Everybody should learn how it works and how to use it, so as to understand its limitations.
I had stable diffusion for a time before deleting it for space on my drive. You need to generate 10 images to find a good seed, and then tweek the prompt to get the pic you want. Lastly, you might need to touch-up some details.
It’s like rolling a dice until you get the number you want.
Everybody should learn how it works and how to use it, so as to understand its limitations.
I had stable diffusion for a time before deleting it for space on my drive. You need to generate 10 images to find a good seed, and then tweek the prompt to get the pic you want. Lastly, you might need to touch-up some details.
It’s like rolling a dice until you get the number you want.
Newer models are much better at this, stable diffusion 1.5 has crazy high variance.