Software Engineering Courses (SECourses)•15mo ago

is there a way of set the diffusion_models folder to another location? in the section "paths: settin

is there a way of set the diffusion_models folder to another location? in the section "paths: settings related to file path" there is not option to that

LLatentnaut is there a way of set the diffusion_models folder to another location? in the se...

Furkan Gözükara SECourses•9/29/24, 7:04 PM

this was asked

Furkan Gözükara SECourses•9/29/24, 7:04 PM

u-net still works too

Furkan Gözükara SECourses•9/29/24, 7:05 PM

but can you ask again here : https://discord.gg/ZAd8fnqY

happy_in_happy•9/30/24, 2:55 PM

Good afternoon! There was a question about Kohya ss. It first installs and then deletes torchvision.

happy_in_happy•9/30/24, 2:55 PM

02:42:25-947966 INFO Kohya_ss GUI version: v24.2.0

02:42:26-515966 INFO Submodule initialized and updated.
02:42:26-517966 INFO nVidia toolkit detected
02:42:29-163644 INFO Torch 2.4.1+cu124
02:42:29-182644 WARNING Torch reports GPU not available
02:42:29-185644 INFO Python version is 3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit
(AMD64)]
02:42:29-186644 INFO Verifying modules installation status from requirements_pytorch_windows.txt...
02:42:29-188644 WARNING Package wrong version: torchvision 0.20.0.dev20240929+cu124 required 0.19.1+cu124
02:42:29-189644 INFO Installing package: torchvision==0.19.1+cu124 --index-url
https://download.pytorch.org/whl/cu124
02:42:32-482224 INFO Installing package: xformers==0.0.28.post1 --index-url https://download.pytorch.org/whl/cu124
02:42:38-043874 INFO Verifying modules installation status from requirements_windows.txt...
02:42:38-046874 INFO Installing package: tensorflow>=2.16.1
02:43:08-250115 INFO Verifying modules installation status from requirements.txt...
02:43:08-252115 INFO Installing package: altair==4.2.2
02:43:10-972027 INFO Installing package: gradio==4.43.0
02:43:14-299235 INFO Installing package: lycoris_lora==2.2.0.post3
02:43:15-852875 INFO Installing package: open-clip-torch==2.20.0
02:43:19-852215 INFO Installing package: pytorch-lightning==1.9.0
02:43:23-240331 INFO Installing package: timm==0.6.12
02:43:36-565241 INFO headless: False
02:43:36-630241 INFO Using shell=True when running external commands...
Running on local URL: http://127.0.0.1:7860

happy_in_happy•9/30/24, 2:56 PM

Why is this happening? Can I fix it?

Hhappy_in_happy 02:42:25-947966 INFO Kohya_ss GUI version: v24.2.0 02:42:26-515966 INFO ...

Furkan Gözükara SECourses•9/30/24, 3:08 PM

well you have to skip verification

Furkan Gözükara SECourses•9/30/24, 3:08 PM

my installer actually does that

Furkan Gözükara SECourses•9/30/24, 3:08 PM

downlaods a start bat file

Furkan Gözükara SECourses•9/30/24, 3:08 PM

that skips validaiton

BlackStar•10/1/24, 4:46 AM

Hello mates

FFurkan Gözükara SECourses well you have to skip verification

happy_in_happy•10/1/24, 10:15 AM

Thank you!

BBlackStar Hello mates

Furkan Gözükara SECourses•10/1/24, 10:45 AM

hello

FFurkan Gözükara SECourses hello

BlackStar•10/1/24, 11:45 AM

I am here

yolkypalky•10/3/24, 3:07 PM

I finally tried fine-tuning a Flux (style) model instead of LoRA and the results are so much better. The illustration style had so many nuances, LoRA learning rates were just too fast, it kept over-cooking before learning everything I wanted it to.Thanks @Furkan Gözükara SECourses . Your config saved me so much time.

yolkypalky•10/3/24, 3:08 PM

I now want to update the SDXL version of the style model as well, did you ever make a config for finetuning SDXL? I've only been following you since Flux came out.

Yyolkypalky I now want to update the SDXL version of the style model as well, did you ever m...

Furkan Gözükara SECourses•10/3/24, 3:11 PM

yes i have SDXL

Furkan Gözükara SECourses•10/3/24, 3:11 PM

i plan to update them even better later but here

Furkan Gözükara SECourses•10/3/24, 3:12 PM

https://www.patreon.com/posts/very-best-for-of-89213064

Patreon

The Very Best Workflow For SDXL DreamBooth / Full Fine Tuning - Res...

Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon

yolkypalky•10/3/24, 3:14 PM

THANK you as always.

daroth•10/3/24, 5:13 PM

wonder if the Flux settings could work with pony training

Ddaroth wonder if the Flux settings could work with pony training

crystalwizard•10/3/24, 5:48 PM

nope. pony is its own animal

Ddaroth wonder if the Flux settings could work with pony training

Furkan Gözükara SECourses•10/3/24, 9:31 PM

yep wouldnt work

daroth•10/4/24, 5:34 AM

I'm confused, how long should a full dreambooth training last (4090)? started one with this (Quality_1_23100MB_14_12_Second_IT), and in 7 hours its only at this point:

daroth•10/4/24, 5:34 AM

that doesn't look right

Ddaroth I'm confused, how long should a full dreambooth training last (4090)? started on...

crystalwizard•10/4/24, 5:35 AM

as long as it has to how much vram do you have?

daroth•10/4/24, 5:36 AM

24gb

crystalwizard•10/4/24, 5:36 AM

and how large is your data set?

daroth•10/4/24, 5:37 AM

crystalwizard•10/4/24, 5:38 AM

okay... AMD or Intel?

daroth•10/4/24, 5:38 AM

the cpu? what does that matter?

crystalwizard•10/4/24, 5:38 AM

yes. AMD likes to try to do the GPUs job

crystalwizard•10/4/24, 5:39 AM

and you don't want any of the processing on your cpu

Ccrystalwizard and how large is your data set?

Error-404-unknown•10/4/24, 6:56 AM

Oh so a smaller dataset will reduce vram? I tried fine tuning 128 images on a 3090 last night but couldn't get it to come under 26.8Gb vram use which gave me an estimate of 184hrs

gave up a used mass compute as it was cheaper than my electricity at that time estimate. But it's good to know for the future to try again with a smaller dataset.

Deleted User•10/4/24, 7:57 AM

When I trained mine, it was 26 images and it still used just 23.1GB VRAM. It shouldn't be that slow for 30 images unless you have memory fallback enabled or you're going over your limit somehow.

EError-404-unknown Oh so a smaller dataset will reduce vram? I tried fine tuning 128 images on a 30...

nicearoni•10/4/24, 1:55 PM

smaller dataset will reduce overall time of training assuming you keep repeats\epochs the same, won't necessarily help with VRAM usage

joshtech•10/4/24, 3:29 PM

@Furkan Gözükara SECourses Can you recommend a NSFW capable Flux Checkpoint for training a Dreambooth Model? Or would you recommend to use the base version?

daroth•10/4/24, 4:14 PM

how would you handle captioning for Flux training, if the character has unusual features?

daroth•10/4/24, 4:15 PM

like her spots

Ddaroth how would you handle captioning for Flux training, if the character has unusual ...

crystalwizard•10/4/24, 4:41 PM

Something like "a woman with sots running down the sides of her face". The caption shoukd tell the AI exactly what its looking at so it understands what it is.

daroth•10/4/24, 4:44 PM

hm, wouldn't mentioning the spots make them a variable? so if I afterwards prompt for her with name, the markings wouldn't be there unless mentioned?

Ddaroth hm, wouldn't mentioning the spots make them a variable? so if I afterwards promp...

crystalwizard•10/4/24, 4:46 PM

Youre training it, not prompting it.

crystalwizard•10/4/24, 4:48 PM

You are creating an object that will update its weights for specific information when used. Put that specific information in you dataset lables

crystalwizard•10/4/24, 4:50 PM

And only use that object if you want it to know that specific information

Ddaroth I'm confused, how long should a full dreambooth training last (4090)? started on...

Furkan Gözükara SECourses•10/4/24, 7:07 PM

it is easy