Software Engineering Courses (SECourses)•16mo ago

has anyone trained a flux lora on a logo? any advice? treat it more like a character or a style? cap

has anyone trained a flux lora on a logo? any advice? treat it more like a character or a style? captions or no? (if yes, really simple captions or joy captions?)

nicearoniOP•9/14/24, 7:19 AM

dxqb•9/14/24, 7:48 AM

anyone trying to use Forge UI for flux? Swarm UI is inefficient with Loras, it apparently combines and uploads them to GPU for each generation. So I was trying forge but I am getting memory leaks I think

dxqb•9/14/24, 8:11 AM

the 8 step hyper Lora also doesn't seem to work with forge, even though other Loras do

dxqb•9/14/24, 8:11 AM

strange

Nnicearoni Click to see attachment

PixelMonkey•9/14/24, 8:19 AM

I just hade a oh F.. moment

thanks!

Luciani•9/14/24, 12:01 PM

Anyone any idea why my Forge doesn´t save images anymore on the output folder I configured? I can´t find the generated images anywhere... The console seems normal and does´t bring any errors

dxqb•9/14/24, 12:05 PM

there is a setting that enables saving of images

Luciani•9/14/24, 12:07 PM

Oh yes -- that was it -- I don´t know how this was not checked anymore! Thanks!

MMaxivi actually nevermind its stuck on intializing for 1 hour now

Furkan Gözükara SECourses•9/14/24, 12:27 PM

i am testing now

Furkan Gözükara SECourses•9/14/24, 12:27 PM

could be that image was getting updated

AAIGambino dr do you have a one click installer for dreambooth models for SDXL

Furkan Gözükara SECourses•9/14/24, 12:28 PM

which models you need?

Nnicearoni has anyone trained a flux lora on a logo? any advice? treat it more like a chara...

Furkan Gözükara SECourses•9/14/24, 12:28 PM

i would say it would be like training a person

LLuciani Oh yes -- that was it -- I don´t know how this was not checked anymore! Thanks!

Furkan Gözükara SECourses•9/14/24, 12:29 PM

lol this should be default

MMaxivi others work except yours

NicB•9/14/24, 12:29 PM

Let me know if this happens again...We were doing some infrastructure changes last night so I am curious if you made a request when we were rolling out some changes.

NNicB Let me know if this happens again...We were doing some infrastructure changes la...

Furkan Gözükara SECourses•9/14/24, 12:31 PM

i just tested

Furkan Gözükara SECourses•9/14/24, 12:31 PM

happened

Furkan Gözükara SECourses•9/14/24, 12:31 PM

Furkan Gözükara SECourses•9/14/24, 2:07 PM

kohya added layered training

Furkan Gözükara SECourses•9/14/24, 2:07 PM

but it is just too confusing so many params not much info

Furkan Gözükara SECourses•9/14/24, 2:35 PM

ok he clarified that it wasnt specific layer

FFurkan Gözükara SECourses ok he clarified that it wasnt specific layer

�

𝗡𝗶𝘁𝗶𝗻 𝗣𝗮𝗿𝗲𝘃𝗮•9/14/24, 2:44 PM

Then what is it?

�𝗡𝗶𝘁𝗶𝗻 𝗣𝗮𝗿𝗲𝘃𝗮 Then what is it?

Furkan Gözükara SECourses•9/14/24, 2:44 PM

rank for each layer

Furkan Gözükara SECourses•9/14/24, 2:44 PM

very specific case

Furkan Gözükara SECourses•9/14/24, 2:44 PM

network lora rank

FFurkan Gözükara SECourses rank for each layer

�

𝗡𝗶𝘁𝗶𝗻 𝗣𝗮𝗿𝗲𝘃𝗮•9/14/24, 2:45 PM

how can it help?

Furkan Gözükara SECourses•9/14/24, 2:46 PM

i dont know i didnt test

Furkan Gözükara SECourses•9/14/24, 2:46 PM

not planning either

FFurkan Gözükara SECourses rank for each layer

�

𝗡𝗶𝘁𝗶𝗻 𝗣𝗮𝗿𝗲𝘃𝗮•9/14/24, 2:46 PM

for example I am training a dress can I specify the rank for the dress layer and will it only learn the dress then?

�𝗡𝗶𝘁𝗶𝗻 𝗣𝗮𝗿𝗲𝘃𝗮 for example I am training a dress can I specify the rank for the dress layer and...

Furkan Gözükara SECourses•9/14/24, 2:47 PM

nope not for dress layer

Furkan Gözükara SECourses•9/14/24, 2:47 PM

he didnt add layers yet

�

𝗡𝗶𝘁𝗶𝗻 𝗣𝗮𝗿𝗲𝘃𝗮•9/14/24, 2:47 PM

Okay

El Matedor•9/14/24, 2:54 PM

Is there anyone who could recommend me guides about the efficiency and optimization considerations to be made when using ComfyUI workflows? Expecially to be used at scale.
Thanks in advance!

EEl Matedor Is there anyone who could recommend me guides about the efficiency and optimizat...

Furkan Gözükara SECourses•9/14/24, 2:55 PM

i use swarmui it handles all automatically

Neo•9/14/24, 5:08 PM

https://education.civitai.com/quickstart-guide-to-flux-1/#rapid-flux-training under 5min training?

Civitai EducationCivitai

Quickstart Guide to Flux.1 - Civitai Education

Updated! Our Quickstart Guide to Flux.1 will help you understand what the model is, what it can do, and how to start using it!

NNeo https://education.civitai.com/quickstart-guide-to-flux-1/#rapid-flux-training un...

Furkan Gözükara SECourses•9/14/24, 5:08 PM

ye low quality

Furkan Gözükara SECourses•9/14/24, 5:09 PM

even their example image very low quality

xlads•9/14/24, 5:44 PM

Hey is there a way to make the gen images 1080 pixels wide by 1920 pixels tall, with a 9:16 aspect ratio?

Xxlads Hey is there a way to make the gen images 1080 pixels wide by 1920 pixels tall, ...

nicearoniOP•9/14/24, 6:21 PM

In SwarmUI you can choose custom resolution and type whatever values you'd like...

xlads•9/14/24, 6:32 PM

whaaa yes thanks you lol

xlads•9/14/24, 6:36 PM

man ive spent hours editing multiple images into video to have them fit into the 1080 x 1920 XD

PixelMonkey•9/14/24, 7:15 PM

Hey everyone i have a few questions and will start with this one ..
I there a way to speed up the generation of flux dev on Sd forge. I feel it is very slow with my RTX 3090 4-5 min sometimes until it starts generating. With flux1-dev-bnb-nf4.safetensors and as well the normal dev version as you can see in the image

PixelMonkey•9/14/24, 7:18 PM

taining related:

I finished a training of myself with 36 images i used epoch 150 and got 10 loars. I used ohwx man and was wondeirng if i have to use ohwx everytime in the prompt to trighger the lora? It seems that it didnt train enough or smething since i start seeing only a bit of likeness starting at 90..

PixelMonkey•9/14/24, 7:18 PM

What did i miss?

PPixelMonkey Hey everyone i have a few questions and will start with this one .. I there a w...

dxqb•9/14/24, 7:33 PM

"I am getting 26 secs on 30 steps, 19 secs on 20 steps, and 14 secs on Hyper 8 steps. That's including the unfortunate Lora handling by Swarm, as I'm using Hyper as a Lora"

dxqb•9/14/24, 7:33 PM

that's on a 4070, I think a 3090 is still fast even though older

dxqb•9/14/24, 7:33 PM

so yes, there should be a way to speed it up, but my numbers are on Swarm not forge

dxqb•9/14/24, 7:34 PM

I tried forge but it did weird things with loading and unloading the model, consuming a lot of system mem (not vmem). maybe you have the same issue and your system is swapping to disk?

PixelMonkey•9/14/24, 7:38 PM

Yeah it does look like it.. Maybe Install swarm then to test. Thanks for the feedback :). Looking at the output ein it seems a issue with the memory loading and unloading. was there a installer from the Dr. or just main page

PPixelMonkey Hey everyone i have a few questions and will start with this one .. I there a w...

Furkan Gözükara SECourses•9/14/24, 10:12 PM

sounds very wrong

Furkan Gözükara SECourses•9/14/24, 10:12 PM

swarm is really fast try it

has anyone trained a flux lora on a logo? any advice? treat it more like a character or a style? cap

Similar Threads

Similar Threads

Similar Threads