Software Engineering Courses (SECourses)•3y ago

here are the results from my subject LoRA. These are ai

00511-number-seed-__photo_of_beautiful_lna_woman_in_the_style_of_hugo_pieter_1.2_wasteland_salt_lake_1.1_red_water_1.1_red_soil_ariz.png

00519-number-seed-__photo_of_beautiful_lna_woman_in_the_style_of_hugo_pieter_1.2_wasteland_salt_lake_1.1_red_water_1.1_red_soil_ariz.png

Mmayen so i dnt even need to worry abt 4 gb files per model

Furkan Gözükara SECourses•5/1/23, 11:19 PM

you need extra extension

rafrafOP•5/1/23, 11:19 PM

Using some images from a studio shoot idid

Furkan Gözükara SECourses•5/1/23, 11:19 PM

to generate lower size files

Furkan Gözükara SECourses•5/1/23, 11:19 PM

with dreambooth extension

Furkan Gözükara SECourses•5/1/23, 11:19 PM

i plan to make a video but havent done yet

Furkan Gözükara SECourses•5/1/23, 11:19 PM

i had been waiting them to update and i think they did

Furkan Gözükara SECourses•5/1/23, 11:19 PM

so i can make a new video

Furkan Gözükara SECourses•5/1/23, 11:19 PM

maybe tomorrow

rafrafOP•5/1/23, 11:20 PM

but I havent managed to get good body outfit mix yet

Rrafraf but I havent managed to get good body outfit mix yet

Furkan Gözükara SECourses•5/1/23, 11:20 PM

i think easier way would be

Furkan Gözükara SECourses•5/1/23, 11:20 PM

generate person with outfit

Furkan Gözükara SECourses•5/1/23, 11:20 PM

and inpaint face

FFurkan Gözükara SECourses i think easier way would be

rafrafOP•5/1/23, 11:21 PM

the inpaint would be using the subject LoRA only. Right. Ok

Rrafraf the inpaint would be using the subject LoRA only. Right. Ok

Furkan Gözükara SECourses•5/1/23, 11:21 PM

yep

Furkan Gözükara SECourses•5/1/23, 11:21 PM

i got to sleep see you tomorrow hopefully

Furkan Gözükara SECourses•5/1/23, 11:21 PM

so many videos i need to make so little time :d

FFurkan Gözükara SECourses i got to sleep see you tomorrow hopefully

rafrafOP•5/1/23, 11:22 PM

thank you so much! Get your well deserved sleep Dr

ZZono50 Click to see attachment

rafrafOP•5/1/23, 11:45 PM

So are you simply using your Face LoRA InPainted?

P1pSl1p•5/2/23, 12:40 AM

still a lot of testing to do. at least I didn't turn into a robot or something halfway through the video. style and structure remained consistent throughout. so it's progress.

Rrafraf here are the results from my subject LoRA. These are ai

P1pSl1p•5/2/23, 12:40 AM

wicked awesome results

PP1pSl1p still a lot of testing to do. at least I didn't turn into a robot or something h...

Khang•5/2/23, 2:08 AM

Look awesome

Chase•5/2/23, 2:42 AM

If all my Dreambooth checkpoint outputs are grainy, what does this indicate?

Chase•5/2/23, 2:42 AM

All outputs on txt2img with the trained model are grainy as well. Running 23GB Vram.

Chase•5/2/23, 2:44 AM

Screen_Shot_2023-05-01_at_10.44.02_PM.png

Chase•5/2/23, 2:57 AM

Does this have to do with the VAE?

Screen_Shot_2023-05-01_at_10.56.51_PM.png

Screen_Shot_2023-05-01_at_10.56.43_PM.png

P1pSl1p•5/2/23, 3:09 AM

could be your model is overbaked

P1pSl1p•5/2/23, 3:10 AM

too much training

CChase Click to see attachment

P1pSl1p•5/2/23, 3:11 AM

this image def looks overbaked

FFurkan Gözükara SECourses i have a very interesting idea working on that right now. using AI to sort most ...

Khang•5/2/23, 3:11 AM

That's a super cool idea, Dr. Furkan! Can we also use it for checking styles and objects? We're kinda stuck because SD can churn out way more images than we can check manually. It doesn't have to be perfect, but if it could rank each image from 1 to 10, it would save us a ton of time and effort

PP1pSl1p this image def looks overbaked

Chase•5/2/23, 3:14 AM

At all sample steps, training images are blurry:

Chase•5/2/23, 3:15 AM

500, 600, 700