Author Here: Just wanted to share some technical details. The code can be found here https://github.com/valentinvieriu/stylegan2 The main code and modification were put together by https://github.com/pbaylies/stylegan2 . I've started with the pretrained ffhq model, and I've trained it using come hand picked 1000 images of cubist Wikiart paintings. The frontend is Vuejs and it's using the amazing https://github.com/Ak…
Very cool. Some of the results reminded me of Max Ernst’s frottage works and futurist compositions. If you are interested in exploring this with other curated collections, reach out to me. I have a MOMA collection of about 10k images.
Show HN: A stream of AI-generated art
181–190 of 191 posts
Re: Show HN: A stream of AI-generated art
#182This is indistinguishable from something that would be hanging in an art museum, and anyone pretending otherwise only believes so because they know a priori that it is generated by ML. Honestly I would consider buying something like this, though I'd never waste my money on something so pretentious from a human. This, however, is an achievement, and there's beauty in that. Edit: in fact, if love to see two tests: a bl…
This is indistinguishable from something that would be hanging in an art museum, and anyone pretending otherwise only believes so because they know a priori that it is generated by ML. I certainly couldn't distinguish an individual images from something I might see in a museum. At the same time, scrolling the images felt very different than scrolling through a series of images by a given artist or images from a given…
https://www.google.com/search?q=composition+abstract+paintin...
I think this is quite expected, since ANN constructs its own view on stuffs, which is basically deconstruction of images into local components, like local color relation and sub-patterns. Global composition and shapes will be destroyed in favor of local composition, which is pretty much what many modern abstract paintings were after.
While ANN can mass-produced stuffs like this, that doesn't mean it's gonna replace artists. It's gonna be the opposite: if artists learn through AIs, each of which trained on different sets of data (which can overlap), we will be able to see works in a different level. It's like go players study AlphaGo for new principles that humankind have missed.
Re: Show HN: A stream of AI-generated art
#183Earlier quoted context omitted.
What I think would be interesting would be to add a human discriminator test: Can you tell genuine cubist art from computer-generated cubist art?
What would the point of that be may I ask? Something to do with a Turing like test? Lets not forget that the would be cubist model is actually fed tons of cubist images and all it does is regurgitating that back. Art done by people is somewhat similar but it does take place in a context that which AI could not yet understand.
I have to admit I'm rather in the first camp. I have spent time in art galleries, but not a huge amount; I have opinions about art, but 90% of what I hear said about paintings seems to me to be nonsense.
I know a lot about music; and so far I've always been able to distinguish 100% of computer-generated musical snippets from genuine classical snippets. And specifically, the key difference is exactly what is claimed by the second group -- that computer-generated melodies are just piles of similar musical ideas mashed together, without any apparent intent or direction.
One of the claims of skeptics of art is that there really is nothing there, and that all the "art critics" are just doing a wetware version of GPT-2. If someone trained in art can really distinguish a cubist painting with artistic intent from one generated with this neural net, then it's 1) proof that there really is something there, and 2) it's maybe worth learning about.
Re: Show HN: A stream of AI-generated art
#184This is awesome. There's a business here somewhere.
Re: Show HN: A stream of AI-generated art
#185Earlier quoted context omitted.
I think you have an opportunity here --- I found myself wondering how much it would cost to get a nice-quality print on canvas or metal. Something I could hang in my office. They look that good. Are they (or could they be) generated in sufficiently high resolution for a large format print?
Tbh, with direct printing to (large) canvas being reasonably priced now, this could be an easy sell Especially if they can be generated at 300dpi+
Re: Show HN: A stream of AI-generated art
#186Earlier quoted context omitted.
I'd bet most of them are quite "low quality art", but interesting anyway... plus, afaik, generally the image itself is not that important, as is the story behind it, if you have a story, a black square is a masterpiece, if you don't, it's junk...
The "story" aspect of art is bullshit, unless the art itself tells that story in a tangible way. If there's a story behind a black square, it's still just a black square.
A piece is more than just itself, there is a history to each one and that story adds to the piece. Some (honestly most) pieces and stories just are not very interesting, maybe even to the artist, and that is fine. And some are interesting to just some people at some times.
Art 'speaks' to each person differently, or sometimes not at all. Art is what you make of it. If you make nothing of some forms of art, that is on you. If you make nothing of all forms of art, you may want to consider why so many other humans feel so strongly the other way.
Though pretentious, ArtAssignment on YT is a great into to the wider world of art. I really think you should give it a try:
https://www.youtube.com/playlist?list=PLdGqz6dgvIzbBxo7XJHnO...
Re: Show HN: A stream of AI-generated art
#187Earlier quoted context omitted.
This is indistinguishable from something that would be hanging in an art museum, and anyone pretending otherwise only believes so because they know a priori that it is generated by ML. I certainly couldn't distinguish an individual images from something I might see in a museum. At the same time, scrolling the images felt very different than scrolling through a series of images by a given artist or images from a given…
The images are somewhat closer to abstract "composition" pieces, which are literally studies on visual composition: https://www.google.com/search?q=composition+abstract+paintin... I think this is quite expected, since ANN constructs its own view on stuffs, which is basically deconstruction of images into local components, like local color relation and sub-patterns. Global composition and shapes will be destroyed in f…
The thing is that I think that art involves using texture to "say" something. The artist puts combines represent of the physical world with reference to previous artists and images with composition. Some artists stress one or the other of these things - so that, yeah, a pure composition painting can definitely be "good enough to be in a museum". But a stream of painting like this definitely doesn't feel a museum or art as a whole.
Re: Show HN: A stream of AI-generated art
#188The further down the page I scrolled, the more I got a kind of despair. To realize that the space was infinite, and I would never reach its end. Each individual image is pretty cool. As a collection, they're sad. What's funny is that I then tabbed back to HN, where I was deep down the comments tree, and momentarily I forgot that real humans make these comments. I thought "I shall never reach the end of the HN comment…
Interesting, so would you say that the infinite aspect helped out the experience or damaged it? I've had doubts first when I've introduced it ( more from caching perspectives and data transfer costs ). But then I've wanted to showcase this amazing feature of those networks of generating an infinite amount of unique items, and also offer an unique experience to every user, the possibility of seeing unique art and book…
I am standing next to the artist. Seeing what she saw, imagining what it must be like to arrive from the other direction, without all the layers of preconception and intention that she baked in. And hopefully, from her direction and from mine, the good housekeeping seal, there is something "good" about this meeting of two perspectives.
Your project throws that good-housekeeping seal out the window. Which is not to say that a given image is bad. Rather, that the social contract between one viewer and one artist, is missing. My viewer's eye keeps assuming it exists, and my brain reminds me -- no, this is all fake, there is no humanity behind the curtain. There is nobody to connect to. Maybe it's still art, but not the kind I... emotionally expect from a thing that looks, at first glance, beautifully painted?
So, the infinite aspect definitely "helps" the experience. The exhaustion helps one feel viscerally how big is the problem space.
Re: Show HN: A stream of AI-generated art
#189Earlier quoted context omitted.
Would you appreaciate these images if you thought they were made by a person?
Machine learning enthusiasts always treat this response as if it's a checkmate, but it's a silly stance, you are depending on either: a) the ignorance or apathy of the person answering the question; and/or b) deception if you are telling them it's created by a human, when it's not. There's less inherent value of this kind of "art" for most people if they know it's created by an algorithm in milliseconds, rather than…
I can't appreciate how much pain something takes if it can be achieved for less. Much like opting to use a bucket for water transport instead of pipes, it's just silly.
Re: Show HN: A stream of AI-generated art
#190This is awesome. There's a business here somewhere.
It would be great if it would be :) At this moment is just an art experiment
- An interface where people select a subset of artists they like. - Build a model on the fly with art just from those artists - Get HQ high res downloads of generated art
Freemium model, with 1 HQ download free and the then $1 per download.