AI Hairball - Stable Diffusion + ChatGPT

This is an example piece from Deeplizard’s Stable Diffusion for beginners (coming soon). The visuals were generated by Stable Diffusion, and the lyrics were generated by ChatGPT (with modifications by Deeplizard). In the course, the visuals from this piece are used to illustrate how images generated by diffusion models change when varying parameters, like the CFG scale and output resolution. 🧠 CHECK OUT OUR DEEP LEARNING COURSES: 🔗 https://deeplizard.com 🤖 Original ChatGPT prompt: “Make a song about an artificially intelligent hairball with many faces that was created by a diffusion model, changing the CFG scale, increasing the output resolution with the euler adaptive sampling method with 20 sample steps. The many faces are of women with blond hair and bangs in the style of a headshot magazine cover. Make it rhyme.” 🎵 Lyrics: Verse 1: An AI hairball with many faces, Built by a diffusion model, in digital spaces. Upped CFG scale, resolution too, Euler adaptive sampling, 20 steps will do. Chorus: Faces emerging from hairballs A sight to see, one to appall A journey from darkness to light AI hairballs, a chilling sight Verse 2: Each face unique, beautiful, and smart, Blond hair and bangs, a work of art. Headshot style, like a magazine cover, An AI hairball, like no other. Chorus: Faces emerging from hairballs A sight to see, one to appall A journey from darkness to light AI hairballs, a chilling sight Verse 3: An AI hairball, marvel to see, Created with care, so intricate and free. So many faces, all unique, A diffusion model and some technique. Chorus: Faces emerging from hairballs A sight to see, one to appall A journey from darkness to light AI hairballs, a chilling sight Outro: Euler adaptive sampling, 20 steps to take, An AI hairball, masterpiece to make. A sight to behold, it will astound, An AI hairball, truly profound. 👩 Meta-analysis - ChatGPT prompt: “Is it weird to write a song about an AI generated hairball?” 🤖 ChatGPT response: It is not