Perhaps due to captioning with the dataset?
Perhaps due to captioning with the dataset?
Fsame i just use ohwx man
Ftest reg images impact too
Frunning a big test for an hard prompt
�oh ok yeah that makes sense now. I just figured probably one of your training images didn't have the glasses described in the caption, but then again the frames of the glasses aren't super bold and noticeable from a distance so that could be another reason
Ddoes anyone know if normal alternating via prompt such as [x|y] where it alternates between x and y each step is still usable in a natural language style model like flux?
Smaybe i will wait more, so the process can make it with a gpu 16 vram, maybe onetrainer make support for flux
Fall images has eyeglasses and only ohwx man :d
Fflux has wildcards
Fif kohya fixes also
Feular normal vs karras
Ftesting all combinations
Fkarras make huge diff and make it more real
Fwill post entire grid on patreon
Scant wait the config files and tutorials. 
Fadding configs already :d
Snow just seen, from all the png kpeg attachments that there is one zip file, .. I downloaded.
I have the latest branch kohya.
now i will start the training.
I have the latest branch kohya.
now i will start the training.
Sany recommendation for 4090 and beginner?
S1e_04_bf16_128_rank_full_bf16 ? maybe?
FLook very top written best ones
HWhat guidance you using
HNo more Motorcycle handles on the T rex back now lol
F3.5 still gonna test it too 
Fye
DIn my first tests I'm still having problems whith concept bleeding
Dworks fine if you train just one person
Fcurrently impact of lora so powrful
Fit mixed me even with woman prompt
Dif you train many people mix them all, is not paying attention to the captions
Dmix person 1 with person two the problem is worse if they are from the same class
Dmust to be the text ecoder
Dthe same issue I hade with sd3
Dyes is only training the unet the text captions are ignored
R@Dr. Furkan Gözükara
Are you doing FLUX Lora training or dream booth training?
And do you need Adetailer for generating images? Is this inside the SwarmUI?
Are you doing FLUX Lora training or dream booth training?
And do you need Adetailer for generating images? Is this inside the SwarmUI?
Fflux lora
Dor looks like the text captions are ignored
Fi use segment feature of swarmui
FYouTubeSECourses
Do not skip any part of this tutorial to master how to use Stable Diffusion 3 (SD3) with the most advanced generative AI open source APP SwarmUI. Automatic1111 SD Web UI or Fooocus are not supporting the #SD3 yet. Therefore, I am starting to make tutorials for SwarmUI as well. #StableSwarmUI is officially developed by the StabilityAI and your m...
Fhuge speed up for rtx 4xxx series for flux
Fhttps://www.reddit.com/r/StableDiffusion/comments/1ex3ol5/comfyui_experimental_rtx_40_series_update/
Reddit
Explore this post and more from the StableDiffusion community
Done test that you can do is training without the class just "ohwx" leve the token man untouched, then trigger ohwx and see if is doing something or is ignored completely, if that its the case thats the reason why concepts bleed and is not learning anything from the captions
HIsnt it late at home? haha
Hwonder if this is true on runpod 4090
V@Dr. Furkan Gözükara possible to show me the json for sdxl lora settings
CI trained 2 person lora (nmwx for myself and ohwx for my wife), all images of me tagged with nmwx, despite prompting simply "nmwx" it returned images of my wife. BUT its very consistent that all "man" return me and all "woman" prompts returns my wife. So it looks like the model just overfitted for all men to look like me and all women look like my wife.
Also "nmwx man, HUGO BOSS advertisement" and "man, HUGO BOSS advertisement" return 99.5% the same image and likeness, just minor details changed.
BTW the 2 character lora works amazing. no bleeding between the two characters for me and can generate both within one image. Probably because its two genders, i'm doubting it would work if its two men though without training the encoder
Also "nmwx man, HUGO BOSS advertisement" and "man, HUGO BOSS advertisement" return 99.5% the same image and likeness, just minor details changed.
BTW the 2 character lora works amazing. no bleeding between the two characters for me and can generate both within one image. Probably because its two genders, i'm doubting it would work if its two men though without training the encoder
U
Ddefine enhance
PDoes kohya full finetunig working or still not? for Flux ?
Freally late :d
