@Dr. Furkan Gözükara so the models command in the terminal automatically fetches this one right?
@Dr. Furkan Gözükara so the models command in the terminal automatically fetches this one right?
Fif you are using our downloaders yes
Fflux 1 dev
POk thanks
Fi hate slowness of runpod
Fyou waste at least 5x time to load and install something compared to massed compute
Fsometimes even more
PYeah massedcompute is really fast,I like it
P@Dr. Furkan Gözükara i want my model to get the body structure also correct so can I put some images in the dataset without cropping ,only resizing
Cif i am trying to create a character for myself like was done in these video, is it best to set this drop down in the top right to LoRa?
Cin the video it was on fine tune
Cand would that effect what settings i would want to use
CI’m also curious about number of images. I have 59 in my dataset. I noticed in the video you did 15 images x 200 epochs = 3000 steps.
I went ahead and did 59 x 100 epochs = 5900 steps and will look extra carefully at the models around epoch 5, I kinda just wanted the slack of the other epochs in case it’s under trained by then. thoughts on this? how many epochs would you go with? seems like with my hardware I’ll hit epoch 100/100 in 4-5 hrs
I went ahead and did 59 x 100 epochs = 5900 steps and will look extra carefully at the models around epoch 5, I kinda just wanted the slack of the other epochs in case it’s under trained by then. thoughts on this? how many epochs would you go with? seems like with my hardware I’ll hit epoch 100/100 in 4-5 hrs
Cdef got 9 bucks worth of info out of the video and I’m not even done yet lol, cheers mate
Fyep
Fyes have more images for that
Fi dont recommend lora
Fdreambooth better
Fthen you can extract lora
P@Dr. Furkan Gözükara I have messaged you please check
�@Dr. Furkan Gözükara Is there a way to save epochs mid training? Like I have set up save evry 25 epoch but while training I want to save the 10th epoch as well?
Fi dont know
Fchecking
TSorry, thought you are training flux dev, my data is unrelated to SDXL training
Cahh gotcha
Cto clarify, the “fine tune” option is dreambooth right? I’m using your patreon configs now
Cand what are your thoughts on this? for higher image counts do we want more steps?
P@Dr. Furkan Gözükara i have messaged you please check
Fyes dreambooth is = fine tuning
Fwhen you dont use reg images
Fwhen you use reg images still fine tuning but in a more specialized way
F5900 steps good
Fsure checking now
M@Dr. Furkan Gözükara where do you place the .txt captioning files when training a flux checkpoint with your latest dreambooth setup?
I have captions made from JoyCaptions, do i just leave them in the input dataset folder?
I have captions made from JoyCaptions, do i just leave them in the input dataset folder?
Fyes same folder
Flike
Fimage1.png
Fimage1.txt
P@Furkan Gözükara SECourses is captioning better while training on flux+dreambooth? does it give better results?
Mi don't understand this error when i hit start training in dreambooth kohya, does anyone know why I have this "data not found error"?
my kohya prepared data directory looks correct to me :/
my kohya prepared data directory looks correct to me :/
Fi prefer activation token + class token
Flike ohwx man
Fdataset error
POk so I shouldn't add captions right?
Fwatch this section of this video
FYouTubeSECourses
Ultimate Kohya GUI FLUX LoRA training tutorial. This tutorial is product of non-stop 9 days research and training. I have trained over 73 FLUX LoRA models and analyzed all to prepare this tutorial video. The research still going on and hopefully the results will be significantly improved and latest configs and findings will be shared. Please wat...
F17:08 Dataset preparation, emphasizing image captioning, cropping, and resizing
F@m3dia_offline
Fyes i dont recommend
P@Furkan Gözükara SECourses i have 4 A6000 total 192 gb ram the training config will automatically distribute the load among all 4 right?