what it doubles de trainig speed?
what it doubles de trainig speed?
Dthanks
G(train_data_dir must be the parent of folders with images) Error
The training guide does not say how to properly name the folder. Please advise
The training guide does not say how to properly name the folder. Please advise
Ni use the dataset preparation section... set instance prompt (token word like ohwx) and class prompt (like man). then set the training images directory to where your source pics are and set repeats to 1. then give a destination directory (where your training images will be copied in the format kohya needs, and where you want your output loras to be written). then click prepare training data. then click copy into to respective fields. that will copy your training images into a new folder for kohya to use, like "1_ohwx man" - then when you start training, kohya pulls the repeats, instance prompt and class prompt from the name of the directory.
Dgreat speed on wsl and rtx 3090
G@Dr. Furkan Gözükara Bro you should add these steps for noobs like us, I had no idea about this step
GTraining is started now
Edoes anyone know the best way to upscale an image while fixing strange artifacts without adding new ones? I've read about using an upscaler followed by img2img with flux.What's a good and efficient way of doing this without spending hours on inpainting?
GUpscale inside Comfy UI. I do 20 steps a 1.5 X, it does an amazing job imo, sharper, more details, fixes most artifacts and increases the resolution size 50%, I save as .JPG and the size is around 500KB instead of 1.5-2MB while maintaining quality
Fhello when I'm on kohya on massed compute I don't have this part
GThis is probably a LORA on img to img with a low denoise, essentialy the photos are stolen from real women and they swap the face bro
GGot it running and still get an error. I dont get it man. I did all the settings for 24gb and 48gb both still fail
Greinstall Kohya, you have a different version, i did that and that reappeared. Also, make sure youre on the LORA tab not Dreambooth
TIt is perfectly possible
GSDXL can do this
GFlux with a Lora and your set
TYes
Gbut her body stays the same
Gits the same thick body
Gits a trained face lora with that body
Gthose dudes make money with fake woman lol
GI think its img to img
TBut freckles vary (on the body) and the size of certain parts also varies 
Gwith that face lora thats why the fingers mess up a buit
Gbit
Gits sdxl before flux
TIt could be img2img or a good editing job for the skin
Gthats what im saying
Gthe denoise will do that
Gthat will happen
Gits img2img of that im sure with a lora face, the body is/are stolen photos of a woman or the woman is known to the person running it, or the woman is the one running it with a fake face to protect identity, you will never know
TIf it's real, the photos are definitely stolen otherwise she would say it's a woman somewhere hahaha
otherwise she would say it's a woman somewhere hahahaGthe face is not real we know that
Gthe body is
Gim having trouble even training my damn lora on FLUX rn haha
TYes, in her fanvue it says that she is a real girl powered by AI
Gits more than likely a dude with stolen photos imo
TWhy?
Gerrors, i did all the settings right, not sure
Gand I have 48GB VRAM and 2 24GB Cards
Glol
TI don't know what to tell you... fanvue almost closed my account because they thought they were stolen (and no, they are 100% AI)
Gif thats true than the body and face are trained in the lora
TI can't help you there, it works perfectly in Massedcompute
GSo why are you asking how its done if you are the one running that page?
G:7036what:
TNo no, it's not my page I wish it was haha
I wish it was hahaG:304fakekid:
HHas anyone experimented with 1 image Lora train on flux ?
Ftorch upgrade but on linux
Fwow really good