you will get better with fine tuning
you will get better with fine tuning
WDoes your recent patreon post contain instructions to do it on runpod? If it does, I will rejoin as a member for a couple months
Fyep. only difference is
Fload config into dreambooth tab
Frest 100% same
Feach checkpoint will be 23.8 gb pay attention to storage
Fi am also preparing a massive tutorial
W23.8gb. Holy crap!
Is it possible to extract a lora from that? is it worth it?
Is it possible to extract a lora from that? is it worth it?
Fit works
Fbut you still need high dimension :d
Fi made article here public
FPatreon
Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon
Wthankyou. I will wait for the runpod tutorial.
FYouTubeSECourses
Unlock the power of FLUX LoRA training, even if you're short on GPUs or looking to boost speed and scale! This comprehensive guide takes you from novice to expert, showing you how to use Kohya GUI for creating top-notch FLUX LoRAs in the cloud. We'll cover everything: maximizing quality, optimizing speed, and finding the best deals. With our exc...
Fstart from 44:38
Fi wont repeat all again 
Fin new tutorial cuz exactly same
Wso this includes the dreambooth tutorial?
Fthis is lora. for dreambooth you load this config into dreambooth tab else is 100% same : https://www.patreon.com/posts/kohya-flux-fine-112099700
Patreon
Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon
Fso follow this tutorial
Fwhen loading config
Finstead of lora
Fload config of dreambooth
Fdreambooth config : https://www.patreon.com/posts/kohya-flux-fine-112099700
lora config and installers (the one used in tutorial) : https://www.patreon.com/posts/110879657
lora config and installers (the one used in tutorial) : https://www.patreon.com/posts/110879657
CWill you do in future tutorial for Fine Tune / DreamBooth?.
Fyep
Fi am fully preparing it right now
Fit will show massive comparison too
Dfor those who were interested in this dreambooth-training here are some more tests. Some with reg data, some without, some with training prompt, some without anything just code. It doesn't seem to make much of a difference.
Dsimpler grids without reg images
Jsome vips are already trained in to some degree, in my experciene no reg, no prompt gives me more then good enough results. and i switched to the competitor product (fgym) because it is so easy to use 
D"no reg, no prompt" will work but change the entire model. Barack Obama will look like your trained subject.
Ji dont train known people, just ohwx works super
Dyou don't get the point, please read the thread that I linked above first.
Jits about the bleeding i know, if it works it could be usefull, but its only for onetrainer right now?
Dyes. SimpleTuner wants to implement, and I've heard kohya too, but nothing ready
Fgreat results but i couldnt get it :d
TI have a few questions. I'm now trying the DB method for flux training, no captions, 30 images, 1 repeat, 150-200 epoch about 5000-6000steps - is that too much. Also since i'm a male but i have long hair, every sample image during training shows as a women, in my case is it better to provide captions saying I'm a man? Joy caption also thinks I'm female. Also i noticed that in any of my trained/extracted lora, i can never add any other person in the picture, because they all look like me, even though I'm using a numeric trigger, and even if I import another celebrity lora, it just blends the 2 lora's together. is there no way to do an AI image of myself and my favorite celeb? ie. "photo of k3v1n with billie elish" (another lora)
Fabout second part yes it bleeds
F@dxqb found a way to lessen it on onetrainer
Fbut i couldnt make it
Fabout first part only use
Fohwx man
Fit should work with sufficient training
Fi will update patreon post with grids and my best epochs
Fbased on number of images
Fand batch size
Tso inpainting is the only way? how does it work in base flux compared to lora? because in base flux you can specify 2 embedded base celebrities, ie. john wick and taylor swift playing in a concert together.
Fthe problem is that since we train on distilled model
Fit bleeds a lot
Fbut famous people were trained before the distillation