-colb
#✨│ai-help
pastel oak
does it usually take long omg
depends on your upload speed
brittle wing
What's so vits svc fork? is it any different from rvc/ rvc forks like mangio?
lament eagle
depends on your upload speed
Dude my upload speed is good
brittle wing
What's so vits svc fork? is it any different from rvc/ rvc forks like mangio?
@sand iris
pastel oak
Dude my upload speed is good
if its colab check if youre still connected, else reupload again
south tree
got this error:
Dataset folder already found. Wiping...
Temporary folder already found. Wiping...
Archive: /content/drive/MyDrive/rvcDisconnected/Gummi.zip
End-of-central-directory signature not found. Either this file is not
a zipfile, or it constitutes one disk of a multi-part archive. In the
latter case the central directory and zipfile comment will be found on
the last disk(s) of this archive.
unzip: cannot find zipfile directory in one of /content/drive/MyDrive/rvcDisconnected/Gummi.zip or
/content/drive/MyDrive/rvcDisconnected/Gummi.zip.zip, and cannot find /content/drive/MyDrive/rvcDisconnected/Gummi.zip.ZIP, period.
Sanitizing...
Dataset Type: Audio Files (Single Speaker)
NameError Traceback (most recent call last)
<ipython-input-14-8b01eb60dc85> in <cell line: 58>()
60 #Move everything to /dataset/experiment_name/.
61 print("Dataset Type: Audio Files (Single Speaker)")
---> 62 expDir=os.path.join(final_directory, experiment_name)
63 os.makedirs(expDir, exist_ok=True)
64 for r, _, f in os.walk(temp_directory):
NameError: name 'experiment_name' is not defined
rare gobletBOT
Ayo? @south tree level 1 !!! 
old oracle
-colab
half cove
depends on your upload speed
if its public gradio then its gunna be slow no matter what
half cove
got this error:
Dataset folder already found. Wiping...
Temporary folder already...
have you ran this cell?
arctic cargo
Does RMVPE use hop length parameter?
If yes, does this translate to the crepe_hop_length parameter in the mainline rvc inference?
south tree
have you ran this cell?
yepp
half cove
If yes, does this translate to the `crepe_hop_length` parameter in the mainline ...
mainline does not have crepe hop length iirc
arctic cargo
mainline does not have crepe hop length iirc
Mangio sorry*
half cove
Mangio sorry*
crepe hop length for mangio crepe
cobalt zealot
Hey, im trying to make an rvc and was wondering how many epochs i should run it for?
lament crest
excuse me help guys, i made ai voice model file but i cant export it for huggingface link...
pastel oak
Hey, im trying to make an rvc and was wondering how many epochs i should run it ...
install tensorboard, it will tell you when your voice model is overtraining and thats when you stop
https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/
lament crest
install tensorboard, it will tell you when your voice model is overtraining and ...
excuse me how can I make huggingface link for voice model?
rare gobletBOT
Ayo? @lament crest level 1 !!!
brittle wing
not retrain but resume training
half cove
not retrain but resume training
1 sec ima do something
half cove
not retrain but resume training
put all file with all the training files in logs then make a new cell with this in it
@title set variables
model_name = "Darwin" # @param {type:"string"}
sample_rate = "40k" # @param ["32k", "40k", "48k"] {allow-input: false}
rvc_version = "v2" # @param ["v2", "v1"] {allow-input: false}
f0method = "rmvpe" # @param ["pm", "dio", "crepe", "crepe-tiny", "harvest", "rmvpe"] {allow-input: false}
hop_length = 128 # @param {type:"slider", min:1, max:512, step:0}
do the same as you did before
then click train with the save things you did before just with higher epoch
short ridge
-colab
civic meteor
this isnt really rvc but uhhh https://discord.com/channels/1159260121998827560/1220397683962024057
brittle wing
put all file with all the training files in logs then make a new cell with this ...
these files are on drive, not in the colab notebook files
half cove
these files are on drive, not in the colab notebook files
make sure its on the colab notebook files
in logs
as the modelname
azure marshBOT
RVC Guides (How to Make AI Cover)
brittle wing
make sure its on the colab notebook files
should I upload them into colab from drive or my pc?
half cove
into colab from drive would be faster
tbh i might revamp the colab more and add import model and shi
brittle wing
I'll say no
half cove
on the left side of your colab
click the file icon
as seen in the image
you should be connected to drive already i assume
spice needle
2024-03-21 18:06:03.242396530 [E:onnxruntime:Default, provider_bridge_ort.cc:1548 TryGetProviderInfo_CUDA] /onnxruntime_src/onnxruntime/core/session/provider_bridge_ort.cc:1209 onnxruntime::Provider& onnxruntime::ProviderLibrary::Get() [ONNXRuntimeError] : 1 : FAIL : Failed to load library libonnxruntime_providers_cuda.so with error: libcufft.so.10: cannot open shared object file: No such file or directory
2024-03-21 18:06:03.242425295 [W:onnxruntime:Default, onnxruntime_pybind_state.cc:861 CreateExecutionProviderInstance] Failed to create CUDAExecutionProvider. Please reference https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirementsto ensure all dependencies are met.
[~] Applying DeReverb to Vocals...
brittle wing
then the RVC_Backup folder
half cove
lily is your model correct?
brittle wing
yes
spice needle
2024-03-21 18:06:03.242396530 [E:onnxruntime:Default, provider_bridge_ort.cc:1548 TryGetProviderInfo_CUDA] /onnxruntime_src/onnxruntime/core/session/provider_bridge_ort.cc:1209 onnxruntime::Provider& onnxruntime::ProviderLibrary::Get() [ONNXRuntimeError] : 1 : FAIL : Failed to load library libonnxruntime_providers_cuda.so with error: libcufft.so.10: cannot open shared object file: No such file or directory
2024-03-21 18:06:03.242425295 [W:onnxruntime:Default, onnxruntime_pybind_state.cc:861 CreateExecutionProviderInstance] Failed to create CUDAExecutionProvider. Please reference https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirementsto ensure all dependencies are met.
[~] Applying DeReverb to Vocals...
someone can help me?
half cove
will it dissappear out of my drive?
yes
half cove
I can't drag and drop it
eh?
brittle wing
half cove
fallow terrace
can i get some help with my rvc model. its my first time making one and i used 40 min data set of talking and it sounds REALLY good on some words and then others, robotic. i tested the audio to see if the data sets were possibly bad but idk. i used 275 epochs so it would be that or the data set like maybe not wide enough to accommodate all the words.
half cove
what? 
half cove
what do you mean by wrong
brittle wing
how could it be training in 1 sec
brittle wing
can we, err, find a way to instruct it to use an already-processed dataset
rare gobletBOT
Ayo? @brittle wing level 6 !!!
brittle wing
@half cove
half cove
can we, err, find a way to instruct it to use an already-processed dataset
just redo the prepocessing again ig
brittle wing
idea: Preprocess the data within the notebook. Let it create the directory /logs/lily. But then replace the D and G paths with the ones which I have already trained
half cove
idea: Preprocess the data within the notebook. Let it create the directory /logs...
just do pre-processing again
should do the same thing
hoary spire
hello guys where do i get voice changer i have now voice model
brittle wing
just do pre-processing again
yeah
fluid matrix
you should be able to just retrain, if you didnt preprocess anything again. and also enter the exact same settings, very important
brittle wing
you should be able to just retrain, if you didnt preprocess anything again. and ...
how? Also, is retrainig the same as resuming training
orchid palm
i followed this tutorial exactly but i just get errors when i try to do anything. Can anyone help?
fluid matrix
no
tender kite
Which is the best MDX-Net/Demucs/VR Arch model in UVR for voice isolation? I don't care about encoding speed, only quality and to remove background noise.
rare gobletBOT
Ayo? @tender kite level 1 !!!
tender kite
i followed this tutorial exactly but i just get errors when i try to do anything...
What errors? And which tool did you use?
tender kite
great, what tool did you use? what google colab or huggingface space, or did you host it...
orchid palm
i hosted
rare gobletBOT
Ayo? @orchid palm level 1 !!!
tender kite
ok send logs then
orchid palm
where might I find those
tender kite
in the console opens
orchid palm
in the console opens
tender kite
what os are you using
orchid palm
windows11
tender kite
orchid palm
https://www.nvidia.com/content/DriverDownloads/confirmation.php?url=/Windows/551...
installing now
fluid matrix
so retraining is not the same as resuming training
id assume it isnt, yes
retraining means doing it from scratch
resuming is picking up where you left of
just training it more
median quiver
barış akarsu
orchid palm
ok its installed
median quiver
nerden bulurum
rare gobletBOT
Ayo? @median quiver level 1 !!!
median quiver
level 2 mı olmam lazım
median quiver
anlamadım kı
rare gobletBOT
Ayo? @median quiver level 2 !!!
brittle wing
anlamadım kı
true
median quiver
ne dogru
azure musk
what is the best RVC can i use?
strange quest
-colab
brittle wing
why isn't my tensorboard updating
even though I am refreshing
nvm I am stupid
I reached my target epochs goal and it stopped training
and if the training isn't live, tensorboard can't pull data
azure musk
-rvc
grand cradle
ok so in a nutshell why do the voice covers sound shit with apparently good models
spring orchid
yes
orchid palm
https://www.nvidia.com/content/DriverDownloads/confirmation.php?url=/Windows/551...
what should i do after ive installed,made an accoiunt, etc?
tender kite
what should i do after ive installed,made an accoiunt, etc?
you dont need an account
all you needed was the driver essentially
orchid palm
ah i see
well
i have one now
regardless
is there anything else i need to do
this is what the error looks like btw
rare gobletBOT
Ayo? @orchid palm level 2 !!!
brittle wing
nvm this
pastel oak
barış akarsu
dreamy talon
Just joined last night and wanna learn how to set stuff up
pastel oak
Just joined last night and wanna learn how to set stuff up
everything you need is in this simple and direct guide
https://docs.aihub.wtf/
if you got a good gpu (at least 2060 ti) and want to run rvc locally, i recommend downloading Mainline
pastel oak
tho, i dont really know what ia am doing, since I got my first pc recently
everything is straight forward on the guide dw
this rvc section is about making voice models and "making AI cover" as most people want it
theres also a AI realtime voice changer in #🔍│help-w-okada if you want that
and some ai art stuff but i dont know anything about that. #🔍│help-ai-art
dreamy talon
okay
rare gobletBOT
Ayo? @dreamy talon level 1 !!!
waxen sierra
Hey
rare gobletBOT
Ayo? @waxen sierra level 1 !!!
waxen sierra
Why is this?
spare shale
Why is this?
- You need to make your own Voice Dataset
You need a Voice Recording from 30 Seconds to 1 Hour
spark citrus
when i run install rvc it works then on the gui thing it says
No dashboards are active for the current data set.
Probable causes:
You haven’t written any data to your event files.
TensorBoard can’t find your event files.
If you’re new to using TensorBoard, and want to find out how to add data and set up your event files, check out the README and perhaps the TensorBoard tutorial.
If you think TensorBoard is configured properly, please see the section of the README devoted to missing data problems and consider filing an issue on GitHub.
Last reload: Mar 21, 2024, 6:23:08 PM
Log directory: ./logs
spare shale
T-T
spare shale
???
spare shale
Sorry I meant it
spark citrus
yes i did that
spare shale
Can you show me the screenshot?
spare shale
when i run install rvc it works then on the gui thing it says
No dashboards ar...
aaaa
aaaaaa
- Click here
spark citrus
the roload?
spare shale
yep
spark citrus
it didint do anything
spare shale
(To be honest I didn't use TB)
spark citrus
do i have to use tb
spare shale
Therefore, it’s better for you to choose between me, a stupid helper, or wait for a professional
spark citrus
can i use something else other than tb
brittle wing
It is greatly recommended to use TB
spare shale
can i use something else other than tb
tb It's just statistics of your Model
brittle wing
otherwise you are training blind
spare shale
otherwise you are training blind
Yes it's me
brittle wing
huh
spare shale
I don't use tb Because I believe that I'm doing everything right
spark citrus
well im slow
spare shale
I've Made A Lot Of Models (Local) And They All Work Very Well
As you can see next to my nickname there is the Word “Lazy” That’s why I don’t want to study Tb 😅
Yes.. I'm a bad Example Helper
spare shale
i have a folder named dataset with the zip of the mp3
You should unzip and send to dataset folder
spark citrus
ok
waxen sierra
1. You need to make your own Voice Dataset
Everything was so easy before
spare shale
Everything was so easy before
And now everything is also very simple
waxen sierra
Okay, I believe you
spare shale
- Open the dataset folder and send a screenshot
spare shale
Works?
cold raptor
Use Rvc disconnected
spark citrus
idk im going througyh the train your mopdel thing
rare gobletBOT
Ayo? @spark citrus level 3 !!!
spark citrus
Use Rvc disconnected
idk how?
spare shale
Use Rvc disconnected
Sorry T-T
spark citrus
Wait what? i did this for nothing
cold raptor
-colab
spark citrus
THTAS SO EASY
but i didint do allat for nothing so next model im using that but for this one im using this
cold raptor
I didn’t use Rvc disconnected That’s why I didn’t recommend it
I'm not using easy gui but still it look fucked up only good feature it have is inference build in
spare shale
I'm not using easy gui but still it look fucked up only good feature it have is ...
To be honest, I don’t know what could be simpler 😅
Sent Files
name*
Staged by Epochs
Pressed 2 buttons and wait
Sorry mistake
low shard
how long does it take to train the thing
Btw it could help you checking our docs maybe https://docs.aihub.wtf
cold raptor
Wait u can't choose sample rate in easy gui ?
spare shale
Wait u can't choose sample rate in easy gui ?
??
thin stump
Yes.. I'm a bad Example Helper
I mean you could have saved him time by linking the docs 
. I just lurked
low shard
Yall
cold raptor
Oh it's 32k by default
low shard
It's better than step by step taking an hour to help someone
low shard
Dw
cold raptor
It's better than step by step taking an hour to help someone
sometimes sending a link is not enough 😭
low shard
sometimes sending a link is not enough 😭
It's better than people getting confused bc an helper says a thing while the other says another
They can just read it
It's not that hard to read it
spark citrus
INFO:TTTPIG:====> Epoch: 215 [2024-03-21 22:57:02] | (0:00:01.891346)
is this normal
its on 215
cold raptor
It's not that hard to read it
Ye but it's hard for some people 💀
spare shale
Ye but it's hard for some people 💀
For example me
spark citrus
im slow
dreamy talon
Just joined last night and wanna learn how to set stuff up
Tbh still confused
low shard
Ye but it's hard for some people 💀
I mean most of our community is brainrot too, if you meant that
Most people join here
After the 9 months old voice e girl tuts, there's a difference from people who have problems reading and people who dont wanna read and want a step by step yt tut
low shard
Tbh still confused
What error u got ?
dreamy talon
The guide kinda does not help me, because I just joined last night and got this pc recently (I am still kinda new to pc)
I do struggle shit on my own
low shard
The guide kinda does not help me, because I just joined last night and got this ...
What pc GPU you got ?
lament crest
Excuse me
I have trouble guys
I made voice model but I don’t know how to make link in huggingface 😦
low shard
well I am new to pc so, I guessing its a GEforce RTX 4060
If you aren't sure, you can do ctrl+shift+esc and then go to performance, it should show you your gpu
lament crest
How can I make link in huggingface?
dreamy talon
NVIDA
rare gobletBOT
Ayo? @dreamy talon level 2 !!!
obtuse stream
how to work this new function in applio? Step saving in the model name for detailed tracking of the training process.
somewhere indicating that the voice is too overtrained. this part will show?
rare gobletBOT
Ayo? @obtuse stream level 6 !!!
low shard
How can I make link in huggingface?
dreamy talon
GPU 0
NVIDIA GeForce RTX 4060
Driver version: 31.0.15.4633
Driver date: 12/6/2023
DirectX version: 12 (FL 12.1)
Physical location: PCI bus 1, device 0, function 0
Utilization 1%
Dedicated GPU memory 3.0/8.0 GB
Shared GPU memory 0.2/15.9 GB
GPU Memory 3.2/23.9 GB
oops
But their
half cove
how to work this new function in applio? Step saving in the model name for det...
dw just took at tensorboard
low shard
GPU 0
NVIDIA GeForce RTX 4060
Driver version: 31.0.15.4633
Driv...
Yea you can do it locally, it's better if you check https://docs.aihub.wtf/ , inside the local folder of the rvc folder, there are 3 versions, i could suggest you mainline or applio, but can't really help you on local as I can't do it myself
lament crest
https://docs.aihub.wtf/essentials/voice-models/#uploading-to-hugging-face
Thanks! It helped me a lot
pastel oak
download mainline, is very simple
you run it and youre pretty much done 👍
low shard
Thanks! It helped me a lot
Yw
low shard
well, Idk what to do <a:skullcry:1159654577960927353>
Either follow this https://docs.aihub.wtf/rvc/local/applio/ or https://docs.aihub.wtf/rvc/local/mainline/ , they both have pros and cons
obtuse stream
dw just took at tensorboard
what can i see
pastel oak
well, Idk what to do <a:skullcry:1159654577960927353>
download link
https://docs.aihub.wtf/rvc/local/mainline/
and then you just do this step by step. dont overthink it its as simple
dreamy talon
Alright 👍
waxen sierra
Very difficult
waxen sierra
What's easier?
spark citrus
uhh i need help
i trained the model and i got the stuff
but idk what to do
what do i put at listen to your model
waxen sierra
uhh i need help
Good luck man
half cove
what can i see
Check loss/g/total
Check when the line starts going back up
Also turn smoothing up
waxen sierra
You know, I was about to give up, but then I decided to try one last time
Thank you
I'm disappointed in the voice model, it's not the one I wanted
But everything is fine
knotty shadow
Hi I have three questions. I am using basic RVC and TensorBoard
- When i want to make sample for training my own voice, how long it should be to be optimal?
- I know TensorBoard is a tool to check how many Epochs i would like to get eventually, but can someone tell me what is average number of epochs to make high quality model?
- Are there any ways to improve quality of model after training it completely?
verbal geyser
azure marshBOT
RVC Guides (How to Make AI Cover)
proven hill
Hi I have three questions. I am using basic RVC and TensorBoard
1. When i want t...
- minimum 5 minutes, the more the better with the right settings
- at least 300
- no
tranquil trail
Hi I have three questions. I am using basic RVC and TensorBoard
1. When i want t...
-
for optimal at least ten minutes, having some singing / being able to pronounce vowels and making sure the quality is good, doesnt have to be incredibly good but decent.
-
300-500 epochs since usually the ov2 pretrain does good
-
if it doesnt sound good, add more data, usually you cant resume training with more dataset unfourtanetly, and if it still sounds weird keep training but thats usually risky since it could over train
subtle loom
Anyone know where I can find some pre-cleaned vocal samples/recordings?
crystal ibex
Is this the chat for AI voices?
proven hill
Is this the chat for AI voices?
yes
proven hill
1. for optimal at least ten minutes, having some singing / being able to pronoun...
ov2 is good for short datasets
rare gobletBOT
Ayo? @crystal ibex level 1 !!!
proven hill
This shit possible on IOS?
yes with ilaria rvc
tranquil trail
ov2 is good for short datasets
yee but if its possible get more for variety tbh
its bc i think still even with this pretrain thats meant for shorter datasets, juss do it with bigger
ngl
and will require training less
which maybe could probably get good decent with 200-300 epochs
hollow vigil
ov2 is good for short datasets
whats ov2? i train locally and trying to train better a little bit
tame current
when i use the google colab easy gui, and import my model via drive it comes up with this error
subtle loom
had the same issue yesterday. I think that fork has been abandoned
tame current
had the same issue yesterday. I think that fork has been abandoned
do u know any others?
rare gobletBOT
Ayo? @tame current level 4 !!!
tame current
https://docs.aihub.wtf/rvc/cloud/ilaria-rvc/
thanks
subtle loom
That one worked great for me but i ended up downloading applio today and running it locally. highly recommend it as it saves all the models and recordings you add and runs fast if you have a solid gpu
tame current
That one worked great for me but i ended up downloading applio today and running...
i mean i have a 1650 not the best so idk
subtle loom
I have a 1650 as well
tame current
does the program run well for you?
timid steeple
modern day standards a 1650 is not a solid gpu
thin stump
@worldly ether send me 30 seconds of your dataset and I could probably identify the problem
subtle loom
yep. Set up was a little confusing for me but i got it up and running in 20 minute
worldly ether
<@749028181398519848> send me 30 seconds of your dataset and I could probably id...
Ok
rare gobletBOT
Ayo? @worldly ether level 1 !!!
subtle loom
modern day standards a 1650 is not a solid gpu
good enough I guess. only 1/100th the power of an A100 but hasn't been a limiting factor yet. Probably couldn't train though
timid steeple
yea
subtle loom
I'm new to GPU's and computers in general. Could I purchase a better GPU and add it to my machine in the future? Or am i stuck with the 1650? On a gaming laptop
timid steeple
gaming laptops you cannot ever upgrade lol
subtle loom
That's what I figured
gentle tree
I pasted a youtube link and for some reason it keeps on saying nothing but python3: can't open file '/content/src/main.py': [Errno 2] No such file or directory
2 help
finite galleon
I used 17 minutes of dataset and it still doesn't look good enough
All depends on quality and the pitch range of the dataset
thin stump
All depends on quality and the pitch range of the dataset
Yeah, it was dirty with some grunts, mic effects, and instrumentals left in. I gave them some tips
proud elbow
when i use the google colab easy gui, and import my model via drive it comes up ...
use the latest colab https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw
proud elbow
I'm new to GPU's and computers in general. Could I purchase a better GPU and add...
get desktop 3060 12 GB
subtle loom
get desktop 3060 12 GB
with a desktop, can i upgrade my GPU and memory if neede?
rare gobletBOT
Ayo? @subtle loom level 2 !!!
proud elbow
with a desktop, can i upgrade my GPU and memory if neede?
should be easy to install & upgrade, find some YT tutorial if not sure
subtle loom
should be easy to install & upgrade, find some YT tutorial if not sure
Sweet thank you for the recommendation
azure marshBOT
candid wind
what purpose do the D and G pth files serve
proud elbow
what purpose do the D and G pth files serve
training snapshot for resuming later
candid wind
OH MY GOD thats actually really useful
queen jolt
🤔
civic onyx
"Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from http://www.nvidia.com/Download/index.aspx"
a colab is asking for a nvida driver bruh
somebody help
aicovergen is being cringe bruh
brittle wing
how do i make my own voice model of myself
civic onyx
@acoustic scarab
civic onyx
Can't help with that
man, i wish money trees were real so i could join the rich people who are making ai covers with thier super good gaming pc gpus
acoustic scarab
Used 3060 is like $200
Maybe even less in your country
Is enough tho
I use one
spring flicker
thats way too expensive
u srs?
spring flicker
used to cost alot more lmao
civic onyx
used to cost alot more lmao
my job is ass, i ain't paying for that shit
acoustic scarab
my job is ass, i ain't paying for that shit
debt
spring flicker
i think my old 3070 before i upgraded was 1k
rare gobletBOT
Ayo? @spring flicker level 1 !!!
spring flicker
my job is ass, i ain't paying for that shit
well you could always use colab
spring flicker
no idk whats happening rn lmao
azure marshBOT
RVC Guides (How to Make AI Cover)
spring flicker
"Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU...
oh
id probably just restart and make a new runtime session
civic onyx
id probably just restart and make a new runtime session
i tried like 50 times
rare gobletBOT
Ayo? @civic onyx level 6 !!!
spring flicker
i tried like 50 times
damn colab is getting greedy
civic onyx
i even used ngrok but same result
acoustic scarab
just to make a character say funny words
- Local Llms
civic onyx
damn colab is getting greedy
colab is a money whore now
acoustic scarab
colab is a money whore now
Is not if you're a researcher
civic onyx
Is not if you're a researcher
how do i become a researcher so i can bypass it
im not giving google a single cent
it can get off my tip
acoustic scarab
how do i become a researcher so i can bypass it
Attend to uni
civic onyx
Attend to uni
oh hell no
civic onyx
i hope whoever made it so you need a nvida driver gets fired
acoustic scarab
i hope whoever made it so you need a nvida driver gets fired
Nvidia made it
civic onyx
Nvidia made it <:trolley:1159468147133395025>
so im guessing i'll never make ai covers ever again in my lifetime unless i whore myself out to google with green paper
acoustic scarab
so im guessing i'll never make ai covers ever again in my lifetime unless i whor...
Or buy a GPU
Well, you can do covers on cpu
But it'll take time
civic onyx
that's incredibly fucking stupid on googles end by limiting its users
therefore making them less money
actual dumbasses
acoustic scarab
that's incredibly fucking stupid on googles end by limiting its users
End user rants, Google swims in cash
civic onyx
Well, you can do covers on cpu
and i really liked colabs
covergen was great while it lasted
acoustic scarab
therefore making them less money
Paid users actually get all of it
civic onyx
having an ai that did the whole song cover for you without you needing to seperate vocals and shit was great
i'll miss it
now if i want to make a cover i'll need to do the incredibly tedious task of seperating it and using a huggingface space that takes like 999999999 hours to generate the cover
and then when i've aged about 20 years ill need to go into audacity
and line them up perfectly
proud elbow
thats way too expensive
or get used 1660 super for $100
though 8-12+ GB vram is more desirable
civic onyx
or get used 1660 super for $100
that is if you can even upgrade a intel laptops gpu
@acoustic scarab still though why was i able to make covers for like.. a year or longer this whole time without needing a nvida but now all of a sudden im forced to get a new gpu to generate one
like literally just yesterday it worked fine
was there some kind of update to colabs
acoustic scarab
civic onyx
pretty sure mods should edit out covergen then
civic onyx
since its techincally not working as intended
ill try and use it one more time
if i dont get anything then i give up
queen escarp
-colab
rare gobletBOT
Ayo? @queen escarp level 1 !!!
queen escarp
-rvc
summer glacier
How many echopocs should I use for 3 minute clips?
rare gobletBOT
Ayo? @summer glacier level 4 !!!
sand iris
since its techincally not working as intended
who made the colab ? Which one is it that you're using
Honestly, blame :
a) nvidia for their wonky linux driver
b) Microsoft for still not making CUDA 12 accessible without manual intervention on onnxruntime
In no case should you need any hardware on your end to use colab
queen escarp
Is it possible to run locally on mobile?
sand iris
No
Technically, you could if your phone has enough memory, but it would be really, really slow
sand iris
You don't need a GPU to simply run it. It's only needed to get acceptable speeds
Else it's
lol
sand iris
Not too up to date on that 😅 Not sure which options there are
queen escarp
oh alr
lament eagle
just use colab in mobile
hard halo
Hey there I want to learn to make singing voice models, I tried looking at the guide in https://docs.aihub.wtf/ but apparently it gives me an error while trying to load the link up to "How to Make Voice Models" saying it doesn't exist.
low shard
Hey there I want to learn to make singing voice models, I tried looking at the g...
retry clicking this https://docs.aihub.wtf/essentials/how-to-make-voice-models/
brittle wing
If the d and g loss are going up at low epoch (150) with a dataset (good quality, and half an hour long), am I undertraining
I refuse to believe I am overtraining, that would imply the modelhas reached its full potential after just 150 with a very good and long dataset
pastel oak
If the d and g loss are going up at low epoch (150) with a dataset (good quality...
idk id keep going
you checking g total right?
brittle wing
you checking g total right?
yeah, g loss total
hot lagoon
Is there a tutorial on how to train my own pre-train models?
hard halo
retry clicking this https://docs.aihub.wtf/essentials/how-to-make-voice-models/
Oh it works,Thank you
rare gobletBOT
Ayo? @hard halo level 1 !!!
low shard
Oh it works,Thank you
yw
late haven
Is it possible to do a batch voice conversion (for example I just select multiple files and it auto converts all of them) on the RVC web
finite galleon
Is it possible to do a batch voice conversion (for example I just select multipl...
Pretty sure theres a batch inference section where you just input the folder containing audios
late haven
Cool thanks! Will check it out later
brisk quarry
can someone make me remember how to train a model
i made one before but deleted everything and now i forgor how to setup
i dont even remember how to install vrc properly
low shard
https://youtu.be/Q8du7n0vgfU
no thats outdated
dont follow almost 1yo yt tuts
suggest people our docs https://docs.aihub.wtf
brisk quarry
dont follow almost 1yo yt tuts
that one blender tutorial from 2008 that still works in 2024:
low shard
how to setup in general
first what gpu you got
low shard
that one blender tutorial from 2008 that still works in 2024:
ai is different, it always changes and grows in better
brisk quarry
first what gpu you got
i made a model before so its not about the gpu
or do you mean if i have cuda
turbid pagoda
Hey quick question what's the recommended algorithm for text to speech. I thought it was harvest but wasnt sure
analog yoke
Can someone help me ? whats the best RVC Software to Convert vocal with AI model for music related stuff
low shard
Can someone help me ? whats the best RVC Software to Convert vocal with AI model...
check https://docs.aihub.wtf
low shard
i made a model before so its not about the gpu
just checking if it was powerful enough
low shard
but what python versin do i need
3.11
umbral acorn
hello. how to make pth and index file ?
brisk quarry
hello. how to make pth and index file ?
uhhhh
low shard
funni
either check this guide https://docs.aihub.wtf/rvc/local/applio/ Or https://docs.aihub.wtf/rvc/local/mainline/ , both versions have pros and cons, cant help much with local cus i use google colab instead online
low shard
hello. how to make pth and index file ?
brisk quarry
C:UsersLukagDesktoprvcRetrieval-based-Voice-Conversion-WebUI>runtimepython.exe infer-web.py --pycmd runtimepython.exe --port 7897
The system cannot find the specified path.
C:UsersLukagDesktoprvcRetrieval-based-Voice-Conversion-WebUI>pause
Press any key . . .
brittle wing
What bitrate/frequency range is RVC output? It may be useless to collect high quality datasets
kind fox
Is there a way to check the health of the model? I've converted my voice. But the sound is still similar to the original.
brisk quarry
i think i quit again bruh
rare gobletBOT
Ayo? @brisk quarry level 2 !!!
brisk quarry
th9is is cursed
brittle wing
hello, how to fix the glitchy chorus ai voice when using rvc interface in hugging?
hot lagoon
How do I exactly use a vocal synthesizer for back vocals?
unique river
Making AI Covers with RVC
azure marshBOT
RVC Guides (How to Make AI Cover)
wet cave
no thats outdated
Thanks . I will look into your guides. But it still works
waxen sierra
Uh, maybe someone can send it to me Google collab Rvs
To be the coolest
For AI covers
tulip mica
So Well, I'm going to Create a Song for AI Cover Versions with using My Capcut, Bandlab, and your Voice.AI, so This is Upcoming Music for AI Cover Versions for the Future.
teal elm
How to create my own models?
brittle wing
should natural reverb in a room be a problem for the dataset?
hot lagoon
Can someone explane why change the hop length for backing vocals a thing?
fierce barn
I am trying to connect remotely to the rvc but the http address is only allowing me to access the client on the PC that is actually running it
sand iris
I am trying to connect remotely to the rvc but the http address is only allowing...
define "the rvc"
The software that allows you to train models & convert files ?
Or the real time voice changer ?
Refer to this if you're using w-okada/voice-changer #🔍│help-w-okada message
fierce barn
this
pastel oak
I asked you in wokada if its rvc or voice changer and you went to this channel even though its voice changer 😂
fierce barn
i'm new i have no idea sorry
sand iris
i'm new i have no idea sorry
you, refer to the message I linked then ! Should have you covered !
fierce barn
the web address only works on the machine that is running the software on another pc on the same home network it gives the error i screenshotted above
hot lagoon
Bve2 on X minutes how is that? Great then UVR. Bve
sand iris
the web address only works on the machine that is running the software on anothe...
you're using the wrong address
127.0.0.1 resolves to localhost; the exact machine you're trying to connect from
There's another address displayed if you sarted with the start_https script
fierce barn
oh so need to replace the first bit with the ip of the other machine?
signal gull
why its not doesnt load
Web Server Launch Exception, Expecting value: line 1 column 1 (char 0)
whats this
fierce barn
you're using the wrong address
thank you working now I'm not very good with computers
signal gull
help me again
.
golden moss
is my model overtraining already?
teal elm
The software that allows you to train models & convert files ?
Or the real time...
What is the name of the software that allows me to train models and how to use it?
sand iris
Software to make models is "Retrieval-Based-Voice-Conversion-WebUI" or "Applio"
docs at https://docs.aihub.wtf/
rare gobletBOT
Ayo? @low shard level 75 !!!
spark citrus
Im not sure what to do.
golden moss
Im not sure what to do.
can't use colab for like more than two hours, use a new account or wait 24 hours
spark citrus
can't use colab for like more than two hours, use a new account or wait 24 hours
thats fucking stupid google
whatever its fine ill just do another account
spark citrus
can't use colab for like more than two hours, use a new account or wait 24 hours
thanks for your help btw
finite galleon
Im not sure what to do.
In short you using googles gpu for free makes google lose money so they limit you
spark citrus
google greedy asf
finite galleon
google greedy asf
Their gpu is very powerful and for free its a steal
You cant find it anywhere else
spark citrus
Ai is crazy though
golden moss
thanks for your help btw
np learned the hard way when i tried to train a model overnight lol
spark citrus
np learned the hard way when i tried to train a model overnight lol
how many epochs should i use?
rare gobletBOT
Ayo? @spark citrus level 4 !!!
golden moss
how many epochs should i use?
put it at like a 1000 and just watch the tensorboard
stop the training if you see it overtraining
pastel oak
stop training if g/total keeps rising and not going lower than the lowest point after ~150 epochs (you have to decide when youre 100% sure)
then use the pth file of before the overtrain
grim bay
Should I keep doing this to make a dataset?
Btw, all of these are from 25 seconds scene
pastel oak
Should I keep doing this to make a dataset?
Btw, all of these are from 25 second...
I recommend dataset of 10-20 minutes
Whether you have your dataset in multiple clips or one big file doesnt matter
spark citrus
I recommend dataset of 10-20 minutes
Whether you have your dataset in multiple...
i think mine is about 12 minutes, do you think that will be good?
grim bay
I recommend dataset of 10-20 minutes
Whether you have your dataset in multiple...
But should I do that for an easier to choose which parts in the Anime the character said?
subtle loom
when training a model on 10 minutes of audio, is a batch size of 8 too low?
pastel oak
i think mine is about 12 minutes, do you think that will be good?
if the dataset is good then yea
pastel oak
when training a model on 10 minutes of audio, is a batch size of 8 too low?
8 is good
pastel oak
But should I do that for an easier to choose which parts in the Anime the charac...
If you can get 10 mins of your character talking thatd be good yes
I dont know if youre asking that
raw wedge
Does any have any inference advice to fix the occasional weird pronunciation/off tune ?
brittle wing
bruh
What bitrate/frequency range is RVC output? It may be useless to collect high quality datasets
finite galleon
What bitrate/frequency range is RVC output? It may be useless to collect high qu...
Depends on the model. If its trained on 32k, then max freq is 32k, 40k 40k 48k 48k
finite galleon
What bitrate/frequency range is RVC output? It may be useless to collect high qu...
You need high quality datasets so that your model doesnt screech or artifact
brittle wing
what about bitrate
finite galleon
Also so that it doesnt have random noise
finite galleon
what about bitrate
If i remember its float16 wav
brittle wing
You need high quality datasets so that your model doesnt screech or artifact
I know, this is 'cleanliness'. But I'm talm bout numbers like bitrate
brittle wing
If i remember its float16 wav
erm, so is this 16 bit?
finite galleon
erm, so is this 16 bit?
Actually not sure you can try it out for yourself
Do inference and check the info on the output
proud elbow
erm, so is this 16 bit?
the infer results are always 16-bit mono
spark citrus
Should i keep the epoch on 500 or what should i put it as???
rare gobletBOT
Ayo? @grim bay level 9 !!!
grim bay
Please help
short ridge
-colab
grim bay
Helpppppp
split frigate
Why I can't open it? I have Python installed already
grim bay
https://drive.google.com/drive/folders/1EqpyDlYRbUOps0AEl4G0_jm1WVJSKWKE
Is this good? Should I continue to use the same method like this to get the audio?
proven hill
someone need help?
viscid moss
<@274566299349155851>
hewwo
grim bay
Oh god helpppp
grim bay
https://drive.google.com/drive/folders/1EqpyDlYRbUOps0AEl4G0_jm1WVJSKWKE
Is this...
Please help me
viscid moss
https://drive.google.com/drive/folders/1EqpyDlYRbUOps0AEl4G0_jm1WVJSKWKE
Is this...
It's okay, you can continue trimming and then extract the audio to use in RVC
proven hill
Please help me
thats a folder of mp4, what you need it for exactly?
viscid moss
The cover is not saved anywhere
NO GPU for today, u need to change to another google account
grim bay
thats a folder of mp4, what you need it for exactly?
I need to trimming some parts of the Anime, am I doing right?
I will change it to wav once I finished
proven hill
making a dataset?
grim bay
Yeah
proven hill
why its online? use it on your pc
split frigate
Why I can't open it? I have Python installed already
can any1 help me with it?
proven hill
Why I can't open it? I have Python installed already
its mainline?
grim bay
Am I doing right?
proven hill
Am I doing right?
i dont get why youre having your dataset there tho
split frigate
its mainline?
idk
grim bay
i dont get why youre having your dataset there tho
To share it for everyone first
grim bay
I want to know am I trimming it right or not
proven hill
make a dataset with reaper or audacity, reaper is better bcause it can display mp4 video files too so youa ctually watch what youre trimming
then clean the audio with UVR5
split frigate
by the name seems so, send me the link where you got it from
yes it is mainline
proven hill
So, am I trimming it in the right way?
technically yes
split frigate
do i have to chnage it
proven hill
yes it is mainline
did you install it from github or huggingface?
split frigate
did you install it from github or huggingface?
hugging face
rare gobletBOT
Ayo? @split frigate level 2 !!!
proven hill
hugging face
so you have everything including the runtime folder
did you extract it correctly? there was some errors?
split frigate
did you extract it correctly? there was some errors?
i will repeat it in case i didnt notice
split frigate
did you extract it correctly? there was some errors?
yeah there was an error in extracting
now its work tysm
how can i create a path and an index files for a model or a voice i recorded?
grim bay
make a dataset with reaper or audacity, reaper is better bcause it can display m...
reaper is a paid app right?
proven hill
reaper is a paid app right?
free too
proven hill
yeah there was an error in extracting
now its work tysm
no problem!
proven hill
how can i create a path and an index files for a model or a voice i recorded?
you have to manually copy a path, for now
brittle wing
is there rvc for amd?
rare gobletBOT
Ayo? @brittle wing level 5 !!!
proper shale
is there rvc for amd?
sadly no, unless you want a convoluted process that involves installing linux and ROCm which I sadly don't know how to do
brittle wing
sadly no, unless you want a convoluted process that involves installing linux an...
damn
someone should make it fr
since amd gpu are cheaper
proper shale
someone should make it fr
actually, not possible rn
proper shale
since rvc needs pytorch, pytorch has horrible compatibility with amd gpus
brittle wing
damn
proper shale
i mean... maybe an onnx version is possible? mayyyybe
@sand iris correct me if im wrong
but yeah its just hard
sand iris
hm ?
Training on onnx, not sure about that
About RVC inference; parts of it are implemented on RVC if I remember correctly, but I haven't delved into it
DirectML (what they use to run code on GPU) was and still remains ass
brittle wing
damn
split frigate
any1 know how to create a path and index files for a voice file?
proper shale
any1 know how to create a path and index files for a voice file?
this tells you how to train a model
split frigate
this tells you how to train a model
i followed the steps till i reached here
after that it says upload path and index, how i even create to upload?
proper shale
this is just for uploading a downloaded model
you didnt need to follow that part, in this case
on the right part of the screen there's the topics, click on the training part, which is what u should focus on
split frigate
on the right part of the screen there's the topics, click on the training part, ...
tysm
tribal oar
make a dataset with reaper or audacity, reaper is better bcause it can display m...
im trying to do this
tribal oar
so does your dataset
need to have splitted audio files?
or can it be one audio file as a whole
proper shale
or can it be one audio file as a whole
yes
proper shale
rvc does the splitting part in preprocessing anyway, so its a waste of time
plus slows down stuff
if its too many files
tribal oar
also does the sample rate matter too much
cause idk if that could be causing this error
tribal oar
this error
well i just have it set to 32k sample rate
this is my first time training, so yeah
proper shale
huh
wet cave
rvc does the splitting part in preprocessing anyway, so its a waste of time
Oh it does? Good to know!
proper shale
never saw that b4
tribal oar
kanka turk ses modellerı yok mu burada
ben modellere bakmiyorum o kadar, özellikle turk modelleri
brittle wing
why is rvc not working on discord?
compact shale
ben modellere bakmiyorum o kadar, özellikle turk modelleri
neye bakıyorsun hocam
rare gobletBOT
Ayo? @compact shale level 1 !!!
proper shale
why is rvc not working on discord?
give us more details, and we'll try our best
tribal oar
huh
idk if im allowed to share a youtube video here which is a tutorial of training a rvc model
i just followed a youtube video
proper shale
idk if im allowed to share a youtube video here which is a tutorial of training ...
you can
proper shale
most of these videos r outdated though, just be careful if they're like 8-10 months old
brittle wing
give us more details, and we'll try our best
i am using voicemeeter with rvc and on discord my normal voice is getting played back
tribal oar
5 months
tribal oar
so he's using mangio rvc and tensorboard
tensorboard probably to analyze the process
maybe mangio might be outdated? idk
proper shale
tensorboard probably to analyze the process
yeah to check the graphs
tribal oar
i'll try applio
proper shale
maybe mangio might be outdated? idk
id suggest trying something else yes
proper shale
i'll try applio
yeah maybe thatll work
https://docs.aihub.wtf/rvc/local/applio/ in case u want a recent tutorial
tribal oar
thanks
tribal oar
dostum senı yormak ıstemem ama ufaktan bır yardım edebılır mısın
to search turkish models?
tribal oar
o zaman yapılmamışdır
compact shale
o zaman yapılmamışdır
youtubedan gelıyorum ve daha oncesınde var gorunuyor kaldırılmıs olabılır tesekkürler
tribal oar
youtubedan gelıyorum ve daha oncesınde var gorunuyor kaldırılmıs olabılır tesekk...
hee true
tribal oar
im trying to do that exactly lol. train a model
compact shale
arkasına sarkı koyup soylettırıcem
tribal oar
go to here, scroll down until you see train
you can take those steps there which im doing rn
brittle wing
is there any way to save tensorboard itself within the log files
compact shale
you can take those steps there which im doing rn
ok I will implement this thank you for your help
brittle wing
or like, data of D and G total loss
tribal oar
there's rmvpe_gpu, interesting
rare gobletBOT
Ayo? @tribal oar level 3 !!!
brittle wing
what are you guys talm bout
tribal oar
im training a model in applio
brittle wing
every training process uses a gpu
compact shale
https://docs.aihub.wtf/rvc/local/applio/
sesı donusturdukten sonrakı işlemde ne yapıyorum bunu anlayamadım
rare gobletBOT
Ayo? @compact shale level 2 !!!
azure marshBOT
RVC Guides (How to Make AI Cover)
main lichen
I need help rn its so urgent XD
Everytime i speak it says vol then goes up but i cant hear myself
bro PLEASEEE
@stoic cloud
CMON THERE WAS JUST A HELPER TYPINGG
viscid moss
<@274566299349155851>
What do u put here and here?
main lichen
What do u put here and here?
YO HELP
low shard
for wokada ask in #🔍│help-w-okada or #1192011222023950368
viscid moss
put an output device on "monitor"
waxen sierra
What do u put here and here?
Good question, give me a second
waxen bison
JSONDecodeError Traceback (most recent call last)
<ipython-input-8-30278fdbb0d3> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:
5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
what can i do?
low shard
---------------------------------------------------------------------------
JSON...
send google colab link you are using
waxen bison
low shard
https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sh...
thats outdated
check our docs instead https://docs.aihub.wtf
low shard
here you can find updated colabs for making models or use them, theres everything u need just writted down
you are welcome
ocean lake
can anybody help? I am new and wanna try it, but after downloading zip file and extracting there is no mp3 or any of that files voice.ai support how can i add voice?
waxen sierra
What do u put here and here?
Well, on to the first one https://drive.google.com/file/d/16YbMLvUvCC7JLiC7zHvvxVhElJXwiLar/view?usp=drivesdk
rare gobletBOT
Ayo? @waxen sierra level 3 !!!
waxen sierra
What do u put here and here?
And on the second https://youtu.be/2yugk-8txBk?si=rhP3js9ZIoCqMKN7
waxen sierra
Ayo? <@510053938955943936> level **3** !!! <:lfg:1159355870119993496>
Thank you, thank you
viscid moss
ok, that´s fine. And the model name?
low shard
didnt know
you can use rvc ai voice models in wokada tho, they are all free and you can make even your owns, or use 10k+ models already made
viscid moss
dir_name and rvc_name needs to be the same
azure marshBOT
-rt
This interaction has expired, use the command
/guides realtimeif you wish to see it again.
viscid moss
dir_name and rvc_name needs to be the same
waxen sierra
Ок
compact shale
audio slicer what deoes it mean
how can i download this
I want to create someone's voice model,
reef frost
Looks like there's a bit of a problem.
unknwon message
If you clear the information being managed by this app, it may be recoverable.
Initialize
Reload without initialize
Error
unhandledrejection
no error stack
TypeError: (0 , Zz.removeDB) is not a function
TypeError: (0 , Zz.removeDB) is not a function
at http://127.0.0.1:18888/index.js:2:3293292
at m (http://127.0.0.1:18888/index.js:2:1536451)
at Generator.<anonymous> (http://127.0.0.1:18888/index.js:2:1537797)
at Generator.next (http://127.0.0.1:18888/index.js:2:1536880)
at e (http://127.0.0.1:18888/index.js:2:1543652)
at s (http://127.0.0.1:18888/index.js:2:1543855)
can osmeone hep me
brittle wing
The generator and the discriminator are supposed to be fighting each other a lot in the beginning right? Later they should both go down
waxen sierra
u need to change that. Put the name of ur model
Hurray, I made a cover
Bad cover, wasted my day on it
But thanks for the help
spark citrus
my thing is stuck on Updating and installing pip packages...
on the rvc v2 disconnected
install Dependencies
brittle wing
my thing is stuck on Updating and installing pip packages...
I got this as well
It took around 6 min for me.
spark citrus
why?
brittle wing
So if it takes longer than this, just refresh and try again
brittle wing
yeah
brittle wing
idk, for me it doesn't take 30s
but yeah, np
Is the d loss supposed to fluctuate so much?
50 epochs in
meanwhile. this is the g loss
pastel oak
meanwhile. this is the g loss
g total looks fine
brittle wing
g total looks fine
how bout d
Although I must say a downwards trend has started now
70 epochs in
so we are looking good
pastel oak
in 2 out of 3 cases, my d total fluctuated a little bit for the first 2k steps and then started an downwards trend
similar to yours yeah
youre fine
brittle wing
rare gobletBOT
Ayo? @brittle wing level 7 !!!
brittle wing
erm, what the flip
mental spruce
help?
iron cobalt
-local
azure marshBOT
iron cobalt
are there any new model making forks local and Colab?
undone yacht
Why can't I take my CPU like in the video? pls
mental spruce
help
tidal shadow
is there any rvc link that works? the one i used to use seems to not function
rare gobletBOT
Ayo? @tidal shadow level 1 !!!
rigid furnace
can anyone send me the code to use to not get disconnected from inactivity on RVC disconnected please.
btw what epoch would you put 32 min clear audio?
split frigate
I have a model that has a male voice.
is there any way to convert it or merge it or anything that will make it a female voice?
rigid furnace
I have a model that has a male voice.
is there any way to convert it or merge i...
you could try getting a female voice model and merge it using ckpt processing.
When i merge models, it usually comes out sounding super robotic, i'm not sure how to fix it. Maybe you'll get better luck than I did.
split frigate
you could try getting a female voice model and merge it using ckpt processing.
where can i find a guide about ckpt processing
rare gobletBOT
Ayo? @split frigate level 3 !!!
rigid furnace
where can i find a guide about ckpt processing
rare gobletBOT
Ayo? @rigid furnace level 1 !!!
rigid furnace
hope that helps
split frigate
hope that helps
i will se tysm
rigid furnace
i will se tysm
yeah, just figured out. Under that video, he pinned a comment of what may be the issue if your merged voice model sounds raspy or robotic. just in case that happens to you.
rare anchor
Hey guys. It's weird, but when I launch the start_http.bat file, it closes immediately. Does anyone know how to help me?
rigid furnace
Hey guys. It's weird, but when I launch the start_http.bat file, it closes immed...
have you tried right clicking the bat file and running it as admin?
worth a shot if you haven't.
rare anchor
have you tried right clicking the bat file and running it as admin?
I think my computer is just bad
I think it's not well optimized. Actually, it's a gaming laptop that I paid €1500 two years ago. I believe the issue is likely with the RAM.
rare gobletBOT
Ayo? @rare anchor level 1 !!!
vapid gust
pastel oak
where is the tuts
pastel oak
everythings pinned
vivid yew
idk if this really regards to this because im not that good at ai stuff but when i want to install the model i keep getting this bug
JSONDecodeError Traceback (most recent call last)
<ipython-input-13-5060be0469c1> 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)
if i could i would send a screenshot
finite crow
I dont know, but i am not sure how to set up my NVIDIA, but it is in NVIDIA, im using a NVIDIA GeForce GTX 1050
vivid yew
bruh
rare gobletBOT
Ayo? @vivid yew level 1 !!!
vivid yew
idk what i do because im not good at that ai stuff
pastel oak
---------------------------------------------------------------------------
JSON...
send the colab link youre using
vivid yew
send the colab link youre using
.
pastel oak
outdated
vivid yew
ohh
mortal lodge
where should i look on there to find the new collab link? ive been having similar issues i believe for similar reasons
finite crow
Shad.
pastel oak
@mortal lodge ^
vivid yew
i cant find the "running on public" text
pastel oak
i dont know anything about colab, just follow the tutorial, theres also a video tutorial linked
https://www.youtube.com/watch?v=wh4cLMLLkj0 for ilaria
hallow ferry
I trained the rvc model in colab using the rmvpe_gpu method for 500 epochs. I have a high quality data set but after finishing the training (I had to buy colab pro) the model does not sound clean...
pastel oak
I trained the rvc model in colab using the rmvpe_gpu method for 500 epochs. I ha...
you probably overtrained, run Tensorboard over it and check where the g/total is at its lowest point
hallow ferry
you probably overtrained, run Tensorboard over it and check where the g/total is...
is this what you mean?
