I bet that with TE training enabled lower ranks will improve a lot, just a theory!! hehe
I bet that with TE training enabled lower ranks will improve a lot, just a theory!! hehe
FI get this error message when I run training on Umbubu
Fyet I gave the right path for the images
�That has to with the folder training directory having to be the main directory above the folder of the images and not the image folder itself, if that makes sense
F,could be but not supported yet :d
Fyou are missing training dataset
�So, Home/Ubuntu/Desktop/ instead of Home/Ubuntu/Desktop/trainingimages/
Fshow screenshots how did you give folder path and what is in folder
Dinteresting how the glasses are missed in all but the very high rank training
Fflux is very powerful
Fit omits eyeglasses it knows it is not me :d
Fi will retest with adding eyeglasses word
Dfor the test prompt?
Fye for test prompt
Fwhen eyeglasses added it perfectly generates
Falso 3 more final trainings are running
Fwith full captions of
Fjoycaption
�Is there any disadvantages to training at 128 rank @Dr. Furkan Gözükara ? I'm assuming longer training time and more VRAM usage perhaps, but what about with inference, same thing?
Fvram usage slightly more 19 mb per rank for 16 bit
Ffile size bigger
Fbut it also leads more overfit. that is why full fine tuning is the best way
�Ok but quality wise, 128 seems to be the sweet spot so it seems kinda worth it
F128 best
�Ok cool thank you Dr!
Finstead of reducing lora rank i will reduce resolution for lower VRAM gpus
Fwill prepare every GPU config
F9.5 GB, 11.5 GB, 15 GB, 18 GB 
FMy images are in the training_imgs folder, isn't that good?
�Image Folder directory should be: home/Ubuntu/Downloads
Fshow inside of it
Fi can see you missing sub folder
Fput images inside sub folder
F1_ohwx man
Fit will be fixed
F
�Yeah just like how Dr. said put it in the subfolder and don't change directory path and it should work
FSo I should put the images in the download folder?
�Create new folder inside your training_imgs folder and put them in there, don't update the directory path in Kohya
FOk thanks
�It looks for Parent folder, that's why
FI've the same err
�You might have to rename your training images a certain way
G@Dr. Furkan Gözükara I am using 48GB script, did Dataset Prep, changed all directories, dled all safetensors, then i get this error. I dont see it bro, do your eyes see the issue? it works on my 24gb 3090ti, but it has been running 10hr and 33 min already doing 32 images 25 epoch, Anyway, any hint to fix this error on my 48gb? Id rather use this system, i lowerd the images to 21 as well from 32.... let me know, i am missing some setting or maybe need to reinstall, I noticed when i do Fix_For_FLUX_Step_2.bat on this 48gb system it breaks, i then have to install kohya again then run it, then start training
G
�I saw these instructions when I was testing Kohya and had that same error and it's what fixed it for me:
The program expect to find a folder like c:\images\[steps]_[instanceprompt] [classname].
So to train a 100 steps of a woman, i would make a folder called c:\images\100_ohwx woman
Then put the images in the "100_ohwx woman" folder.GDr said folder should be 1_ohwx right? I also have it as 1 not 100
Galso i dont use the word ohwx, i use another word
�Those instructions were from a couple years ago lol so might be outdated but I think you should be fine
Gdataset prep puts it at 1 when it creates the folders