Software Engineering Courses (SECourses)•16mo ago

Hello together, have anyone trained multiple concepts in one or more dreambooth training(s)? I want

Hello together,
have anyone trained multiple concepts in one or more dreambooth training(s)?
I want to train 2 different characters with unique tokens into my model and face the issue that the one concept is always bleeding into the other.
I tried to remove the class token and made several tests with a combined training set and single trainings. Unfortunately no success. Each character has a versatile dataset of 35 high quality images.
Any idea to reduce the bleeding?

FFino Hello together, have anyone trained multiple concepts in one or more dreambooth ...

Furkan Gözükara SECourses•9/2/24, 10:49 AM

i just replied on youtube but it may not be possible since FLUX has internal text encoding. so it knows both are man. however you can try this. only ohwx and bbuk as rare token and use clip text encoder training that i added yesterday

Furkan Gözükara SECourses•9/2/24, 10:49 AM

still may bleed. fine tuning may solve though not sure

FinoOP•9/2/24, 10:53 AM

Sorry for the missing information. I talk about a SDXL Model

FFino Sorry for the missing information. I talk about a SDXL Model 😅

Furkan Gözükara SECourses•9/2/24, 11:01 AM

yes people said they do very good. are you doing lora or fine tuning?

FinoOP•9/2/24, 11:03 AM

I fine tune the model with character images.
Lora training works great with your guides.

FFino I fine tune the model with character images. Lora training works great with you...

Furkan Gözükara SECourses•9/2/24, 11:06 AM

are you using onetrainer?

FinoOP•9/2/24, 11:06 AM

I used OneTrainer (with your config) and dreambooth from kohya_ss also with your configuration

FFino I used OneTrainer (with your config) and dreambooth from kohya_ss also with your...

Furkan Gözükara SECourses•9/2/24, 11:07 AM

i see you tested on both

Furkan Gözükara SECourses•9/2/24, 11:07 AM

and still bleed?

FinoOP•9/2/24, 11:10 AM

Yes, onetrainer was much worse than kohya for some reason.
I made also dataset tagging with following schemas:
-> <unique identifier>
-> <unique identifier > <class token>
-> <unique identifier>, <describing image without character attitudes >

FinoOP•9/2/24, 11:10 AM

To test if the tagging is influencing common model tokens like men or women

FFino Yes, onetrainer was much worse than kohya for some reason. I made also dataset t...

Furkan Gözükara SECourses•9/2/24, 11:20 AM

are you using reg images?

Furkan Gözükara SECourses•9/2/24, 11:20 AM

reg makes huge impact on onetrainer

Furkan Gözükara SECourses•9/2/24, 11:20 AM

and which base model you are training sdxl base?

FinoOP•9/2/24, 11:22 AM

It’s a customized pony based model for a client.
I didn’t use reg images.

FinoOP•9/2/24, 11:22 AM

Did also a run on the base pony to verify model issues on my side.

FFino It’s a customized pony based model for a client. I didn’t use reg images.

Furkan Gözükara SECourses•9/2/24, 11:25 AM

pony

Furkan Gözükara SECourses•9/2/24, 11:25 AM

is the reason i bet

Furkan Gözükara SECourses•9/2/24, 11:25 AM

it is very different

Furkan Gözükara SECourses•9/2/24, 11:25 AM

try on real vis XL 4

FinoOP•9/2/24, 11:28 AM

O.k I didn’t think that could make such big problems.
I will make a training on sdxl base model and realvis.
Thank you for your help. It let me hope to find the problem source

Sandstormleader•9/2/24, 11:34 AM

what is the best upscaler for hires.fix?

SSandstormleader what is the best upscaler for hires.fix?

Furkan Gözükara SECourses•9/2/24, 11:41 AM

supir is best

FFurkan Gözükara SECourses supir is best

Sandstormleader•9/2/24, 11:41 AM

hmm is that an extension i have to download?

SSandstormleader hmm is that an extension i have to download?

Furkan Gözükara SECourses•9/2/24, 11:43 AM

it is different

Sandstormleader•9/2/24, 11:43 AM

Furkan Gözükara SECourses•9/2/24, 11:43 AM

https://youtu.be/OYxVEvDf284

YouTubeSECourses

Complete Guide to SUPIR Enhancing and Upscaling Images Like in Sci-...

In this video, I explain how to 1 click install and use the most advanced image upscaler / enhancer in the world that is both commercially and open source available. The upscaler that I am going to introduce you is open source #SUPIR and the model is free to use. SUPIR upscaler is many times better than both paid Topaz AI and Magnific AI and you...