How can I tell when it's overtraining?
#✨│ai-help
onyx shale
haughty pulsar
-overtrain
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
onyx shale
So when it gets wavy that's when you know?
onyx shale
It's using words I don't understand
haughty pulsar
It's using words I don't understand
https://rvc-docs.github.io/Docs/rvc-resources/epochs-overtraining--tensorboard/ might be more user friendly
onyx shale
Thanks
snow hazel
onyx shale
<https://rvc-docs.github.io/Docs/rvc-resources/epochs-overtraining--tensorboard/...
So from what I see, if it gets down far and then goes back up, it's time to end it?
Does this count?
snow hazel
I don’t know how to get rid of that pop up
onyx shale
@haughty pulsar Is this OT?
haughty pulsar
Does this count?
you can just let it train for a while then check when it overtrains and choose the model closest to the point where it overtrains
make sure you check this before you train
haughty pulsar
onyx shale
The guide didn't say anything about that
I'm sure it's not the biggest deal
If it wasn't in the guide and it's important then that's an issue with the guide
rare gobletBOT
Ayo? @onyx shale level 4 !!! 
haughty pulsar
i should prob add that to the guide
rare gobletBOT
Ayo? @haughty pulsar level 8 !!!
haughty pulsar
ask em to
onyx shale
So what now? Do I just have to restart the whole process?
rain schooner
Hello does anyone remember the name of a app that does all the voice changes all in one program downloaded on a computer, i have it downloaded I just cant remeber the name
haughty pulsar
just let it train tbh
onyx shale
Okay
shut willow
Why do people keep saying “*** epochs”
haughty pulsar
Why do people keep saying “*** epochs”
epochs is the amount of times it goes over the dataset
shut willow
Huh ok
haughty pulsar
more does not mean better
shut willow
Really?
onyx shale
how many epochs did you set?
1000 was the highest I could set it, the guide said 2000 but it's okay
haughty pulsar
1000 was the highest I could set it, the guide said 2000 but it's okay
what are you on right now?
and what save frequency
haughty pulsar
what epoch are you at right now?
onyx shale
Idk
Alright I have to go afk what do I do? @haughty pulsar I don't wanna overtrain it
haughty pulsar
Alright I have to go afk what do I do? <@1200898742283407372> I don't wanna over...
you could stop training then restart training with the option ticked ig
haughty pulsar
if you have the option ticked it saves a model for each checkpoint
so yea
just check tensorbored
find the closest to overtraining
then use that
onyx shale
Okay
onyx shale
find the closest to overtraining
Right now or when I get back?
haughty pulsar
when you get back
haughty pulsar
i guess you trained 1000 epoch already
haughty pulsar
it should stop already
haughty pulsar
the pth will be saved in weights
index in logs
if you clicked train index
also training files will be in logs
onyx shale
Where is it?
haughty pulsar
weights is in assets folder
onyx shale
Where's the AI voice model?
haughty pulsar
you use the pth and index
onyx shale
I found the .pth so how do I index it?
hoary ravine
does anyone know why i cant use crepe or harvest
haughty pulsar
I found the .pth so how do I index it?
did you click train index?
haughty pulsar
does anyone know why i cant use crepe or harvest
what fork are you on
onyx shale
did you click train index?
When and where?
haughty pulsar
what does that meamn
what are you using
hoary ravine
the latest
onyx shale
should be on the left of train model
Well I did it if it's on the guide, otherwise no
I really have to go
hoary ravine
i only got crepe tiny and rmvpe onnx
hoary ravine
what are you using
MMVCServerSIO_win_onnxdirectML-cuda_v.1.5.3.8a.zip
haughty pulsar
your on amd right?
hoary ravine
yea
haughty pulsar
directml can only use those ones
onyx shale
click this
I already closed it
hoary ravine
directml can only use those ones
damn
rare gobletBOT
Ayo? @hoary ravine level 1 !!!
haughty pulsar
I already closed it
just put same settings and instead of train model do train index
haughty pulsar
i also hear a lotta static for no reason
hoary ravine
directml can only use those ones
is
MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.17b.zip
better?
umbral rose
hey me again
haughty pulsar
MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.17b.zip
what gpu do you have?
umbral rose
is there any way to make my own voice
hoary ravine
2080max
brave garnetBOT
⠀
Download for Nvidia GPUs 
Version 17b cuda
Download for AMD GPUs 
Version 17b directml
Download for Intel GPUs 
Version 17b directml
Download for Mac 
Version 17b Mac
⠀
Made with love by Ilaria 💖
haughty pulsar
use nvidia
haughty pulsar
is direct mxl bad
its for amd and intel
spiral osprey
yes, but if I make a voice with 100 epouchs will it still be good?
dawn nebula
yes, but if I make a voice with 100 epouchs will it still be good?
rvc very bad with long datasets, try choose the best clips and the dataset should be no longer than 30 mins
you can experiment with 2h dataset, i wouldnt reccomend tbh
shrewd siren
when ever i try to download a model and it isnt working. im very stupid so its prob just me being an idiot.
this is what it reads
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.
You may still be able to access the file from the browser:
https://drive.google.com/uc?id=1HvOv52-z7R4uUpwhXUtrZzvRMe3YDk_6
Error: Could not find a valid .pth file within the extracted zip.
If there's an error above this talking about 'Access denied', try one of the Alt URLs in the Google Sheets for this model.
dawn nebula
when ever i try to download a model and it isnt working. im very stupid so its p...
try a different url i think
shrewd siren
try a different url i think
that is the download link that the site uses the actual imputed link was this https://drive.google.com/file/d/1HvOv52-z7R4uUpwhXUtrZzvRMe3YDk_6/view
and i cant find a different url
dawn nebula
download it to your device, extract then upload individually
shrewd siren
uploaded to what?
rare gobletBOT
Ayo? @shrewd siren level 1 !!!
dawn nebula
uploaded to what?
upload to colab
rugged kindle
whats the best settings for 3090?
dawn nebula
whats the best settings for 3090?
okada?
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 💖
dawn nebula
yes
shrewd siren
how do i uplaod the model to colabs
?
ok so i made my own google drive with this link
https://drive.google.com/drive/folders/1GdADJGoDsfIqxbP4O-kXPFumJ9PWXIhA?usp=drive_link
and now im getting this error
ModuleNotFoundError Traceback (most recent call last)
<ipython-input-3-1bad439f8678> in <cell line: 4>()
2 #@markdown Link the URL path to the model (Mega, Drive, etc.) and start the code
3
----> 4 from mega import Mega
5 import os
6 import shutil
ModuleNotFoundError: No module named 'mega'
fucking hell imma go make some fucking mac n cheese
finite ferry
LOL
@shrewd siren how is your mac n cheese
how do i ... use a model that ive downloaded from the QuickWick Music AI sets? am i able to use it in a DAW like ableton or logic or do i need a separate program to use it
is there a good tutorial i could watch or something, i just have the .json and .pth file from the download
molten pecan
is there a good tutorial i could watch or something, i just have the .json and ....
Uhmm, you don't use voice models on DAWs
You use the AI app call RVC, input the vocals, input the model, do the inference and then put the output audio to your DAW to mix it
finite ferry
purr thank u so much that’s exactly the info i needed
rare gobletBOT
Ayo? @finite ferry level 1 !!!
static tapir
I was dreaming to be a Model Maker for AI Voices. But I'm having trouble with the approving thing cause I have to be a Model Maker QC in order to do it.
spiral osprey
rvc very bad with long datasets, try choose the best clips and the dataset shoul...
dataset with 30 minutes the voice still becomes robotic. I was recommended to do 1 hour or 1 hour and 30 minutes.
molten pecan
dataset with 30 minutes the voice still becomes robotic. I was recommended to do...
Jesuschrist
Who recommended that?
If the voice is robotic is prob due to bad training or bad dataset
brittle wing
help when i try to use ov2super on applio rvc it says OSError: [WinError 17] The system cannot move the file to a different disk drive: 'C:\Users\aaavi\AppData\Local\Temp\gradio\939079a537714b9bdbf476cf48a02d83de8e1806\f0Ov2Super40kG.pth' -> 'D:\Applio-3.0.2\rvc\pretraineds\pretraineds_custom\f0Ov2Super40kG.pth'
molten pecan
help when i try to use ov2super on applio rvc it says OSError: [WinError 17] The...
@peak tusk
peak tusk
<@744968865481752626>
no soy atención al cliente de applio
molten pecan
no soy atención al cliente de applio
Quien lo es?
peak tusk
nadie
brittle wing
bruh
rugged kindle
whats the difference between rmvpe and rmvpe_onnx
also which one should i choose considering rmvpe is suggested for rtx series
waxen niche
can i train voice in german still would there be any problems with doing other languages in the ai?
brittle wing
can i train voice in german still would there be any problems with doing other l...
no
wide ingot
File "/home/thetrustedcomputer/Software/Git/RVC-Applio/rvc/infer/infer.py", line 10, in <module>
from rvc.lib.infer_pack.models import (
File "/home/thetrustedcomputer/Software/Git/RVC-Applio/rvc/lib/infer_pack/models.py", line 5, in <module>
from . import modules
File "/home/thetrustedcomputer/Software/Git/RVC-Applio/rvc/lib/infer_pack/modules.py", line 8, in <module>
from torch.nn.utils.parametrizations import weight_norm
ImportError: cannot import name 'weight_norm' from 'torch.nn.utils.parametrizations' (/home/thetrustedcomputer/Software/Git/RVC-Applio/python-3.10/lib/python3.10/site-packages/torch/nn/utils/parametrizations.py)
Applio 3 with ROCm doesn't work for me anymore; it breaks inference. GPU is an 8GB 5500 XT. Normal install works. After installing the ROCm build of PyTorch, this error popped up trying to infer audio.
GUI shows up and everything; I can't create AI covers, train models, and all that.
finite ferry
Im using Ilaria RVC and i successfully downloaded the model, but no matter how many times i refresh on the inference page, the model wont come up
harsh plaza
hey uh i have a data set but i think its low quality how do i fix that
rare gobletBOT
Ayo? @harsh plaza level 1 !!!
finite ferry
harsh plaza
thanks but does it talk about cleaning up the data set
molten pecan
thanks but does it talk about cleaning up the data set
Yeah absolutely
harsh plaza
ok
now the thing is theres multible speakers in the audio how do i isolate only one voice
molten pecan
now the thing is theres multible speakers in the audio how do i isolate only one...
I explained how
Please read thoroughly before asking questions
harsh plaza
ok
finite ferry
Weird. Did you try rebooting the colab space?
yea it still didnt work even after i closed it and restarted the system
strange tinsel
Which colab link can I use to clone the voice of an acapella I already have on my cell phone with the voice of an artist?
rare gobletBOT
Ayo? @strange tinsel level 2 !!!
molten pecan
yea it still didnt work even after i closed it and restarted the system
does it only happen with that model?
can you send me the link?
molten pecan
Which colab link can I use to clone the voice of an acapella I already have on m...
Ilaria RVC
there's a guide for it here
strange tinsel
Thank you
strange whale
With rejekts easyGUI training on google colab, is there some sort of timeout I should be looking out for with a free version of google colab?
As in, a max training time I should be aware of?
finite ferry
can you send me the link?
molten pecan
https://huggingface.co/QuickWick/Music-AI-Voices/resolve/main/Doja%20Cat%20163.2...
does it only happen with this model?
finite ferry
i havent tried any other models bc this is the only one im interested in using, but let me see
yea it happens with other models as well
downloads on the system but then doesnt show up when i go to inference
molten pecan
downloads on the system but then doesnt show up when i go to inference
weird
what are you naming the models?
do they contain spaces/special characters?
are you clicking the upper pink Refresh button?
wide ingot
From what I read, PyTorch versions 2.0 and above will work with RX 5000 cards if compiled against ROCm 5.2. There seems to be a regression starting ROCm 5.3. The stable precompiled ones are for PyTorch 1.13. The nightly ones are for 2.0.0.
sand iris
From what I read, PyTorch versions 2.0 and above will work with RX 5000 cards if...
You should be able to work around that... If you get a local build of the entire toolchain
This is the important bit : https://github.com/ROCm/Tensile/commit/efbe0c0caeedcf03776ae4e6a0067d88e38da211
This very patch is what allows it to work again
finite ferry
do they contain spaces/special characters?
nope, all one word as simple as possible and yes clicking the upper pink refresh button before trying
rare gobletBOT
Ayo? @finite ferry level 2 !!!
timber girder
Hey guys I have a curious question about making ai covers.
I was just wondering, if someone might know...
Let's say I record my voice, singing a song.
Should I add pitch correction before or after converting my voice to the ai character?
I wonder which one will give me the best-sounding results.
wide ingot
You should be able to work around that... If you get a local build of the entire...
So what should I do? Firstly, I grabbed the sources of ROCm 5.2.3 from GitHub. Now I need the sources of PyTorch, right?
gusty kestrelBOT
molten pecan
Hey guys I have a curious question about making ai covers.
I was just wondering,...
before, tbh
add pitch correction before
wide ingot
I'm gonna try to build them from source since Applio 3 broke GPU acceleration for me.
azure marshBOT
RVC Guides (How to Make AI Cover)
compact moon
Anyone have the rvc download
eager ruin
brittle wing
I can’t hear myself is that normal ?
When making a dataset do you want one big file for example say it's 30 minutes, should it be one big 30 minute .wav file or should it be multiple different clips? I find a lot of confusion with that. Is there a limit on how long you should make the dataset?
half cove
When making a dataset do you want one big file for example say it's 30 minutes, ...
It can be one big file
Rvc splits it anyway
brittle wing
Thank you
drifting birch
anybody got easy collab ?
half cove
anybody got easy collab ?
Inference or training
half cove
Inference or training
Making covers or making models
drifting birch
Inference or training
inferrence just to use models
half cove
You can try the ila rvc colab then ig
drifting birch
You can try the ila rvc colab then ig
hers dont work
half cove
half cove
hers dont work
What part?
half cove
half cove
I can’t hear myself is that normal ?
What do you have the monitor option set as
I'm assuming your using real-time
spiral osprey
Who recommended that?
a moderator from Brazil who I think left moderation.
spiral osprey
If the voice is robotic is prob due to bad training or bad dataset
the bigger the voice model the better it is. It depends if it's worth 1 hour.
half cove
the bigger the voice model the better it is. It depends if it's worth 1 hour.
The quality of data matters more than length
timber girder
before, tbh
alright
thanks
molten pecan
the bigger the voice model the better it is. It depends if it's worth 1 hour.
not entirely true.
- quality is a big determining factor too
- usually anything further than 40 doesn't add much
spiral osprey
The quality of data matters more than length
then. I use the universal vocal remover and use kim vocal 2 to separate the instrumental from the vocal then I take the vocal and make the voice model.
molten pecan
then. I use the universal vocal remover and use kim vocal 2 to separate the inst...
https://rvc-docs.github.io/Docs/ here I included a guide for cleaning datasets & using audio isolation apps
half cove
then. I use the universal vocal remover and use kim vocal 2 to separate the inst...
For separating instruments I'd use mdx23c
My Internet being bozo rn
wispy lion
when i train any sample in the end they always sound the same , i tried to train different samples (characters) and in the end its always that same default voice , idk if i'm making any mistake , i'm using The Mangio RVC Fork to train them
i would appreciate if anyone knows how to fix my problem
fast nymph
So I have the .zip folder for the voice I wish to use, but I do not have the link that ewnables me to download the model onto RVC GUI, how do I get my previous voice back> They updated the EasyGUI and my voice disappeared
spiral osprey
For separating instruments I'd use mdx23c
Can I use it by voice?
half cove
Can I use it by voice?
By voice?
rare gobletBOT
Ayo? @spiral osprey level 8 !!!
fast nymph
how can I download a model via link if the model hasnt been uploade to AI Hub?
copper sierra
@molten pecan so i finally began to do the guide
https://rentry.org/ilarvc_inf_guide
what if i wanted to run it locally?
still osprey
why cant i change the the "GPU" selection to my GPU it only have CPU
sonic kiln
People who, when opened to download a voice, are saying that the username or password invalidates what I have to do?
pale nimbus
what's the best and easiest RVC converter I can use without downloading the application?
molten pecan
<@791013455251374151> so i finally began to do the guide
https://rentry.org/ilar...
That one is outdated. And you can't run Ilaria RVC locally
Here's the up-to-date one
bleak sparrow
:Step 3. Start the GUI, then open the public URL." Link: https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sharing#scrollTo=F6zbsuNs6xZF Hello, I'm a beginner in the field of AI. I'm following the steps to create, but I can't understand "Step 3", can someone explain how to do it?
snow hazel
Ok now UVR isn't even trying to open on my mac. It just bounces for a bit on the bottom thna it stops.
real fable
what words is best for making my microsoft sam model?
copper sierra
That one is outdated. And you can't run Ilaria RVC locally
oh ok, thanks (:
pale nimbus
what's the best and easiest RVC converter I can use without downloading the appl...
anyone?
shrewd siren
<@1053168040029720576> how is your mac n cheese
it was good. sorry for the delay got kicked.
rare gobletBOT
Ayo? @shrewd siren level 2 !!!
brittle wing
what software do i use in order to use the ai voice files
so could i use them for tts
rare gobletBOT
Ayo? @brittle wing level 1 !!!
covert turret
How do I install rvc in Paperspace?
molten pecan
https://rvc-docs.github.io/Docs/
@brittle wing check the Ilaria TTS page
covert turret
Check the paperspace page
Thanks, but I'm a bit confused after doing "a", do I use "make install.py" after the wget?
molten pecan
Thanks, but I'm a bit confused after doing "a", do I use "make install.py" after...
Yes
Idk much about Paperspace
So if you get stuck, ask @odd shale
brittle wing
could i get a custom ai voice in ilaria tts
molten pecan
I don't use Paperspace so idk
covert turret
lol 
molten pecan
Wait for Leo to go online tomorrow
brittle wing
where is the page at, i am having trouble finding the page
molten pecan
where is the page at, i am having trouble finding the page
Which one?
molten pecan
Ilaria TTS
Did you use the search bar already?
novel pike
okay so using the google collab ive encountered an issue, ive trained my own model and used it successfull a few times with the same interface, now its saying that i may have an access denied error but i have it set to anyone with link any ideas?
molten pecan
brittle wing
what search bar, is this in the discord server
brittle wing
thanks
molten pecan
yw
novel pike
plz help
rare gobletBOT
Ayo? @novel pike level 1 !!!
novel pike
Yoo I just came from the last server
Could anyone help me troubleshoot this access denied error
brittle wing
how do i get the custom voice now
north moss
novel pike
I switched to mega nz as the upload source and now this is happening wtf
I ran it again and I think it worked so hopefully this helps
covert turret
Wait for Leo to go online tomorrow
Could make it run, thanks
rare gobletBOT
Ayo? @covert turret level 2 !!!
molten pecan
Could make it run, thanks
Phew, you are welcome. There's a mangio guide too btw
covert turret
Phew, you are welcome. There's a mangio guide too btw
I ended up using the instructions over at the Mangio RVC Fork github
Only thing I still want to know is how to get the OV2 pre-train there (if it is not already default as is with the colab)
pale nimbus
what's the best and easiest RVC converter I can use without downloading the appl...
can someone please help me
spiral osprey
When using mdx23C in the uvr I can't select the gpu conversion, is there a problem when rendering the audio?
My audio is already vocal, do I need to use the mdx23C to remove the instrumental or can I skip this part.
rare gobletBOT
Ayo? @brittle wing level 2 !!!
real fable
could i upload
you can prob upload it to #1159514067187277865
brittle wing
i am using Ilaria TTS
spiral osprey
When using mdx23C in the uvr I can't select the gpu conversion, is there a probl...
pls help-me.
brittle wing
how do i get a custom voice
molten pecan
how do i get a custom voice
https://rvc-docs.github.io/Docs/ everything you need to know is in the site
Be sure to read it thoroughly before asking pls
brittle wing
i did, but it did not say anything about uploading custom voices
molten pecan
i did, but it did not say anything about uploading custom voices
Check the Ilaria RVC page
And the Voice Models page
brittle wing
is it that
spiral osprey
When using mdx23C in the uvr I can't select the gpu conversion, is there a probl...
Can anyone help me understand?
molten pecan
My audio is already vocal, do I need to use the mdx23C to remove the instrumenta...
Uhh mdx23c is for separating vocals/instrumentals yeah
If your sample is purely vocals only, then you don't need to use mdx23c
spiral osprey
If your sample is purely vocals only, then you don't need to use mdx23c
ok thanks
split grail
cris mj?
spiral osprey
after UVR-BVE do I need to use UVR-DeNoise?
summer tiger
does anyone know what to do when my stuff just keep saying the date and then connet sid? it just keeps repeating and i cant use any of my models
dawn nebula
what stuff
spiral osprey
after UVR-BVE do I need to use UVR-DeNoise?
pls help-me
summer tiger
what stuff
my mmvcserversio
dawn nebula
after UVR-BVE do I need to use UVR-DeNoise?
what dataset are you extracting
spiral osprey
what dataset are you extracting
adam gontier vocalist for three days grace
spiral osprey
what dataset are you extracting
Could you answer my question?
dawn nebula
Could you answer my question?
sorry i was busy
dereverb, denoise, deecho
like you can able to hear those
and then choose the corresponding
one
ashen brook
how do I use premade zip files for the easy gui colab thing
spiral osprey
dereverb, denoise, deecho
what?
dawn nebula
there are models to dereverb, denoise and deecho
spiral osprey
like you can able to hear those
Just tell me if I should use UVR-DeNoise? or not.
dawn nebula
if there is noise then yeah
spiral osprey
if there is noise then yeah
what kind of noise?
dawn nebula
idk, i havent see ur dataset
** Please read this tutorial before asking any questions**
spiral osprey
idk, i havent see ur dataset
I'm using exactly this site.
real fable
what is the thing for rvcDisconnected v2 so i can sumbit
dawn nebula
I'm using exactly this site.
sry that was just a pin, not directing to you
spiral osprey
sry that was just a pin, not directing to you
so sorry
spiral osprey
after UVR-BVE do I need to use UVR-DeNoise?
someone help-me?
real fable
submit what
this. im using rvcDisconnect v2 https://dorker.lifes-a-me.me/6qMmNTavl
(if the embed will load)
dawn nebula
this is a pitch extraction algorithm
dawn nebula
not disconnected
real fable
so how do i get that
dawn nebula
you put sth when you train
idk what you put
probs rmvpe? idk
did u change any settings
-colab
real fable
nope i only followed this guide: https://docs.google.com/document/d/1XuxQYiqEhYrdYeCZRRLrmV_ciMKo0bV-jTCGHu_-5Cc/edit?pli=1
dawn nebula
then u choose rmvpe
azure marshBOT
RVC Guides (How to Make AI Cover)
spiral osprey
someone help-me?
?
thin stump
someone help-me?
That really depends on your audio. If you used BVE and still hear harmonies, it could either be inadequate aggression settings. Denoise can strip away some things if you aren't careful
I make covers only btw
sand iris
So what should I do? Firstly, I grabbed the sources of ROCm 5.2.3 from GitHub. N...
The 5.2 build should work. Try not having rocm installed on the host side at all, apparently it can override what the torch build gets shiped with
thin stump
what are epochs?
Julia compiled a useful guide for everyone https://aihubdocs.github.io/en/rvc-resources/epochs-overtraining--tensorboard/#epochs-overtraining--tensorboard
dense pagoda
where can i change the epochs on the voice changer?
and where can i change batch size and dataset lenght
polar mirage
i cant hear myself
gray condor
@odd shale
odd shale
<@628603171735994379>
?
gray condor
can u help me w okada?
odd shale
can u help me w okada?
I will try.
gray condor
im usin google to run it
nd also
these r my settings
how do i start it
bc i clicked start
chose a model
but i cant hear myself
on discord
do i need to change my input settings on discord?
odd shale
do i need to change my input settings on discord?
Yes you need.
gray condor
i have vb-cable audio bc one of the helpers told me i need it
gray condor
Yes you need.
to what?
rare gobletBOT
Ayo? @gray condor level 3 !!!
gray condor
odd shale
i have vb-cable audio bc one of the helpers told me i need it
And that's true.
gray condor
okay
ive been tryna fix this issue myself for like 2 days but i cant work myself around it
so i js came now to ask 4 actual help
what do i do w the vb audio cable
odd shale
what do i do w the vb audio cable
Maybe you should select it on Discord audio settings of your account
gray condor
okay
done
do i need to do something inside of the vb audio cables settings
or do i js leave it
odd shale
or do i js leave it
Leave it like that but play around with input and output device settings till it works for you
gray condor
i started it now
i set the input to my main mic
but the input in discord is gonna be vb audio cable
or do i need to set the input for both chrome and discord?
half cove
or do i need to set the input for both chrome and discord?
On okada put input mic output vb cable then on discord put input vb cable output headphones
gray condor
On okada put input mic output vb cable then on discord put input vb cable output...
so in discord ill set my input to vb audio cable and also set my output in discord as vb cable too?
@half cove
half cove
so in discord ill set my input to vb audio cable and also set my output in disco...
On discord only input is vb cable
flint geyser
hmm...
gray condor
i did everything they said but it isnt hearing anything
azure marshBOT
🎙️ Realtime Voice Changer (Local)
flint geyser
https://rentry.co/VoiceChangerGuide try following this guide.
if this does not work, ping me again.
shrewd siren
more of a technical question, im using the desktop version, what does crepe hop do?
gray condor
@flint geyser
i did everything but
instead of picking up the sounds from my microphones its just picking up the sounds of my computer
T
WAIT
ITS WORKING NOWW
flint geyser
alr good
flint geyser
more of a technical question, im using the desktop version, what does crepe hop ...
The crepe hop is like the amount of time it takes to switch from one pitch to another, so if you have a shorter crepe hop, its more early when trying to hop from pitch to pitch, but slower in terms of generation.
rare gobletBOT
Ayo? @flint geyser level 9 !!!
flint geyser
DAMN you have a tesla t4 in your computer??
sand iris
DAMN you have a tesla t4 in your computer??
that would be colab, no ?
sand iris
somebody help meh
what do you need ?
flint geyser
that would be colab, no ?
wait there is w-okada colab??
sand iris
yeah
flint geyser
oh
crystal rover
Like it doesnt speak at all
crystal rover
Im terrible at explaining things 😔
flint geyser
its ok
just try following the guide in depth and then message just in case it isnt working
thin stump
command fail
stiff helm
Lol
thin stump
I don't have the perms lol
thin stump
Yeah, what's the question?
stiff helm
Ok, thank you
Im a youtuber, who wants to transition to a new AI voice. I was previously using ElevenLabs
I want to switch for two reasons. (1) elevenlabs is expensive, and (2) I would like something more human
Are these things I can reasonably find on this server?
sand iris
What were you using elevenlabs for ?
Here you'll find mainly voice conversion, and maybe just a little bit of TTS
stiff helm
What were you using elevenlabs for ?
TTS, for my yt channel
sand iris
Oh. There's two ways to go about this :
- Use any existing TTS solutions, and then run voice-conversion to the desired voice. This will not mimic the way the person talks unfortunately
- https://github.com/RVC-Boss/GPT-SoVITS this exists, but everything is being ironed out, and we haven't seen it much yet
stiff helm
Oh. There's two ways to go about this :
- Use any existing TTS solutions, and th...
Which solution do you think would have a more human sounding voice?
sand iris
The pipeline way can work great on some voices, but especially the non-expressive ones.
GPT-SoVITS might sound more "alive", but I haven't extensively tested it
stiff helm
The pipeline way can work great on some voices, but especially the non-expressiv...
Thanks. Is there anywhere I can go to here an example pipeline voice over?
sand iris
I guess by installing the thing ?
Not sure there are demos out there
stiff helm
Lol ok
Got it
@sand iris also wanted to mention that I’m looking for something that is reliable (never randomly freaks out)
warm sparrow
well if non expressive sounds better the corpse husband voice i just made might be perfect
stiff helm
What TTS is used for these? I know tortoise sometimes has issues
sand iris
https://github.com/IAHispano/Applio/
Here's the voice-conversion program which has builtin TTS
sand iris
<@310840858767523846> also wanted to mention that I’m looking for something that...
In earlier versions there were multiple TTS backends available, not sure how it is today, haven't tried 3.0. As long as the input TTS model behaves correctly, and the voice conversion model is good, it should work as expected
I've only done testing against OpenAI's TTS and then running voice-converison, and it worked alright. Really monotone, but sounded great.
stiff helm
I've only done testing against OpenAI's TTS and then running voice-converison, a...
Are there options if I want something more expressive?
sand iris
I wouldn't know
stiff helm
I wouldn't know
I see I see, thank you. This discord is mainly for RVC, right?
sand iris
Yeah
sand iris
RVC does voice conversion
Speaker A input audio -> RVC -> Speaker B converted output audio
stiff helm
RVC does voice conversion
Speaker A input audio -> RVC -> Speaker B converted ou...
I see, so this would be useful in making one voice sounds like another
But for mimicking natural human speech, a good TTS as based is needed
near thorn
Downloading a converted vocal but i got this saying cant download
peak urchin
Downloading a converted vocal but i got this saying cant download
.htm is a file type that download the entire page
Not the audio itself
near thorn
ahh i see
autumn laurel
uh hey how u use the voice changer in discord call?
near thorn
how do i fix it? i click the three dots and then clicked download like the youtube tutorial said
autumn laurel
and why when i use thae voice changer my res is 0 ms
near thorn
It is a shortform of .html if you wondered
ohh i see
autumn laurel
anyone?
autumn laurel
ok ill just see a yt vid
peak urchin
uh hey how u use the voice changer in discord call?
Go the Discord setting and search for microphone setting
Change it to the one w-okada convert in.
autumn laurel
huh i use virtual cable
autumn laurel
hey
why when i click start
the voice changer just show some japanese writing
and the rest age is white
peak urchin
Screenshot it
autumn laurel
i cant
near thorn
i tried again from beggining and it works now thx alot for ur time harbs! i appreciate it!
danke sehr
autumn laurel
how?
autumn laurel
that
it just turn white
after i click start
i use kobo
then i click start
should i re start the voice changer or smth?
peak urchin
I think I also encountered it before.
autumn laurel
so what do i do?
peak urchin
What file did you click to start the voice changer
autumn laurel
file explorer
then i click start_http
then i wait for a few sec
then the voice changer appear
peak urchin
Oh god i forgot what is inside the file of w-okada
💀
It been a while since I last used w-okada
near thorn
lol
autumn laurel
is there other way to start except pressing the start_http?
peak urchin
Through Terminal / cmd
peak urchin
Hmm... It's not the file issue
I am wrong 
I think the next thing it may caused is that
autumn laurel
should i just re start the voice changer?
peak urchin
Pytorch and ONNX thing
peak urchin
should i just re start the voice changer?
Yeah you should try
autumn laurel
ok
near thorn
ich sehe du lernen deutsch und japanisien harbs?
wie gut ist dein duetsch darf ich frage
autumn laurel
it works
rare gobletBOT
Ayo? @autumn laurel level 3 !!!
autumn laurel
let me try starting it
peak urchin
ich sehe du lernen deutsch und japanisien harbs?
Beginner, 90% of time focusing on Japanese
autumn laurel
still turn white
near thorn
Beginner, 90% of time focusing on Japanese
ahhh alles klar kumpel viel gluck!
sand iris
see if anything is visible at http://localhost:18888/
Sometimes the user interface just freaks out like that for no reason at all
peak urchin
I know Indonesian / Malay so I am fine
autumn laurel
oh okay
sand iris
Can you scroll up to the top of the cmd window and re-screenshot ? Actual error might be hidden there
undone pond
hi when i open the http it opens voice-changer-native-client on a blank screen any idea on how to fix?
peak urchin
still turn white
@undone pond like this?
autumn laurel
or is it ur whole scrren?
undone pond
<@888542293303377971> like this?
yea but at the top it says voice-changer-native-client
autumn laurel
same then
peak urchin
Er. Whole screen becoming blank is a bigger issue than just your cilent application
autumn laurel
Can you scroll up to the top of the cmd window and re-screenshot ? Actual error ...
that?
peak urchin
I gonna gone for tuition
autumn laurel
ok
peak urchin
So yeah I am not gonna answer soon
autumn laurel
@sand iris
sand iris
Do you know your computer specifications ? (what's inside of it ?)
near thorn
what specifications does ur pc have? wat intel is it wat gpu n etc
rare gobletBOT
Ayo? @near thorn level 3 !!!
undone pond
autumn laurel
what specifications does ur pc have? wat intel is it wat gpu n etc
um inte; core i3 8th gen
is this enough?
is it that potato to use voice changer or smth?
sand iris
It is probably too potato yeah
autumn laurel
oh ok i guess🥲
so i delete it only?
i use almost a day to download it but it just turn out i cant use it ;-;
tight leaf
Google Colab Free Create AI cover
sand iris
i use almost a day to download it but it just turn out i cant use it ;-;
There's always this if you want to give it a shot. Instead of running on your PC, it's running on one of Google's servers
autumn laurel
huh
neon forge
Can AMD graphics cards be used?
copper pebble
Can AMD graphics cards be used?
yh
solar jacinth
mangio /logs
craggy brook
Which application was used to change the sound and where to download it?
brittle wing
heyy so i tried it with the original pretrain and the graph looks the same 
its so flat
how am i supposed to know if its overtraining now
its worse than yesterday's i fear
azure marshBOT
RVC Guides (How to Make AI Cover)
brittle wing
i am having some issues installing RVC-Boss GPT-SoVITS
azure marshBOT
snow hazel
So it finally showed up in privacy setting in Mac but idk why it’s still not showing up.
zinc nacelle
Is there Applio 3.0.4 colab?
rare gobletBOT
Ayo? @zinc nacelle level 1 !!!
proper shale
Is there Applio 3.0.4 colab?
https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb should be up to date
snow hazel
Idk why my mac is doing this
twin cave
Quick question. With Applio, where do I put my index and my .pth file for a voice model I want to use for covers?
fast agate
how do i use a voice model in real time
solar jacinth
https://colab.research.google.com/github/iahispano/applio/blob/master/assets/App...
omg i didnt know that there is a appilo colab
thanks
snow veldt
Q: wtf are thse tags
zinc anchor
Q: wtf are thse tags
I recognise you from the Source Separation server
snow veldt
I recognise you from the Source Separation server
you do 😛
zinc anchor
But yeah, those were made to help browse the models
snow veldt
But yeah, those were made to help browse the models
be great if the uploaders would use the right tags, in what world is that an artist or celeb
zinc anchor
Oh yeah that's true
Thanks for pointing it out
I don't think a Pikmin 4 character is an artist
snow veldt
np, im finding it hard to just look at music artists and will def point out more if i see it cuz yeah, there are ways to take that stuff out of the listings but not if they're doing that
winter eagle
im trying to train a model and i have the model in tensorboard and i tried training the model 2 times but nothing shows up in tensorboard
help
proper shale
@winter eagle check cmd
covert turret
im trying to train a model and i have the model in tensorboard and i tried train...
you have to click refresh i think
proper shale
does it have any tracebacks or anything
winter eagle
<@757619446763487378> check cmd
sr 40 something doesnt match sr 32
i think its because i changed the thingy to 32
because i was accidentaly training
on
i think 40 because it says that
proper shale
then try training on 40
winter eagle
sample rate
winter eagle
then try training on 40
ok
well training i dont see anything in tensorboard
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
Traceback (most recent call last):
File "multiprocessing\process.py", line 315, in _bootstrap
File "multiprocessing\process.py", line 108, in run
File "C:\Users\kryst\AppData\idk\RVC1006Nvidia\infer\modules\train\train.py", line 233, in run
net_g.module.load_state_dict(
File "C:\Users\kryst\AppData\idk\RVC1006Nvidia\runtime\lib\site-packages\torch\nn\modules\module.py", line 2041, in load_state_dict
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for SynthesizerTrnMs768NSFsid:
size mismatch for dec.noise_convs.0.weight: copying a param with shape torch.Size([256, 1, 80]) from checkpoint, the shape in current model is torch.Size([256, 1, 64]).
size mismatch for dec.ups.0.weight_v: copying a param with shape torch.Size([512, 256, 16]) from checkpoint, the shape in current model is torch.Size([512, 256, 20]).
size mismatch for enc_q.pre.weight: copying a param with shape torch.Size([192, 1025, 1]) from checkpoint, the shape in current model is torch.Size([192, 513, 1]).
what does this mean
uh
proper shale
honestly idk
winter eagle
why does this gotta be so complicated, cant this be automated
half cove
hmm
solar jacinth
really never seem that before
winter eagle
really never seem that before
guess i cant make the model
rare gobletBOT
Ayo? @winter eagle level 4 !!!
winter eagle
Ayo? <@757619446763487378> level **4** !!! <:lfg:1159355870119993496>
shut the fuck up
proper shale
u can do via Colab
winter eagle
u can do via Colab
No
half cove
im really confused with that error
solar jacinth
im really confused with that error
yeah never seem that before
winter eagle
and why not
i think its better localy
proper shale
im really confused with that error
maybe they changed configs whule middle of training?
half cove
maybe they changed configs whule middle of training?
yea thats what im thinking
the keys dont match
winter eagle
like sample rate
well i only set it to v1 and then to v2 back
half cove
while training?
winter eagle
while training?
no, after i did the index and stuff
and after 2 attempts of training
proper shale
no, after i did the index and stuff
yea that might've been the problem
winter eagle
try using rvc2 before doing all that lol
i already closed everything
also what rvc2
im using mainline
or smth
half cove
model arch
proper shale
im using mainline
yea that has rvc v2
covert turret
Two questions
- How do I use OV2 pre-trains while running the regular Mangio-RVC-Fork?
- How do I select rmvpe for pitch extraction? (when using the cli at "go extract-feature" I don't see the option and obviously using rmvpe as an argument won't work)
glad zealot
Two questions
1. How do I use OV2 pre-trains while running the regular Mangio-RV...
- On the gui just change the path of the original pretrains to be the 0v2
- Idk
half cove
Two questions
1. How do I use OV2 pre-trains while running the regular Mangio-RV...
put the pretrains in your pretrained_v2 folder then here put the names
covert turret
1. On the gui just change the path of the original pretrains to be the 0v2
i'm using the cli lol, I tried installing the ov2 and selecting OV2 as an argument seems to work
half cove
mangio cli?
covert turret
yep
half cove
ive never seen it
covert turret
the reason i decided to use the cli is because the gui confused me more lol idk why
covert turret
put the pretrains in your pretrained_v2 folder then here put the names
what i did essentially was download the ov2 pre-trains using git pull in their own "pretrained_ov2" directory
but afaik the pretrain-type is different from the model architecture
glad zealot
Is there anything in the cli command that have .pth?
covert turret
no?
glad zealot
Hmm ye not sure I never seen it ngl
covert turret
everything it asks are arguments that go like this e.g.
extract feature = [Model Name] [GPU Slot] [GPU Threads To Use] [Has pitch guidance?] [f0 Method (harvest, dio, crepe)] [crepe hop length] [version of pre-train models]
half cove
you could maybe just swap the names and use ov2 as the base pretrain ig https://cdn.discordapp.com/emojis/1141554800471834756.webp?size=48&name=RatCar&quality=lossless
covert turret
lol
so i guess i should replace the v2 downloads with the ov2?
but keep the v2 directory?
rare gobletBOT
Ayo? @covert turret level 3 !!!
half cove
i guess that would work
spiral osprey
someone help me?
half cove
maybe change the names to the normal ones
proper shale
someone help me?
https://aihubdocs.github.io/en new link
half cove
someone help me?
they changed name i think
half cove
https://aihubdocs.github.io/en new link
yep
covert turret
maybe change the names to the normal ones
ig that would work, now i have to figure out how to use rmvpe instead of crepe (:
half cove
does it not have an rmvpe option
spiral osprey
Does anyone know if I need to select something to remove the backing vocals?
covert turret
does it not have an rmvpe option
no lol idk why, the options are: pm, harvest, dio, crepe
proper shale
Does anyone know if I need to select something to remove the backing vocals?
should already do it by default ig. but u need to check the outputs
half cove
no lol idk why, the options are: pm, harvest, dio, crepe
where did you even download it from
do you have a github link?
covert turret
Does anyone know if I need to select something to remove the backing vocals?
the vocals only option are the main vocals and instrumental are the backing vocals usually
shadow flax
hey, im using okada AI-voices and im kinda having trouble (AMD), my gpu barely gets used, cpu gets maxxed and its kinda working but a bit choppy
proper shale
hey, im using okada AI-voices and im kinda having trouble (AMD), my gpu barely g...
have u converted to onnx?
solar jacinth
helppppp
shadow flax
have u converted to onnx?
onnx? sorry im not quite sure i know what that is
proper shale
the vocals only option are the main vocals and instrumental are the backing voca...
nope, vocals only would be background vocals in this case.
solar jacinth
my wifi got unplugged while im trainning a model
half cove
https://github.com/Mangio621/Mangio-RVC-Fork
like how do you use as cli
covert turret
nope, vocals only would be background vocals in this case.
lol, confusing, i also used uvr and there was a model that was the opposite lmao
proper shale
onnx? sorry im not quite sure i know what that is
you need to convert to onnx, as that's better for AMD GPUs.
this section walks ya through converting to onnx: https://rentry.co/VoiceChangerGuide#uploading-models-amdintel-gpus-and-fixing-its-laggy-issue
spiral osprey
should already do it by default ig. but u need to check the outputs
the outputs are correct, I'm saying the boxes to select, do I need to select them all and leave only the vocal just deselect?
solar jacinth
poopy sukuna san
proper shale
you need to convert to onnx, as that's better for AMD GPUs.
this section walks ...
do keep in mind that DirectML version will still use CPU
half cove
interesting
solar jacinth
i got the d222.pth and d22.pth
proper shale
the outputs are correct, I'm saying the boxes to select, do I need to select the...
just keep them as is
spiral osprey
just keep them as is
So I don't need to select the instrument in this case?
covert turret
interesting
i liked applio's gui but this "regular" fork's gui is so confusing i gave up and used the cli
proper shale
So I don't need to select the instrument in this case?
if you only wanna have the leads sure, u can select that
better to have both anyway
covert turret
So I don't need to select the instrument in this case?
have both and check which one is the one you want after it finished inferring
half cove
i liked applio's gui but this "regular" fork's gui is so confusing i gave up and...
tbh its not that confusing after you use it a few times
shadow flax
you need to convert to onnx, as that's better for AMD GPUs.
this section walks ...
ok ty im at it
spiral osprey
have both and check which one is the one you want after it finished inferring
I have to wait 48 hours.
covert turret
tbh its not that confusing after you use it a few times
maybe, but i was used to the rvc disconnected colab that i couldn't figure it out lol
covert turret
I have to wait 48 hours.
for 1 song?
half cove
did you read our guides?
spiral osprey
if you only wanna have the leads sure, u can select that
better to have both an...
so I can only mark the instrumental to remove the backing vocals right?
covert turret
did you read our guides?
yup
proper shale
so I can only mark the instrumental to remove the backing vocals right?
it will remove the backings regardless, just use the instrumental output
proper shale
dont need to check it
half cove
its also made to be user friendly
spiral osprey
for 1 song?
yeah
covert turret
so I can only mark the instrumental to remove the backing vocals right?
the inference time is the same regardless, so why not keep both
half cove
just leave it as it is lol
u will have both audios
delete the one you dont want after
covert turret
it should give enough information
it does, but the thing is that the fork's ui is a bit different but the cli has the same terms as the guides, and i'm used to cli, so it's not like i really need the gui
spiral osprey
In this case, I leave them all marked, in this case the instrumental and the sample mode 30s, since it says to leave them all marked except vocals, correct?
half cove
sample mode is to test
proper shale
In this case, I leave them all marked, in this case the instrumental and the sam...
no, u can keep them all unmarked lol
half cove
just leave it as it was before
spiral osprey
no, u can keep them all unmarked lol
hum,ok thanks.
proper shale
yw :)
cobalt trout
yall now how to use text to speech with a custom model
proper shale
Sadly there is no direct TTS with RVC models, so you'll have to do TTS elsewhere and convert that with RVC
glad zealot
ilaria rvc has tts coversion
cobalt trout
have a link?
old hawk
i was using this colab to make ai covers and it was working fine 5 minutes ago but now it randomly says that no nvidia gpu driver is detected, i have reinstalled my driver and restarted my pc but it doesnt seem to fix the issue
colab link: https://colab.research.google.com/github/ardha27/AICoverGen-NoUI-Colab/blob/main/CoverGen_No_UI.ipynb#scrollTo=B3BqnUoePVGd
console logs v
molten pecan
i was using this colab to make ai covers and it was working fine 5 minutes ago b...
Try using it in another browser
glad zealot
covergen colab doesnt use your gpu or your drive
molten pecan
Maybe you ran out of usage time
proper shale
i was using this colab to make ai covers and it was working fine 5 minutes ago b...
check your runtime type on the top right
glad zealot
also you might be out of gpu time thtas why its saying those
old hawk
check your runtime type on the top right
i reconnected various time but it doesn't seem to work, one thing i found odd was that it didnt ask me to run some of the cells but i was confused considering i exited the colab before reconnecting to them
oh wait
sorry i misread
my runtime is set to t4 gpu
hot lagoon
Shoud I covert both the back and led for this one?
old hawk
Maybe you ran out of usage time
how can i check my usage time
glad zealot
you sadly cant
old hawk
oh yeah my usage time ran out
is it like daily or permanently
daily usage or lifetime usage
glad zealot
12 hours
glad zealot
aight
molten pecan
i was gonna goto sleep anyway ill take it
You can switch to another google acc though
hot lagoon
Shoud I covert both the back and led for this one?
Also wich one would be beter for dataset
old hawk
You can switch to another google acc though
oh
rare gobletBOT
Ayo? @old hawk level 5 !!!
old hawk
thanks ill do that tomorrow if i runout again
crystal lion
is anyone bored and wants to help me clean up this mp3? so the voice is clearer
maybe someone knows better what to do i tried with uvr but idk
proper shale
is anyone bored and wants to help me clean up this mp3? so the voice is clearer
try using bandit plus on that
its in mvsep btw
crystal lion
try using bandit plus on that
what and where is that
crystal lion
its in mvsep btw
you got a link? for a colab or sum?
crystal lion
thank youuuu
proper shale
yw
crystal lion
yw
im doing something wrong i put the file in there 😭
can you do it for me rq if you got a minute time
nvm i got it
azure marshBOT
RVC Guides (How to Make AI Cover)
young halo
Yo, how can I fix the artifacting on my model? I've been always told to clean my dataset when I try to upload my models but I really don't know what that means lol
young halo
https://aihubdocs.github.io/en/vocal-isolation--datasets/datasets--how-to-clean-...
tysm!
half cove
tysm!
id go to the uvr/mvsep part for things like removing noise
or separating instrumentals
glacial loom
yo can someone give me a collabs link
google docs
cause I don't know where is my old one
wide ingot
The 5.2 build should work. Try not having rocm installed on the host side at all...
So I need to uninstall it from the package manager?
rare gobletBOT
Ayo? @wide ingot level 13 !!!
sand iris
So I need to uninstall it from the package manager?
I guess you can try both. If it works with it still installed there's no harm
wide ingot
I do know that AMD supports multiple installs of ROCm.
sand iris
Yet they shove everything along with the torch releases to avoid issues
young halo
What's the recommended amount of Epochs for a 10 minutes long dataset so it's not overtraining?
snow hazel
Ok I've decided to do MVSEP one last time and I've been on queue for 3 hours now and it hasn'r changed a number.
hushed osprey
I'm having a bit of trouble with a Virtual Audio Cable, the one by Muzychenko.
Tried uninstalling the Trial version (because of the constant TRIAL voice playing in the background) and it didn't uninstall properly.
Now when I try to install another version of the product, the lite version, it comes up with an error.
Any help?
azure marshBOT
cobalt trout
@glad zealot how to upload models
rare gobletBOT
Ayo? @cobalt trout level 1 !!!
worldly finch
Need help
I want to make voice ai with my own voice
So i m gonna use easygui
But cant go futher
Need help
@half cove
i m stuck here
worldly finch
yes
half cove
-rvc
half cove
original rvc
worldly finch
that messes stuff up
shall i try on hard drive
azure marshBOT
half cove
shall i try on hard drive
if its on one drive it will mess things up
half cove
yea
finite ferry
is .json an index file to put in the logs folder of Mangio or should i have a different file type ?
bc the download for the doja cat ai model i found on hugging face only had a json and pth file
storm kernel
The link is public and was working before how can I fix this please ?
finite ferry
can anybody who can actually use mangio or ilaria successfully pls dm me ? i literally need like a 4 second audio clip converted with an ai model (i can send the link) but it won’t work for me no matter what i do and i’ve used macbook, desktop, EVERYTHINGGG 😭😭😭
worldly finch
I rezipped in harddrive
storm kernel
-rvc
tepid rampart
there's a way to use huggingface models on elevenlabs?
rare gobletBOT
Ayo? @tepid rampart level 1 !!!
copper sierra
how do you do rvc voice changing?
tepid rampart
there's a way to use huggingface models on elevenlabs?
?
worldly finch
<@936673139419664414>
i m trying to train from these localhost
half cove
i m trying to train from these localhost
use the one i recommended before
this doesnt have rmvpe
storm kernel
-help
worldly finch
use the one i recommended before
which one did u recommended?
half cove
which one did u recommended?
-links
azure marshBOT
Suggestions for @worldly finch
half cove
original rvc
half cove
what gpu?
half cove
what gpu exactly
storm kernel
hello @acoustic scarab I'm running RVC in local, do you know where do I have to put my models at ? They're in the models folder but nothing shows up
worldly finch
what gpu exactly
nvidia gforce gtx
rare gobletBOT
Ayo? @worldly finch level 4 !!!
worldly finch
thats only what i know...
half cove
go task manager
acoustic scarab
hello <@362876249032359939> I'm running RVC in local, do you know where do I hav...
models in weights, indexes in logs
if its rvc webui
acoustic scarab
have geforce gtx nvidia
right click the start menu, then go to device Manager, open GPUs and
and tell us what's inside
half cove
dont think gtx cards can train
storm kernel
models in weights, indexes in logs
it worked, thank you
half cove
unless im being goofy
acoustic scarab
dont think gtx cards can train
starting from GTX 750 ti they can
half cove
starting from GTX 750 ti they can
understandable
cobalt trout
How to download model huggingface links dont work
acoustic scarab
it's all up to CUDA compute capability which must be higher than 3.7 iirc
half cove
right click the start menu, then go to device Manager, open GPUs and
he could just go task manager lol
worldly finch
nvidia gforce gtx 1650 @half cove