✨│ai-help

[somber valve]

wait how do I download

nvm

[rare goblet]

Ayo? @somber valve level 1 !!!

[cursive flare]

hello guys, I have forked https://github.com/fumiama/Retrieval-based-Voice-Conversion-WebUI, everything is fine and I am getting errors related to fairseq when running the python

I showed error output to chat gpt, it says

The error message indicates that there's an issue with how default values for mutable types are being set in a dataclass in the fairseq library. Specifically, it’s saying that a mutable default value is being used where an immutable type is expected, which is causing a ValueError.

This is likely a problem in the fairseq library's code, rather than an issue with your own code. Here’s a breakdown of how you might address this:```

im using py 3.11, pip 24.0

i have tried

if config.dml == True:

    def forward_dml(ctx, x, scale):
        ctx.scale = scale
        res = x.clone().detach()
        return res

    import fairseq

    fairseq.modules.grad_multiply.GradMultiply.forward = forward_dml```

of course but didnt work

.modules is highlighted red

I have used from fairseq.modules.grad_multiply import GradMultiply too

[red kayak]

well its good for removing parts u dont like

[red kayak]

https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006Nvidia.7z

download this

and run the go-web.bat

[cursive flare]

fumiama claims it is no longer supported by the dev

[agile gale]

no 32k

[red kayak]

you misinterpreted fumiamas message. They were implying that the updates RVC boss will make will be slow and it wont be updated as often

[red kayak]

switch back and forth to v2

[cursive flare]

okay, thank you

oh

I fixed it

i basically needed to install fairseq she mentioned in py311.txt

[rare goblet]

Ayo? @cursive flare level 4 !!!

[agile gale]

where do i drop the audio?

[red kayak]

did you do as i told u?

refer to this^^

[patent pasture]

is it overtrain point?

[rugged crag]

Hey, I’m new here and a bit lost. I installed RVC on my Mac, but it doesn’t seem to be working properly. Are there any guides for the correct installation or troubleshooting that I can refer to? Thanks!

[low shard]

for Mac you can only inference locally, you can't train models, it would be better you use a cloud way

[rugged crag]

i m not trying to train models, so it should be fine for now.

[pseudo raptor]

Hey I want to make an rvc can I ask where to start

[odd shale]

You can start reading the docs buddy.

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[pseudo raptor]

Which should I read M

[low shard]

https://docs.aihub.wtf

[pseudo raptor]

K

[low shard]

if you got a good pc GPU do it locally, else google colab or kaggle

[rugged crag]

cool, thanks! i ll see if i can find answers here

[rare goblet]

Ayo? @rugged crag level 1 !!!

[pseudo raptor]

Any specification

[low shard]

Wdym?

[pseudo raptor]

Specification of pc

[low shard]

what's your PC GPU? you can find it in task manager via ctrl+shift+esc -> performance tab -> GPU

[pseudo raptor]

Ok

Thanks

[low shard]

yw, so whats your pc GPU?

[rare goblet]

Ayo? @pseudo raptor level 1 !!!

[low shard]

alright thats good for locally both inference and training, from the docs https://docs.aihub.wtf you can find the local section where you can choose if to download applio or mainline

[pseudo raptor]

Ok

Thanks

[low shard]

yw, for any issues ask here

[agile gale]

@red kayak no export

@low shard

can uguys help

[rare goblet]

Ayo? @agile gale level 4 !!!

[agile gale]

i dont see this

[red kayak]

set smoothing to 0

[red kayak]

export audio

...

can you not see export audio?

it should be there

[gleaming scarab]

Can someone help me ?

[red kayak]

sure go on

[agile gale]

its not there

lol

[gleaming scarab]

When I click on start_http nothing happens

[agile gale]

im crazy

nvm

[red kayak]

are you trying to run w okada

[agile gale]

@red kayak this dif from the one on website

[red kayak]

not really

regardless select mono, set encoding to 32bit float and export

[rugged crag]

i downloaded rvc on my mac and tried to convert an audio file with the inference model. the result just sounds like radio static. what am i doing wrong? does anyone has the answer?

[red kayak]

okay first and foremost

[red kayak]

share a screenshot of ur gui

[patent pasture]

oh, I have already closed the colab

[red kayak]

oh well then

[gleaming scarab]

Before I dont have to download it with start http ?

[red kayak]

i need context

what are you trying to do

[red kayak]

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[gleaming scarab]

For run voice changer I have to click on start_http for download it but when I click on Nothing happens

[patent pasture]

I simply read somewhere that model becomes overtrained when falling trend stops?

[red kayak]

mac isnt my forte but im assuming it has to do with cuda and such. Really though, you shouldn't be using mangio fork. its outdated

[rugged crag]

what should i use?

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

Don't forget to thank Ilaria if you find it useful! 💖

[pastel oak]

this

Free, online, fast

[rugged crag]

thanks, i ll look into it

[red kayak]

omg @proven hill this command is a slay

[knotty moth]

set smoothing 0 and it actually looks rather flatline with so many collapses

btw I'd suggest the audio labeling method for the dataset to help mitigating the collapses
https://rentry.co/RVC-dataset-RX11#noise-gating-and-audio-labeling

[agile gale]

erroring

[gleaming scarab]

Pleaase

[modern surge]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• 🆕 Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[tough fiber]

guys theres a option. it says special pre-trained model. this new model says it can be a guide when training, should I use it?

@pastel oak

[odd shale]

Nope, don't use that.

[tough fiber]

oh okey thanks

shad sorry for mention -*-

[tough fiber]

is there any option should i turn on

[odd shale]

I don't think you should touch anything.

[tough fiber]

im not sure should i enable for this

is it working?

anyway applio is only saving last epochs so

i cant even choose older saved pth files

if i can i will check tensorboard and i can choose best epoch

i think its work idk

[odd shale]

Nope, don't use the overtraining detector.

[tough fiber]

why

whats difference normal one

even applio only saving last epochs

doesnt older epochs

[odd shale]

Just check the Tensorboard.

[tough fiber]

..

if its saving last epochs whats matter to check tensorboard

even i cant choose older epochs if its already over trained

[odd shale]

If you don't check tensorboard, you'll won't know which is the lowest point on the graph.

The lowest point means the best epoch.

[tough fiber]

thats why im saying if lowest point is older epochs

already is over

[odd shale]

I don't really understand, sorry..

[tough fiber]

so applio saving

best epochs

i didnt too because i didnt use applio before

i was using mangio rvc fork

when i started to train models it was creating files 10th epoch 20th epoch 30th epoch. in folder

and when i check tensor board after 500 epoch

i can say oh thats 300 epoch is lowest point

after i was deleting other epochs and getting 300 one

so idk applion how its working

it didnt save every epochs

@red kayak litsa help

[odd shale]

Yeah, because you probably set 10 epochs on save frequency

[tough fiber]

yes i was talking about this

but when i do same thing on applio

applio is only saving LAST EPOCHS

it was delete old epochs

after 500 epoch if i check tensorboard and if i saw 300 epoch is lowest point

i cant choose because applio wasnt save

i mean if i still training a model and it reached 532 epoch. when i check applio folder i can see xxxx_532.pth. not others. xxxx_300.pth isnt exist there so

[analog obsidian]

whats your save frequency rate?

if is not 1 then you'll not get every epoch

[tough fiber]

i dont want to get every epoch.

i want to get 5-10-15-20-25 ....

epochs

because i set " save every 5 epochs "

[analog obsidian]

oh, and is not doing that?

[odd shale]

El dice que no.

[tough fiber]

yea applio only saving last epochs only

[analog obsidian]

maybe applio bugged
i advise to use rvc mainline instead (original, unedited rvc) as lacks any weird issues

and is actually easier to use

[tough fiber]

when i was using mangio everyone said use applio

now im using applio and everyone saying mainline lol

[odd shale]

You can use RVC Mainline instead.

[tough fiber]

its idk

[analog obsidian]

applio is more laggy and bloated than mainline

[tough fiber]

where can i download

ill delete applio

if is there link

for mainline

[analog obsidian]

https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/releases/tag/2.2.231006

[tough fiber]

thanks sir

[analog obsidian]

no problem, hope this fixes the issue

[odd shale]

Gracias por ayudarlo, estaba teniendo dificultades para ayudar a ese tipo ya que no comprendía bien su ingles.

[analog obsidian]

no hay de que

[odd shale]

🐢 👍

[gleaming scarab]

Someone Can help me ?

[odd shale]

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[gleaming scarab]

Yes sorry I copy my message

For run voice changer I have to click on start_http for download it but when I click on Nothing happens

[tough fiber]

@analog obsidian rvmpe or rvmpe_gpu or harvest

which one is good

harvest is better quality but idk

also i dont know whats differrence rvmpe_gpu and rvmpe

[proven hill]

harves is NOT better quality

[analog obsidian]

rmvpe_gpu

harvest is worse quality than rmvpe

[proven hill]

rmvpe and rmvpe_gpu damn i wonder what could be the difference

[tough fiber]

oh really

idk whats says here lol

[knotty moth]

you will get insta rejected in #1159514067187277865 for using harvest

[analog obsidian]

is for CLI usage (two gpu)

[rare goblet]

Ayo? @knotty moth level 30 !!!

[analog obsidian]

leave it as is

[tough fiber]

well i always used rvmpe generally

but someone said harvest is good so lol

hes wrong

should i turn on last ckpt file save?

[proper shale]

decent but rmvpe better and uses GPU

[tough fiber]

i think ytea

[proper shale]

yeah sure, it's probably better

[analog obsidian]

harvest is very outdated, even more outdated than crepe
however it has good bass, so it can kinda help isolated vocals have better bass volume, but rmvpe already has a decent bass anyways

and sometimes harvest just doesn't work and generates distorted audio

like a demon speaking lol

[tough fiber]

ohh get it

[proper shale]

[analog obsidian]

it does that even on utau (not ai-based) lol

[proper shale]

yes, no (unless u have VRAM to spend and it meets requirements), yes

[analog obsidian]

just save the latest ckpt: yes
final small model to weights: yes
cache dataset to gpu: no

[tough fiber]

bet tx guys

[red kayak]

rocket science this rvc thing is owo

[safe quartz]

can i use youtube link for ilaria rvc? or just downloaded songs?

[low shard]

you can't directly paste a yt url link in ilaria rvc, you can just download it yourself and upload it

[gleaming scarab]

Please help me

[finite pumice]

hello, i tried running w okada using google colab but i cant start it.

[finite pumice]

@acoustic scarab

[acoustic scarab]

sorry, i dont use colab at all and hence cant help with it

local only, as it says in the name

[finite pumice]

who uses it

[acoustic scarab]

https://cdn.discordapp.com/attachments/1224878046251585626/1250313326903296063/xryu4ewo.gif

[acoustic scarab]

unfortunately i don't know

[finite pumice]

ok

[brittle wing]

Hello guys i need help with elevenlabs (TTS) i need voice MrBeast. Link me pls or dm

[tough fiber]

@odd shale

how can i open tensorboard

for mainline rvc

[gleaming scarab]

Now it's working but my voice doesn't change

[rare goblet]

Ayo? @gleaming scarab level 3 !!!

[vale osprey]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• 🆕 Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[rare goblet]

Ayo? @vale osprey level 3 !!!

[pastel oak]

Screenshot of voice changer please

[gleaming scarab]

It's picture I dont have discord on this computer sorry

[pastel oak]

F0 det: rmvpe

You didnt select GPU, do you have one?

You dont have virtual cable, you need virtual cable to use on discord and games

[gleaming scarab]

I have downloaded Voice changer on a small computer so I dont have is it necessary ?

[pastel oak]

For good performance you need GPU, cpu will take 10 seconds to change voice and maybe damage PC if you play game next to it

[gleaming scarab]

On my gaming PC start_http doesn't download voicechanger

[native granite]

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[tough fiber]

@pastel oak

shad

how can i open tensorboard on rvc mainline

theres no bat file for this

[pastel oak]

download tensorboard bat file

from docs guide

[gleaming scarab]

Dont ignore me 👉👈

[pastel oak]

saying doesnt download doesnt help me bro

i need details error log anything even a video of u trying it

i can check it out tmr

[gleaming scarab]

Ok I will send you it tmr is it okay for you ?

[proven pecan]

What's the best Text to speech AI software atm that is free for making my own ai audio books?

[proven hill]

free: gpt sovits
paid: elevanlabs

[brittle wing]

guys i need help with elevenlabs (TTS) i need voice MrBeast. Link me pls or dm

[brittle wing]

gpt sovits can use another voice models?

[cloud eagle]

the rvc dosent open

C:\RVC1006Nvidia>runtime\python.exe gui_v1.py
Traceback (most recent call last):
File "C:\RVC1006Nvidia\gui_v1.py", line 3, in <module>
from dotenv import load_dotenv
ModuleNotFoundError: No module named 'dotenv'

C:\RVC1006Nvidia>pause
Press any key to continue . . .

[rare goblet]

Ayo? @cloud eagle level 1 !!!

[cloud eagle]

if u know how to solve this dm me

[proven hill]

you can use gpt sovits models

[brittle wing]

i need mrbeast voice

[rare goblet]

Ayo? @brittle wing level 2 !!!

[proven hill]

check if it was made in voice models

[unkempt pasture]

anyone mind looking at my settings and tell me if i have them setup wrong or to optimoze real time voice changer i feel like it sounds very robotic and cracks alot

[native granite]

can someone very kind train a model for me? i can't get it to work :(

it's a short dataset

[rare goblet]

Ayo? @native granite level 1 !!!

[peak radish]

i cant download any files from colab

Failed to fetch
TypeError: Failed to fetch

[low shard]

#1159289738314919936 or #1191429836321849435

[low shard]

huh? what google colab link

[peak radish]

Applio

[low shard]

Be sure you are using the T4 daily free gpu and that its connected, and don't close the google colab

[modern surge]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• 🆕 Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[rare goblet]

Ayo? @modern surge level 8 !!!

[merry wharf]

-rvc

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[merry wharf]

-links

[azure marsh]

🖥️ Local stuff

• 🍏 Applio, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub
• GPT-SoVITS, by RVC-Boss GitHub

Credits to Faze Masta and Antasma for compiling these links.

More commands: -audio, -colabs, -overtrain, -realtime, -rvc, -help

[carmine hearth]

🥺 https://discord.com/channels/1159260121998827560/1202754985255764060

[supple turret]

Can i use rvc on windows 7?

[torpid wasp]

How to keep a voice model permanently on the list in ilaria RVC?

[tough fiber]

i didnt find

bat file

[pastel oak]

https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/

[tough fiber]

bro theres too many docs link

which one

thanks

i see now

[tough fiber]

ups sorry

im not sure which step is lowest point lmao

[pastel oak]

Did you set smoothing to 0.987

[tough fiber]

yea

[thin stump]

You'll notice the values go up as it ascends the slope. Lowest value = best step count

[tough fiber]

i know whats mean lowest point

[pastel oak]

It shows the step number when you hover over the graph

so look for the lowest number something with 31.xx judging from the graph

[rotund zephyr]

hello folks, is there any way i can make a rvc model sing? like they use in samples

[pastel oak]

Thats called Inferencing / making an AI Cover

https://docs.aihub.wtf/essentials/how-to-make-ai-cover/

If you have a good gpu, use the Local option Mainline.

If you dont want to bother installing a program and learning it, theres an easy way to do it online fast from a trusted person in this server with Ilaria RVC Zero

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

Don't forget to thank Ilaria if you find it useful! 💖

[pastel oak]

^ this for making it online

Would still recommend reading the docs thing I sent regarding "how to extract vocals"

[rotund zephyr]

okay thanks bro

[rotund zephyr]

i tried downloading it but it says error in status bar

this model only has the .pth file btw if that does anything

[pastel oak]

I dont remember fully if index is mandatory @proven hill can u have a look 🥺

[quick jungle]

https://cdn.discordapp.com/attachments/1160576561636966491/1165372018414342284/warning.gif

[rotund zephyr]

can i just upload a different index file from a different model and just not use it?

[pastel oak]

I suppose that works if you dont want the accent of the voice model

At least it works in local use

[rotund zephyr]

ok ill try it

[rotund zephyr]

it works now

thanks again bro

[barren dragon]

did u fix it?

[pastel oak]

Delete folders pretrain and model_dir then launch start_http again

[barren dragon]

thx

i fixed it

[gleaming scarab]

Ok I have succeeded to download it on my PC but I don't hear my voice change

[carmine hearth]

monitor

[gleaming scarab]

Will I hear my voice change ?

[carmine hearth]

Yes, you can. Change monitor to your sound device.

[gleaming scarab]

I did

i still not hear

[knotty moth]

bruh passthru should be green, not red (click it to change)

[gleaming scarab]

even green is not working

[proven hill]

yes

[pastel oak]

Two things:

passthru has to be green, this is how it works

If you mean you cant hear your voice on discord or games, you need virtual cable for output

The rest of your settings seem fine

[solar bridge]

Subject: Issue with Batch Processing of Audio Files in Python Script rvcgui.py batch adio editing #needhelp

Hello everyone,

I'm currently facing a challenge with a Python script designed for batch processing audio files. The script is supposed to take multiple audio files from a specified folder, process them using a separate script (rvcgui.py), and save the processed files to another folder.

However, I'm encountering issues where the script sometimes fails to create a log file, and it doesn't catch errors effectively. As a result, I’m unable to determine why the processing stops or where it fails.

Key Points:
Input Folder: Contains multiple .mp3 files.
Processing Script: Uses subprocess.run() to call rvcgui.py for each audio file.
Output Folder: Should store processed audio files.
Log File: Intended to capture the processing status and any errors.
Problems:
The script occasionally exits without creating the log file.
Errors are not being captured, making it difficult to troubleshoot.
I need a reliable way to process audio files in batches and handle any potential errors.
Questions:
How can I ensure that all audio files are processed correctly, even if an error occurs with one of them?
What are some best practices for logging errors and statuses during batch processing in Python?
Are there any libraries or methods that could improve the robustness of my script?
Any insights or suggestions would be greatly appreciated!

Is it able to receive such inputs?
rvc_gui_path,
"--input", input_path,
"--output", output_path,
"--model", model_path,
"--index", index_path,
"--f0_method", "harvest",
"--f0_pitch", "-10"

Thank you!

[gleaming scarab]

I can't hear my voice on the logiciel without virtual cable ?

[pastel oak]

Whats logiciel

[gleaming scarab]

voice changer

[pastel oak]

You can hear your voice with your headphones, but for OTHER PEOPLE to hear, you need virtual cable

[gleaming scarab]

okk thank you

[finite pumice]

hey guys what is the best settings for google colab rvc gpu T4 on python 3?

[cerulean creek]

Is it normal for your voice to sound warpy when hearing it under the AI model?

Seems like every voice I use, they all sound warpy and not normal. Can't even tell what I'm saying. My mic is a Blue Yeti.

[pearl lagoon]

how to disable echo?

he is off

But when I say any word, it repeats exactly 2 times

[pine bolt]

sorry if this is a common issue but do yall know how to fix a blank white interface after loading it up?

[pastel oak]

What does the console window show, any errors or weird thing

[pine bolt]

a few wait updates

[pastel oak]

Delete stored_settings.json then try again

[pine bolt]

didnt seem to work

[rare goblet]

Ayo? @pine bolt level 1 !!!

[pine bolt]

YUHHH

heres what it looks like

[pastel oak]

Just to be safe in case this is a rare double case, if the file got created again, delete it again lmao

[pine bolt]

good idea but the file wasnt created again

[pastel oak]

There was a report of huggingface being down 7 minutes ago, it could be connected to that (it needs huggingface to be running to work)

[pine bolt]

ok good point

my pc blue screened last night which was unrelated but i got a notif from windows saying it removed a program for causing problems, but i reinstalled the voice chancer and overwrote everything so idk

[pastel oak]

Should not be the error

Wokada works for me, so im not sure what the issue is unfortunately

[proper shale]

how does command prompt look like?

[proper shale]

someone NEEDS to remove the thing where it needs to check the sample list or something, and that's linked to hf

[pine bolt]

this good?

[proper shale]

yeah

hmmm

[pastel oak]

i dont know about any 2nd fixes to the wait web server, aside from maybe activating a vpn?

[pine bolt]

ok bro

i deleted the whole thing again and reinstalled it again and its booting up fine now

mustve been a random corrupted file after a bluescreen maybe?

bc i didnt fully delete the whole thing the first time i just overwrote the files

[pastel oak]

Alrightyy as long as it works:)

[odd shale]

🐢

[pine bolt]

legit

[empty fjord]

After downloading the Nvidia001 file from this guide and following the instructions to extract it, I encountered two errors. Is this normal?
https://rentry.co/ForkVoiceChangerGuide

[wispy lodge]

Did you also download 002 file?

[empty fjord]

Yeah

[wispy lodge]

Is it in the same folder? Maybe zip files didn't download completely

[empty fjord]

I downloaded it into the same folder.

[wispy lodge]

Then try redownloading zip files. Maybe the download was interrupted which happens sometimes with github links

[empty fjord]

I’ve re-downloaded it many times, but it’s always the same issue

[inner sparrow]

hey i have a question, when it says RVC v2 that means the second version of Okadas software correct?

and what the heck is the epochs number like, the higher the better?

[rare goblet]

Ayo? @inner sparrow level 1 !!!

[inner sparrow]

SHUSH'

[pastel oak]

Do you have an antivirus program?

[pastel oak]

No thats related to RVC model architecture version, nothing to do with okada

[pastel oak]

epochs is the cycle amount of how many times the model has been trained. higher epochs does not necessarily mean better model since theres an overtrain point and its unique to all models

[inner sparrow]

ok, and about the Epochs?

[inner sparrow]

you gorgeous human

thank you very much'

now i can sound like daddy mark

but abt the V2 of okadas Software

do you recommend that i use the V2 or the first

cause the V2 is still in alpha correct?

[pastel oak]

Yea its in alpha, dont recommend it

First is also kinda meh considering theres someone who made a modified version of okada that runs better, so id recommend the modified one

https://rentry.co/ForkVoiceChangerGuide

[inner sparrow]

WAAAAAAAAAAAAAAAAA

TYSMMMM!!!!!!!

TEHEHE

also are there like universal recommended settings?? other than the tone (which i know changes from model to model)

[pastel oak]

I mentioned all of that on the guide

[inner sparrow]

ooo theres a guide

okiii

lemme look for it

[pastel oak]

i linked the guide

the rentry.co thing

[inner sparrow]

oh kk

thanks alot shad!

appreciate ya <3

[pastel oak]

npnp:)

[inner sparrow]

whenver i choose my mic it shows this...?

[wispy lodge]

Try Microsoft Edge or Google Chrome

[inner sparrow]

dear god why....

ill try disabling an extension or two

[inner sparrow]

huh... this worked?

thanks

engi

chrome uses alotta ram ig, is there a way to cap this?

[rare goblet]

Ayo? @inner sparrow level 2 !!!

[violet heron]

Try using ungoogled chromium and see if that uses any less ram then normal chrome

[bitter tiger]

hey i need rvc google collab

[odd shale]

-rvc

[azure marsh]

Suggestions for @bitter tiger

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[odd shale]

If what you want to do is use Applio to test models, there you have too:

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb

[eternal galleon]

the voice changer its working perfectly when i try it out, but on discord it cuts out even with Krisp off, and im using Vcable, any idea why?

[faint island]

i wanna download the voice changer but i can't find the fil

[violet heron]

What’s your GPU

[violet heron]

use VAC Lite instead of VB Audio

[low cedar]

Rvc de miguel simon por favor?

[ionic fossil]

I added the Tensorboard batch file to my RVC mainline folder and installed it, but opening it leads me to nothing, just white. I had it working on another computer before, but I can't figure it out now. Does anybody know why that could be?

[alpine cobalt]

does anyone know what to do when it says Error: No available storage method found.

[ionic fossil]

I figured it out. In case anyone else ever has the same problem, follow these steps.

1. Open WIn+R regedit
2. Navigate to H_KEY_CLASSES_ROOT.js
3. Change the data of "Content Type" to "text/javascript"

[alpine cobalt]

mine for some reason does not have "H_KEY_CLASSES_ROOT.js"

[sinful prawn]

is there any free rvc training websites without any errors or problems, google colab or github, ill take it

[thin stump]

-kaggle

[azure marsh]

Kaggle Notebooks

• UVR5 NO UI by Eddy
• How to use RVC Mainline Kaggle by Cauthess
• 🍏 Applio RVC by IA Hispano
• ✨ RVC Mainline by Hina
• 🆕 Original W-Okada's Voice Changer Kaggle
• 🆕 Modified W-Okada's Voice Changer Kaggle

Note: Kaggle limits GPU usage to 30 hours per week.

More commands: -colabs, -hf, -local, -rvc, -help

[sturdy garden]

so im tryna use the voice changer (ive used it b4 just fine) but now im getting this error
'NoneType' object has no attribute 'host_api'

[devout star]

can someone help when i use ai in game it's go perfectly but in discord my friend saying it's laggy and cutting sometimes what should i do?

[rare goblet]

Ayo? @young belfry level 2 !!!

[young belfry]

how do these settings look?

they often tell me that they dont understand what im saying or it sounds like im trying to sneeze all the time

[devout star]

try to make chunk little more u dont need that much lowe u youtubers using it like that because they have 4090 super if u have u can use like that too

and my advice dont touch in and out

https://cdn.discordapp.com/attachments/1266761438446485565/1274219060166398064/image.png?ex=66c17479&is=66c022f9&hm=9464421375e0d385266223fcf7da3911322db3d05470d845a62df49509da7759&

u can use like that

[rare goblet]

Ayo? @devout star level 1 !!!

[cursive flare]

what the hel is this graph

[storm crag]

Bitcoin

[knotty moth]

NFT (in the graveyard)

[knotty moth]

• that is a mode collapse
• when resuming training, the batch size should be same as before

[cursive flare]

bitcoin be stepping in future timeline

[cursive flare]

I used 5 batch to get lagless from ~299 epochs to 300 epochs on discord voice call while sreaming, but the main batch size was 10, i did put it back. Does it affect voice negatively?

[knotty moth]

refer to this comment for choosing the batch size #📑│making-models message

[dire ingot]

yo is there any way to extract the samples of a ai model? like the audios used for training

[tame mica]

no

[dire ingot]

damn

[cursive flare]

thank you, but my gpu probably wotn be able to handle 16 batch size due to 8gb vram(rtx 4060). I can try tho

[young belfry]

it keeps sounding like im holding back tears

[knotty moth]

use batch size 8, or 4 if the dataset is short

[dire ingot]

bcuz a friend made a ai model of a song's raw vocals but wont give me the raw vocals directly, just the model

[cursive flare]

oki

[stiff rampart]

how do i download it

[snow hazel]

How do I remove harmonies?

[coral portal]

.

[jaunty shale]

not a help, but more of a question

is this normal?

usually the graph goes down

[sharp glacier]

is this correct I always forget cuase i'm to stupid.

and then what do I need to set these to

i fixed it. i had things right nvm.

[knotty moth]

too few epochs to judge

[jaunty shale]

I put 400 epochs for a 2 minute dataset then looking for the one that's overtrained, using rvc disconnected.

It went back to how it should be

[magic rivet]

Does anyone know if RVC is working with Kaggle?

I ran into issues

[thin stump]

Besides ngrok kicking me out for bandwidth limit, no

[magic rivet]

damn... because it keeps telling me this

do you know of any RVC forks that allow me to use a model uploaded to a google drive?

at this point it doesn't even need to be strictly RVC, just has to do the same thing essentially

[rare goblet]

Ayo? @magic rivet level 2 !!!

[thin stump]

for inferencing?

https://docs.google.com/document/d/1YbXcLFPaGjhOdG5NFkK3QrucCEpHZBwFUxkeMO8aB18/edit

[magic rivet]

I think so! I'm kind of a noob still. Thought Ilaria's was down

[thin stump]

it goes down at random

[magic rivet]

Yeah that would make sense. I think I struggled figuring how to upload a model there too, does it still work when hitting the "share" button on a model in google drive?

[thin stump]

just download the pth file and index and upload thsoe individually

[magic rivet]

gotcha! thanks!!

[thin stump]

thats what i do. too lazy to get huggingface

[magic rivet]

always meant to ask, does it worth without an index file?

[thin stump]

I thought it didnt let you upload a model without an added index so im not sure

[magic rivet]

it did but it sounded like ass

[thin stump]

Depends on the model. Most of the accent is baked in the added index apparently

because file size

[magic rivet]

ahhh that would make sense. I don't think you can make a model on the llaria though, right? is there still a way to do that?

[thin stump]

thats a different one https://colab.research.google.com/drive/1mHKTGH5e3SAyDSBss1KtiYRbDdQzwSMs. But we gave you kaggle anyways so just use that

for training models

[magic rivet]

tyyy

[magic rivet]

I actually ran into an issue with kaggle but couldn't figure it out

[thin stump]

Idk, could be button jamming etc. Nothing is broken though

[magic rivet]

it kept saying the resource already exists when i hit install so I thought maybe i created something but had no idea how to get to it

[silver pawn]

why it keep crashing

[magic rivet]

I figured it out, i had "use the latest version" ticked off!

[sinful prawn]

uhhh first time for me to train a model and i got this error: CUDA ran out of memory

can someone find out how to fix this

[tough fiber]

@pastel oak shad is it over trained right

its weird graph

[silver pawn]

it keeps crashing

wtf

[low shard]

are you running it locally? What's your PC GPU as it seems like its not enough

[pastel oak]

Yea the OT is somewhere between 31-34k

[knotty moth]

delete and copy again the original notebook, then reconfigure everything
(I'm prob talking about hina's mainline kaggle)

[crimson depot]

@viscid moss check this out

2024-08-17 09:13:30.804773104 [E:onnxruntime:Default, provider_bridge_ort.cc:1992 TryGetProviderInfo_CUDA] /onnxruntime_src/onnxruntime/core/session/provider_bridge_ort.cc:1637 onnxruntime::Provider& onnxruntime::ProviderLibrary::Get() [ONNXRuntimeError] : 1 : FAIL : Failed to load library libonnxruntime_providers_cuda.so with error: libcudnn.so.9: cannot open shared object file: No such file or directory

2024-08-17 09:13:30.804804822 [W:onnxruntime:Default, onnxruntime_pybind_state.cc:965 CreateExecutionProviderInstance] Failed to create CUDAExecutionProvider. Require cuDNN 9.* and CUDA 12.*. Please install all dependencies as mentioned in the GPU requirements page (https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), make sure they're in the PATH, and that your GPU is supported.

[rare goblet]

Ayo? @crimson depot level 4 !!!

[wispy lodge]

Ah yes, onnxruntime 1.19 update

[edgy flare]

i downloaded

but its not changing my voice

@pastel oak

like i hear my normal voice

[pastel oak]

If passthru is red, click on it to disable it so its green

Else send a screenshot of your app

[edgy flare]

i hear myself only when i turn on pass through

without pass throught it wont work

[pastel oak]

Make sure its green either way and send a screenshot

[edgy flare]

here

[pastel oak]

Chunk: 112
Extra: 16384

On discord, did you select "Input: Cable Output"?

[edgy flare]

yes i did

[rare goblet]

Ayo? @edgy flare level 2 !!!

[pastel oak]

Is there any error message or anything about Pipeline in the console window

[edgy flare]

no

the vol doesnot go up either while its on

[pastel oak]

select Start on Windows ,then select Settings > Privacy > Microphone . Select Change, then turn on Allow apps to access your microphone.

[edgy flare]

okay

is alr allowed

[pastel oak]

100% sure you dont see anything about Pipeline not initiated on console window? thats usually the problem

Select the Beatrice model and check if it works

[edgy flare]

Beatrice is working

it sounds awfull tho

[pastel oak]

Then your installation failed

Delete the folders pretrain and model_dir , then launch start_http again

[edgy flare]

alr

[edgy flare]

reinstalled still does not works

[pastel oak]

I dont know what the issue is then. Considering the other wokada version had issues installing for you aswell, I would suggest there is something on your PC blocking it from working

With no errors displayed I cant help

[edgy flare]

oh

[edgy flare]

oh

[random spindle]

My ai voice on Collab is lagging H A R D

Any ideas?

Been working fine all the other times

[empty fjord]

I only use the Windows firewall.

[rancid stone]

I'm not sure if it's client side, however #1175430844685484042 the threads arent appearing for me.

[pastel oak]

Download from the drive then

https://drive.google.com/file/d/1UwsMJTzKOMPEmrF17iq7cjZpH1A-ppr2/view?usp=drive_link

should only be this one zip I believe

[random spindle]

My ping is constantly fluxuating between 100ms every second

[timid crow]

would there any repair tools in rx11 that could increase my infered audio quality?

[grizzled moon]

when im using voice blender in applio colab it keeps saying ValueError: too many values to unpack (expected 2)
what to do?

i only put like two models

[patent pasture]

what version do you use?

[grizzled moon]

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb#scrollTo=-7cQtXouqpQi

[broken inlet]

hello

how much time do i need to wait for 400 epoch training ai model ?

[grizzled moon]

I dont use local btw

[pastel oak]

depends on your gpu, the length of the dataset, batchsize etc nobody can answer that for u

[broken inlet]

ok thank you very much sir

[patent pasture]

applio may have some bugs

[odd shale]

I don't think so.

[timid crow]

aw

[odd shale]

That will depend on your GPU and training settings like batch size and dataset length.

[grizzled moon]

any alternative colabs? (with voice blender)

[patent pasture]

Ilaria Mainline, but I don't remember if there is blender

[grizzled moon]

nope

[versed linden]

hey I just trained a model and I cant seem to find it in my logs folder, im using Applio local

[odd shale]

your model's pth files are located on assets/weights

[versed linden]

I do not have a /weights folder

these are all

[knotty moth]

it is named model/ckpt fusion, also there in mainline RVC

also you'd better take a look on this guide to know how useful it is

[pastel oak]

I said that too already 😅 docs also say its in logs folder, but apparently it isnt for Noahcrow

Does Applio do something differently or did the training not actually take place, would you know about that Leo?

[odd shale]

Haha i didn't see your message.

[odd shale]

Some folder paths are different on Applio.

I thought the pth's would be on assets/weights since Applio's folder path is a bit similar to Hina's mainline.

Maybe i got confused.

[knotty moth]

in Applio that I knew, it goes to the model's logs folder

[odd shale]

Then maybe Noahcrow did something wrong.

Maybe training for him didn't even start properly.

[viscid moss]

AICoverGen?

[vale osprey]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• 🆕 Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[wispy shuttle]

why are all my voices so choppy nd dont sound real at all

[brittle wing]

-help

[azure marsh]

Command has expired

[brittle wing]

-local

[azure marsh]

🖥️ Local stuff

• 🍏 Applio, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub
• GPT-SoVITS, by RVC-Boss GitHub

Credits to Faze Masta and Antasma for compiling these links.

More commands: -audio, -colabs, -overtrain, -realtime, -rvc, -help

[drowsy cobalt]

does anyone know how to convert pth file to onnx or where to get onnx voice models

[ionic fossil]

There's an option in W-Okada to convert the pth file to onnx if I recall correctly.

[pastel oak]

^ yeah wokada has a convert feature, though any reason youd need onnx models?

[devout loom]

hey guys i made a little model but i dont know how to make the sample with a preset audio any tips?

[tame wharf]

i have a problem with ngrok

It tells me that something has already expired.

[thin stump]

data transfer out capped. You need a new ngrok and phone number

[drowsy cobalt]

Im using AMD graphic card and it's not very compatible with pth files

[rare goblet]

Ayo? @drowsy cobalt level 1 !!!

[hot lagoon]

Y’all know where to find the best parts for build PC that include the 4090 rtx

[drowsy cobalt]

For some reason it cannot detect my graphic card

[ionic fossil]

Did you make sure to install the correct version of W-Okada? Just making sure because AMD has a separate version, I think.

[drowsy cobalt]

I did but its way to slow

[ionic fossil]

It does detect the GPU?

[drowsy cobalt]

Even i convert it into onnx it still works slower than last time

[drowsy cobalt]

Thats the problem

I cant really tell if its connected since its only showing the gpu list with numbers

[pastel oak]

If you have gpu(dml) with a bunch of buttons, you are on the AMD wokada version

If it doesnt have buttons, its the nvidia one

The right gpu button you can find on task manager > performance

Usually 0 or 1

Put chunk to a very high number, then export to onnx (original wokada is bugged) this may process faster

[untold steeple]

is w okada a new version for amd gpus?

[pastel oak]

I dont undertsand the question, wokada has versions for nvidia, amd, mac, linux etc. so yes amd is one of them

for amd gpus its recommended to use this version of wokada:
https://rentry.co/ForKVoicechangerguide

[untold steeple]

Nm it was s dumb question, thnaks for anwering!

[magic rivet]

Are there any recommendations for what settings to use when training a model with the kaggle?

also, does it reference all wav files I put into the dataset glob folder? Or do I need to merge all the separate wav files into one big wav?

[thin stump]

I think you have batch size covered in the guide, maximum 1000 epochs, save frequency depends on you. You still need to delete epochs to maintain storage

2. we split the files using audio labeling so there's less chances of a mode collapse, and introduce more stable graphs https://rentry.co/RVC-dataset-RX11#noise-gating-and-audio-labeling

[magic rivet]

thanks!

[median delta]

how do i make a voice model for rvc?

[devout loom]

An error occurred extracting the index: need at least one array to concatenate
If you are running this code in a virtual environment, make sure you have enough GPU available to generate the Index file.

does anyone know how to fix this, it appeared when i tried to generate the index at applio

[rare goblet]

Ayo? @devout loom level 1 !!!

[median delta]

i already have the audios ready

[violet heron]

-docs

[azure marsh]

Suggestions for @median delta

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• How to use RVC Mainline Colab by Cauthess
• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[fathom geyser]

I have an error when I try to download the output of my mp3 audio files in Applio. What I did wrong? Can anyone help me?

[finite prawn]

if i have more data i'd like to use for an existing voice model, am i able to update the model with the new data, or do i have to create an entirely new model with the new and old data?

[odd shale]

You would have to make a entirely new model using the new and old data.

[finite prawn]

figured. thanks! :)

[gray dagger]

can i train a model on amd?

[fathom geyser]

How can I fix this???

I have this problem and I don't know why. Does anyone know why this is happening to me???

[fading lodge]

any way i can make my index to 0.75 instead of 0.1?

[compact iron]

whats the best way to make a model? ive got a bit over 10 minutes of data

[sturdy sage]

You will likely want a disconnect prevention method to keep Colab from timing you out for "inactivity" during training.
I won't be providing one here, but they're not hard to come by. Search around the Colab AI community and you'll find one. ;)

sooo where do i find it?

[thin stump]

read the section on idle timeout https://docs.google.com/document/d/1Pr-AZndodmWgsbOeuHQU4IrgbatFgYc1ChOq_ZAf_5s/edit or just use kaggle

[thin stump]

-kaggle

[azure marsh]

Suggestions for @sturdy sage

Kaggle Notebooks

• UVR5 NO UI by Eddy
• How to use RVC Mainline Kaggle by Cauthess
• 🍏 Applio RVC by IA Hispano
• ✨ RVC Mainline by Hina
• 🆕 Original W-Okada's Voice Changer Kaggle
• 🆕 Modified W-Okada's Voice Changer Kaggle

Note: Kaggle limits GPU usage to 30 hours per week.

More commands: -colabs, -hf, -local, -rvc, -help

[sturdy sage]

thanks

[halcyon anchor]

hi i was just trying out this program and watching a youtube video i seem to not have the free models and also the upload button
https://gyazo.com/66be7aaa7431ddc7a6002f51b66d5a02

[rare goblet]

Ayo? @halcyon anchor level 1 !!!

[eternal turtle]

lookin pretty good

[halcyon anchor]

i keep getting this error message
VC PROCESSING!!!! EXCEPTION!!! 'NoneType' object has no attribute 'get_processing_sampling_rate'
Traceback (most recent call last):
File "voice_changer\VoiceChanger.py", line 401, in on_request_sola
AttributeError: 'NoneType' object has no attribute 'get_processing_sampling_rate'

[eternal turtle]

can anyone give critique on my ai voice model? the source came from noisy audio but i cleaned that up real good

[molten marlin]

is rx 570 8gb oc good for the voice chabger?

[pastel oak]

Not the best but itll run decent on the fork, wouldnt recommend original wokada for amd
https://rentry.co/ForkVoiceChangerGuide

[molten marlin]

thanks but may i ask how many delay would i get

[pastel oak]

I havent had anyone with an RX 570 before, so the value I put in the guide is an estimate

Assuming you are able to run 200 ms chunk, this would be 200 + 100 crossfade + whatever natural delay comes from other devices like you speaking into your microphone before it goes into the voice changer

Assuming you use wasapi for inputs/outputs (which reduces 50-80ms), anything between 0.4 - 1.0 seconds total delay is possible

[molten marlin]

Thankss

[harsh dome]

What is the best site currently for make AI cover?

[pastel oak]

To make AI covers online, currently the best one I'd recommend is Ilaria RVC Zero

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

Don't forget to thank Ilaria if you find it useful! 💖

[pastel oak]

^

[harsh dome]

TYSM ❤️

[halcyon anchor]

https://gyazo.com/7dc57aeca6867a79f4fcc946953b7196
is this thing meant to start onto a desktop server type? i thought it's an application type

and it's not showing any of my input and output options

[pastel oak]

Yes

[pastel oak]

Can you screenshot the options?

[halcyon anchor]

sure!

[pastel oak]

But in either case, you can also choose Audio: SERVER with the prefix [MME] and see if the options show up there

(You can also use [WASAPI], which provides less delay, but you cant use the noise suppression functions if theyre important to you)

[halcyon anchor]

server options works thank you very much!

for some reason it shows up now

this is a lot more work than dubbing ai but it does sound better

[pastel oak]

Only the setup may take some time, afterwards its all good:)

[pastel oak]

Also just realised using any of audio:server disables the noise supp, so may aswell take wasapi

Wasapi may crash if you havent configured your audio devices, which takes like a minute to change in case it crashes. Heres the quick explanation in case you need it:
https://rentry.co/LessDelayWasapi

[barren mortar]

guys what kinda voice changer should i use i actually need it

[pastel oak]

https://rentry.co/forkvoicechangerguide

[gusty kestrel]

@marble mountain, I have found 2 results that match your search!

spanatz

Omen [Valorant]

• Uploaded: <t:1692057600:d>
• Server: AI Hispano
• Likes: 0
• Lang: Spanish

Epochs

300

Technology

RVC

Algorithm

Crepe

Requested by deus_fps · Aug 18, 2024, 5:40 PM

@marble mountain, I have found 2 results that match your search!

spanatz

Chamber [Valorant] - Latino

• Uploaded: <t:1696695746:d>
• Server: Unknown
• Likes: 0
• Lang: Spanish

Epochs

500

Technology

RVC

Algorithm

Crepe

Requested by deus_fps · Aug 18, 2024, 5:43 PM

@marble mountain, I have found 2 results that match your search!

l0s3s0brads

Gekko [Valorant]

• Uploaded: <t:1706302851:d>
• Server: AI Hispano
• Likes: 0
• Lang: Spanish

Epochs

480

Technology

RVC

Algorithm

Rmvpe

Requested by deus_fps · Aug 18, 2024, 5:44 PM

[patent pasture]

what have I to do next?

[halcyon anchor]

for some reason the rvc keeps dying on discord and noo ne can hear me but i can hear myself fine from the rvc server

[rare goblet]

Ayo? @halcyon anchor level 2 !!!

[pastel oak]

Are you using VB-CABLE or Line 1/VAC?

And send a screenshot of your settings too

[red kayak]

model has overtrained early on pretty much, lacking new diverse outputs

take these 2 low points and compare them

if u have them saved that is

[timid olive]

@cyan depot

[cyan depot]

applio gives this error

[timid olive]

we couldn't figure it out

[patent pasture]

ok, if I set Save Every Epoch to 20 and those 2 points weren't saved, do I need to take other point with low value?

[fading lodge]

every time when i use any voice, it always makes it sound very wierd before the end of each sentince

and ive even reinstalled to the right version, it didnt help

[red kayak]

well in that case take something around 1k and around 2.7k and compare

[patent pasture]

thanks for explaining

[halcyon anchor]

https://gyazo.com/c0c130f3aba4c15ff7f4388f84eeb98e

[pastel oak]

First, select F0 det: rmvpe

It may also help to not use sup1 and echo. You can use sup2, but sup1 is not preferred. Echo might be the reason words are cutting off

[halcyon anchor]

that works great discord can finally hear it thank you

[pastel oak]

If you want, you can increase Extra to 2.7s for more quality. And since your perf number on the screenshot is around 109, you can also reduce the chunk (delay) to something like 150

[halcyon anchor]

that sounds slightly more crisp! thank you!

[cosmic tide]

why its not making the audio?

[signal oracle]

hi i have this problem with RVC And SD... do someone help me pls!

[rare goblet]

Ayo? @signal oracle level 2 !!!

[violet heron]

-gui

[azure marsh]

https://cdn.discordapp.com/attachments/1122285248844144733/1203460490475343953/caption.gif?ex=65d12cec&is=65beb7ec&hm=bd2fb8d010006dd7c6e3c1c67d3ae846fd1478e1a3124c544c31b43086fe54aa&

[verbal widget]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• 🆕 Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[amber charm]

Hey, I got a 18 mins dataset in 48k to train recorded by my boyfriend of him singing and talking using a studio mic with a pro quality, And i got a 4070ti RTX card with a good processor, what batch size should i put? I'm gonna use crepe

Ik batch size 4 gives a really good quality but it's so so slow

[magic rivet]

I was trying out rvc disconnected and got over halfway trained, but got disconnected. Is there a way to pick up where I left off? Trying to hit "import model from drive to notebook" says it has missing parts

[rare goblet]

Ayo? @magic rivet level 3 !!!

[magic rivet]

I know it saved after epochs

[low shard]

@marble mountain there's #🔍│find-models & #🤖│bots btw

[eternal turtle]

8

[amber charm]

Thanks

[rare goblet]

Ayo? @amber charm level 1 !!!

[crude flame]

Don't use normal crepe use mangio crepe with a hop length of 128 or lower

[amber charm]

I use Applio and I don't have Mangio Crepe, how can I get that?

[crude flame]

Oh NVM you got Mangio crepe

Applio renamed Mangio crepe to just crepe

[amber charm]

Oh period, thanks 😊

[crude flame]

Also make sure you are using fp32

Makes it sound a bit better

[amber charm]

Oh thanks for the advice!

[amber charm]

I'll try that now

[quaint dock]

hey, can I ask for weights.gg model training help here?

[fading lodge]

can someone please help me

[low shard]

😭

[proven hill]

why

[spare shale]

hey @low shard What are the currently popular colabs to create a voice model 😅

[crude flame]

Simpler and caused less issues

[spare shale]

https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw
Doesn't work or I'm stupid idk

[proven hill]

strongly disagree

[low shard]

• Google Colabs (4 hours of daily gpu for free, not much hours, but easy to use):
• Applio (ui)
• Mainline (UI)
• RVCDISCONNECTED (no ui)
• Kaggles (a bit harder to use and needs phone number but gives 30 hours weekly of better gpus):
• Mainline (UI)

[low shard]

easygui google colab's creator isnt active since month and its broken #📰│dev-updates message

[low shard]

a 'fix' would be adding !pip install pip==23.1 at the top, but i looked over the code and there's other outdated things, its overall not good to use as theres other errors, using the colabs i mentioned are just better

[fossil vigil]

yo is there a version of rvc gui for amd?

[low shard]

if you meant https://github.com/Tiger14n/RVC-GUI, its outdated, and no RVC doesn't have AMD support

Its better you use cloud ways, are you looking for inference (use models) or train (make models) ?

[fossil vigil]

just for fun as long as its better than the cpu encoding

[low shard]

there are different cloud ways depending if u wanna do inference or training

thats why i asked u which one u wanna use

[fossil vigil]

ah ok

dont want to go into a loop hole or whatever

do i download the outdated or use cpu encoding

[low shard]

if u want i can give u the cloud guides

[fossil vigil]

im good thanks

[low shard]

dont use the outdated rvc gui, the updated mainline and applio are in the docs https://docs.aihub.wtf but they don't have amd support soo, and yea u can (slowly) inference on cpu but not train

[fossil vigil]

ah ok thanks

[low shard]

yw

[knotty moth]

if in case the model doesn't sound good or lacks generalization due to a lot of mode collapses, try the audio labeling method instead of the old truncate method for your dataset https://rentry.co/RVC-dataset-RX11#noise-gating-and-audio-labeling

[magic rivet]

why do you recommend mangio crepe and not crepe?

[crude flame]

mangio has adjustable hop length and normal crepe doesnt

crepes default hop length is like 312

[knotty moth]

you prob meant mangio crepe in the og mangio (or rvc disconnected), versus the adjustable "crepe" in applio

[magic rivet]

yeah that one!

[brave garnet]

RVC Colabs and Spaces

⠀

HuggingFace Spaces 🤗

⠀
Ilaria RVC
EasyGUI port with some improvements, by Ilaria.

RVC-HFv2
Applio port, by r3gm.

AICoverGen
AICoverGen port, by r3gm.

Advanced RVC Inference
Extended version of the GUI with advanced settings, r3gm.
⠀

Made with love by Ilaria 💖

RVC Colabs and Spaces

⠀

HuggingFace Spaces 🤗

⠀
Ilaria RVC
EasyGUI port with some improvements, by Ilaria.

RVC-HFv2
Applio port, by r3gm.

AICoverGen
AICoverGen port, by r3gm.

Advanced RVC Inference
Extended version of the GUI with advanced settings, r3gm.
⠀

Made with love by Ilaria 💖

RVC Colabs and Spaces

⠀

Google Colabs

⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.

AICoverGen-NoWebUI
Useful for making covers, doesn't inclued a UI, by Ardha, by Eddy, Hina and Gdr.

RVC Disconnected
To train new voice models, by Kit Lemonfoot.

EasyGUI
The OG interface, by Rejects.
⠀

Made with love by Ilaria 💖

[brittle wing]

Hello guys! I want to use only one specific voice, but in some sentences it goes super robotic often and I want to know if there is anything I can do to fix it or improve it.

I'm using deiteris' modified W-Okada RVC Client with an 4070ti, and F0 Det.: rmvpe, always on green perf value at around 25-35ms. Changing chunk value or extra never fixes it. The other settings are pretty much untouched.

If you can give me some tips to make this specific voice sound better it'll be greatly appreciated because it is the only one that interests me (Yusuf Dikec)

[rare goblet]

Ayo? @brittle wing level 1 !!!

[soft stratus]

But you told me easygui is broken, How will be working

[fading lodge]

hey

have u went to ur taskmanager

and changed audiodg?

@brittle wing

[knotty moth]

try the non-onnx rmvpe or try another better model to compare

[brave patio]

im on a great pc but the rvc is still bad quality and i cant get it to sound realistic

everyone can instantly tell its a vc

[spring socket]

Can I ask guys,
When I download the voice model, the huggingface website is telling me:
"Repository not found"
What should I do to get the model? Thankyou!

https://huggingface.co/SeoulStreamingStation/IU-Voice-MultiLanguage-V1/resolve/main/KoreanGirlPro_Vol2_KLMv7s.zip?download=true

[knotty moth]

what kind of model are you looking for?
speaking of the KLM pretrain, the latest available ones are 4.2 for 40k sample rate, and 4.1 for 32k & 48k https://discord.com/channels/1159260121998827560/1265083834039533588
so it might be the reason of the model availability

[spring socket]

I'm very new to RVC and doesn't understand any of that (KLM/Pretrain/SampleRate etc) :<
I just want to test the voice, download and upload to Okada like any other models that I've test.
My apologies before 🙏

[rare goblet]

Ayo? @spring socket level 1 !!!

[brittle wing]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• 🆕 Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[low shard]

Seems like the model has been deleted or privated, where did you find that?

Oh found it here https://discord.com/channels/1159260121998827560/1229631472315535391

Should tell the creator

[brave garnet]

RVC Colabs and Spaces

⠀

HuggingFace Spaces 🤗

⠀
Ilaria RVC
EasyGUI port with some improvements, by Ilaria.

RVC-HFv2
Applio port, by r3gm.

AICoverGen
AICoverGen port, by r3gm.

Advanced RVC Inference
Extended version of the GUI with advanced settings, r3gm.
⠀

Made with love by Ilaria 💖

[spring socket]

Thankyou for helping me!

So it's deleted? So that's why :")

Want to try it so much hahaha

Thankyou mods

[low shard]

Either deleted, moved somewhere else or privated, it's the author choice so I asked him about it as the post is still up and other may not know about this

For now, I think it's better you try another model

[low shard]

Your welcome

[noble wyvern]

why it keep stating illegal combination of I/O device

its really wierd

[pastel oak]

Are you using RVC Realtime? Make sure both inputs and outputs are (MME), it should be telling you to do that on the GUI aswell.

[knotty moth]

ah well, try the client audio devices first, otherwise the MME or WASAPI ones

[noble wyvern]

ight

[brittle wing]

I was using the first one always and all models I try have the same issue. Robotic words randomly sprout in every 2 or 3 phrases.

[brittle wing]

Yep, it did nothing. It is not a crackle it is full on robotic

It is maybe the rvc I'm using. I've been told it is the best one for Nvidia but maybe I can try others. I tried another one with a different interface, more rustic. RVCNvidiaLatest is how the folder is named

Are there any other options available?

[pastel oak]

deiteris Fork is the best option, the RVC Realtime voice changer wont make much of a difference, itll have the same results

The problem is: not every voice model is perfectly usable. You will have to try to fit the energy/tone/speech of the person you are trying to imitate if the voice model is not suitable for realtime (since most models are made with inferencing (AI Covers) in mind) - in your case, you might have that kind of voice model if its consistently robotic. Nothing to be done, it would be advisable to train the model yourself if you know how to properly set up the voice model for realtime

I am assuming you are trying this voice model? https://thevoicemodels.com/preview?id=1272604151087628328

The audio sample shown there I am 99% sure is inferenced and not used in realtime

If its that voice model, the creator of that model is also in this server. You could reach out to him and ask him to train this model with mangio-crepe pitch algorithm instead of rmvpe. People have said training with that algorithm "smoothens" the voice more, makes it softer in realtime usage

[knotty moth]

unless you set index rate more than 0, go find better model, or u can try comission in https://discord.com/channels/1159260121998827560/1191429836321849435

[brittle wing]

Could it be possible to do it myself?

[brittle wing]

Changing the index makes it sound even more robotic

[pastel oak]

You need to prepare the dataset yourself then. How to do that and how to clean it is explained on the docs
https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[knotty moth]

yeah I mean keep it 0

[brittle wing]

I'll give it a try then, thanks guys

[plush compass]

Index Algorithm
Auto
Faiss
KMeans What are these three options?

[pastel oak]

what do you mean does not contain chinese 💀 post a screenshot

Use this
https://rentry.co/ForkVoiceChangerGuide_tr

[surreal magnet]

100% safe

Shad is the person you can trust on this cord

[rare goblet]

Ayo? @brittle wing level 2 !!!

[surreal magnet]

function wise, if you want to "voice changer" realtime then yeah

no problem, also i recommend the deiteris fork of w-okada

gave better performance than rvc realtime

[pastel oak]

the turkish guide i linked is atm better than normal rvc, it has more features

Theyre all false dw

False positives because it doesnt know what file it is so it thinks its virus, happens all the time. Code is open source, experts went through the code before, its safe

[surreal magnet]

w-okada is easier imo, they gave you this feature that tells you whether the voice changer is lagging or okay

also i have it installed without problem

try this link
https://rentry.co/ForkVoiceChangerGuide_tr#nvidia-i̇ndirme

cuda001-b2300 and cuda002-b2300

do you want forked w-okada like in the guide or normal w-okada ?

https://huggingface.co/wok000/vcclient000/tree/main

[pastel oak]

https://github.com/deiteris/voice-changer/releases

I wrote the guide, so you can trust it but i recommend this wokada version instead of normal. heres the github

If you dont care about anything we are saying (respectfully) then here https://github.com/w-okada/voice-changer

[chrome sphinx]

hey, but does anyone know how to make ai models whisper or growl, like in vocaloid?

[autumn dust]

anyone can help with ilaria model download error?

[proven hill]

whats the error

[autumn dust]

I get error when trying to download model

"failed to unzip file"

although I have pth and index in the zip

[proven hill]

theres only a pth and index or theres something else?

[autumn dust]

no only those 2

[rare goblet]

Ayo? @autumn dust level 1 !!!

[proven hill]

they should not be in a folder

[autumn dust]

oh just straight in zip?

[proven hill]

yes

thats your problem

[autumn dust]

thank you !

[proven hill]

no problem!

[autumn dust]

same error for some reason, anything to do with file names?

[proven hill]

show

[autumn dust]

does the huggingface thing have to be public?

[unborn owl]

guys i think easygui is broken

[proven hill]

seems to be normal

[proven hill]

yes

[proven hill]

yes its broken

[autumn dust]

I guess thats another reason

[autumn dust]

now it works lol, thanks

[low shard]

#📰│dev-updates message yea it is

use an alternative

[proven hill]

lmao

[strange orbit]

hello I am new to this wanted to ask if they're is something better than this https://github.com/Tiger14n/RVC-GUI/releases/ I got from a yt vid some more better for voice quality?

[pastel oak]

For online use Ilaria RVC Zero

[brave garnet]

Ilaria RVC: CLICK HERE 🤗

Guide on how to use it: CLICK HERE 📝

Don't forget to thank Ilaria if you find it useful! 💖

[strange orbit]

anything I can download into my pc?

[pastel oak]

https://docs.aihub.wtf/rvc/local/mainline/

[strange orbit]

nice I'll take a look I appreciate the help man

[vale osprey]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab
• UVR5 UI, by Eddy Google Colab
• 🆕 Modified W-Okada's Voice Changer Google Colab

ℹ️ Note

While the Colab free plan provides up to 12 hours of daily usage, the GPU is typically available for only about 4 hours each day on average.

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[placid heath]

i have a question, how do people end up making like wet samples vs dry samples

like how do you get it to sound like wet samples

[odd shale]

You can make wet samples by just adding echo and reverb to your sample made with a model made by yourself/someone else

[fading lodge]

sorry if this will be a dumb question. where do i start making voicemodels?

[pastel oak]

https://docs.aihub.wtf/essentials/how-to-make-voice-models/

[fading lodge]

thanks man

[fading lodge]

@pastel oak ive been reading this article and i have a small question

where can i download videos for sampling and listening to datasets with?

im already downloading UVR

[odd shale]

You can download MediaHuman, cobalt.tools site or y2down.

[fading lodge]

thanks

i think this is gonna be the hardest thing im going to work on yet

😭

from what i see, voicemaking looks very difficult

[odd shale]

Nah building datasets after a lot of practice isn't that hard.

[fading lodge]

how long did it take you to get the hang of it?

[odd shale]

Various months.

[fading lodge]

and can you teach me some tips/basics on it if your avalible now or later?

i just need to get the hang of it

then ill be good

[odd shale]

Actually on some things you must learn them by yourself.

[fading lodge]

i just need to know how to use a few apps

[odd shale]

If you wanna know where to start, just download a voiceline video of any character

[fading lodge]

i can learn the others on my own

[fading lodge]

oh

alright

thanks

should i use local or webui?

does either matter/

?

[odd shale]

Do you mean RVC?

[fading lodge]

yeah

im not sure how to explain sorry

[analog obsidian]

start by doing game characters as the cleaning process for those are extremely minimal and they have good quality by default (they're only a bit compressed and thats it)

[odd shale]

local if your pc is strong enough.

[fading lodge]

yeah

i have a RTX 3070

[odd shale]

Deja que yo me encargo Lyery porfis.

[fading lodge]

im getting a 4060 in a few months

i should be good

[odd shale]

Nice.

[fading lodge]

alright

[odd shale]

Which is your current processor?

[fading lodge]

i7 12700F

16gb ram

[odd shale]

You'll be good on local then. And vram?

[fading lodge]

ive been using okada for a while now i wanted to make some voicemodels

[fading lodge]

vram?

not sure yet

one moment

16gb

i belive

wait

no

thats memory

[odd shale]

Yep, that's RAM.

[fading lodge]

ohh

you said vram, got confused

my bad

[odd shale]

Actually i asked for your VRAM

hehe, sorry for the confusion.

[fading lodge]

like ddr3 or ddr4?

[odd shale]

I mean video ram.

[fading lodge]

oh

[odd shale]

Every graphic card has X amount of VRAM if it's a dedicated one.

[odd shale]

Una preguntita Lyery

Cuanta Vram tienes?

Se que tienes 32 GBs de ram

[analog obsidian]

8gb, lo maximo que te deja usar mainline sera 8 batch size

[odd shale]

Nice.

[fading lodge]

any sources on where i should get game voices?

[analog obsidian]

rip them from the game

[odd shale]

Youtube videos or look around on internet.

[fading lodge]

alright

[fading lodge]

ehh

that takes a bit too much time

i can understand making voices takes a bit but

[analog obsidian]

yes but youtube compresses the audio even more

[fading lodge]

is there any guides on using this?

[analog obsidian]

but alr get them from yt

[fading lodge]

idk what to do for imput/output

is it for my mic or for an audio file?

[odd shale]

Put the path to the audio file you wanna isolate on input

[odd shale]

Audio file.

[fading lodge]

i figured

lol

im definetly going to start doing this alot

[odd shale]

Output is for the place where the isolated file will appear.

[fading lodge]

so do i keep that untouched?

or do i put the isolated file

sorry, just confused on this stuff

[odd shale]

Actually you can select a folder or any place on your Drive where you want to get your isolated file.

You can't let that on blank

[fading lodge]

my desktop?

[odd shale]

Yep, or any folder

[fading lodge]

alright

let me make onje then

is there any other settings i should do for this?

[odd shale]

Para un principiante si lo que quiere es simplemente probar cosas dudo que lo que le importe sea la calidad Lyery.

Así que Youtube es un buen lugar para empezar.

[fading lodge]

like what should i check off and stuff

[analog obsidian]

pero si no dije nada malo

si yo igual uso youtube a veces sjdf

[odd shale]

No malinterpretes pequeñito, solo corrijo humildemente.

No te estoy regañando.

[analog obsidian]

yay

[odd shale]

Yo uso mediahuman para agilizar.

Cualquier cosa menos Y2mate

[analog obsidian]

yo yt-dlp con el comando -x para conseguir el .opus

[fading lodge]

does mediahuman go quicker then others?

[analog obsidian]

yt-dlp literally takes one command and thats it

[odd shale]

Use any tool you want.

[analog obsidian]

yt-dlp.exe -x url

[fading lodge]

damn

never knew that

lol

[analog obsidian]

xD

[odd shale]

Jamas supe como configurar el yt-dlp ese.

[analog obsidian]

te enseño?

[odd shale]

No sé, creo que mi laptop tiene algo que no me permite ejecutarlo.

[analog obsidian]

el exe de github no se tiene que abrir normalmente, tienes que hacerlo por el cmd

[odd shale]

Si quieres me enseñas por privado cuando te lo pida

[analog obsidian]

oki no hay problema, es facil y fast

[odd shale]

Tqm Lyery.

[fading lodge]

@odd shale one question

how long would u think it would take for me to make genuine decent models in the span of 3-4 months?

would that depend on how long i work on each model or how long i look at like tips or tricks on guides, etc

[odd shale]

Both.

Both things.

You can use any dataset length for any model.

[analog obsidian]

if u need help don't be afraid of asking here or #📑│making-models

[fading lodge]

alright, thanks

[odd shale]

What lyery said too :))

[fading lodge]

so like

for instance

i was reading the guide

and it said

if i do models over 40min

its unessocary

what if i do like

1 hour

would it make a diffrencf

?

[odd shale]

Actually you don't need to use exactly 1 hour of dataset.