Software Engineering Courses (SECourses)•3y ago

has anyone tried training multiple concept? I noticed that it basically doubled the steps, so it see

has anyone tried training multiple concept? I noticed that it basically doubled the steps, so it seems like it's simply just training one concept and than continuing to the second concept. Is it the same as just doing two training?

Zono50•6/5/23, 3:19 PM

where does dreambooth generate the sample images?

ZZono50 where does dreambooth generate the sample images?

felixsanz•6/5/23, 3:32 PM

you specify everything in the config

felixsanz•6/5/23, 3:32 PM

path, number of samples, etc

Zono50•6/5/23, 3:32 PM

oh yeah i totally forgot about that, thanks for reminding me

felixsanz•6/5/23, 3:35 PM

@Dr. Furkan Gözükara did you find if those high quality reg images improve the training results? or not yet

Zono50•6/5/23, 3:42 PM

Hey @Dr. Furkan Gözükara any updates onn how to use the new optimizers in dreambooth?

felixsanz•6/5/23, 3:56 PM

@Dr. Furkan Gözükara roop on a qemu VM works wonderful, because you can assign half of your cpus and use script with max-cores, so your system doesn't collapse. everything's fast and smooth

Zono50•6/5/23, 5:05 PM

I have a question that's been bugging me for a while. How do we know what to set our lora strengths to?

Hazmuut•6/5/23, 5:20 PM

Hi everyone

Hazmuut•6/5/23, 5:20 PM

Just joined the server

Zono50•6/5/23, 5:32 PM

welcome

ZZono50 I have a question that's been bugging me for a while. How do we know what to set...

felixsanz•6/5/23, 5:33 PM

trying

playereckksdee•6/5/23, 5:55 PM

@Dr. Furkan Gözükara How can one have both kohya and automatic1111 without them interfering each other?

Pplayereckksdee @Dr. Furkan Gözükara How can one have both kohya and automatic1111 without them ...

ashley•6/5/23, 6:07 PM

1) Create e different venv for each
2) Run them on different ports

Aashley 1) Create e different venv for each 2) Run them on different ports

playereckksdee•6/5/23, 6:08 PM

Does kohya not automatically make a venv folder for windows installation?

Pplayereckksdee Does kohya not automatically make a venv folder for windows installation?

ashley•6/5/23, 6:09 PM

I think so, what is your issue you're having with them interfering with each other?

playereckksdee•6/5/23, 6:10 PM

python version installed on my machine is 3.10.6 and kohya is requiring 3.10.9

Pplayereckksdee python version installed on my machine is 3.10.6 and kohya is requiring 3.10.9

ashley•6/5/23, 6:11 PM

Thats strange Kohya works on Linux with 3.10.6

playereckksdee•6/5/23, 6:11 PM

windows !

playereckksdee•6/5/23, 6:11 PM

ashley•6/5/23, 6:11 PM

Yes I know I was just saying

playereckksdee•6/5/23, 6:11 PM

ashley•6/5/23, 6:12 PM

Can't you just use 3.10.9 for both?

playereckksdee•6/5/23, 6:12 PM

I can try but this should not be needed because of venv folder

Pplayereckksdee python version installed on my machine is 3.10.6 and kohya is requiring 3.10.9

rafraf•6/5/23, 6:51 PM

There's a workaround m. A weird one Sr Furkan helped me with. 1) setup 3.10.9, set env path. Install Kohya. 2) make copy of 3.10.9, uninstall original one. The problem is when you install Python system wide it basically wipes parts of the precious phthon install. Has worked for me. Have auto1111 running with 3.10.6 set as system wide Python. Then give 3.10.9 in env path for Kohya startups

rafraf•6/5/23, 6:51 PM

I recommend you become familiar with environment paths. The doctor discusses it on one of his videos about python

Aashley Can't you just use 3.10.9 for both?

rafraf•6/5/23, 6:52 PM

Auto1111 is pretty strict about 3.10.6

Rrafraf Auto1111 is pretty strict about 3.10.6

ashley•6/5/23, 7:04 PM

Ah, wonder why A1111 and Kohya are like that on Windows but not Linux, strange

Aashley Ah, wonder why A1111 and Kohya are like that on Windows but not Linux, strange

rafraf•6/5/23, 7:05 PM

I think I saw 3.20.6 as requirement somewhere

rafraf•6/5/23, 7:05 PM

Yea strange

Rrafraf I think I saw 3.20.6 as requirement somewhere

ashley•6/5/23, 7:07 PM

Think you mean 3.10.6. I was able to run A1111 on 3.10.10 on RunPod, haven't tried Kohya... but clearly different requirements for Linux and Windows

rafraf•6/5/23, 7:08 PM

Yea sorry, 3.10.6

playereckksdee•6/5/23, 7:21 PM

3.10.6 works for both a1111 and kohya

playereckksdee•6/5/23, 7:21 PM

but for some reason all training is now worse quality compared to a couple of days ago

playereckksdee•6/5/23, 7:22 PM

both runpod and local machine gave same results

Zono50•6/5/23, 7:53 PM

Does lion not work with lora in dreambooth? cause the results I got look nothing like my model

ZZono50 Does lion not work with lora in dreambooth? cause the results I got look nothing...

ashley•6/5/23, 8:00 PM

What did you set the learning rate to?

Zono50•6/5/23, 8:02 PM

5e-7

Zono50•6/5/23, 8:02 PM

and 3e-7 for tenc

Zono50•6/5/23, 8:02 PM

i'm doing lora extended

ZZono50 i'm doing lora extended

ashley•6/5/23, 8:05 PM

Was it overtrained? Usually the LR for Lion is 1e-7

Zono50•6/5/23, 8:08 PM

i don't think so, typically overtrained it looks burnt I thought.

Zono50•6/5/23, 8:11 PM

i've been trying to train a model of my wife for ages now, and I get close, but never really there. I've followed every video I've come across, but still can't get it perfect

Yume_Origins•6/5/23, 8:53 PM

help !! this error on dreambooth 1111 i'ts new (again ) PermissionError: [WinError 5] Accès refusé: 'C:\Users\UTILIS~1\AppData\Local\Temp\tmp8_yihns4\text_encoder\model.safetensors'

playereckksdee•6/5/23, 10:01 PM

I'm unable to have both python 3.10.6 and 3.10.9, when I try to install 3.10.9 it is asking to upgrade from 3.10.6

Pplayereckksdee I'm unable to have both python 3.10.6 and 3.10.9, when I try to install 3.10.9 i...

ashley•6/5/23, 10:26 PM

Can't you use something like this?

https://github.com/pyenv-win/pyenv-win

GitHub

GitHub - pyenv-win/pyenv-win: pyenv for Windows. pyenv is a simple ...

pyenv for Windows. pyenv is a simple python version management tool. It lets you easily switch between multiple versions of Python. It's simple, unobtrusive, and follows the UNIX tradition ...

Pplayereckksdee @Dr. Furkan Gözükara How can one have both kohya and automatic1111 without them ...

Furkan Gözükara SECourses•6/5/23, 11:01 PM

very easy. each one have their own venv

Ffelixsanz @Dr. Furkan Gözükara did you find if those high quality reg images improve the t...

Furkan Gözükara SECourses•6/5/23, 11:01 PM

i think they improve significantly because i train on 768x1024

Furkan Gözükara SECourses•6/5/23, 11:01 PM

i did a training for @eggbom

Furkan Gözükara SECourses•6/5/23, 11:01 PM

and you cant believe what raw quality we get

has anyone tried training multiple concept? I noticed that it basically doubled the steps, so it see

Similar Threads

Similar Threads

Similar Threads