CivArchive
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined

    This model replaces the previous one, Model Link, delivering the ultimate 'Asian Beauty' in a Semi-Realistic style.


    Origin: A Merge for Beautiful Faces with a Semi-Realistic Goal

    This model is the outcome of a meticulous merging operation dedicated to actualizing the most beautiful Asian female face in a semi-realistic aesthetic. Instead of encompassing diverse facial types, it was born from a journey of selectively extracting and blending the characteristics of the 'most ideal and beautiful face'.

    The result goes beyond mere diversity, focusing on a captivating semi-realistic beauty that immediately draws the eye.


    Prompt Strategy for Optimal Results

    The greatest strength and defining characteristic of this model is that it yields the best results when Positive and Negative prompts are minimized.

    • โœจ Minimal Intervention: The model itself has a powerful, internal definition of a 'beautiful semi-realistic face'. Complex or lengthy prompts may actually interfere with the model's aesthetic judgment.

    • ๐Ÿ–ผ๏ธ Extraction of Pure Beauty: Use only the necessary keywords and reduce the input of Positive and Negative prompts. The model's inherent vision of the most beautiful Asian semi-realistic face will emerge in clear, high-quality images.


    ์ด ๋ชจ๋ธ์€ ๊ธฐ์กด์˜ ๋ชจ๋ธ ๋งํฌ์„ ๋Œ€์ฒดํ•˜๋ฉฐ, ๋ฐ˜์‹ค์‚ฌ(Semi-Realistic) ์Šคํƒ€์ผ๋กœ ๊ถ๊ทน์˜ '์•„์‹œ์•„ ๋ฏธ(็พŽ)'๋ฅผ ์„ ์‚ฌํ•ฉ๋‹ˆ๋‹ค.


    ํƒ„์ƒ ๋ฐฐ๊ฒฝ: ์˜ค์ง '์•„๋ฆ„๋‹ค์šด ์–ผ๊ตด'์„ ์œ„ํ•œ ๋จธ์ง€(Merge)์™€ ๋ฐ˜์‹ค์‚ฌ์˜ ๋งŒ๋‚จ

    ์ด ๋ชจ๋ธ์€ ์ตœ๊ณ ์˜ ์•„๋ฆ„๋‹ค์šด ์•„์‹œ์•„ ์—ฌ์„ฑ ์–ผ๊ตด์„ ๋ฐ˜์‹ค์‚ฌ ์Šคํƒ€์ผ๋กœ ๊ตฌํ˜„ํ•˜๊ธฐ ์œ„ํ•œ ์ •๊ตํ•œ ๋จธ์ง€(Merge) ์ž‘์—…์˜ ์‚ฐ์ถœ๋ฌผ์ž…๋‹ˆ๋‹ค. ๋‹ค์–‘ํ•œ ์–ผ๊ตดํ˜•์„ ํฌ๊ด„ํ•˜๊ธฐ๋ณด๋‹ค๋Š”, '๊ฐ€์žฅ ์ด์ƒ์ ์ด๊ณ  ์˜ˆ์œ ์–ผ๊ตด'์˜ ํŠน์ง•๋งŒ์„ ์„ ๋ณ„์ ์œผ๋กœ ์ถ”์ถœํ•˜๊ณ  ์œตํ•ฉํ•˜๋Š” ์—ฌ์ • ์†์—์„œ ํƒ„์ƒํ–ˆ์Šต๋‹ˆ๋‹ค.

    ์ด ๊ฒฐ๊ณผ๋ฌผ์€ ๋‹จ์ˆœํ•œ ๋‹ค์–‘์„ฑ์„ ๋„˜์–ด, ๋ณด๋Š” ์ด์˜ ์‹œ์„ ์„ ์‚ฌ๋กœ์žก๋Š” ๋งคํ˜น์ ์ธ ๋ฐ˜์‹ค์‚ฌ์  ์•„๋ฆ„๋‹ค์›€์— ์ดˆ์ ์„ ๋งž์ถ”๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.


    ์ตœ์ ์˜ ๊ฒฐ๊ณผ๋ฅผ ์œ„ํ•œ ํ”„๋กฌํ”„ํŠธ ์ „๋žต

    ์ด ๋ชจ๋ธ์˜ ๊ฐ€์žฅ ํฐ ํŠน์ง•์ด์ž ๊ฐ•์ ์€ '๊ธ์ •๊ณผ ๋ถ€์ •์„ ์ตœ์†Œํ™”ํ–ˆ์„ ๋•Œ ์ตœ์ƒ์˜ ๊ฒฐ๊ณผ'๋ฅผ ๋ณด์—ฌ์ค€๋‹ค๋Š” ์ ์ž…๋‹ˆ๋‹ค.

    • โœจ ์ตœ์†Œํ•œ์˜ ๊ฐœ์ž…: ๋ชจ๋ธ ์ž์ฒด์— '์˜ˆ์œ ๋ฐ˜์‹ค์‚ฌ ์–ผ๊ตด'์— ๋Œ€ํ•œ ์ •์˜๊ฐ€ ๊ฐ•๋ ฅํ•˜๊ฒŒ ๋‚ด์žฌ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. ๋ณต์žกํ•˜๊ฑฐ๋‚˜ ๊ธด ํ”„๋กฌํ”„ํŠธ๋Š” ์˜คํžˆ๋ ค ๋ชจ๋ธ์˜ ๋ฏธ์  ํŒ๋‹จ์„ ๋ฐฉํ•ดํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

    • ๐Ÿ–ผ๏ธ ์ˆœ์ˆ˜ํ•œ ๋ฏธ(็พŽ)์˜ ์ถ”์ถœ: ํ•„์š”ํ•œ ํ‚ค์›Œ๋“œ๋งŒ ์ตœ์†Œํ•œ์œผ๋กœ ์‚ฌ์šฉํ•˜๊ณ , ๊ธ์ •(Positive) ๋ฐ ๋ถ€์ •(Negative) ํ”„๋กฌํ”„ํŠธ์˜ ๊ฐœ์ž…์„ ์ค„์ด์„ธ์š”. ๋ชจ๋ธ์ด ์ž์ฒด์ ์œผ๋กœ ๊ตฌํ˜„ํ•˜๋Š” ๊ฐ€์žฅ ์˜ˆ์œ ์•„์‹œ์•„ ๋ฐ˜์‹ค์‚ฌ ์–ผ๊ตด์ด ์„ ๋ช…ํ•˜๊ณ  ๊นจ๋—ํ•œ ๊ทธ๋ฆผ์œผ๋กœ ๋‚˜ํƒ€๋‚  ๊ฒƒ์ž…๋‹ˆ๋‹ค.


    Description

    ๐Ÿ…’ Build 20260515a โ€“ 2.5D Allure & Deep Gaze

    • Leans closer to a 2.5D style, offering a perfect blend of high-end character art and realistic textures.

    • Features a calm yet seductive, deep gaze and expressive gestures that add narrative weight to the subject.

    • Lighting is atmospheric and moody, accentuating the depth of the eyes and the fluidity of the posture.

    • Feels like: sophisticated 2.5D realism ยท alluring aesthetic ยท expressive character gesture ยท premium stylized art.

    FAQ

    Comments (12)

    TianDi200May 19, 2026
    CivitAI

    Hello, will the "b" model be updated?

    innofree
    Author
    May 31, 2026

    Do you have a desired direction for the update?

    TianDi200Jun 2, 2026

    @innofreeย This version has exceptional aesthetic appeal, and its anatomical accuracy and detail rendering are also excellent. However, it seems that all the characters have the exact same face. While this face is very beautiful, it becomes aesthetically fatiguing after prolonged viewing. I think it would be beneficial to appropriately add other beautiful facial features, presenting a diverse range of stunning faces rather than just one single beautiful face.

    Ryan83May 27, 2026
    CivitAI

    i am frustrated anyone pls give me workflow for this model with vae and loras

    909059952May 28, 2026

    The base model of this model is Illustriou, and you can use sd-webui-forge-one, so you can use lora and vae without a workflow. I'm not very familiar with workflows, but quite a few workflows should support models that use Illustriou as the base model.

    Ryan83May 29, 2026

    @909059952ย thanks but i some how managed to use vae and loras

    innofree
    Author
    May 31, 2026ยท 1 reaction

    Have you tried opening the image in a new tab and downloading it? All images include a workflow. Also, the images I uploaded likely didn't use Lora. I only uploaded images that can be represented with a single checkpoint.

    Ryan83Jun 8, 2026

    @innofreeย i don't understand what you said and how works

    akringerzekrom656Jun 12, 2026ยท 1 reaction

    @Ryan83ย if you are using this model on ComfyUI then just drag and drop the images uploaded in this model onto the comfyUI and you will get the entire workflow. All images uploaded in this model by the creator contains meta data with in-built workflows. This method will work for ComfyUI

    Ryan83Jun 12, 2026

    @akringerzekrom656ย oh thanks

    770020801Jun 21, 2026
    CivitAI

    Is it effective to use your model to train LORA?

    ws928Jul 9, 2026

    fku

    Checkpoint
    Illustrious

    Details

    Downloads
    3,087
    Platform
    CivitAI
    Platform Status
    Available
    Created
    5/19/2026
    Updated
    8/21/2026
    Deleted
    -

    Files