGen and combined with the instrumental version to complete the song.
The problem is:
The sound quality of the finished song is significantly lower than that of the original song. What is certain is that there were no problems with the audio files, at least as far as I listened, up to the point where I separated the vocals with UVR. After inferring the song, the sound quality is very poor compared to the original, and strange mechanical sounds and echoes are mixed in.
I tried to infer again by adjusting various parameters in the GUI, but it did not solve the problem. I would like to know how to improve this.
Thank you for reading my long comment. I searched for a solution to this all day, but couldn't find it. I really want to find a solution
wispy burrow
it does have assets but no weights il try and look through rq lol but i have no ...
Your ModelName.pth files are in the logs folder.
There is no weights folder anymore in Applio...
rigid dirge
ight thanks
also btw do you train the index first or the pth ;-;
or does it matter
subtle holly
Is there a way to do the ai cover thing on mobile because one of my friends showed me that he did it all on mobile but i was a bit confused. Is there a website i need to visit?
rare gobletBOT
Ayo? @subtle holly level 1 !!!
rigid dirge
just wondering cuse i dont think i did it right lmfao
proud elbow
also btw do you train the index first or the pth ;-;
doesnt matter but index training is much shorter than the actual training
wispy burrow
also btw do you train the index first or the pth ;-;
It doesn't matter but training the Index doesn't take long, So I do that first.
rigid dirge
doesnt matter but index training is much shorter than the actual training
thanks cause i did train the index first lol
yeah idk reading this i think i did it wrong 😭
cause its not apearing at all lmfao
wispy burrow
cause its not apearing at all lmfao
If you're talking about the index training, yeah it doesn't show anything for some reason.
rigid dirge
no the index apears fine
its just the pth files
just the pth files that dont apear lmfao
wispy burrow
yeah idk reading this i think i did it wrong 😭
Also try to avoid spaces in the model Name.
wispy burrow
no the index apears fine
I meant nothing is shown in the cmd.
rigid dirge
sorry had to crop it at the top but these are the parts the index files are shown
and here
wispy burrow
just the pth files that dont apear lmfao
Your .pth should in the logs.
Not Logs/ModelName.
wispy burrow
sorry had to crop it at the top but these are the parts the index files are show...
That's not training the index.
That's generating the files for training the index.
wispy burrow
but the index was made i think right this is it 😭
Yes, the added_IVF* is your Index.
wispy burrow
Hmmm...
wispy burrow
Can you check in downloads/Applio3.1.1
rigid dirge
its just this unfortunaly, tbh i did think cause its in downloads it might not work (idk why but maybe lol)
thats the folder i mean lol
wispy burrow
Where in the world are your .pth files.
You should have a bunch of them.
rigid dirge
thats what i be saying lmfao
i did try to edit this actually lol but it said smth about permissions so idk if i want to mess with this
but this is the only thing i can think off
tbh im just gonna hope this helps this is just before the whole index part
rigid dirge
Where in the world are your .pth files.
You should have a bunch of them.
btw do you use apollio or smth diffrent cause whatever your using i might just switch to that if its easier to use in the way to make model cause i just wanna make this model lmfao😭
faint bear
tbh im just gonna hope this helps this is just before the whole index part
I'm getting this exact same error for DAYS
Was just about to post about this error page. Again, it's been like this for days.
Like, everything goes smoothly and from the second I click training, the thing works for under 20 seconds before showing this message and stopping. It doesn't ever turn red. It has a checkmark as if it did its job
rigid dirge
oh really
rare gobletBOT
Ayo? @rigid dirge level 5 !!!
rigid dirge
tbh i left it afk so i have no idea how long it took actually
maybe we have the same issue
ionic venture
the first time i download and it run normally but now this seccond time it just white windows screen on the app
how can i fix it ?
i can't send image to show the proplem
wispy burrow
the first time i download and it run normally but now this seccond time it just ...
You need to be level 1 before you can add images here.
faint bear
Heres a screenshot.
I literally never came to this part of the Discord because I always assumed things would fix themselves after a few days, and they did, but now it's just... always like this
rare gobletBOT
Ayo? @faint bear level 1 !!!
ionic venture
You need to be level 1 before you can add images here.
how chat more ?
wispy burrow
how chat more ?
Indeed.
proper shale
but to anyone who uses appolio does anyone wknow where the pth files go to
assets > weights folder iirc
wispy burrow
assets > weights folder iirc
Applio no longer has a weights folder.
rare gobletBOT
Ayo? @ionic venture level 1 !!!
wispy burrow
its called models then?
They appear in logs
rancid aurora
Guys is there anyway i can make ai cover in mobile
wispy burrow
tbh i left it afk so i have no idea how long it took actually
Could you train something new for 1 Epoch?
rigid dirge
Could you train something new for 1 Epoch?
il try
rancid aurora
Guys is there anyway i can make ai cover in mobile
Please can someone answer
low shard
Guys is there anyway i can make ai cover in mobile
yes, check in the cloud folder inside the rvc folder in our docs https://docs.aihub.wtf, you will find some google colab to do it on mobile
rancid aurora
yes, check in the cloud folder inside the rvc folder in our docs https://docs.ai...
Why
rancid aurora
Nvm nvm it works
rare gobletBOT
Ayo? @rancid aurora level 3 !!!
low shard
yea, read inisde of any of the cloud one and see the pros and cons depending on your needs
brittle wing
Is there a way i could make my ai cover better and correct it.
rare gobletBOT
Ayo? @brittle wing level 2 !!!
brittle wing
what nvidia gpu generate ai cover in less than 20 seconds?
like the minimum
the cheapest nvidia gpu that generates them less than 20 seconds
hard gyro
hello i have nvidia gpu wich one i use
tardy elbow
hey, anyone can help me? I can't download my model at step 2
I'm getting this error
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
https://drive.google.com/file/d/1ljl5bcdKZfyeHhm7sktIVHw6SYY9J9fP/view?usp=sharing
proud elbow
hello i have nvidia gpu wich one i use
gpu-cuda
proud elbow
hey, anyone can help me? I can't download my model at step 2
I'm getting this er...
outdated colab
vast fractal
What rvc fork should I use for a Nvidia gpu, does anyone know
low shard
hey, anyone can help me? I can't download my model at step 2
I'm getting this er...
Use only the updated Colabs inside the cloud folder inside the rvc folder in our docs https://docs.aihub.wtf
low shard
What rvc fork should I use for a Nvidia gpu, does anyone know
I mean any forks you want as long as the Gpu is good, i suggest either applio or mainline
azure marshBOT
RVC Guides (How to Make AI Cover)
tardy elbow
Use only the updated Colabs inside the cloud folder inside the rvc folder in our...
Got it , Thank you
low shard
yw
brittle wing
i have a question where do i set Epochs?
i have been looking around the app but nothing named epochs there
or is it just something that is said normally
proper shale
i have a question where do i set Epochs?
You can't set epochs - the only one that can do so is the model creator.
brittle wing
ohhh
rare gobletBOT
Ayo? @brittle wing level 1 !!!
subtle holly
Is there a way to do the ai cover thing on mobile because one of my friends showed me that he did it all on mobile but i was a bit confused. Is there a website i need to visit?
proper shale
Is there a way to do the ai cover thing on mobile because one of my friends show...
yeah, u can do ai covers via colab on phone
just follow the stuff in https://docs.aihub.wtf :)
subtle holly
I tried colab but it fails when i put in the huggingface rcv link
The .zip links
I get to there and it wont work
I will try again just to be sure.
proper shale
I tried colab but it fails when i put in the huggingface rcv link
Maybe you were using an old colab
meager ingot
subtle holly
Maybe you were using an old colab
I run Gradio but no links appear/open
proper shale
subtle holly
Ok!
low shard
no russian version sadly
Btw shouldnt we delete the other languages docs from the kalo bot
Anyways only the english and spanish one eixst now
rigid dirge
Could you train something new for 1 Epoch?
tbh i dont think its working lol, what do you use btw i'll just use that
rigid dirge
lets hope this works lol
wispy burrow
tbh i dont think its working lol, what do you use btw i'll just use that
I have used multiple ones.
older versions of of Applio.
Don't recommend those.
I've used a couple of the regular RVC ones.
And Mangio.
But that is a bit old and outdated.
There's also Ilaria-RVC.
But that's a beta version.
Haven't trained with that yet.
shadow lynx
anyone have a tutorial for how can i make someone voice an ai model
rare gobletBOT
Ayo? @shadow lynx level 1 !!!
rigid dirge
I have used multiple ones.
older versions of of Applio.
Don't recommend those.
...
aw thank you pookie wookie epic 💀
lmfao anyway thanks il try rvc locally and see how that works
vagrant bolt
it still wont work on discord
vague cypress
past dew
I entered it thinking that Turks had it, but when I looked, no sound model appeared.
rigid dirge
i clicked train index is this normal lol
rich crest
so im trying to download a model onto the doc but it keeps giving me an error help would be much appriciated
allright nevermind i was just using an old colab
steady depot
how to fix this?
JSONDecodeError Traceback (most recent call last)
<ipython-input-6-6026005ebaf8> 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)
dont mind my pfp i will change it later
i found it on youtube
ye i think no one will help
rare gobletBOT
Ayo? @steady depot level 1 !!!
rigid dirge
just wondering where does the pth go
i just clicked train model
and its done finally lol
but this is what happened lol
idk if its right lol
😭
subtle holly
https://colab.research.google.com/drive/1mHKTGH5e3SAyDSBss1KtiYRbDdQzwSMs try th...
The page crashed and then said that there is no inferenced running
proper shale
The page crashed and then said that there is no inferenced running
Hmmmmm... maybe try the Applio colab then;
https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb
proper shale
is that normal or not just wondering cause its stopped training
if u did that manually then yeah
else idk
it seems to have started well tho
subtle holly
How do i acsess the rcv thing step be step
Bc i cant remember what i did the first time
proper shale
How do i acsess the rcv thing step be step
in Applio's case: https://docs.applio.org
actually nvm no colab guide there
but eh, everything's pretty straightforward so
tired badger
yo
hexed ocean
i need help
tired badger
i downloaded the ai voice changer but when i extract it it dont show what the video showed to me
hexed ocean
it tells me i need input device
whats that
can anyone go in a call wiht me and i do live stream
and u guys show me
hexed ocean
zyft go ina call wiht me qand tell me why this sisnt wornking
rare gobletBOT
Ayo? @hexed ocean level 1 !!!
rigid dirge
if u did that manually then yeah
else idk
yeah i did it manually instead of one step thing
but then it did one espoch then stopped
tired badger
i downloaded the ai voice changer but when i extract it it dont show what the vi...
.
proper shale
huh, weird
low shard
it tells me i need input device
i think thats the wokada realtime voice changer, ask in #1192011222023950368 or #🔍│help-w-okada for that pls
rigid dirge
i just wanna make a ai model bruh lol
tired badger
someose help me 😭
low shard
i downloaded the ai voice changer but when i extract it it dont show what the vi...
i think thats the wokada realtime voice changer, ask in #1192011222023950368 or #🔍│help-w-okada for that pls
help rvc is for rvc which is the tool for voice conversion of an audio file, mostly used for ai covers
tired badger
i think thats the wokada realtime voice changer, ask in <#1192011222023950368> o...
no one answer in these chats, ill just give up
low shard
and unfortunately i cant help yall cus i dont use voice changers cus i never call
proper shale
i downloaded the ai voice changer but when i extract it it dont show what the vi...
u can copy paste the stuff that appears in command prompt if thats the case
and, u can check the W-Okada guide in https://rentry.co/VoiceChangerGuide too
frank root
rigid dirge
btw for making models what type of mp3 flac or wav file should the audio be
and it should be in one folder right
the file
low shard
LMAO
proper shale
btw for making models what type of mp3 flac or wav file should the audio be
wav is better but anything works
rigid dirge
ok cause when i followed the tutorial it was ment to be a flac
if anyone would be willing to help me for the model thing considering it obv isnt working lol
ive tried appolio and that did the same as well
faint bear
is absolutely no one else having the isuue i reported earlier today?
proper shale
is absolutely no one else having the isuue i reported earlier today?
yeah I don't think so, I'd suggest reaching out to Kit Lemonfoot in that case tbh
rigid dirge
is absolutely no one else having the isuue i reported earlier today?
whats your issue again
faint bear
Like that one ou posted just seconds before I did. Something about me clicking on train and it not even starting out properly
I'll get the image again
proper shale
it's in RVC disconnected, so i doubt it is the same issue
faint bear
This is so weird
At this point I just think google's angry at me or something
rigid dirge
you think it being in downloads could have smth to do with it
like would it need to be somewhere else maybe
idk
proper shale
you think it being in downloads could have smth to do with it
ehhh idk
rigid dirge
thats all i can think off
faint bear
rigid dirge
cause the index works but its just the pth
faint bear
Yep
proper shale
cause the index works but its just the pth
MIGHT HAVE FOUND THE ISSUE
SPACE ON DATASET FOLDER NAME
😭
rigid dirge
SPACE ON DATASET FOLDER NAME
WOULD THAT ACTUALLY BE IT
ISTG IF THATS IT IM ENDING IT ALL LMFAO
proper shale
I THINK SO???
proper shale
I'd suggest doing the whole preprocessing/feature extraction part again after u rename the folder
rigid dirge
yeah dw
proper shale
I hope it's that, bcuz there's nothing else wrong
rigid dirge
im just gonna relaunch it
rigid dirge
I hope it's that, bcuz there's nothing else wrong
dw i hope so
proven hill
I hope it's that, bcuz there's nothing else wrong
what’s happening
wispy burrow
WOULD THAT ACTUALLY BE IT
I doubt it.
While RVC doesn't like spaces or special characters.
RVC no longer uses your audio files in your Dataset in your dataset after step 2a (process data).
from that point onwards it uses the audio files it created from it in your logs/ModelName folder.
proper shale
I doubt it.
While RVC doesn't like spaces or special characters.
RVC no longer u...
But if it can't preprocess the files in the first place... it screws up
Feature extraction pointed out that some files were NaN, meaning it didn't even feature extract properly
rare gobletBOT
Ayo? @lusty parrot level 1 !!!
lusty parrot
hey listen, I'm currently making a model of Lamb's voice from the game Cult of The Lamb. I downloaded an 8-minute video from YT, and I'm currently removing music, noise, etc. I wouldn't have written this if it weren't for the fact that there are different voices in this recording, but I want there to be only one. If someone could help me, I would be extremely grateful. Additionally, I will send additional information
proper shale
But idk tbh. Maybe it could also be GTX 1660 and its CUDA issues...
wispy burrow
But if it can't preprocess the files in the first place... it screws up
Didn't notice that!
proper shale
hey listen, I'm currently making a model of Lamb's voice from the game Cult of T...
That's tough tbh, if there isn't any clips with one single voice
proven hill
Feature extraction pointed out that some files were NaN, meaning it didn't even ...
fun fact thats actually why ilaria rvc haves one click training now
proper shale
There isn't exactly separation of multiple voices (that actually works rn)
proper shale
fun fact thats actually why ilaria rvc haves one click training now
wha
proven hill
hey listen, I'm currently making a model of Lamb's voice from the game Cult of T...
use Karaoke 5 HP
it separates back vocals from main vocals
or at least try to do the job, only tool
proven hill
not perfect but get the job done
lusty parrot
There isn't exactly separation of multiple voices (that actually works rn)
You can simply remove and separate other voices from Lamb's voice (i.e. sometimes he speaks and sometimes someone speaks), but I don't have much time and I don't have that much experience in it because I just started my adventure with it
proven hill
for this you need time, a lot
proper shale
ye
proven hill
so prepare yourself
lusty parrot
use Karaoke 5 HP
Thanks, I'll use it if necessary
sinful nexus
-colab
flat pond
whats a good voice swapper i can use really quickly free
rigid dirge
But idk tbh. Maybe it could also be GTX 1660 and its CUDA issues...
you think so?
im just about to test the new way lol
rigid dirge
I doubt it.
While RVC doesn't like spaces or special characters.
RVC no longer u...
or this ;-;
proper shale
im just about to test the new way lol
let's pray it isn't
odd shale
looking good so far lol
Which batch size are you using?
Don't use a high one since your GPU isn't that powerful
inner scaffold
in discord when i say something it does my voice than does voicew changer how can i fix it
rare gobletBOT
Ayo? @inner scaffold level 1 !!!
inner kayak
im trying to use the changer in discord, but its coming across real static-like, any help?
hard shadow
So how does one get like a personlized code to not get time out for RVC disconnect?
rigid dirge
why not colab?
Last I checked google cracked down on that and everyone had to use local lol
rigid dirge
why not colab?
is it faster btw and if so is there a tutorial for it
rare gobletBOT
Ayo? @rigid dirge level 6 !!!
pastel oak
is it faster btw and if so is there a tutorial for it
Check pinned message, open docs ai hub
rvc disconnected should be the one for training models on colab
rigid dirge
Check pinned message, open docs ai hub
ight il check now thanks
rigid dirge
rvc disconnected should be the one for training models on colab
big help thanks cause its working its just slow af 😭
pastel oak
check under the category "cloud"
rigid dirge
i just want a faster one ngl
wispy burrow
I'm liking Ilaria RVC Beta so far.
The training tab is compact and straightforward.
azure marshBOT
RVC Guides (How to Make AI Cover)
mighty crane
interested in using Tekken 4's voice announcer
but i don't know how to
work out the kinks
so im unsure how to work this out
im not even sure if im in the right place to ask questions.
wispy burrow
so im unsure how to work this out
That is a GPT-SoVITS Model.
It's TTS (Text-to-Speech)
You are going to need GPT-SoVITS to be able to use that.
There is a section about GPT-SoVITS in this guide.
https://docs.aihub.wtf/tts/gpt-sovits/
mighty crane
Hmm, thanks.
rare gobletBOT
Ayo? @mighty crane level 1 !!!
mighty crane
there's like
different beta
@wispy burrow which one, 0217.7z, beta0306.7z, or beta0306.zip?
rocky radish
How can I make an AI voice cover for a song? I have already gotten a vocals track seperated from the intruments without echo or reverb, but I don't know how to add the AI audio onto it.
wispy burrow
<@548973417458696212> which one, 0217.7z, beta0306.7z, or beta0306.zip?
Hmm...I think 0306 is the newest version.
mighty crane
there's two
.7z and .zip
it did say in the GPT website to "download the prezip of the latest version"
"advisable to use 7-zip"
and i assume .7z is 7-zip.
wispy burrow
and i assume .7z is 7-zip.
Not sure if they were referring to the 7-zip program or the 7z (7-Zip) file.
tawny quarry
is there any cypher voice i can get?
wispy burrow
this thing says "one day remaining" when it says extract
One day!? Wow.
Is it that large!?
loud solar
yello i need some assistuance
rvc keeps sending me this error message, it worked fine a week ago but not anymore. yes the files are set to be accessible by anyone with the link
`Access denied with the following error:
Cannot retrieve the public link of the file. You may need to change
the permission to 'Anyone with the link', or have had many accesses.`
normal isle
using the first link here: https://canary.discord.com/channels/1159260121998827560/1163883017446629508
brittle wing
AI voice changer echoing everything that plays in the background, even when other people speak, it plays back their voice in the AI. This applies for everything that contains audio, is there any way how to fix this?
faint moon
real time i face delay voice chnager help me
rare gobletBOT
Ayo? @faint moon level 2 !!!
turbid crescent
This keeps happening on step 2
JSONDecodeError Traceback (most recent call last)
<ipython-input-3-4cb009f9c861> 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)
faint moon
https://discord.com/channels/1159260121998827560/1159290139609137264/12253285608...
real time i face delay voice changer help me
azure marshBOT
RVC Guides (How to Make AI Cover)
low shard
real time i face delay voice changer help me
Ask in #🔍│help-w-okada or #1192011222023950368
low shard
How can I make an AI voice cover for a song? I have already gotten a vocals trac...
Check in our docs https://docs.aihub.wtf
rocky radish
ok thanks!
low shard
how do you all make ai covers? 😭
Check https://docs.aihub.wtf you can find info here on how to
If smt is not understood js ask
latent kettle
how do you all make ai covers? 😭
use RVC to do that
winter shell
I can't hear my voice can some1 help me w/ that?
latent kettle
I can't hear my voice can some1 help me w/ that?
are you using real-time voice changer?
winter shell
are you using real-time voice changer?
I already fixed it
brittle wing
anyone willing to help me I am a bit confused xd
steady depot
https://discord.com/channels/1159260121998827560/1159290139609137264/12253285608...
thanks
winter shell
I can't turn on my microphone on discord is it because CABLE output?
rare gobletBOT
Ayo? @winter shell level 1 !!!
winter shell
I can't turn on my microphone on discord is it because CABLE output?
@latent kettle Hey can u help me with this?
pastel oak
I can't turn on my microphone on discord is it because CABLE output?
on wokada:
Input: Microphone
Output: Cable input
on discord:
Input: cable output
Output: Headphones
If you have monitor to your headphones and you hear something on the app but nothing on discord, then thats VB-Cable being bad again and you need the other virtual cable
pastel oak
anyone willing to help me I am a bit confused xd
just ask
brittle wing
setting up stuff I would like to go over stuff but idk where to start xd
winter shell
on wokada:
Input: Microphone
Output: Cable input
on discord:
Input: cable outpu...
I don't have headphones menu on my output
pastel oak
I don't have headphones menu on my output
or the default or the speakers or earphones, same thing
pastel oak
setting up stuff I would like to go over stuff but idk where to start xd
rvc or wokada
winter shell
btw what's wokada?
pastel oak
realtime voice changer
theres also a voice changer in rvc
i assumed you use wokada but correct me if kot
not*
brittle wing
rvc
rare gobletBOT
Ayo? @brittle wing level 1 !!!
brittle wing
@pastel oakno clue what I am doin just trying to get shit to work xd
winter shell
I alr resetted my voice settings on dc but it's still not working
pastel oak
I alr resetted my voice settings on dc but it's still not working
you still havent answered if you can hear yourself through the app when you set "monitor" to your speakers
pastel oak
It is plain simple and structured
I cant help you if youre not willing to do the bare minimum
pastel oak
Then send me a screenshot
rugged stirrup
I ran into this error in the new beta the Ov2 32k pretrain is there in the pretrained v2 folder
zinc sleet
why whenever i speak it feels like it is speaking with itself lol
keen vault
hi
low shard
why whenever i speak it feels like it is speaking with itself lol
i think thats for realtime wokada voice changer, maybe you should ask im #🔍│help-w-okada or #1192011222023950368
keen vault
in easygui , I downloaded my model via drive but regardless how many times I refresh the model voice My model does not show in there