I love the content posted here. As someone that has recently started dabbling in AI image Gen, it would be an amazing learning tool to see a) which model was used to generate the image and b) which prompt was given. With a) being most interesting imo.

Is there any reason this is not required?

(hope it’s OK to post this here, didn’t see anything in the rules about it)

  • Thelsim@sh.itjust.worksMEnglish
    12·
    10 months ago

    As the sidebar says, it’s recommended to include your model but not a rule. I like to think this is first and foremost a community to share the images you’ve generated. I’d hate to scare off potential posters by putting up strict requirements.

    That said, of course I wholeheartedly endorse the sharing of models and prompts. For my own posts I like to put the model in the title and include the prompt in the body. It’s easy to do for Midjourney because the prompts are usually quite small, but I can imagine it can become quite a chore for some of the more complex workflows.
    Though I tend to skip prompts and models when it’s just a one-off comment in someone else’s post.

    On a side-note, this my 1000th comment! 🎉

    Midjourney prompt

    a tall fancy cake with “Thelsim 1000” spelled out in burning candles --v 6.0

  • Usernameblankface@lemmy.worldEnglish
    62·
    10 months ago

    I have noticed that including what model I used and at least the prompt if not the story of how I got the results I wanted gets a better response than throwing a picture out there with a title alone. I have a keyboard shortcut for “In Bing Image Creator, DAL-E 3 Prompt:” and I’ve gotten used to copy-pasting my prompt from Bing after that.

    So, I think that asking about the model and prompt if it’s not included boosts the community feeling here, but if it were required it would be frustrating and would chase people away.

    • Firipu@startrek.websiteOPEnglish
      1·
      10 months ago

      Yeah, I use that one a lot. But some of the more outlandish designs don’t appear to be from models from civetai. I’m mostly interested in those tbh. Making humans is easy :). Making cool non human things is more interesting of a challenge to me

    • Thelsim@sh.itjust.worksMEnglish
      31·
      10 months ago

      Something I tried before is to ask Chat-GPT for a quick summary of the conversation afterwards. It can create quite a nice overview of how you got to the image. Though that’s more of a workflow overview and not so much the final prompt.
      I think Chat-GPT attaches a generated prompt to the image itself?

      • Last@lemmy.worldEnglish
        1·
        10 months ago

        Good idea. I asked for one of these to include with my Community Challenge submission just now.

        They used to add a prompt to the filename, but I believe it was limited to 255 characters.