How AI image generators make a picture from your words
Latent space: what an AI image generator really works on
A picture of 512 by 512 pixels has more than 260,000 pixels. Cleaning every pixel at each of 50 denoising steps would be a lot of work. In fact, the image generator does not clean the pixels of your picture. So what does it clean?
-
1
Look at your fox up close: it is made of tiny squares of color called pixels, 512 across and 512 down.
-
2
The square the diffusion model cleans is nowhere near that big. Still, your fox comes out with every pixel in place, sharp down to the snowflakes.
The rest of this episode is for subscribers
Subscribe to read every premium episode. You can cancel when you want.
Paid yearly: $99 instead of $119.88