CivArchive
    IDK Hair Dynamics - v2.0
    NSFW
    Preview 136395090
    Preview 136395091
    Preview 136395092
    Preview 136395093
    Preview 136395096
    Preview 136395097
    Preview 136395098
    Preview 136395099
    Preview 136395100
    Preview 136395104

    # IDK Hair Dynamics V2

    ## About

    Adds realistic hair physics — wind-blown strands, wet/soaked hair, floating zero-gravity effects, and light halos. Turns static hair into something alive.

    PT-BR: LoRA focada no conceito 'Hair Dynamics' para modelos SDXL baseados em Illustrious. Reforca a geracao de floating hair, wet hair.

    ## Trigger words

    ### Primary triggers

    ```

    floating hair, wet hair

    ```

    ### Supporting tags

    Frequent thematic tags from the training set - combine them with the primary triggers to steer the concept:

    ```

    long hair, wet, black hair, brown hair, short hair, blonde hair, very long hair, photo (medium), realistic, nude, lips, water, photorealistic, nipples

    ```

    Also works with the V1 triggers: floating hair, zero gravity hair, windblown hair, soaked hair, wet hair, hair light halo

    ## Example prompts

    ```

    masterpiece, best quality, absurdres, dynamic hair in motion, windblown flowing hair, running through field, summer dress, cinematic, <lora:IDK_hair-dynamics_V2:0.9>

    ```

    ```

    masterpiece, best quality, absurdres, dynamic hair in motion, hair dramatically swirling around, passionate embrace, nude with hair as only covering, cinematic, <lora:IDK_hair-dynamics_V2:0.9>

    ```

    ## Preview generation settings

    | Setting | Value |

    | --- | --- |

    | Checkpoint | IDN_Illustrious_V10_T |

    | Size | 1280x720 + hires 1.5x RealESRGAN 4x+) = 1920x1080 |

    | Sampler | Euler a |

    | Steps | 28 (+14 hires) |

    | CFG scale | 5.5 |

    | Denoising strength | 0.4 |

    | Seeds | 1000-1009 |

    Negative prompt: worst quality, low quality, bad anatomy, bad hands, extra digits, fewer digits, missing fingers, jpeg artifacts, signature, watermark, username, blurry, lowres, bad proportions, deformed, mutated

    ## Recommended weight

    Use 0.8-1.1. This file is a V1+V2 merge (rank concatenation), which slightly dilutes each source - it tolerates and often benefits from weights a bit above 1.0.

    ## Changelog V2

    - New, larger and more diverse training dataset.

    - 1080p preview images.

    - Merged with the V1 LoRA, so V1 behavior is preserved.

    Base model: Illustrious SDXL IDN_Illustrious_V10_T)

    Training pipeline: [IDK-Trainer](https://github.com/NOTcisLol/IDK-Trainer)

    File: IDK_hair-dynamics_V2.safetensors

    Description

    FAQ

    LORA
    Illustrious

    Details

    Downloads
    108
    Platform
    CivitAI
    Platform Status
    Available
    Created
    7/11/2026
    Updated
    7/12/2026
    Deleted
    -

    Files

    IDK_hair-dynamics_V2.safetensors

    Mirrors