CandyLaser
Adventurer
Found an interesting edge case where we can see some of the limitations of the model in understanding the prompts. The prompt in this case was: "a giant pink fairy armadillo being ridden by a gnome down a forest path, antique children's book illustration." Pink fairy armadillos are a real thing; this is what they look like:
But DALL-E reads "pink fairy" as adjectives modifying armadillo, giving images like these:
They're not bad, but they're not what I was going for.
But DALL-E reads "pink fairy" as adjectives modifying armadillo, giving images like these:
They're not bad, but they're not what I was going for.