CivArchive
    WKB-Krea2-Illustrious - Real_v3.0
    NSFW
    Preview 120279253
    Preview 120356916
    Preview 120356714
    Preview 120357138
    Preview 120356692
    Preview 120279259
    Preview 120279235
    Preview 120356877
    Preview 120356931
    Preview 120357600
    Preview 120377121
    Preview 120377261
    Preview 120377427
    Preview 120377485
    Preview 120377716

    [WKB-Krea2-Booru-v1.0]

    A Krea2 checkpoint trained on booru tags, built to understand both natural language and tag-style prompts.

    ๐Ÿš€ Early Access News

    This model will be in generation-only mode until 31st Dec. Alternatively you can download this model from my Ko-fi store โค๏ธ(https://ko-fi.com/s/e6926a9648).

    โœจ About This Model

    WKB-Krea2-Booru is a Krea2-based checkpoint fine-tuned on booru-tagged data. Write prompts the way you like: plain sentences, booru tags, or a mix of both.

    Highlights

    • ๐Ÿ“ท Photorealism: natural skin, lighting, and detail

    • ๐ŸŽจ Anime and illustration: clean, expressive results with strong style range

    • ๐Ÿง Pose control: responds reliably to pose and composition tags

    • ๐Ÿ”ž NSFW capable: handles mature content when prompted

    • ๐Ÿ‘ค Character LoRA support: stays consistent with character LoRAs

    • ๐Ÿ–Œ๏ธ Style LoRA support: works well with style LoRAs without losing coherence

    • ๐Ÿ’ฌ Dual prompting: natural language, booru tags, or both

    ๐Ÿ“ Prompting Guide

    Booru style:

    1girl, solo, long hair, looking at viewer, outdoors, sunset, cinematic lighting, masterpiece, best quality

    Natural language:

    A young woman with long hair standing on a beach at sunset, soft golden light, photorealistic, shallow depth of field.

    Hybrid (recommended):

    1girl, solo, standing on a beach at sunset, wind blowing her hair, photorealistic, detailed skin texture, cinematic lighting

    ๐Ÿ’ก Tip: Booru tags control subject, pose, and attributes precisely. Natural language is good for scene, mood, and lighting. Combining them gives the best results.

    ๐Ÿ”Œ LoRA Compatibility

    • Character LoRAs work well at standard weights (0.6 โ€“ 1.0)

    • Style LoRAs blend cleanly; lower the weight if the style overpowers the base

    • Stacking multiple LoRAs is supported; reduce individual weights as you add more

    โš ๏ธ Notes

    • This model can generate NSFW content. Use responsibly and follow Civitaiโ€™s content rules.

    • Please donโ€™t use it to depict real people without consent.

    • Results vary with prompt, sampler, and LoRA combination.

    โค๏ธ Support

    If you enjoy the model, consider supporting me on Ko-fi. Please also share your generations in the gallery and leave a review. Your feedback helps shape future versions!

    [WKB-Illustrious-v3.0] - Real Style

    Follow my workflow(included in Images) to get the best results.

    • Sampler: DPM++2M + Karras

    • Steps: 25-30

    • CFG Scale: 4

    • Clip Skip: 2

    • VAE: No Need

    • HiRes-Fix(Recommended): 10 + 10 steps (Dual Sampler) with 4x-Ultrasharp.pth or whatever you prefer and denoise of 0.4-0.6.

    Prompting Tip V3.0:

    Positive: masterpiece, best quality, amazing quality, 4k, very aesthetic, high resolution, ultra-detailed, absurdres, newest, scenery, realistic {prompt} BREAK, detailed background, blurry foreground, bokeh, depth of field, volumetric lighting

    Negative: modern, recent, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra fingers, extra digits, fewer digits, cropped, very displeasing, (worst quality, bad quality:1.2), sketch, jpeg artifacts, signature, watermark, username, (censored, bar_censor, mosaic_censor:1.2), simple background, conjoined, bad ai-generated

    [WKB-Illustrious-v2.0] - Pixar Style

    This version improves on the character anatomy and adds a Pixar style. Create better eyes, better faces even without upscaling or hires-fix.

    • Sampler: Euler A + Karras

    • Steps: 35โ€“40

    • CFG Scale: 5

    • Clip Skip: 2

    • VAE: No Need

    • HiRes-Fix(Recommended): 15 steps with 4x-Ultrasharp.pth or whatever you prefer and denoise of 0.4-0.6.

    Prompting Tip V2.0:

    Positive: HDR, 8K, masterpiece, best quality, detailed skin, detailed hair, Smooth Quality, {Your Prompt}

    Negative: missing fingers, extra fingers, poor resolution, poor quality, poor detail, poor anatomy, poor eyes, poor hands, watermark, signature, artist's name, lineart, anime,

    [WKB-Illustrious-v1.0] - Disney Style

    This is a fine-tuned checkpoint designed to replicate the signature 3D animation aesthetic of modern Disney and Pixar films. It excels at creating high-fidelity characters and environments with soft lighting, expressive features, and tactile textures.

    Key Features:

    • Stylized Realism: Balances cinematic lighting with exaggerated, appealing character proportions.

    • Texture Depth: Renders complex surfaces like subsurface scattering on skin, fluffy fur, and realistic fabric weaves.

    • Vibrant Color Science: Tuned for the saturated yet harmonious palettes typical of big-budget animation.

    • Versatility: Works seamlessly for portraits, wide-angle cinematic shots, and whimsical creature design.

    • Sampler: DPM++ 2M Exponential or Euler a Karras

    • Steps: 25โ€“30

    • CFG Scale: 4-7

    • Clip Skip: 2

    • VAE: Baked-In (or "sdxl.vae" if you want)

    • HiRes-Fix(Recommended): 12-15 steps with 4x-Ultrasharp.pth and denoise of 0.4-0.5


    Prompting Tip V1.0:

    Positive: 3d render, octane render, cinematic lighting, masterpiece, disney animation, subsurface scattering, highly detailed, BREAK, {Prompt}, volumetric lighting, detailed background, living room background, skindentation, night, from angle,

    Negative: bad quality, worst quality, worst detail, sketch, censor, anime, simple background,

    Description

    Realist Model. Follow my provided workflow for gorgeous and detailed results.

    FAQ

    Checkpoint
    Illustrious

    Details

    Downloads
    1,217
    Platform
    CivitAI
    Platform Status
    Available
    Created
    2/7/2026
    Updated
    11/22/2026
    Deleted
    -

    Files

    WKB-Real-v01.safetensors

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.