Sign In

[IL] Karis - Golden Sun: Dark Dawn / カリス - 黄金の太陽 漆黒なる夜明け

54

345

319

12

Verified:

SafeTensor

Type

LoRA

Stats

111

57

29

Reviews

Published

Jun 17, 2025

Base Model

Illustrious

Trigger Words

Karis_GS, green hair, ponytail, bangs, asymmetrical hair, sidelocks, purple eyes, triangle earrings,
capelet, blue capelet, small breasts, mandarin collar, coat, white coat, ribbon trim, gold trim, undershirt, pink shirt, pink sleeves, crossed belts, miniskirt, pink skirt, white gloves, elbow gloves, bracelet, black pantyhose, white boots, knee boots

Hash

AutoV2
A04EDEC298

Illustrious conversion of this LoRA, which will let you make the best part of Golden Sun: Dark Dawn (I actually have a better opinion of the game than most, and still think she's the best part). I’ve actually got almost everything I need to make a Himi LoRA, but don’t have any way to get good pics of her model thanks to all the Djinn floating around on the status screen once she joins (I got around it for Karis by capturing pics of her before she got Djinn). If anyone has a solution (model rip?), let me know.

Caution: Use of a VAE is recommended due to significant use of monochrome images in dataset. Colors will come out weird otherwise.
Tags:
Trigger: Karis_GS
Prompt: green hair, ponytail, bangs, asymmetrical hair, sidelocks, purple eyes, triangle earrings, capelet, blue capelet, small breasts, mandarin collar, coat, white coat, ribbon trim, gold trim, undershirt, pink shirt, pink sleeves, crossed belts, miniskirt, pink skirt, white gloves, elbow gloves, bracelet, black pantyhose, white boots, knee boots
Other tags in data you might be interested in: fighting stance, magic, holding, holding weapon, weapon, staff, sword
Negatives: monochrome, greyscale, manga illustration, flat coloring, 3d, low poly, pixel art, speech bubble, blank speech bubble, (the first four are quite critical)

This conversion uses data identical to my V2 for PDXL. Another case where the base is recognized pretty instantly (first or second epoch) and the rest is refining the details

Update: Prodigy version uses a different sampler in training. Part of a series of tests to determine if the sampler is why a few LoRAs based on mainly black and white images had weird colors in Illustrious but not Pony. Since it worked for Mana, I'm trying it here too.