CivArchive
    K-Pop Poses - v1.0 Qwen
    NSFW
    Preview 109433010
    Preview 109433018
    Preview 109433025
    Preview 109433048
    Preview 109433061

    K-Pop Poses

    Trigger keyword: KPopPosesDaal

    Bring dynamic, idol-style choreography into your generations!
    K-Pop Poses captures the signature look and energy of multi-member idol groups: expressive stances, synchronized dance lines, confident formations, and that polished, editorial K-pop aesthetic.
    Unlike most LoRAs that focus on single-subject posing, this one is designed to create multi-character compositions perfect for generating trios, quartets, or full stage lineups.

    This makes it both powerful and a little chaotic: it handles complex body placement surprisingly well, but when it misses… it really misses. Expect the occasional extra limb, misplaced face, or avant-garde “experimental anatomy.” Consider it part of the charm. 😅


    Model Behavior

    • Available for Flux, Qwen, and Illustrious.

    • Recommended strength:

      • Flux & Qwen: 1.0 for clean adherence; 0.5–1.5 for variation.

      • Illustrious: slightly softer around 0.8, depending on checkpoint.

    • Lower strengths = subtle integration and natural rendering.

    • Higher strengths = stylized, pop-art, sometimes surreal results.


    Use Cases

    • Group dance scenes

    • Fashion editorials or music-video stills

    • Pop idols, cosplay troupes, or digital stage shoots

    • Concept art for choreography or performance design


    ⚠️ Caution (with a smile)

    Due to the multiple characters per frame, the chance of “Cronenberg Monster Idol” moments is higher than usual. The LoRA cleans it up about 80% of the time — the rest are nightmare fuel or happy accidents. Proceed with humor and curiosity.


    Show Support

    If you enjoy this LoRA:
    Give it a thumbs-up on CivitAI!
    Post your first image to the model page — it keeps the buzz alive and helps me decide whether to create K-Pop Poses v2.
    Your feedback and uploads are the best way to inspire more dance-floor chaos and refinement.

    Description

    Qwen does well with photorealistic, non, anime and others. Less monstrosities. Occasionally you will need a face detailer which is not normal for Qwen but it seems to fuzz at random times.

    FAQ

    Comments (5)

    egpieperNov 12, 2025· 1 reaction
    CivitAI

    In Flux for only the dynamic poses and removing the illustrious cartoon style I used strength 0.5 on the K-Pop poses. Works pretty well! Good job!

    Daalis
    Author
    Nov 13, 2025· 1 reaction

    Very glad you found it useful and found settings that work for you. Thanks for sharing the tips.

    RisingVNov 14, 2025
    CivitAI

    This should be category concept instead of poses.

    Daalis
    Author
    Nov 15, 2025

    I was going for poses. Though concept or poses would fit. Unfortunately, you can only pick one and I felt poses was closest since concept implies more additional style, topics, etc. (At least to me ☺️ )

    RisingVNov 16, 2025

    Hey, thanks for responding. Actually in terms of loras anything is a concept except for styles (or characters) according to this guide and that guide. That means any lora that does not fit another category should be categorized as concept. The pose category is for specific poses like this one or that one. While your lora is for grouping people in a certain way, as much as I can tell, it does not show them in a specific pose. Therefore it should be in the conecpt category.

    LORA
    Qwen

    Details

    Downloads
    253
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/8/2025
    Updated
    5/18/2026
    Deleted
    -
    Trigger Words:
    KPopPosesDaal