I am admittedly kind of a noob at this, but if you enjoy my humble work, any donations are greatly appreciated: https://ko-fi.com/insult_to_ninjary
Your generosity helps me stay motivated to keep improving and to keep trying new things!
At a Glance:
Base model: Illustrated (but you can experiment with other kinds!)
Use a VAE?: Yes (baked or your choice)
Clip skip: 1 or 2 (experiment with both!)
LoRA weight: generally from 1 - 0.7 (start at 1 and adjust down as needed)
Hires. fix: use non-default upscaler; denoising strength 0.5 - 0.4
Trigger Words:
woodcut, monochrome - these are both necessary to achieve the intended effect; "woodcut" gives you the proper art style and "monochrome" gives you an appropriate color palette
If you use no trigger, you'll get sort of a retro, sepia effect on your image, but it's not very strong.
If you use only "woodcut" by itself, you can sometimes get a really neat vintage look to your picture, but it usually won't look like actual print art.
If you use only "monochrome" by itself, you'll get a monochrome image (as expected), but again, it probably won't really look like print art.
You may need to weight the trigger words with numerical values or parentheses if the effect is not strong enough. I generally put the trigger words inside 3 sets of parentheses, like so: (((woodcut, monochrome))).
Description:
This LoRA will give your art a general vintage woodcut/etching/engraving print art style. Obviously, those three types of art are not quite the same thing, and this LoRA leans more towards woodcut style (probably because of the trigger word).
Speaking of the trigger word, I effed up a bit when making this LoRA. "Woodcut" was not intentionally trained in as a trigger word. I thought I could get away with making this as a style LoRA with no trigger word, but it didn't work. Rather than re-make it, I tested a bunch of words related to the intended art style to see if anything the AI already knew could get the effect to trigger. "Woodcut" worked the best out of everything I tried, but there could be other terms that work too, maybe even better. For example, "etching" and "engraving" do also work sometimes, but not as consistently well as "woodcut" does. If you find any other triggers that work well, leave the info in a comment!
Troubleshooting:
It has a bit of trouble with fine details. It's probably why it defaults to more of a woodcut style, as woodcut often uses thicker lines compared to etching and engraving, which often use thinner lines with finer detail. I think this is just a product of being trained on SD1.5, so I don't know what can be done about it.
Related to the trouble with fine detail, it can have anatomy issues at times as well, especially with hands and fingers, and sometimes with facial features. Just try your best with all the usual fixes - different prompting, try a different base model, use a VAE, try clip skip 2, lower the LoRA weight, use adetailer, etc.
I sometimes got some unexpected foreground or background blur that you definitely wouldn't see in print art. I never found a way to consistently avoid that, but it only happened here and there, and it's not really a major dealbreaker IMO.
Outro:
If you make anything awesome, please consider posting them to the gallery! That goes for all of my LoRAs! I really like seeing what kind of stuff you all are able to make!
If you have any questions, feel free to ask! I will try my best to help! Also, if you have any ideas for future LoRAs, please don't hesitate to share them with me!