Weight 1 works well. Prompting; if, green eyes, hair bow, works great for her general appearance, i.e. the hair structure and bow? placement. Add in brown hair, long hair, if needed. However, it seems to not, as the preview data shows I didn't need it. Mileage may vary.

For her basic outfit; black tank top, bike shorts

For her full outfit; blue jacket, fur trim, open jacket, oversized clothes, black tank top, bike shorts, socks, kneehighs, boots, cross-laced footwear

Adjust for the shot. No need for legwear in a portrait/profile/upper body prompt.

This LoRA has a Net Dim of 18 with an Alpha Rank of 9, slightly heavier like my Minazuki LoRA. Larger size allows for picking out details, but also a drawback as it lets in more unwanted data. So you'll have to do some general prompting for a non-solid color background, and to get her to stand in a full body prompt.

Trained on 27 images, most of them received varying levels of editing.

Still not sure if her leaf thing is a bow, ribbon, or a (leaf) hair ornament, but the LoRA responded much better to hair bow. So that's what ended up in the dataset.