swarmui notebook updated : https://www.patreon.com/posts/106650931
Patreon
Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon
YI am trying now many tests now to find best possible way for generation
Y@Dr. Furkan Gözükara Saw sdxl lastest release
Y?
Y3 models
Yblur canny i guess and another one
Y@Sebastianblood
Ylooks good ?
VI guess I cannot finetune flux there? 
YThis is with artifacts :
YThis is optimized see jewelery :
VWas not able to watch all parts of the video for fine tuning flux yet. only massed compute and some parts of other previous tutorial video.
I didnt see yet the part where is discussed quality difference between 15 images and 250 images for training. I am curious how big or in what range is the difference in quality?
I didnt see yet the part where is discussed quality difference between 15 images and 250 images for training. I am curious how big or in what range is the difference in quality?
A@Dr. Furkan Gözükara where can I find what every new flux model does ?
T
YThey say it has been leaked
YOpenAI sora
Ycheck code source turned out they use api but maybe didnt dig enough
Y
YThey gave artists access to check it out
Ysomeone took the api access and put it to the public
Ylol
TxD
V@Dr. Furkan Gözükara ah sorry if previous messages sounded rude or sarcastic. I realized they might appear in such a way. But rest assured, if I ask such questions, its because I never used swarmui or kaggle before. I would say first one is a UI and the second term I dont know yeah I guess I need to watch the videos 
yeah I guess I need to watch the videos Va different question: I heard in the news there are problems with scaling or with training new models. Does somebody know more about this?
Fyou cant find you have to search internet follow repos
Fread patreon post there are more grids
Fopen ai scrwd them
Fno only api
Sface skin is not as vivid as the other pictures you shared earlier
GAre you saying they only had the API? So no source code was leaked?
Fyep
Fflux redux test
YFinal version for my workflow !
YWhat do you guys think of output please ?
Treally good
Ftesting flux redux on my fine tuned model
CVery impressive firearms notoriously dont come out right and that looks spot on to my eye
Fredux restyle image
Fthis is from grid
Fstill testing
F
VAs requested just for fun an outPOST Analysis...
Text was unreadable and too long when pasted here. So I had to put it on Notion page:
https://candy-tarsier-5d6.notion.site/EU-Female-Personnel-in-Villa-in-Tropical-Setting-Lora-Flux-OutPOST-Analysis-14a49fe68df28004a100c5de95bccba8
Text was unreadable and too long when pasted here. So I had to put it on Notion page:
https://candy-tarsier-5d6.notion.site/EU-Female-Personnel-in-Villa-in-Tropical-Setting-Lora-Flux-OutPOST-Analysis-14a49fe68df28004a100c5de95bccba8
Notion von David W on Notion
LoRA
A@Dr. Furkan Gözükara why does my onetrainer save my .safetensors in two different spots? It's the last safetensor made that is the name of the keyword, the rest are named something else for some reason
ZCan anyone tell me how to use Clip-L and t5xxl windows separately for prompting in SwarmUI? Is it possible? I couldn't find such an option... The thing is, it gives more accurate results when I worked in ComfyUI.
Fit is a known issue
Fdeveloper didnt fix
For rather he doesnt see it as a problem
Fwhat you mean by that?
Fyou can pick them from advanced prompting
F
Fmerging 3 images