Software Engineering Courses (SECourses)•3y ago

full real pictures classification dataset

Furkan Gözükara SECoursesOP•6/3/23, 8:37 PM

coming to the patreon

FFurkan Gözükara SECourses very weird

ashley•6/3/23, 8:37 PM

Yeah, I don't understand it either, because I am pretty sure you said you were using that GPU type in RunPod

FFurkan Gözükara SECourses full real pictures classification dataset

ashley•6/3/23, 8:38 PM

This is awesome, will it working better than the usual classification images?

Furkan Gözükara SECoursesOP•6/3/23, 8:38 PM

Aashley This is awesome, will it working better than the usual classification images?

Furkan Gözükara SECoursesOP•6/3/23, 8:39 PM

i am hoping so

Furkan Gözükara SECoursesOP•6/3/23, 8:39 PM

from very high resolution

Furkan Gözükara SECoursesOP•6/3/23, 8:39 PM

i will post different aspect ratios

Furkan Gözükara SECoursesOP•6/3/23, 8:39 PM

512x512 , 768x1024 , 1024x768 , 640x960, 960x640

Furkan Gözükara SECoursesOP•6/3/23, 8:39 PM

512x768, 768x512

Furkan Gözükara SECoursesOP•6/3/23, 8:39 PM

this can be used to fine tune as well

Sequa•6/3/23, 8:41 PM

saw a reddit post somewhere by someone that swears by using real images for classification images, instead of SD generated images

Sequa•6/3/23, 8:41 PM

I wonder how much of a difference it makes

Sequa•6/3/23, 8:41 PM

I think they mentioned using unsplash images as well

Sequa•6/3/23, 8:44 PM

@Dr. Furkan Gözükara there is also this https://github.com/NVlabs/ffhq-dataset

GitHub

GitHub - NVlabs/ffhq-dataset: Flickr-Faces-HQ Dataset (FFHQ)

Flickr-Faces-HQ Dataset (FFHQ). Contribute to NVlabs/ffhq-dataset development by creating an account on GitHub.

Furkan Gözükara SECoursesOP•6/3/23, 8:44 PM

i also collected from unspash

Sequa•6/3/23, 8:45 PM

would be interesting seeing the results

Kasaihyo•6/3/23, 9:03 PM

do you know how to find 2400 Photo Of classification images for women ?

Kasaihyo•6/3/23, 9:03 PM

@Lecturer - Main Admin

Kasaihyo•6/3/23, 9:03 PM

@Dr. Furkan Gözükara

Furkan Gözükara SECoursesOP•6/3/23, 9:19 PM

@dbai64 sure

Furkan Gözükara SECoursesOP•6/3/23, 9:19 PM

do git checkout commit_hash

KKasaihyo do you know how to find 2400 Photo Of classification images for women ?

Furkan Gözükara SECoursesOP•6/3/23, 9:19 PM

i dont know have that

felixsanz•6/3/23, 9:46 PM

wow, my article is already on the second gogle page for the search "how stable diffusion works"

felixsanz•6/3/23, 9:46 PM

felixsanz•6/3/23, 9:47 PM

almost on page 1

Ffelixsanz wow, my article is already on the second gogle page for the search "how stable d...

ashley•6/3/23, 9:52 PM

That's awesome, congratulations!

Furkan Gözükara SECoursesOP•6/3/23, 10:33 PM

dmn this aspect ratio took a lot of time

Ffelixsanz wow, my article is already on the second gogle page for the search "how stable d...

Furkan Gözükara SECoursesOP•6/3/23, 10:33 PM

awesome

Furkan Gözükara SECoursesOP•6/3/23, 10:33 PM

finally fixed the script

Furkan Gözükara SECoursesOP•6/3/23, 10:33 PM

aa.zip4.92MB

Furkan Gözükara SECoursesOP•6/3/23, 11:13 PM

i had been struggling to figure out like 4 hours. constantly debugging haha :d dmn

Aashley New RunPod template available for Stable Diffusion Web UI: https://runpod.io/gs...

ashley•6/4/23, 12:16 AM

Template has been updated:

Updates:

Automatic1111 Web UI version 1.3.1.
Dreambooth extension 1.0.14 (no longer dev branch).

Additions:

Kohya_ss added.
Tensorboard added.

Furkan Gözükara SECoursesOP•6/4/23, 12:43 AM

5 instance processing class images

Furkan Gözükara SECoursesOP•6/4/23, 12:57 AM

spent 2 days for this dataset :/

ashley•6/4/23, 1:00 AM

By the way, you don't need to kill the web ui process to view the logs in my new image, you can simply tail the logs at

/workspace/logs

/workspace/logs

to see what is happening. You only need to kill it and restart it to resolve issues.

Furkan Gözükara SECoursesOP•6/4/23, 1:21 AM

computer crashed :/

FFurkan Gözükara SECourses currently i am working on a new workflow. until that time i find dreambooth + re...

julius•6/4/23, 1:51 AM

Awesome, I'm really looking forward to that new workflow! Thanks for replying

Furkan Gözükara SECoursesOP•6/4/23, 2:20 AM

super slow training without xformers :/

Furkan Gözükara SECoursesOP•6/4/23, 2:20 AM

Furkan Gözükara SECoursesOP•6/4/23, 2:21 AM

i am going to sleep hopefully see you tomorrow

Micah•6/4/23, 5:25 AM

Does anyone know where to get this ui for Controlnet

Micah•6/4/23, 5:25 AM

Micah•6/4/23, 5:30 AM

Nevermind, i figured it out. needed to update CN

ashley•6/4/23, 8:39 AM

What is the best commit to use for Kohya? the master branch seems different from the videos and can't get it to work

ashley•6/4/23, 9:39 AM

I find it interesting to watch the learning rate of the Facebook D-Adaptation optimizers

m3dia_offline•6/4/23, 9:40 AM

@Dr. Furkan Gözükara I tried a dreambooth training with 2 concepts

1 of a female character and 1 of aesthetic

However, I feel my aesthetic is not trained well as it does not really generate my trained female model in that aesthetic, it is a hit and miss, I am asking here as I was watching one of SE courses videos and that part of style training was skipped.

What kind of settings are appropriate for style training?

And should I train my style and character separately and then merge the checpoints?

Furkan Gözükara SECoursesOP•6/4/23, 11:33 AM

results are amazing quality

Furkan Gözükara SECoursesOP•6/4/23, 11:33 AM

my workflow looks like worked great

Furkan Gözükara SECoursesOP•6/4/23, 11:33 AM

i will train myself now :d

Aashley I find it interesting to watch the learning rate of the Facebook D-Adaptation op...

Furkan Gözükara SECoursesOP•6/4/23, 11:35 AM

nice graph

full real pictures classification dataset

Similar Threads