It works fine with the default flux dev, but not this other finetuned dev ccheckpoint
It works fine with the default flux dev, but not this other finetuned dev ccheckpoint
�Does this mean we can now resume training where we left or if I want to train again with more epochs can I just add it and start from where it last trained?
Fyou need to report to kohya
Fyes you can
Fyou can also use latest checkpoint
Sdr Furkan i hope everything is ok , the full tutorial about make drambooth with flux is in process right?
Fyep editing
F
Show this process make dreambooth models intead of loral model make better models thna make the loras? only to undertand how is work
Ffine tuning is training full model
Flora is optimization
Fthus fine tuning better
D@Furkan Gözükara SECourses not sure if kohya and/or you still see this on a closed PR, so here too:
@bmaltais can we add this to gui so many people want to test
all my proof-of-concept samples were at weight 1.
Yes, loss of reg steps is then very low compared to train steps, but this is a function of the reg step prediction already being very close to the target.
you can try at weight 10 or even 1000, but I'd expect that the regularization then overwhelms the training steps, and the model doesn't learn anything anymore.
@bmaltais can we add this to gui so many people want to test
- Specify a large value for --prior_loss_weight option (not dataset config). We recommend 10-1000.
- Set the loss in the training without using the regularization image to be close to the loss in the training using DOP.
all my proof-of-concept samples were at weight 1.
Yes, loss of reg steps is then very low compared to train steps, but this is a function of the reg step prediction already being very close to the target.
you can try at weight 10 or even 1000, but I'd expect that the regularization then overwhelms the training steps, and the model doesn't learn anything anymore.
Vnope it works, just tested it thoroughly, weight 10 works fine just takes longer to train
Deven at 1 I think it dampens training of the concept you do want to change, so even if 10 works I'd be careful and/or try different values
DI was more thinking in the other direction, doing less reg steps than 1:1, because even at 1:1 with weight 1, regularisation is strong
Dbut currently working on sth else so I'd probably not get around to more tests any time soon
Fi am waiting bmaltais to add
Dit's merged in sd3 branch
VYou could always do this if you used the save-state options. You would then paste the path of the folder in the resume from state box. However, the issue is that if you trained for 100 epochs, and the setting was set for 200 epochs, resuming from the trained state of 100 epochs will then train for another 200 epochs, so 300 in total. The epochs in the setting would have to be adjusted if you want max 200.
The new setting is to change the maximum amount of save-state folders you can have. Normally it would save either at the end, which is not great if something bad happens, or at every time you save a copy of the model. This can be a lot of big files, so with this option you can have it set to say a max saved-state amount of 1, so that each new saved-state folder overwrites the older one. Great for saving HD space.
The new setting is to change the maximum amount of save-state folders you can have. Normally it would save either at the end, which is not great if something bad happens, or at every time you save a copy of the model. This can be a lot of big files, so with this option you can have it set to say a max saved-state amount of 1, so that each new saved-state folder overwrites the older one. Great for saving HD space.
WHi @Furkan Gözükara SECourses I try to use your Quality_1_23100MB_14_12_Second_IT.json to train checkpoint but it's very difficult to get 23.1 GB Vram. Windows mostly eat 1.2 - 1.3 GB VRam. How do you limit VRAM use by windows. Since I have 24gb VRAM choosing Quality_1_15500MB_32_Second_IT.json is double the training times and not ultilize my VRAM.
CHello, I want to try to finetune flux dev on a style rather than a character? Any advice on what is different? For example the captioning, I only captioned with my trigger word + class for the character, is that the same for style? If you have any documentation or training (even if not from you) that's be very much appreciated!! love
VFor styles you can either go with a trigger word for the entire image, fully caption everything, or no captions. Characters require captions for everything except the things about the character, like their hair, clothing etc. Things that are captioned can be changed later, things not captioned are static and will be reproduced when prompted.
Fyes in that case
Fjust increase number of block swaps
Fshown in coming video
Fyes i have a full guide and research
Fplease read fully
CThanks mate
Shi, do you know when the Fine Tune tutorial video will be online?
Fhopefully today
Fcurrently generating subtitles
F
Ccant wait, already prepared 4 sets of photos for training
SDr Furkan do you have some flux base model favorites for training?
VSdr Furkan do you planing to do a titorial about trainiga style? i just saw the last video what i must change for a trining of a style?
Ssorry dr furkan you already mentioned in your last video my quesiton about style
Fyes hopefully
Fwe hjave it already
F
VStyle will be your most popular video. I recommend doing a walkthrough on it. You can pick an art style that is copyright free if that was your concern.
VObelisk Art History
As much as possible, Obelisk Art History uses images of artworks and writings that are in the public domain. Works in the public domain are those whose...
Fwe literally have it there
Feverything
Fbut i can make a video too later
VPeople don't want to read, they want a video walkthrough lol. 
VIf you do make a video choose an art style that is interesting, will get more views.
