Software Engineering Courses (SECourses)•15mo ago

@Dr. Furkan Gözükara Ok here's my theory, I train a Dreambooth with a character I call X for example

@Dr. Furkan Gözükara Ok here's my theory, I train a Dreambooth with a character I call X for example, then I train this model with another character I call Y, always in the same style. With the captions well detailed, the goal is to have a final model that would allow me to make X and or Y appear via the prompt, is this consistent? or is it better to go through lora? btw I can have 256 images per character, if I win against my laziness lol

Ffadounav @Dr. Furkan Gözükara Ok here's my theory, I train a Dreambooth with a character ...

Furkan Gözükara SECourses•10/15/24, 9:16 PM

for SDXL it can work

Furkan Gözükara SECourses•10/15/24, 9:16 PM

for FLUX they bleed

Furkan Gözükara SECourses•10/15/24, 9:16 PM

also you may train all at once

Furkan Gözükara SECourses•10/15/24, 9:16 PM

assign differnt unique toknes

Furkan Gözükara SECourses•10/15/24, 9:16 PM

ohwx man

Furkan Gözükara SECourses•10/15/24, 9:16 PM

bbuk man

Dazzastrous [4090]•10/15/24, 9:17 PM

just merge sdxl and flux together lol Fluxl

DDazzastrous [4090]just merge sdxl and flux together lol Fluxl

fadounavOP•10/15/24, 9:18 PM

Haha xcellent

FFurkan Gözükara SECourses assign differnt unique toknes

fadounavOP•10/15/24, 9:21 PM

interesting, I could note the token for each character in the captions, but then I'd also have to note it separately in the training parameters? Is there a risk of it bleeding?

Ffadounav interesting, I could note the token for each character in the captions, but then...

Furkan Gözükara SECourses•10/15/24, 9:27 PM

risk of bleeding exists

Furkan Gözükara SECourses•10/15/24, 9:27 PM

perhaps

Furkan Gözükara SECourses•10/15/24, 9:27 PM

ohwx man

Furkan Gözükara SECourses•10/15/24, 9:27 PM

bbuk woman

Furkan Gözükara SECourses•10/15/24, 9:27 PM

bbuk person

Furkan Gözükara SECourses•10/15/24, 9:27 PM