Software Engineering Courses (SECourses)•3y ago

yes but before this new update to the patreon

ElisoliOP•10/12/23, 7:10 PM

virginia•10/12/23, 7:10 PM

@Dr. Furkan Gözükara also got it, let me try again

FFurkan Gözükara SECourses please use my kohya installer on runpod it has instrucitons too : https://www.pa...

Jas•10/12/23, 7:14 PM

I am using the Ashleykza template which has Koyha installed, but I was worried that the trainings are not currently good, due to the update you posted on Patreon.

Or am I getting confused and this does not affect trainings?

Tt800se I'm getting this error now 😦

Furkan Gözükara SECourses•10/12/23, 7:16 PM

use v4 file

Furkan Gözükara SECourses•10/12/23, 7:16 PM

redownload please

EElisoli Click to see attachment

Furkan Gözükara SECourses•10/12/23, 7:16 PM

please try reinstall

Furkan Gözükara SECourses•10/12/23, 7:16 PM

i am testing too

FFurkan Gözükara SECourses use v4 file

t800se•10/12/23, 7:17 PM

will do now, thank you

EElisoli yes but before this new update to the patreon

Furkan Gözükara SECourses•10/12/23, 7:17 PM

you are using all_files.zip right?

Furkan Gözükara SECourses•10/12/23, 7:18 PM

preparing runpod version of captioner

JJas I am using the Ashleykza template which has Koyha installed, but I was worried t...

Furkan Gözükara SECourses•10/12/23, 7:19 PM

i am not sure. i havent tested. but i am still searching for a better workflow for sd 1.5

FFurkan Gözükara SECourses i am not sure. i havent tested. but i am still searching for a better workflow f...

Jas•10/12/23, 7:23 PM

ok, I figured out what to do, just had to change PUBLIC=false to PUBLIC=true in the setup.sh file

t800se•10/12/23, 7:43 PM

Ok, I am not getting that error with the v4, thank you @Dr. Furkan Gözükara . I see there is a new cell in this version, when shall I execute that one?

Tt800se Ok, I am not getting that error with the v4, thank you @Dr. Furkan Gözükara . I ...

Furkan Gözükara SECourses•10/12/23, 7:44 PM

it is optional

t800se•10/12/23, 7:44 PM

ahh good to know

Furkan Gözükara SECourses•10/12/23, 7:54 PM

runpod installer added

Furkan Gözükara SECourses•10/12/23, 7:54 PM

https://www.patreon.com/posts/90744385

Furkan Gözükara SECourses•10/12/23, 7:54 PM

still testing

SuperTurboHero•10/12/23, 9:20 PM

i must be doing something wrong

it's RealisticVision5.1, Dreambooth trained. on do you guys know any good tutorials where I can learn how to get good looking realistic photos?

SuperTurboHero•10/12/23, 9:25 PM

the best results so far I got from Automatic1111 v1.4.0 interface, because there are these options "restore faces" and "Hires.fix"

t800se•10/12/23, 9:26 PM

Hey @Dr. Furkan Gözükara , very silly question probably but I am new in all of these

I just started the training process in Kaggle Notebook by watching your videos. I didn't relise that it will take 7-9 hrs

Do I need to leave wifi on and my pc on while it trains? or because it's in the cloud, so it doesn't matter?

SSuperTurboHero i must be doing something wrong 😄 it's RealisticVision5.1, Dreambooth trained. ...

Furkan Gözükara SECourses•10/12/23, 11:21 PM

looks like training errors

Furkan Gözükara SECourses•10/12/23, 11:21 PM

what is your gpu 12gb?

Tt800se Hey @Dr. Furkan Gözükara , very silly question probably but I am new in all of t...

Furkan Gözükara SECourses•10/12/23, 11:22 PM

yes you need to leave browser open on kaggle

Furkan Gözükara SECourses•10/12/23, 11:22 PM

on runpod you dont have to

Jas•10/12/23, 11:40 PM

If i train a dreambooth SDXL model on a style, is it weird to use regulisation images?

Jas•10/12/23, 11:42 PM

i am making a model of images like this

Jas•10/12/23, 11:43 PM

so it would be strange to have regularisation images of portrait photos of people?

JJas If i train a dreambooth SDXL model on a style, is it weird to use regulisation i...

Furkan Gözükara SECourses•10/12/23, 11:51 PM

you dont have to use

Furkan Gözükara SECourses•10/12/23, 11:51 PM

if you dont have good ones

Furkan Gözükara SECourses•10/12/23, 11:51 PM

i would do 2 trainings and compare

FFurkan Gözükara SECourses you dont have to use

Jas•10/12/23, 11:59 PM

thanks for quick reply, from what I understand the regularisation images should be images that you would expect to see when using the class prompt?

So if i use the word "style" as a class prompt then should I have regularisation images of different styles?

Really the images will all be kind of macro photos of colourful water and such like this...

What kind of regularisation images would be best?

JJas thanks for quick reply, from what I understand the regularisation images should ...

Furkan Gözükara SECourses•10/13/23, 12:13 AM

the logic is like this

Furkan Gözükara SECourses•10/13/23, 12:13 AM

we are leveraging a class info knowladge

Furkan Gözükara SECourses•10/13/23, 12:13 AM

reg images should be as much as possibly like those class images used in the original training of model

Furkan Gözükara SECourses•10/13/23, 12:13 AM

so it depends on your class word

Furkan Gözükara SECourses•10/13/23, 12:14 AM

lets say you use style

Furkan Gözükara SECourses•10/13/23, 12:14 AM

then you need to collect images similar to original images used during initial training that contained style word

Furkan Gözükara SECourses•10/13/23, 12:14 AM

or use whatever the model generates easier

virginia•10/13/23, 12:17 AM

regularization images are reference of corrections? or what kind of function they offer?

Vvirginia regularization images are reference of corrections? or what kind of function the...

Furkan Gözükara SECourses•10/13/23, 1:08 AM

https://arxiv.org/pdf/2208.12242.pdf

Furkan Gözükara SECourses•10/13/23, 1:08 AM

Furkan Gözükara SECourses•10/13/23, 1:08 AM

read that paper

virginia•10/13/23, 1:09 AM

Hi Furkan, I have a question. for example, if I want to train panda dataset. could I use any panda image as regulation images?

Vvirginia Hi Furkan, I have a question. for example, if I want to train panda dataset. cou...

Furkan Gözükara SECourses•10/13/23, 1:20 AM

yes

virginia•10/13/23, 1:22 AM

dataset and regularization should look like the same, or for data sets have a big circle on the belly, but regularization have no circle or dataset and regularization images can be the same?

Furkan Gözükara SECourses•10/13/23, 3:06 AM

https://twitter.com/GozukaraFurkan/status/1712665890259447831

Spicy Chicken Sandwich Deluxe•10/13/23, 5:07 AM

Would you say Dall-e is much better than other models?

EElisoli Click to see attachment

Metaphysix•10/13/23, 5:54 AM

Those who have such problem running on windows do the following:

start cmd in venv\scripts
type activateactivate then
pip install clip_interrogator, gradiopip install clip_interrogator, gradio

These packages are missing from the virtual environment

Gerdes of Earth•10/13/23, 6:58 AM

Anyone have any experience merging LoRa's into models with Kohya? Having a hard time with the LoRa overriding the model too much but requiring a higher % to make sure I preserve much of the LoRa's strengths.

yes but before this new update to the patreon

Similar Threads

Similar Threads

Similar Threads