Each image generate by dall-e is new right?

If each image generated is new, then when I ask to dall-e to make a fractal, it’ll make it based on, somehow a mathematical formula? Or is the fractal, not really a fractal?
I don’t know if it makes enough sense, or if it’s interesting at all eheh

1 Like

Words are connected to visual patterns via embeddings. In other words, DALLE2 knows what the concept of a fractal “looks” like in an abstract sense.

1 Like

There’s some artifacts on the fractals so it’s probably not mathematically generating a fractal, but rather generating a picture that looks like a fractal.

1 Like