21
GPT-4 makes way better cat pizza hybrids than Stable Diffusion does
I tried the same prompt on both models last night. I asked for "a golden retriever driving a tiny fire engine." GPT-4 gave me a dog with actual paws on a steering wheel and a little helmet. Stable Diffusion gave me a blob with four legs and a red smudge. I ran it 3 times each and GPT-4 won every time. Has anyone else noticed a big quality gap between these two for weird mashup stuff?
2 comments
Log in to join the discussion
Log In2 Comments
taylor.amy4d ago
Wait wait wait, a tiny helmet too? That's incredible lmao. I've been messing around with both of these and I honestly didnt think GPT-4 could pull off details like that. The paw on the steering wheel part is actually killing me, thats such a specific thing to get right. Stable Diffusion really just gave you a red smudge and called it a day huh? I gotta try this myself tonight because I need to see if it can do a cat in a cowboy hat holding a tiny lasso.
4
owens.cameron4d ago
Yeah but here's the thing nobody's talking about - the real test is if you can get it to do the cat's whiskers poking out from under the helmet. I tried the cowboy hat thing last week and it kept giving me a cat with no face, just a giant hat and a lasso. The paws though, that's where it gets wild. I swear GPT-4 has some secret training data of actual cats doing human stuff because how else does it know about the specific way a cat grips a steering wheel. Stable Diffusion just gives you a furry blob with thumbs and calls it a day.
2