Shiro-sama_NAI-V4.5-Curated (シロ様)

High-fidelity character LoRA for Shiro-sama (シロ様) from The Helpful Fox Senko-san, for use with Krea 2 Turbo.

This LoRA was built from a high-quality synthetic dataset of 598 captioned images generated with NovelAI 4.5 Curated. It is designed for txt2img character generation and img2img, with an emphasis on character likeness, clean linework, outfit flexibility, and descriptive natural-language prompting.

Trigger word

shiro_sama

Include shiro_sama in your positive prompt, followed by a description of the outfit, activity, setting, and visual style you want.

What this model does best

• Recognizable identity: Shiro’s reddish-pink eyes, long white hair, fox ears, short fluffy fox tail, and pink cheek markings.

• Outfit flexibility: Her canonical white-dress ensemble, jackets, casual clothing, beach outfits, and other wardrobe choices.

• Natural-language prompting: Supports descriptive prompts through Krea 2 Turbo’s Qwen3-VL-4B text encoder. Describe clothing, poses, expressions, lighting, and composition in clear sentences.

• Scene flexibility: Indoor and outdoor scenes, traditional and modern settings, and different times of day.

• Fast generation: Uses Krea 2 Turbo’s 8-step generation workflow for quick iteration.

Recommended generation settings — ComfyUI

Use an updated ComfyUI installation with native Krea 2 support.

• Base model: Krea 2 Turbo

• Text encoder: Qwen3-VL-4B

• Text encoder loader type: krea2

• VAE: Qwen Image VAE

• Trigger word: shiro_sama

• Prompt style: Descriptive natural language

• LoRA weight: Start at 1.0; adjust within 0.85–1.0

• Sampler: Euler (euler)

• Scheduler: Simple (simple)

• Steps: 8

• CFG scale: 1.0 in ComfyUI

• Denoise: 1.0 for txt2img

• Resolution: Start at 1024 × 1024; Turbo also supports generation up to 2048 × 2048

• Negative prompt: Leave empty at the recommended CFG

Guidance setting: ComfyUI’s CFG 1.0 disables additional classifier-free guidance. In Krea’s official inference code or Diffusers, use CFG 0.0 / guidance_scale=0.0 for the equivalent setting.

For img2img, adjust denoise according to how much of the source image you want to preserve.

Adjusting LoRA strength

Start at 1.0 and adjust in small increments.

• Try 0.85–1.0 to balance character identity with outfit and style flexibility.

• Lower the weight if character details overpower the requested clothing or composition.

• Try 1.1–1.2 for stronger character influence, checking for exaggerated features or reduced flexibility.

Prompting

Use descriptive natural language. Start with shiro_sama, then explain the image in clear sentences.

• Describe her clothing and expression explicitly.

• Specify her pose or activity, including what her hands are doing.

• Include the setting, camera angle, framing, and lighting.

• State the intended style, such as anime illustration, clean linework, or soft cel shading.

• Keep instructions consistent and prioritize the details that matter most.

For dynamic scenes, describe the direction of movement and how her body, hair, clothing, and tail respond. For precise compositions, explain her position relative to the camera and any foreground objects.

Default outfit and clothing changes

Shiro’s default outfit is her white strapless dress and red shawl ensemble: a white dress with side slits and a gold crescent motif on the bodice, a deep-red shawl draped around her upper arms, black elbow-length gloves, black thigh-high stockings, and a broad black choker with a red diamond-shaped ornament.

For the default wrap, use red shawl rather than red scarf. A miko, kimono, or hakama outfit is an alternate costume.

If shiro_sama alone does not recover her default outfit, reinforce it with:

white dress, strapless dress, side slit, red shawl, black gloves, elbow gloves, black thighhighs, black choker

These are garment descriptions; shiro_sama remains the character trigger.

For alternate outfits, explicitly describe the replacement clothing. Retain descriptions only for the default pieces you want to keep.

When adding a jacket, omit strapless dress and describe the jacket directly. To keep the underlying dress and black accessories, retain white dress, side slit, black thighhighs, black choker, then add the jacket description.

Use short tail to reinforce the intended tail length. Avoid detailed eyes if it causes her black pupils to disappear.

Generation examples

The following examples were generated with Krea 2 Turbo v1.0, using 15 steps rather than the recommended 8-step starting configuration.

Shared settings: txt2img, LoRA weight 1.0, Euler (euler), CFG 1.0, and 1024 × 1024 resolution.

Example 1: Sunglasses above her forehead

Prompt

anime, shiro_sama on a sunny airport tarmac near Miami, caught mid-step as she begins pulling her suitcase toward a small passenger airplane. Full-body view at chest height, fluffy white fox tail just above her luggage, her body still facing the viewer, one foot stepping diagonally toward the boarding area and the opposite heel lifting, just before she turns. Her raised right hand holds slightly squared plum sunglasses by their outer corner just in front of her forehead, caught as she begins lowering them into place. The entire frame and both lenses remain above her eyebrows, leaving her eyes completely unobstructed. Her face points toward the viewer, both eyes making direct eye contact, with slightly lowered upper lids, relaxed eyebrows, a subtly raised chin, and a confident, playful smile. Her left hand pulls the extended suitcase handle, the case tilted onto its rear wheels and just beginning to roll behind her. She wears an ivory short-sleeved camp-collar blouse with narrow plum piping, loosely tucked into high-waisted sea-glass-teal ankle-length trousers, a slim tan belt, and ivory low-top sneakers with tan soles. Her hair tips and blouse hem lift slightly with her movement and a light breeze, her tail following behind without obscuring the luggage. The compact warm-yellow suitcase faces the viewer, displaying a cute curled-up fox sticker, a smaller blue wave sticker, and a coral travel stamp. Behind her stands a well-kept pearl-white twin-engine propeller passenger plane with a navy-and-teal stripe, oval cabin windows, stationary propellers, and short boarding stairs. A few distant passengers approach and ascend the steps with their carry-ons. Palm trees stand off to the left beside a low terminal, beneath a vivid blue sky with scattered white clouds. Upper-left sunlight peeks through palm fronds in a small sunburst, casting warm highlights across her sunglasses, shoulder, and suitcase, with soft blue-violet shadows on the pale tarmac. Keep her eyes and smile clearly visible, the background gently softened, and the sunny atmosphere warm and breezy without harsh glare.

Workflow: txt2img. Checkpoint: Krea 2 Turbo v1.0. LoRA weight: 1.0. Steps: 15. Sampler: Euler (euler). CFG scale: 1.0. Seed: 813028192. Size: 1024 × 1024.

Example 2: Looking above the sunglasses

Prompt

anime, shiro_sama on a sunny airport tarmac near Miami, caught mid-step as she begins pulling her suitcase toward a small passenger airplane. Full-body view at chest height, fluffy white fox tail just above her luggage, her body still facing the viewer, one foot stepping diagonally toward the boarding area and the opposite heel lifting, just before she turns. Her right hand guides slightly squared plum sunglasses onto her nose, holding the outer frame corner with the lenses still below her eyes, looking directly at the viewer above the rims. Her chin is slightly raised, eyebrows relaxed, and mouth curved into a confident, slightly asymmetric smile. Her left hand pulls the extended suitcase handle, the case tilted onto its rear wheels and just beginning to roll behind her. She wears an ivory short-sleeved camp-collar blouse with narrow plum piping, loosely tucked into high-waisted sea-glass-teal ankle-length trousers, a slim tan belt, and ivory low-top sneakers with tan soles. Her hair tips and blouse hem lift slightly with her movement and a light breeze, her tail following behind without obscuring the luggage. The compact warm-yellow suitcase faces the viewer, displaying a cute curled-up fox sticker, a smaller blue wave sticker, and a coral travel stamp. Behind her stands a well-kept pearl-white twin-engine propeller passenger plane with a navy-and-teal stripe, oval cabin windows, stationary propellers, and short boarding stairs. A few distant passengers approach and ascend the steps with their carry-ons. Palm trees stand off to the left beside a low terminal, beneath a vivid blue sky with scattered white clouds. Upper-left sunlight peeks through palm fronds in a small sunburst, casting warm highlights across her sunglasses, shoulder, and suitcase, with soft blue-violet shadows on the pale tarmac. Keep her eyes and smile clearly visible, the background gently softened, and the sunny atmosphere warm and breezy without harsh glare.

Workflow: txt2img. Checkpoint: Krea 2 Turbo v1.0. LoRA weight: 1.0. Steps: 15. Sampler: Euler (euler). CFG scale: 1.0. Seed: 915615285. Size: 1024 × 1024.

Shared negative prompt

Both examples contain the following negative prompt in their Civitai generation metadata:

tears, crying, cry, sweat, rain, worst quality, low quality, lowres, jpeg artifacts, bad anatomy, bad hands, extra fingers, missing fingers, fused fingers, extra arms, extra legs, disconnected limbs, multiple tails, detached tail, cat ears, human ears, photorealistic, 3d, cgi, realistic skin, painterly, gradient shading, monochrome, muted colors, underexposed, blurry face, multiple girls, crowd, child, loli, petite, text, watermark, signature, long tail, belt

For the recommended ComfyUI CFG 1.0 workflow, leave the negative prompt empty because additional classifier-free guidance is disabled. The text above is retained as a record of the examples’ submitted generation settings.

Dataset and caption curation

The dataset contains 598 tagged synthetic images generated with NovelAI 4.5 Curated. Caption curation combined PixAI auto-tagging with manual review, custom tooling, and AI-assisted policy refinement. The tagging dictionary and baseline rules were carried over from my previous Senko-san LoRA projects, then revised and refined for Shiro.

Tag curation used a customized version of toshiaki1729’s dataset editor, modified to support AI-assisted dataset review and tag management. The editor applied the active policy to the raw auto-tagger output and produced curated captions.

AI-assisted review surfaced edge cases, normalized inconsistent terminology, and classified tags as Keep, Reject, or Unknown. The policy was refined and reapplied across the dataset to reduce caption noise and concept bleed.