CivArchive
    Natsumi Arai | The Edible Season | Character LoRA - v1.0
    NSFW
    Preview 147316308
    Preview 147316329
    Preview 147316306
    Preview 147316313
    Preview 147316315
    Preview 147316318
    Preview 147316324
    Preview 147316325
    Preview 147316327
    Preview 147316328
    Preview 147316335
    Preview 147316333
    Preview 147316336
    Preview 147316337
    Preview 147316338
    Preview 147316340
    Preview 147316343
    Preview 147316344
    Preview 147316345
    Preview 147316348
    Natsumi Arai is an original fictional adult Japanese character LoRA for Krea 2. She is artisan wagashi sculptor and seasonal confection colorist, built around a distinct human, professional, and visual lane rather than generic glamour.
    
    Her identity is light neutral-beige skin with fine natural texture, softly oval face with a delicate jaw, wide-set very dark-brown monolid eyes, low straight brows, small refined nose with a rounded tip, full lower lip, one tiny beauty mark beneath the anatomical right eye, and glossy black hair gathered into a loose low knot with two straight temple strands. Her body anchor is adult 5-foot-2 petite slender pear-shaped frame with graceful narrow shoulders, subtle waist, softly rounded hips, compact strong legs, and a small natural B-cup bust.
    
    Her defining professional styling begins with fully opaque fitted plum high-collar blouse with three-quarter sleeves, high-waisted cream trousers, indigo cross-back confectioner's apron, and closed dark studio shoes, and her core environment begins with empty contemporary wagashi studio in soft morning light with one low preparation table, natural food-color bowls, wooden shaping tools, seasonal leaves arranged as references, and no customer, assistant, tea ceremony guest, readable label, logo, brand packaging, anime styling, or generic bakery display. The validated adapter also supports modern fashion, lifestyle, movement, formal, night, profile, and out-of-role scenes without losing the woman.
    
    The approved twenty-frame trained gallery covers close identity, full-body role, waist-up, seated, front and rear three-quarter, exact profile, walking, dynamic movement, rain-lit night, cool light, studio editorial, and occupation-free lifestyle compositions. Seven frames preserve professional and role-adjacent behavior; thirteen neutral, lifestyle, lighting, and out-of-dataset frames prove flexibility.
    
    **Trigger:** `natsumi_arai_vn`
    
    **Recommended strength:** `1.0`
    
    **Useful range:** `1.0 tested; other strengths unverified`
    
    For identity prompting, include:
    
    `fictional adult Japanese woman age 28, light neutral-beige skin with fine natural texture, softly oval face with a delicate jaw, wide-set very dark-brown monolid eyes, low straight brows, small refined nose with a rounded tip, full lower lip, one tiny beauty mark beneath the anatomical right eye, and glossy black hair gathered into a loose low knot with two straight temple strands, adult 5-foot-2 petite slender pear-shaped frame with graceful narrow shoulders, subtle waist, softly rounded hips, compact strong legs, and a small natural B-cup bust`
    
    For wagashi scenes, use exactly one tray, confection, shaping tool, or color bowl and keep food count and hand ownership explicit.
    
    Keep exactly one adult subject, one readable action, concrete camera distance, countable props, explicit ownership/contact, and complete required anatomy. Avoid crowds, duplicate or reflected people, overloaded hands, vague abstract actions, branded elements, readable text, real-person likenesses, franchise references, transparent clothing, and childlike styling.
    
    The exact epoch-10 adapter passed six unseen scenes covering close identity, full-body role fidelity, exact profile, dynamic professional action, cool lighting, and out-of-dataset flexibility. The twenty-frame trained gallery then passed direct owner review for identity, clearly adult age/body, anatomy, role/lifestyle separation, scene intent, diversity, duplicate control, and gallery quality.
    
    Natsumi Arai is fictional, clearly adult, and not based on an actual person.

    Description

    Initial Krea 2 release.
    
    The fifteen-image cloud-training package used one concept role anchor, three neutral diagnostics, six occupation-free lifestyle frames, and five professional frames. Krea cloud training produced ten checkpoints;
    the exact epoch-10 adapter passed the established six-scene unseen validation. This launch gallery contains twenty new trained-adapter renders.

    Details

    Downloads
    20
    Platform
    CivitAI
    Platform Status
    Available
    Created
    10/30/2026
    Updated
    10/30/2026
    Deleted
    -
    Trigger Words:
    natsumi_arai_vn

    Files

    natsumi-arai-vn-krea-2-base-epoch-10.safetensors