No elephants: Breakthroughs in image generation
oneusefulthing.org
No elephants: Breakthroughs in image generation
1–10 of 373 posts
Re: No elephants: Breakthroughs in image generation
#2That proper "no elephants" first image is hilarious. Another key point here is the generative AI's meme game is getting rather strong.
Which isn't a small thing, humour is an advanced soft skill.
Re: No elephants: Breakthroughs in image generation
#3This is a before/after moment for image generation. A simple example is the background images on a ton of (mediocre) music youtube channels. They almost all use AI generated images that are full of nonsense the closer you look. Jazz channels will feature coffee shops with garbled text on the menu and furniture blending together. I bet all of that disappears over the next few months.
On another note, and perhaps others are feeling similarly, but I am finding myself surprised at how little use I have for this stuff, LLMs included. If, ten years ago, you told me I would have access to tools like this, I'm sure I would have responded with a never ending stream of ideas and excitement. But now that they're here, I just sort of poke at it for a minute and carry on with my day.
Maybe it's the unreliability on all fronts, I don't know. I ask a lot of programming questions and appreciate some of the autocomplete in vscode, but I know I'm not anywhere close to taking full advantage of what these systems can do.
Re: No elephants: Breakthroughs in image generation
#4I had a reasonable intuition for how the "old" method works, but I still don't grok this new approach.
"in multimodal image generation, images are created in the same way that LLMs create text, a token at a time"
Is there some way to visualise these "image tokens", in the same way I can view tokenized text?
Re: No elephants: Breakthroughs in image generation
#5This is a before/after moment for image generation. A simple example is the background images on a ton of (mediocre) music youtube channels. They almost all use AI generated images that are full of nonsense the closer you look. Jazz channels will feature coffee shops with garbled text on the menu and furniture blending together. I bet all of that disappears over the next few months. On another note, and perhaps other…
I've never used a stock photo site before, so I suppose it's no surprise I have no real use for "generate any image on demand".
Re: No elephants: Breakthroughs in image generation
#6[dead]
Re: No elephants: Breakthroughs in image generation
#7Looking at the example where the coffee table is swapped, I notice every time the image is reprocessed it mutates, based on the previous iteration, and objects become more bizarre each time, like chinese whispers.
* The weird-ass basket decoration on the table originally has some big chain links (maybe anchor chain, to keep the theme with the beach painting). By the third version, they're leathery and are merging with the basket.
* The candelabra light on the wall, with branch decorations, turns into a sort of skinny minimalist gold stag head, and then just a branch.
* The small table in the background gradually loses one of its three legs, and ends up defying gravity.
* The freaky green lamps in the window become at first more regular, then turn into topiary.
* Making the carpet less faded turns up the saturation on everything else, too, including the wood the table is made from.
Re: No elephants: Breakthroughs in image generation
#8> The results are not as good as a professional designer could create but are an impressive first prototype.
I like to look at how far we've come since the early days of Stable Diffusion. It was fascinating to play with it back then, but it quickly became apparent that it was "generic" and not suited for "real work" because it lacked consistency, text capabilities, fingers! and so on... Looking at these results now, I'm amazed at the quality, consistency and ease of use. Gone are the days of doing alchemy on words and adding a bunch of "in the style of Rutkovsky, golden hour, hd, 4k, pretty please ..." at the end of prompts.
Re: No elephants: Breakthroughs in image generation
#9I had a reasonable intuition for how the "old" method works, but I still don't grok this new approach. "in multimodal image generation, images are created in the same way that LLMs create text, a token at a time" Is there some way to visualise these "image tokens", in the same way I can view tokenized text?
Imagine you cut the image into 32x32 pixel blocks. And then for each block, you can chose 1 out of 128,000 variations. And then a post-processing step smoothes out the borders between blocks and adjusts small details. That's basically how a transformer image generation model works.
As such, the process is remarkably similar to old fixed-font ASCII art. It's just that modern AIs have a larger alphabet and, thus, more character shapes to choose from.
Re: No elephants: Breakthroughs in image generation
#10This is a before/after moment for image generation. A simple example is the background images on a ton of (mediocre) music youtube channels. They almost all use AI generated images that are full of nonsense the closer you look. Jazz channels will feature coffee shops with garbled text on the menu and furniture blending together. I bet all of that disappears over the next few months. On another note, and perhaps other…
I was wondering yesterday how AI is coming along for tweening animation frames. I just did a quick search and apparently last year the state of the art was garbage:
https://yosefk.com/blog/the-state-of-ai-for-hand-drawn-anima...
Maybe this multimodal thing can fix that?