明日から本気出す

LoRA model for Hatsuyuki from Kantai Collection.

Trained using 190 images. Many thanks to all the artists. k2 where

Recommended settings

• LoRA network weight around 0.8-1.0

• Please generate using highres fix for best effect.

• For eye expression it can help to use specific lora such as: hanme1 by

  2vXpSwA7, jitome by ManaMomo, or other similar

For LoRA Block Weights:

• OUTD (1,0,0,0,0,0,0,0,1,1,1,1,0,0,0,0,0) seems pretty safe

Suggested prompts

Base appearance (no rigging)

1girl, hatsuyuki \(kancolle\), 
blunt bangs, hime cut, black hair, long hair,
serafuku, teal sailor collar, short sleeves,
pleated skirt, blue skirt,
black socks, anchor symbol,

There's only one main outfit baked in, so you can usually get away with less tags than this. Also you can try "(blue) sailor collar" but it might not turn very blue.

Rigging

smokestack,
machinery,
rigging,
torpedo launcher,
thigh strap,
backpack,
holding turret,
cannon,
armored boots,

Rigging is quite gacha. Try controlnet or other methods to at least limit the search space.

You might need to add some tags if your base model draws her a bit too big (flat chest, young, etc.)