So in https://perchance.org/ai-text-to-image-generator (and other t2i-based image generators) you could choose an art style and then a few dropdowns come below the prompt description textbox allowing you to choose any additional effects (or built-in modifiers, I should say) to be added to the prompt.

Well, I’ve found a bug on these. When I choose one that has a dropdown menu (for example, “Professional Photo”), the dropdowns did come out, but when I reload the page, the option I chose earlier retained, however the dropdowns did not come out as if the default option was selected.

I also tested if this bug happened on other t2i-based AI image generators too, such as https://perchance.org/furry-ai and https://perchance.org/ai-image-lab, and it also did the same, so this is most likely a bug within the t2i framework.

  • VioneT@lemmy.worldMEnglish
    2·
    5 months ago

    Saw this while looking at the code:

    ...
// NOTE: triggering change/input events like this doesn't trigger addEventListener handlers - like those added via initHandlers, above, but
// for now this is a happy accident because it's not for modifiers to not show up until you've actually typed something - reduces UI
// complexity on first impression.
...

    So, I think it is not intended to not show the modifiers on load, but it gives a good UI.
    The modifiers only shows upon user input on the description after the reload.

    • TAPgiles@lemmy.worldEnglish
      2·
      5 months ago

      Though the situation is that the user has already used the generator a bit, and are coming back to the page–so such “simplification” shouldn’t be involved here, I think.

      (Also personally I don’t buy the whole “reduce complexity on first load” = good UX idea. But that’s a different topic.)

      • allo@lemmy.worldEnglish
        1·
        4 months ago

        lol. well there tho definintely can be too much complexity on first load (see initial forms of both my and adcomfortable’s generators which lost like 50%+ of new users via sheer initial appearance of complexity). But ya. There should be a check that the user is returning, possibly onload check if text in textarea, and then show necessary stuff if so. no philosophical excuse not to rlly

        • TAPgiles@lemmy.worldEnglish
          2·
          4 months ago

          having looked at the t2i code, it looks like it does actually track how many images have been generated by the user, and reveal more when they’ve generated more. So if you come back, it should show it all again anyway. Seems like a bug with it anyway.

          I’m working on my own image-gen page, and just putting the advanced stuff behind an easily togglable section. Job done. No gatekeeping, no preventing advanced users using advanced features 👍

    • BluePowerOPEnglish
      1·
      5 months ago

      Oh yep, I didn’t ever thought of that before.