CivArchive
    Auditorium - v1.0
    Preview 1137405
    Preview 1137407
    Preview 1137406
    Preview 1137404

    中文简介在后方

    Base Model: RealisticVision_v20

    Tirrger tag is auditorium, Use prompts like view to stage, view to seat, side view, blue chairs, end stage/central stage/black box theatre, cool/warm/dark/light colors, dark colored flooring, stripe ceiling, hanging equipment, light, spotlight.... to guild this lora.

    Controlnet is allways reccomended if you wish to use it in real life design, much more efficiency in generating the right spacial structure, I personally have tested it and get some interesting feedback. tip: Use lineart or scrible model depend on how bad you drawing is, which should catch the basic assumption of the stage, stalls, balconies, and something that looks like chairs.

    The model is trained on 200+ images of built or unbuilt aditoriums.

    Please rate the model or post you image if you have tried this, really appreciate any feedback.

    底模选用 RealisticVision_v20,提示词见上方。

    使用 Controlnet 可以有效提高生成图片空间关系正确的概率,尤其当你是用来进行真实项目设计的时候。我试用过得到的反馈还是挺有意思的。tip:你的手稿应该要反映舞台、池座、楼座的大关系,座位可以用一些看起来像那么回事的线条表示。

    训练使用了两百多张观众厅的照片或渲染图,主要是照片。

    试用过的网友欢迎评论打分哟,期待你们的反馈。

    Description

    This version knows how to draw black box, but tends to add more fatastic effect when weight is high.

    FAQ

    Comments (5)

    saeedkhaliliDec 29, 2023
    CivitAI

    hi there its a good model thank you.. wanted to ask how many photos did you use?

    ZEZIA
    Author
    Jul 18, 2024

    Thanks for comment and sorry for the late reply, I used 10~70 images for each view angle (view to stage, view to seat, side view, black box...) in the first round, and about a half in the second round.

    saeedkhaliliAug 7, 2024

    @ZEZIA wow! I am new and learning about modeling. I wanted a model for a specific tank. how do you suggest me to do it ? just recording around it would sufice?

    ZEZIA
    Author
    Aug 29, 2024· 1 reaction

    @saeedkhalili I kind of foucs on modeling about space and architecture, so I am not sure my experience would apply. But as far as I know, it takes only about 10~20 images to model a person, and most people starts by simple feeding a few images to the algorithm before getting quiet good results.

    saeedkhaliliSep 25, 2024

    @ZEZIA thanks for replying, I once tried on an architecture model and failed there too, I wanted to model Persian mosques architecture.

    LORA
    Other

    Details

    Downloads
    569
    Platform
    CivitAI
    Platform Status
    Available
    Created
    6/14/2023
    Updated
    5/11/2026
    Deleted
    -
    Trigger Words:
    Auditorium

    Files

    Auditorium_v10.safetensors

    Mirrors

    CivitAI (1 mirrors)

    Available On (1 platform)

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