• 520@kbin.socialOP
      link
      fedilink
      arrow-up
      2
      ·
      8 months ago

      I had to put “more than 5 fingers” as a negative prompt! stable diffusion is a lot less…assuming of what you want compared to online AIs

      • WhatsHerBucket@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        8 months ago

        I need to get better about my prompts. I always have an idea of what I want and it never seems to come out that way lol

        • 520@kbin.socialOP
          link
          fedilink
          arrow-up
          2
          ·
          edit-2
          8 months ago

          Sometimes the model just doesn’t have the data to do what you want. I tried to make a cute anime drawing of my gf and her dog, but her dog is a rare local breed, with zero depictions in anime. It rarely ever got close.

          Other than that, it’s all about being specific to a massive degree and then hit the retry button a lot lol

      • Ziggurat@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        1
        ·
        8 months ago

        In general my negative prompts include

        nsfw, extra fingers, extra limbs, low quality

        This would resolve a lot of issue.

  • MajorHavoc@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    ·
    8 months ago

    Whatever that is, I can probably still trust it and respect it more than the pathetic sniveling murdering worm and LOTR character that it’s based on.

    • 520@kbin.socialOP
      link
      fedilink
      arrow-up
      3
      ·
      8 months ago

      When it starts referring to other countries as ‘my precious’, then you gotta worry!

  • tal@lemmy.today
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    8 months ago

    If you have a shot and you want two people in it, Stable Diffusion will tend to merge them; I don’t know if your original goal was a “bride and groom” shot, based on the “married to” prompt.

    If you install the Regional Prompter extension, you can tend to do better with two people, because you can have part of the prompt apply to only part of the image. I had an example up yesterday:

    https://lemmy.today/post/2916131?scrollToComments=true

    Using the same approach here:

    bride and groom ADDCOMM putin ADDCOL gollum

    Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 0, Size: 1024x1024, Model hash: ebf42d1fae, Model: realmixXL_v15, Token merging ratio: 0.5, RP Active: True, RP Divide mode: Matrix, RP Matrix submode: Horizontal, RP Mask submode: Mask, RP Prompt submode: Prompt, RP Calc Mode: Attention, RP Ratios: “1,1”, RP Base Ratios: 0.2, RP Use Base: False, RP Use Common: True, RP Use Ncommon: False, RP Change AND: False, RP LoRA Neg Te Ratios: 0, RP LoRA Neg U Ratios: 0, RP threshold: 0.4, RP LoRA Stop Step: 0, RP LoRA Hires Stop Step: 0, Version: v1.5.1

    Or switching roles:

    bride and groom ADDCOMM gollum ADDCOL putin

    Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 0, Size: 1024x1024, Model hash: ebf42d1fae, Model: realmixXL_v15, Token merging ratio: 0.5, RP Active: True, RP Divide mode: Matrix, RP Matrix submode: Horizontal, RP Mask submode: Mask, RP Prompt submode: Prompt, RP Calc Mode: Attention, RP Ratios: “1,1”, RP Base Ratios: 0.2, RP Use Base: False, RP Use Common: True, RP Use Ncommon: False, RP Change AND: False, RP LoRA Neg Te Ratios: 0, RP LoRA Neg U Ratios: 0, RP threshold: 0.4, RP LoRA Stop Step: 0, RP LoRA Hires Stop Step: 0, Version: v1.5.1