vqgan+clip tests with some post processing // music: keyboard - “world view”
input: chimpanzee / windowsxp / hospital / funeral / shame / td bank