Inline images in GPTs between text

Prompting
, ,
1

I’m struggling to find a way to get GPT to generate text, then an image and then text again.
It seems impossible with prompting alone. So far, the best thing I got was an image at the beginning or at the end, but not in between.
Did anything work for you all?

2

Welcome to the community!

something like this?

image
image952×1752 170 KB

You can instruct the models to do things in sequence. What have you tried so far?

3 Likes
3

Thanks! Not sure why it’s not working. My use case is a bit different, but what I tried was:

Your answer has the following strucutre:

  • Text
  • Generate and wait for image 1
  • Text
  • Generate and wait for image 2
  • End

I guess the rest of my prompt throws it off then… I give style instructions and other information on the image to generate.

It just says stuff like:

Generating the first highlight image: a detailed acrylic painting of the sanctuary room filled with friends, food, and drinks, capturing a moment of joy amidst chaos.

or dalle generates image

but continues right with the text

4

Took a little coercing

image
image964×1768 159 KB

here’s my prompt:

image
image1005×1572 94 KB

as you can see it took a few attempts to get it going and it’s not perfect, but it’s a start, needs some more work :muscle: :laughing:

here's one of its failure modes:

image
image1168×532 24.6 KB

I suspect that with a little more work we might be able to get it to work without the tags :thinking:

do you think you can pick it up from here? would be amazing if you could share your results if you get it to work :smiley:

2 Likes
5

Thanks a lot!
I was not able to get around this limitation yet. I really have to go back to the drawing board on this one and start with the basics and then expand from there.
If I find something reliable, I’ll update

1 Like