yep overfit
yep overfit
WActually I send an image here for reference but somehow discord remove it. It's the same person about same age but different make up, hair, outfit, the pose also different, some images will be lying pose 768x512, some images are portrait 1024x1024, some images are standing pose 512x768. Do you think it is insconsistent data
WMaybe I will run mass compute by myself to test with caption and give you 10% fee to make it clearer right 
?
?
Fyep good idea
XI’m going to take a 1-minute clip from a movie and use Joy Caption to generate captions for each frame. Then, I’ll edit those captions into prompts that fit my AI model to hopefully optimize my workflow.
Fsounds like a plan
Dyes but maybe the key is the class on this case, it learned and associated the token to the background and not me, may be training T5 I must specify the class man, so it learns me rather than the background, is like it needs guidance to what to learn, just a theory.. hehe
Sin the <segment:face,0.7> part what happens if I don't reduce it to 0.7? Won't it bring better result?
Xso my caption joy just complete batch processing progress and says that the output folder is same as the input folder but when i check i can see anything
Xstatred processin progress again this time with oput folder pasted in ui
VAnyone has any idea why this workflow is not working? It gives me an error, but it was fully working with the original Flux-Dev(8fp)
V
Ntheres 17 gpu aliviable
M@Dr. Furkan Gözükara is the lasest kohya broken for you?
LI have exactly the same issue with vertical lines appearing on the results. One point is to restrict generating to 1024 and do not go above. However, often the lines are still visible on that smaller images. Did you find a solution yet? You may also send me a direct message!
3Is there a way to use 2 loras of 2 different people together without them blending?
Fi,t is 0.6 by default
Fas far as i know
Fhe fixed it
Myep i saw he replied to me
Mseems you can only fix the bleeding with lokr
Mbut the face quality is not good
Fi almost never get them with 1024
Fbut when i go 1280 1536 very noticeable
DI'm beginning to suspect that overtraining might be the cause, but I'm not entirely sure anymore. Flux training can be quite unpredictable, especially in the early stages, and the guides often give conflicting advice. In my personal photo set, vertical lines keep showing up, even after three attempts. I’m starting to think there may be an issue with my training image sources as well. Whether I train for 1,000 steps or 3,500, the lines persist. Oddly enough, I trained two other Loras using different subjects, and those turned out perfectly fine, which makes the situation even more confusing.
Flooks like hugging face having huge internal issues
Fi like your new photo :d
Phaha
Pis was an sd 1.5 photo realistic trainning
Fnice
Pbut was try and error, because i never generated other like haha
Pso i not have spend much time with training
Phi guys! i and @Dr. Furkan Gözükara is working in a new web ui for improve sdxl image generations, it will bring the most desired feature InstantID with InstantStyle, so stay tuned!
Fwe are #1 in the sub 
Pthis nitro features is cool
PBravo, I think few people are doing training like this in Flux.
Fyep
M@Dr. Furkan Gözükara how many steps do you normally train for and how long does it take you?
Fnormally my experiments are targeting 150 epoch but i train up to 200 epochs with my 15 images dataset
Ffor last 256 images i trained up to 200 epoch again but 80 epoch was best
Fused 8x gpu
Fduration totally depends on config and gpu
Fso normally 2250 steps - 15 images
F256 images were 256 * 80 / 8 steps for best checkpoint
FYouTubeSECourses
These images are generated with FLUX LoRA training for free. Full workflow is as follows:
Medium article 
https://medium.com/@furkangozukara/ultimate-flux-lora-training-tutorial-windows-and-cloud-deployment-abb72f21cbf8
Windows main tutorial
https://youtu.be/nySGu12Y05k
Cloud tutorial for GPU poor or scaling
https://youtu...
Medium article https://medium.com/@furkangozukara/ultimate-flux-lora-training-tutorial-windows-and-cloud-deployment-abb72f21cbf8 Windows main tutorial https://youtu.be/nySGu12Y05kCloud tutorial for GPU poor or scaling https://youtu...Mwont it get copyrighted for the music haha
Mhave you tried multires training yet with flux shift timestep?
Fit is loyalty free
Fnope
Fbut when i increase inference resolution