✨ Mizuha Hiiragi from Osananajimi wa Seihoushi Toban
— She has that dangerous childhood-friend charm where the gentle smile feels sweet at first, then slowly starts looking a little too confident. She gives off a mix of strict schoolgirl, teasing upperclassman, and “I know exactly what I’m doing” energy, which makes her expressions fun to generate in both wholesome and more suggestive scenes. Her sporty side also fits surprisingly well, especially with gym uniform or athletic showcase prompts.
621059 & 637520 & 622005
If you know, you know~
>Doujinshi COLLECTION<
⚠️Turn off your R/X/XXX Filter to see hidden JUICY Loras in >HERE<
🖼️ How I Generate My Images
I made a full article that explains how I generate my images — including my settings, prompting tips, and even my potato PC setup 😅
👉 Check it out here if you’re curious or want to try the same method.
🔞 NSFW gallery (18+)
Check my account here to see the spicy showcase that I can't show here, if any.
🎨 Prompt Tips
These are optional starting points for handling a few behaviors of this version. You may still need some retries depending on the checkpoint, prompt, and LoRA strength.
🖤 Manga-like result
For a result closer to the original doujin’s greyscale manga appearance, avoid positive tags that strongly push anime coloring, vivid colors, or an anime screenshot style. It is also better to avoid character or outfit color tags, since they may encourage the checkpoint to colorize the image.
Recommended:
✅ Positive prompt: greyscale, high_contrast
The high_contrast tag helps reproduce the stronger black-and-white contrast found in the original source.
For a softer greyscale result with less intense blacks and whites, you can remove high_contrast from the positive prompt or place it in the negative prompt instead:
🚫 Negative prompt: high_contrast
🦺 Removing unwanted vest logos
Some checkpoints may add a school logo, crest, or other decorative symbol to her sweater vest even though it is not part of the intended design.
To reduce this behavior, try:
🚫 Negative prompt: logo, emblem
This is mainly useful when the unwanted symbol appears because of the checkpoint’s existing school-uniform bias.
🏃 Track outfit color stability
Her track-and-field outfit is tagged with:
✅ Positive prompt: two-tone_sports_bra, two-tone_panties
However, the outfit colors may still be inconsistent because the dataset contains very few colored references. Depending on the checkpoint, using the two-tone_* tags can sometimes encourage it to invent its own color combination rather than follow the intended design.
Because of that, try both approaches:
Use
two-tone_sports_bra, two-tone_pantieswhen you want to reinforce the outfit’s two-color structure.Remove those tags if the generated colors keep drifting. In some cases, leaving them out can produce more accurate colors.
If the outfit repeatedly turns black, try:
🚫 Negative prompt: black_sports_bra, black_panties
You may also need to specify the intended colors directly in the positive prompt, since the greyscale-heavy dataset cannot provide strong color information by itself.
🎨 Controlling the cover-page style
This LoRA was trained with both textless_cover_page and cover_page.
These tags refer to the colored doujin cover images, which have a slightly different rendering and coloring style from the greyscale interior pages.
To steer the result toward that colored cover style, try:
✅ Positive prompt: textless_cover_page
To reduce the influence of the colored manga-cover style and allow another coloring direction, such as a more anime-like appearance, try:
🚫 Negative prompt: textless_cover_page, cover_page
These tags mainly influence the visual direction rather than guaranteeing one exact coloring style.
🧥 Default outfit details appearing on other outfits
When a default outfit detail keeps appearing in a different outfit, add the exact leaking detail to the negative prompt.
For example, if the pleated details on her skirt appearing when you do not want it:
🚫 Negative prompt: pleated_skirt
Use the specific leaking tag instead of adding many unrelated negative tags.
⚖️ LoRA strength and outfit flexibility
A weight of 1 can still work with a very different outfit, especially when the new outfit is clearly different from her default one.
Only lower the LoRA weight gradually when outfit details keep leaking even after using clear outfit tags and negative prompts. Lowering it too much may also weaken her face, hair, and overall character identity.
📊 Lora Project
If you want to know what characters that I'm covering next you can check these link :
👉 Personal Project
👉 Commision Project
If this model sparked something in you, show some love with a like, review, or buzz.
Now go—
Be wild
Description
📖 Mostly greyscale source material
This LoRA was trained mainly using greyscale images from the original doujin, with only three colored source images available as additional references.
Because the colored coverage is very limited, some colors may not remain perfectly consistent between generations. The checkpoint and prompt may have a stronger influence on the final coloring than they would with a dataset containing more colored references.👕 Two main outfits
The dataset contains two primary outfits:
- Her school uniform with a sweater vest
- Her track-and-field gym outfit
Both outfits are included in the training, but the school uniform has a somewhat stronger influence on the LoRA.🎓 Stronger school-uniform influence
Her school uniform may remain noticeable when prompting alternative clothing, especially the shirt and vest structure.
I used the dress_shirt tag for the uniform shirt. Depending on the checkpoint and the rest of your prompt, this can occasionally blend with dress-related clothing concepts or affect the structure of other outfits.
You may need to suppress the school-uniform pieces through the negative prompt when trying more significant outfit changes.As usual, the result can still depend on your checkpoint, prompt, LoRA strength, CFG scale, and generation settings.