Tools that I used to make this LoRA: LosslessCut (To capture PNG images from source), XnView MP (For image cropping), DatasetEditor (for captioning), LoRA_Easy_Training_Scripts (for training).

Tags:

Style:

Remove this tag, or decrease its weight if you need better style flexibility

anime screencap,

Appearance:

ankouguisumv, 1girl, solo, black eyes, brown hair, round eyewear,

Haircuts:

Short ponytail:

medium hair, sidelocks, short ponytail, folded ponytail, 

Long ponytail:

medium hair, sidelocks, ponytail, 

Hair down:

medium hair, hair down, 

Short hair:

short hair, 

*There is a scene in restaurant where her hair color is of a different brown shade, I tagged them with alternate hair color tag, it doesn't have strong effect in positive prompt, but adding it to negative prompt helps in getting more consistent hair color

Outfit:

brown coat, open coat, 

OR

brown jacket, open jacket, 

+

grey shirt, turtleneck shirt, shirt tucked in, black skirt, partially fingerless gloves, brown footwear, high heel boots,