ILLUSTRIOUS VERSION:
Trained on 195 HD images with heavily curated tags of Arthur Rackham's golden age illustrations.
ALL TAGS -> GDrive <-
Booru-style prompts - (e.g., 1girl, 1boy, old man, size difference, age difference, giant, wrinkled skin, long nose, nymph, gnome, faux scenery, etc.)
When prompting for two+ subjects try adding a slash between the subjects: 1girl/1boy
Depending on the style you want (ink, watercolor, etc.), it helps to include the desired medium in the prompt and place any unwanted mediums in the negative prompt.
For example - Positive: watercolor_(medium) Negative: ink_(medium)
That said, it's not a strict rule! Some mediums can work well together, so feel free to experiment~
SD1.5 VERSION
Trained on 235 HD/upscaled images of Arthur Rackham's work.
STRENGTH/WEIGHT: 0.6 - 1.2
Trained on base Stable Diffusion 1.5 (with anime/booru prompts; Yet flexible)
You might have to shout at it for specific tags, for example: (old, man, wrinkled skin:1.6) <- depending on the model and strength of the lora
Generates humans sometimes when asked not to.. so add a higher weight to negative prompts for 1girl, 1boy, human, etc.
Wild west tagging (auto-tagged)
Activation tag is a mash-up of "fairytale" and "illustration" with a number instead of an "i" - fa1l