//=time() ?>
you know this means I had to feed these models some typography
having fun feeding diffusion models thresholded patterns as init images. also really interesting to see which init images work and which ones don’t
Hey everyone bugging me to do a Disco Diffusion tutorial for animation...
can’t believe I actually did this audioreactive part in touchdesigner. prob could have done it faster in python and opencv. (warning, lots of fast moving/flickering)
fish update: better fidelity overall, still getting lots of jumps