Can't AI recognize the image it created?
I use Chat gpt with dall-e, and it didn't seem to work. I had it create five coloring pages of dwarves. Then I wrote: "Please, the last one, with thinner lines." Then it created a dwarf with thinner lines, but it was a different one.
Apparently, Chat gpt can't replicate an already created image, and for example, can't recognize anything in it? So, for example, "The same dwarf, but with a red hat instead of a blue one" probably doesn't work either?
(1 votes)
Loading...
No, that’s not possible. ChatGPT simply uses the same prompt, so the image will look different every time. To edit images with Prompts, you need an image editing program with AI functions.
This is simply not provided. The instructions
a) by text and not by image
b) after trained patterns.
Try to put creatures together from two others, or to generate an ugly, fat fairy, or to put a magician on a cylinder hat instead of a pointed hat!
The one with the ugly fat fee I have just tried, and it has worked very well 🤣🥲 I am traumatized 😵 💫
Then why are they getting thin and cute with me?
Well, if you show her.
Download the picture and then upload or what do you mean?
Like that, Google Lens.