Shinomiya Kaguya – Flux.2-klein-9B Character LoRA

(四宮かぐや from Kaguya-sama: Love is War)

Trained on Flux.2-klein-9B base (Modelscope version) and intended for use with the distilled/KV variant.

Current status

• This is a checkpoint from 10 000 steps, training is completely finished.

Recommended settings

• Strength: 1.3 – 1.8 → my personal favorite — 1.5 (best balance of likeness and flexibility)

• Steps: 8 – 16 → my sweet spot — 12

• Sampler: Euler / Euler simple

• CFG: 1.0 – 1.2 (1.0 usually gives the cleanest and most natural results)

• Resolution: 1024×1024 or similar aspect ratios

Trigger & base prompt
Main trigger: 4n1meG1rlSh1n0m1ya, anime girl with long black hair in a ponytail tied with a ribbon and red eyes

Side variant (fewer examples in the dataset, may have slightly lower responsiveness and different hairstyle): 4n1meG1rlSh1n0m1ya, anime girl with long loose black hair and red eyes

Classic outfits from the Anime

• Standard / winter uniform: She wears a Shuchiin Academy girls' uniform, a black long-sleeved dress with a white collar.

• Summer uniform: She wears a Shuchiin Academy girls' uniform, a black short-sleeved dress with a white collar.

After the trigger, simply describe clothing, hairstyle, emotion, and scene in natural language — the model responds very well to narrative text without booru tags.

Dataset

• 78 images

• Mostly anime screencaps from the two seasons of Kaguya-sama

• A small amount of high-quality fanart and booru generations for better angle and expression variety

Notes

• This is a non-commercial experimental LoRA.

• Not monetized, not for sale — created purely for personal interest and to test Flux Klein's capabilities on anime characters.

Link to all training epochs: compatibility with my previous LoRA of Hayasaka Ai : [ However, very little research has been done on weight interaction and potential LoRA bleeding — use with caution and adjust weights individually.

PS:

If you like the idea of creating images with anime characters on Flux.2 and want to help develop it further into a full proper model (including Klein versions), feel free to reach out!

I'm especially interested in collaboration if you:

• Have more experience with high-quality dataset curation and captioning

• Want to help expand the dataset

• Are open to adding tasteful NSFW images to the training set

You can write me a DM on Civitai. I can share my current images and their captions with you so we as a community can work together on improving the next versions.