Software Engineering Courses (SECourses)•17mo ago

I am unsure how epoch works. I see your settings always has 200, why change it? Is 1000 images too

I am unsure how epoch works. I see your settings always has 200, why change it? Is 1000 images too much to train a style?

Why touch or change epochs? Thats how steps are determined?

GomezBroOP•9/1/24, 2:42 AM

Can I get that coupon and website please? @Dr. Furkan Gözükara

GGomezBro I am unsure how epoch works. I see your settings always has 200, why change it?...

druhl•9/1/24, 5:21 AM

From what I understand (correct me, if I am wrong, Doc), an epoch equals how many times an image in the input dataset is trained.
So suppose you have an input dataset of 15 images, and you select an epoch of 200, each image will be trained 200 times, so the No. of steps = 15x200 = 3000 (3000 steps is a trade-off figure b/w time and quality that the Doc likes to use in his videos).

Now, the time taken for each step will depend upon your GPU. So, say a 12GB GPU trains with Flux at 12 second/ iteration. Then, the total time =
3000 x 12 = 36,000 secs (approx. 10hrs.)

druhl•9/1/24, 5:23 AM

The above is when you use repeat = 1,
So no. of steps = epoch x repeats x no. of images in input dataset

druhl•9/1/24, 5:25 AM

And time taken (in hrs.) = (no. of steps x time taken for each iteration by the GPU)/3600

druhl•9/1/24, 5:28 AM

What I do not understand (yet!), however, is:
1) what difference it makes if you use 3000 steps versus, say 7000 steps (increase the epoch for the same no. of input images. My assumption is it takes more time, but the quality will be better to a certain extent, before it gives diminishing returns and then peaks).
2) Use less images & more epoch, versus more images and less epoch (However, the no. of steps is maintained the same. Eg. 15x200=3000, and 50x60=also3000. Here, I make no assumption and would like to hear from Doc).

dxqb•9/1/24, 6:45 AM

1) largely correct, but there is no peak, there is a cliff. you can overfit diffusion models quite easily

dxqb•9/1/24, 6:50 AM

2) there is no clear answer, it's an art. Less images tend to give you an inflexible result, more images tend to be harder to train until the model picks up the details. You often need more steps in total for many images than for a few, meaning not reduce the epochs as much as you have increased the number of images. Larger batch sizes can help with (many) more images. Adjustments in LR might be necessary.

dxqb•9/1/24, 6:52 AM

and to make it even less predictable, it depends on the dataset, especially how similar the images are

druhl•9/1/24, 6:54 AM

Yes, think I understand.
Good to see I got the basics right.

druhl•9/1/24, 6:54 AM

Thanks for explaining further.

Neuromasters•9/1/24, 7:54 AM

it writes errors, do I have to put in an experimental file?

Thereiver•9/1/24, 8:51 AM

Hey, does anyone have a prodigy config for flux? Back in the day we’d set the warmup steps to 100% and have loras ready in 30-45 mins or so depending on the amount of images in the dataset.

I like the results that your rank 3/4 config provides, it’s really high quality. But I’d need something a bit faster than 35 hrs for a 300 image dataset - any tips would be really appreciated

MMaxivi i think for linux xformers works fine because of the torch version

Furkan Gözükara SECourses•9/1/24, 10:14 AM

you were right i see improvment in the training

Furkan Gözükara SECourses•9/1/24, 10:14 AM

looks like bmaltais fixed something for sure

Furkan Gözükara SECourses•9/1/24, 10:14 AM

prev vs new

Furkan Gözükara SECourses•9/1/24, 10:14 AM

clothing much better definitely same settings

Thereiver•9/1/24, 10:22 AM

What do we need to do to update? Just run the step 2 bat?

Furkan Gözükara SECourses•9/1/24, 10:27 AM

posted after bug fix training results : https://www.reddit.com/r/StableDiffusion/comments/1f6b948/kohya_gui_fixed_a_bug_yesterday_and_after/

From the StableDiffusion community on Reddit: Kohya GUI fixed a bug...

Explore this post and more from the StableDiffusion community

TThereiver What do we need to do to update? Just run the step 2 bat?

Furkan Gözükara SECourses•9/1/24, 10:27 AM

yes

Furkan Gözükara SECourses•9/1/24, 10:28 AM

and retrain of course

Furkan Gözükara SECourses•9/1/24, 10:28 AM

but you can wait me to update configs as well

Furkan Gözükara SECourses•9/1/24, 10:28 AM

because i am gonna add text encoder training

Furkan Gözükara SECourses•9/1/24, 10:28 AM

which certainly improves quality

GGomezBro Can I get that coupon and website please? @Dr. Furkan Gözükara

Furkan Gözükara SECourses•9/1/24, 10:28 AM

coupon is SECourses

Furkan Gözükara SECourses•9/1/24, 10:28 AM

select our image

FFurkan Gözükara SECourses select our image

GomezBroOP•9/1/24, 10:29 AM

Whats the website?

GomezBroOP•9/1/24, 10:29 AM

And since you are here. I want to train 1000 NFTs to grab the art style that I made. Is that too many? I see on my a6000 it would take like 450 hrs!

GGomezBro Whats the website?

Furkan Gözükara SECourses•9/1/24, 10:29 AM

let me show you

GomezBroOP•9/1/24, 10:30 AM

its 1000 too many to train a style?

Furkan Gözükara SECourses•9/1/24, 10:30 AM

https://youtu.be/XFUZof6Skkw

YouTubeSECourses

How to Use SwarmUI & Stable Diffusion 3 on Cloud Services Kaggle (f...

In this video, I demonstrate how to install and use #SwarmUI on cloud services. If you lack a powerful GPU or wish to harness more GPU power, this video is essential. You'll learn how to install and utilize SwarmUI, one of the most powerful Generative AI interfaces, on Massed Compute, RunPod, and Kaggle (which offers free dual T4 GPU access for ...