help

[slow hull]

total_epochs whats average to use ?

[warped galleon]

keep at 500, check graphs for overtraining

[slow hull]

i dont understand anything on tensorboard

how to check graphs

[warped galleon]

-overtrain

[slender vault]

Hey, @slow hull!

👇 Here are some resources to help you identify if your model is overtraining

How do I know if my model is overtaining?

All-In-One Guide on how to make a good model

This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models

Credits: LUSBERT

Automated Overtraining Detection (AOD)

Will be available soon in #1159513888199540817

Credits: grvyscale

[warped galleon]

all in one guide should explain

[slow hull]

i dont understand thing

if i keep 500 can it sound good

this is example

[warped galleon]

i mean the quality IS okay but...

why is it trimmed like that

[slow hull]

because he takes lot breath takes

some1 told me thats bad

so i have 30 clips

of 2-3 sec

[warped galleon]

you didn't need to trim it like that though

could have just removed them with a noise gate and kept everything in one audio

[warped galleon]

yeah?

[warped galleon]

and breaths are fine if they're not excessive

[slow hull]

ok thx will keep tht in mind and can i play games and keep browser on or i need be on site

while it trains

[warped galleon]

id recommend to watch how the training is going

and stop training after 1/2 hours

[slow hull]

why stop

[celest spire]

google colab may disconnect you due to gpu usage limit

[warped galleon]

^^^

[slow hull]

this for pause

[warped galleon]

if you're using RVC Disconnected especially, you must export model or else ya lose everything

[warped galleon]

yeye

[slow hull]

👍

i runned Feature Extraction

[warped galleon]

i think ya screwed up on the dataset formatting

The dataset must be formatted like this;

• folder with your model's name (must be the same as experiment_name you will set on the Colab, no spaces too)
• Audios here (no spaces here too)

Also, make sure you don't use .rar and instead use .zip

[slow hull]

its same name and zip

does experiment_name need to be same as folder name

[celest spire]

yes

[slow hull]

i need restart everyhing?

-colab

[slender vault]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface Spaces

• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

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

[slow hull]

can numbers be in name

[warped galleon]

yes!

but make sure its the same in the folder name

[slow hull]

same error its same name now

zip file name and folder in it can have same name?

[warped galleon]

yeh

also make sure nothing has a space or anything

[warped galleon]

might be ur audios then

[slow hull]

irs 70 seconds long

training done in 10 min

[warped galleon]

i suggest putting it all on audacity and re exporting as a single wav file

[slow hull]

ok thx

[slow hull]

it worked now thanks so much

[celest spire]

mark this as solved

[slow hull]

btw wich one of those 2 pth i download

klinja is one i named

[celest spire]

d and g are training

the one you need is [modelname].pth

so id assume last one

[slow hull]

👍

[thorny talon]

Ayo? @slow hull level 5 !!!

[slow hull]

how to get sample voice of model