Urusan on Nostr: What if the main reason SD1.5 is more chaotic and "vibey" is simply that the ...
What if the main reason SD1.5 is more chaotic and "vibey" is simply that the resolution of the input noise is lower?
I was thinking about this because I realized that the denoising balance between the latent input image and the prompt determine whether the prompt or image are more important. Noise reduces how much information is in the input image.
While pure noise is maximally low influence, larger txt2img generations will have a lower randomness because there's less chance of variation.
I was thinking about this because I realized that the denoising balance between the latent input image and the prompt determine whether the prompt or image are more important. Noise reduces how much information is in the input image.
While pure noise is maximally low influence, larger txt2img generations will have a lower randomness because there's less chance of variation.