CivArchive
    Preview 2029756
    Preview 2029758
    Preview 2029765
    Preview 2029767
    Preview 2029768
    Preview 2029770
    Preview 2029771
    Preview 2029773
    Preview 2029774
    Preview 2029775
    Preview 2029776
    Preview 2029777
    Preview 2029778
    Preview 2029779
    Preview 2029780
    Preview 2029781
    Preview 2029786
    Preview 2029787

    model for asians

    Description

    1. t3xx3 (b09-ver3) 1st model out of 9 (3~4 to be released)

    2. t3 model integration + color correction

    3. Low-end PC models, high-end options available

    4. Dynamic thresholds available (recommended 2-5)

    5. See prompts and setting values in some pictures.

    FAQ

    Comments (16)

    franek111Aug 14, 2023
    CivitAI

    Can we expect Diffus for that model?

    darksagaq12926
    Author
    Aug 15, 2023· 1 reaction

    yes

    franek111Aug 15, 2023
    CivitAI

    Why when creating with that models (and a few other as well) i mostly get results where a person has additional parts of body or there are like 2 heads at the image. How to fix that?

    darksagaq12926
    Author
    Aug 15, 2023

    Please write a prom or post a picture below

    Blackened15Aug 15, 2023

    create your picture at 512 x 512 or somewhat close to that

    darksagaq12926
    Author
    Aug 16, 2023

    I'll make it in the afternoon. I'll post a photo after I check it out.

    darksagaq12926
    Author
    Aug 16, 2023

    It's strange. I seem to be doing well no matter what I do. Picture added.

    franek111Aug 16, 2023

    @darksagaq12926 What prompt and negative prompt do you use? And what other parameters do you use and how you set them? I tried with prompts and other parameters as in examples and got weird results. I tried mainly on 1080x1350

    darksagaq12926
    Author
    Aug 16, 2023

    Right-click on the picture I uploaded and save it to my pc, then upload the picture to png information in Stable Diffusion and click Send to txt2img to see all the proms I wrote. Every time I upload a model, I keep leaving it in about this version. You can refer to the prom in the picture

    Blackened15Aug 16, 2023

    @franek111 1080x1350 <-- That is your problem, SD1.5 model are not made for high resolution, start with half that, and then upscale it

    brokoliesAug 21, 2023· 1 reaction

    try 512 x768 and use hiRes fix set to 2

    guting123Apr 6, 2024

    @brokolies Negative prompts

    wlan001Aug 18, 2023· 4 reactions
    CivitAI

    I set it all according to your parameters, except for the positive keyword,Why is it that every image generated is the same looking girl?

    How to set it, in order to generate pictures are different looking girls?

    darksagaq12926
    Author
    Aug 18, 2023

    Change the seed to a random or frequently used one, or adjust the CFG scale value.

    alvinfengAug 22, 2023· 2 reactions
    CivitAI

    一直觉得这个模型的版本名字很难理解,很奇怪……

    darksagaq12926
    Author
    Aug 23, 2023

    t3 (3rd among test models) b09 (9th model among numerical adjustment a to d among the third) ver3 (3rd among weight adjustment ver1 to 3 among b09)

    Checkpoint
    SD 1.5

    Details

    Downloads
    846
    Platform
    CivitAI
    Platform Status
    Available
    Created
    8/14/2023
    Updated
    5/14/2026
    Deleted
    -

    Files

    t3_B09Ver31.safetensors

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)
    ShakkerAI (1 mirrors)

    Available On (1 platform)

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