LoRA in Stable DIFFUSION -something you MUST LEARN- (fooocus)

learn how to use LoRA in your artworks. This can change everything and take you to the next level.
civitai.com/models
email: [e-mail]

[@alpaykasal2902]

Yooo, I am so here for "SuperBatMan"!!! A vigilante detective from Krypton? Take my money, please.

[@sikliztailbunch]

That little funnel icon to filter for lora als lets you filter for SDXL only. Then you don´t get loras shown that look cool but won´t work. I made the same mistake in the beginning. But you should really make a video about training own loras. It´s so easy on civitai. I bet a lot of people are interested in that

[@MasterDiffusion]

Yeah, you’re correct. Actually plan on doing that video when I have the time to sit down and do one. But yes, it’s a great idea.

[@yelagin]

I trained a girl's face through Lora1.5 and it doesn't work in Fooocus at all. In your case, it wasn't Lora, it was the application of comic book styles.

[@MasterDiffusion]

You are correct. It didn’t notice it was a 1.5. The point of the video was how to add a LoRA. I will pay closer attention. Thanks

[@Slice.]

thank you Sir 😊

[@MasterDiffusion]

I hope the video helps!

[@isreal8339]

Can you go over inpainting with loras?

[@Paulo-ut1li]

Isn't that Jim Lee Lora SD 1.5? It shouldn't work with SDXL on Foocus.

[@MasterDiffusion]

Yikes. Nice catch. I was trying to show how to add a LoRA and didn’t even notice it was a 1.5. Embarrassing yes, but the idea is the same for installing.

[@SomoneTookMyName]

@@MasterDiffusion The 1.5 loras will work. with fooocus. I have been spending a fair amount of time with fooocus and adsorbing everything I can. currently there seems to be very little documentation on fooocus. I did read that 1.5 loras will work and did test them as well and found that they will in fact will work. Where you will run into issues, basically a model set not working is with the checkpoints. Those in fact have to be SDXL.

[@yelagin]

no, I tried some works, lora15 doesnt works@@SomoneTookMyName

[@ShadowKiller71]

...mind sharing that technique?@@SomoneTookMyName

[@phalxor]

@@SomoneTookMyName thanks for any actual info. my experience just bumping into loras is that 1.5 loras work just fine of course pairing xl with xl is probably going to give better results but i've seen people outright just say do not use 1.5 lora with sdxl and I don't understand that. it works great. so if you are picky about only using xl loras but maybe one doesn't exist for what you need, don't be afraid to try 1.5

[@bbrother92]

What is lora????

[@MasterDiffusion]

I explain it in the video. Simply put, it is an add on that has a specific data set that has been trained for a specific type of art. such as trained on 100 car pics then it will reference the cars to make your pic

[@bbrother92]

@@MasterDiffusion So lora is for customizing image creation? Is it pretrained or I need to train it?

[@MasterDiffusion]

@@bbrother92 pre trained and ready

[@lanzer22]

If the title reads “ - something you must learn”, it implies there is something about LORA that you must learn. If this is an intro to Lora, use a comma instead of a hyphen.

[@MasterDiffusion]

if someone does not know how to add a Lora in, is that not something they don't know and need to learn? Sorry for any misleading info. Im just a guy, trying to put stuff out to help people learn. Have a good one!

[@lanzer22]

@@MasterDiffusion The comment is simply on the grammatical error in your title. By no means am I somehow saying that sharing of information is not worth the effort. Although there are countless LORA tutorials out there for months, I was just surprised that there is yet another intro guide out yesterday. You can see why I was assuming there's more to it.