Software Engineering Courses (SECourses)•17mo ago

I have a 3090

FFurkan Gözükara SECourses it still generates my image :d

J2_Nedo•8/27/24, 1:36 PM

try this lora now with a different prompt like photo of a ohwx woman, it should work, because “man” was not associated with your training photos. I am currently testing this stuff and it seems to be working.

SSharinel For someone with a 4090, what is the best model to use in a speed/quality tradeo...

Furkan Gözükara SECourses•8/27/24, 1:49 PM

use FP8 directly

HHorrid I have a 3090

Furkan Gözükara SECourses•8/27/24, 1:49 PM

you should get like 5.5 second it with slow

Furkan Gözükara SECourses•8/27/24, 1:49 PM

on windows

Furkan Gözükara SECourses•8/27/24, 1:49 PM

but i get like 7.7 atm

JJ2_Nedo try this lora now with a different prompt like photo of a ohwx woman, it should ...

Furkan Gözükara SECourses•8/27/24, 1:49 PM

yep it is about flux internal text encoding

Furkan Gözükara SECourses•8/27/24, 1:49 PM

i am preparing video tutorial patreon post at the moment

Furkan Gözükara SECourses•8/27/24, 1:49 PM

so starting tutorial recording hopefully today

FFurkan Gözükara SECourses you should get like 5.5 second it with slow

HorridOP•8/27/24, 1:50 PM

why you getting 7.7 it if its suposed to be 5.5

HorridOP•8/27/24, 1:50 PM

juat wondering

HHorrid why you getting 7.7 it if its suposed to be 5.5

Furkan Gözükara SECourses•8/27/24, 1:50 PM

windows vs linux

Furkan Gözükara SECourses•8/27/24, 1:51 PM

linux is 5.5 windows is like 8 second it

HorridOP•8/27/24, 1:52 PM

do you know whats causing windows to be slower

HorridOP•8/27/24, 1:52 PM

for 3090

HHorrid do you know whats causing windows to be slower

Furkan Gözükara SECourses•8/27/24, 2:00 PM

i wish i knew

Furkan Gözükara SECourses•8/27/24, 2:00 PM

one of the libraries for sure

Furkan Gözükara SECourses•8/27/24, 2:00 PM

i compared versions they are same with linux

Furkan Gözükara SECourses•8/27/24, 2:00 PM

simplified flux lora training post published for video tutorial : https://www.patreon.com/posts/110879657

Patreon

Kohya FLUX LoRA Training Full Tutorial For Local Windows and Cloud ...

Get more from SECourses: Tutorials, Guides, Resources, Training, FLUX, MidJourney, Voice Clone, TTS, ChatGPT, GPT, LLM, Scripts by Furkan Gözü on Patreon

HorridOP•8/27/24, 2:01 PM

i have another question

HorridOP•8/27/24, 2:01 PM

did u figure out how to make a 1 image lora?

HorridOP•8/27/24, 2:01 PM

i saw your comment on a reddit post of someone who did it with flux u said youd make a tutorial just wondering where that stands

FFurkan Gözükara SECourses i am waiting bmaltais

dsienra•8/27/24, 2:02 PM

where is published? can be testes adding a custom argument in here?

SSharinel For someone with a 4090, what is the best model to use in a speed/quality tradeo...

TOPUI•8/27/24, 2:02 PM

Do you think a higher batch size has a big impact on quality? But a higher batch size does increase training speed. @Dr. Furkan Gözükara

HHorrid did u figure out how to make a 1 image lora?

Furkan Gözükara SECourses•8/27/24, 2:07 PM

havent tested yet

TTOPUI Do you think a higher batch size has a big impact on quality? But a higher batch...

Furkan Gözükara SECourses•8/27/24, 2:07 PM

higher batch size = lower quality

Furkan Gözükara SECourses•8/27/24, 2:07 PM

that is certain

Furkan Gözükara SECourses•8/27/24, 2:07 PM

it prevents overfitting if your LR is high but our LR already optimized

Ddsienra where is published? can be testes adding a custom argument in here?

Furkan Gözükara SECourses•8/27/24, 2:08 PM

you cant use yet with gui

FFurkan Gözükara SECourses you cant use yet with gui

dsienra•8/27/24, 2:11 PM

ok thanks, but was added to the script?

Ddsienra ok thanks, but was added to the script?

Furkan Gözükara SECourses•8/27/24, 2:13 PM

yes

Furkan Gözükara SECourses•8/27/24, 2:13 PM

kohya ss main repo

dsienra•8/27/24, 2:18 PM

cool, who made the announcement?

Ddsienra cool, who made the announcement?

Furkan Gözükara SECourses•8/27/24, 2:18 PM

kohya

dsienra•8/27/24, 2:20 PM

where? sorry for my ignorance but where the last news are published?

dsienra•8/27/24, 2:20 PM

on github I suppose but were

Ddsienra where? sorry for my ignorance but where the last news are published?

Furkan Gözükara SECourses•8/27/24, 2:20 PM

i follow github discussions and his tiwtter

dsienra•8/27/24, 2:20 PM

Furkan Gözükara SECourses•8/27/24, 2:20 PM

yes

dsienra•8/27/24, 2:21 PM

thanks

dsienra•8/27/24, 2:44 PM

we have to wait then... I'll be waiting your first CLIP_L tests

pat._.fenis•8/27/24, 3:56 PM

Is anyone aware of a quality video upscaler?

Ppat._.fenis Is anyone aware of a quality video upscaler?

Furkan Gözükara SECourses•8/27/24, 4:09 PM

covers most of the ai models

Furkan Gözükara SECourses•8/27/24, 4:09 PM

https://github.com/AaronFeng753/Waifu2x-Extension-GUI

GitHub

GitHub - AaronFeng753/Waifu2x-Extension-GUI: Video, Image and GIF u...

Video, Image and GIF upscale/enlarge(Super-Resolution) and Video frame interpolation. Achieved with Waifu2x, Real-ESRGAN, Real-CUGAN, RTX Video Super Resolution VSR, SRMD, RealSR, Anime4K, RIFE, I...

Furkan Gözükara SECourses•8/27/24, 4:09 PM

but nothing good as supir

Websteria•8/27/24, 4:27 PM

Using the flux swarmui provided, normal flux generates an image using dev-fp8 in 20 seconds. If I add a lora, trained on flux.dev it takes several minutes per 5 steps. Any idea what causes this or if a full dev Lora can be quantized to work faster with the fp8 model?

Thanks!

WWebsteria Using the flux swarmui provided, normal flux generates an image using dev-fp8 in...

nicearoni•8/27/24, 4:44 PM

what GPU do you have? maybe adding lora causes you to have to use shared vram?

WWebsteria Using the flux swarmui provided, normal flux generates an image using dev-fp8 in...

Furkan Gözükara SECourses•8/27/24, 5:04 PM

not it doesnt have to be

Furkan Gözükara SECourses•8/27/24, 5:04 PM

what is your gpu

Websteria•8/27/24, 5:04 PM

4090

Furkan Gözükara SECourses•8/27/24, 5:04 PM

lora may be overloading your vram

I have a 3090

Similar Threads