CivArchive
    AIArtist-CADS2.0 XL Commercial Advertising Assistance System - Exploded view
    Preview 1
    Preview 2
    Preview 3
    Preview 4
    Preview 5
    Preview 6
    Preview 7
    Preview 8
    Preview 9
    Preview 10
    Preview 11
    Preview 12
    Preview 13
    Preview 14
    Preview 15
    Preview 16
    Preview 17
    Preview 18

    Description

    Miracle-Explodedview Industrial Design Assisted LORA Model XL Version


    File name: Miracle-Explodedview.safetensors or Miracle-Explodedview_v20240103.safetensors


    New pit, testing with SDXL+ industrial design, generating exploded detail display of products (auxiliary composition).

    The current version is still in testing. Common components have been defined, combined with XL's powerful composition ability,

    simulating the explosive form of industrial design products, often used to output materials for recombination,

    not a one-time forming, but enough for advertising images. Planting a seed, looking forward to seeing more ways to play.


    Trigger words: Explodedview, reinforced effect words reflection, 3d, shadow, science fiction


    Style words: simple background, white background, black background

                  gradient background, grey background defined above


    Common elements include but are not limited to: analog clock, 3d, magic, circle, wheel, speaker, circuit board

                              battery, wristwatch design, handheld devices, instrument

                              electric drill, vehicle, car, motorbike, instrument, glass panel,

                              camera lens, screw, Electronic devices etc., decompose or explode any object


    Usually recommended, not limited to element words, with the powerful understanding ability of XL, natural language description is sufficient.


    Recommended LORA weight: 0.65, can transform any decomposition form between 0.5-0.8 XYZ view does not accurately represent


    Recommended sampling steps: 25-50   CFG scale: 5-8 Reducing the ability to zoom in on the base model, increasing the tendency toward LORA content


    Recommended sampling method: for realistic models, DPM++ 2M Karras, for drawing models, Euler a

    Feel free to try other sampling methods, solutions like LCM Turbo have not been tested,

    FP8 reasoning can produce images normally, with a VRAM usage of only 6G and no overflow.


    Recommended resolution: any resolution


    For image enhancement, recommended setting: 4x-UltraSharp


    For other questions, please leave a message. For collaborative learning about AIGC, please join the group for discussion.

    ————————————————————————————————————————————————


    For test questions, please leave a message, for business cooperation, check the personal homepage  +V Zyuan980

    AiARTiST | Metaverse China·Shandong Digital Human/AiGC/Metaverse Direction


    ————————————————————————————————————————

    FAQ

    Details

    Downloads
    19,069
    Platform
    ShakkerAI
    Platform Status
    Available
    Created
    1/3/2024
    Updated
    3/14/2025
    Deleted
    -
    Trigger Words:
    Explodedview

    Files