3 ms·
Like most image generators, it didn’t pass the piano keyboard test. (Black keys are wrong.) https://aistudio.google.com/app/prompts?state=%7B%22ids%22:%5B%221t
by skybrian 1y ago
Like most image generators, it didn’t pass the piano keyboard test. (Black keys are wrong.)
https://aistudio.google.com/app/prompts?state=%7B%22ids%22:%5B%221tWprDfMdTpOYRm08ugIJsvN15mF09vB9%22%5D,%22action%22:%22open%22,%22userId%22:%22114156500057804356924%22,%22resourceKeys%22:%7B%7D%7D&usp=sharing https://aistudio.google.com/app/prompts?state=%7B%22ids%22:%...
- mikepurvis 1y agoInteresting! I feel like that's maybe similar to the business of being able to correctly generate images of text— it looks like the idea of a keyboard to a non-musician, but is immediately wrong to someone who is actually familiar with it at all. I wonder if the bot is forced to generate something new— certainly for a prompt like that it would be acceptable to just pick the first result off a google image search and be like "there, there's your picture of a piano keyboard".
- joombaga 1y agoWhat is the piano keyboard test? Your link requires granting AI Studio access to Google Drive, which I do not want to do.
- raincole 1y agoJust ask it to generate a correct piano keyboard. It's something the current gen of image generator AIs fail at.
- ZiiS 1y agoDo most humans pass?
- adzm 1y ago2-2-1-2-2-2-1
- polynomial 1y agoI still feel like most humans would fail, haha.
- twodave 1y agoMaybe, but anyone who knows what a chromatic scale is should be able to reason it out. E# == F, B# == C, so no black keys between those.
- phainopepla2 1y agoPresumably most humans with a camera do
- raincole 1y agoMost humans fail at 4 digits multiplication, or drawing a cube in perspective.
- vunderba 1y agoAnything that is heavily periodic can definitely trip up image gen - that being I just used Flux Kontext T2I and got a got pretty close (disregard the hammers though since thats a right mess). Only towards the upper register did it start to make mistakes. https://imgur.com/a/fyX42my https://imgur.com/a/fyX42my
- psbp 1y agoDoesn't pass the analog clock test either.
- Workaccount2 1y agoThe selling point of this model really seems to be it's consistency between generations rather than it's raw generating ability. for instance: https://aistudio.google.com/app/prompts/1gTG-D92MyzSKaKUeBu236pJTaU3pNvg4 https://aistudio.google.com/app/prompts/1gTG-D92MyzSKaKUeBu2...
- skybrian 1y agoI can’t see it. You probably need to set permissions to “anyone with the link can access.”
- carimura 1y agoor my "hands with palms facing down" test.... no matter how hard I try it just can't get open hands, palms down.
- pbhjpbhj 1y agoI guess the vast majority of images have the palms the other way, that this biases the output. It's like how we misinterpret images to generate optical illusions, because we're expecting valid 3D structures (Escher's staircases, say).
- vunderba 1y agoYes - it's the same reason generating a 5-leaf clover fails - massive amounts of training data that predisposes the model against it.
- vunderba 1y agoIt's probably just a matter of rerolling a few times. I was able to get it around 25% of the time. https://imgur.com/a/H9gH3Zy https://imgur.com/a/H9gH3Zy
- carimura 1y agothat's pretty good. I was using a cartoon girl as an example of a dance move for kids. https://g.co/gemini/share/0e0de0d42029 https://g.co/gemini/share/0e0de0d42029
- cubefox 1y agoLike most image models, except GPT-4o, it also didn't pass the wooden Penrose triangle test. (It creates normal triangles.)
- pbhjpbhj 1y agoAre their models that have vector space that includes ideas, not just words/media but not entirely corporeal aspects? So when generating a video of someone playing a keyboard the model would incorporate the idea of repeating groups of 8 tones, which is a fixed ideational aspect which might not be strongly represented in words adjacent to "piano". It seems like models need help with knowing what should be static, or homomorphic, across or within images associated with the same word vectors and that words alone don't provide a strong enough basis [*1] for this. *1 - it's so hard to find non-conflicting words, obviously I don't mean basis as in basis vectors, though there is some weak analogy.
- heyjamesknight 1y agoHow would you encode those ideas?
- pbhjpbhj 1y agoI don't know, in part that's why I asked ... I wonder if there's a way to provide a loosely-defined space. Perhaps it's a second word-vector space that allows context defined associations? Maybe it just needs tighter association of piano_keyboard with 8-step_repetition??
- conception 1y agoFailed my horizontal text test as well.