Been speaking to @therealchefjeff for a few months about how to use text-to-image machine learning tools to generate art to pair with music. You imagine a visual to pair with the track, input parameters into an engine, and ask the computer to draw the image from "memory". The more abstract the prompt the more creative, but wanted to share the first one from the interlude to the new EP w/ the parameters we used 😌 Style data set: @artstationhq ---- sound: murakami daydream (snippet) text prompt: manhattan at night width: 500 length: 500 model: vqgan_imagenet_f16_16384 images_interval: 10 init image: N/A target images: N/A seed: 42 max_iterations: 1500