10 ms·
Hey! this is Joel, the maker of Artbreeder (originally called ganbreeder). The project was inspired by interactive evolutionary algorithms and novelty search (h
by joelS 6y ago
Hey! this is Joel, the maker of Artbreeder (originally called ganbreeder).
The project was inspired by interactive evolutionary algorithms and novelty search (https://www.youtube.com/watch?v=dXQPL9GooyI&t=1s https://www.youtube.com/watch?v=dXQPL9GooyI&t=1s)
The goal was to make a tool that reflected the collaborative and explorative aspects of creativity. Also to make GAN's (and high-dimensional spaces more generally) accessible to everyone in a fun
way.
Let me know if you have any questions :)
- j4nt4b 6y agoAwesome work! I love this concept, but I'm a writer who's interested in collaborative generative writing. Do you know of any efforts to create something similar for writing, such as poetry or micro-fiction? If not, I'm curious if you could point me in the right direction for putting something like that together.
- joelS 6y agoHey, thanks! I dont know about anything collaborative and for writing. https://aidungeon.io/ https://aidungeon.io/ Maybe collaborative to play. And then there are a lot of writers exploring fine-tuning text models like GPT2 for personal use. https://twitter.com/MagicRealismBot https://twitter.com/MagicRealismBot is one of many examples. https://towardsdatascience.com/how-to-fine-tune-gpt-2-so-you-can-generate-long-form-creative-writing-7a5ae1314a61 https://towardsdatascience.com/how-to-fine-tune-gpt-2-so-you... Maybe you should make it!
- j4nt4b 6y agoThank you!
- kleer001 6y agoIf you check out the archives of NaNoGenMo over at github you'll likely find what you're looking for: https://nanogenmo.github.io/ https://nanogenmo.github.io/ Check out the yearly resources.
- gwern 6y agoIf you'd like to write poetry or fanfiction, I provide GPT-2-1.5b models for both on https://www.gwern.net/GPT-2 https://www.gwern.net/GPT-2 - I don't provide a slick interface for writing (although Astralite has an upcoming web interface which is very snazzy), but they could be plugged into any such interface.
- MattRix 6y agoSomething like https://talktotransformer.com/ https://talktotransformer.com/ ?
- Jack000 6y agoa direct transfer of this concept to text isn't possible because the SotA for text generation is based on transformers, which do not work using latent space sampling. The directly analogous version of stylegan in the nlp world would be vq-vae, but currently it doesn't match transformers in generation quality. There are other ways to do generative writing though, mostly with prompts.
- archibaldJ 6y agoGreat work mate! Any plans of openning up an API for a subscription fee or something? :))
- joelS 6y agoThanks! No plans at the moment but it’s possible!
- sillysaurusx 6y agoHi Joel, Given that nVidia has a pretty clear license saying that their model and all derivates may only be used non-commercially (https://github.com/NVlabs/stylegan2/blob/cec605e0834de5404d5c7e5cead7053bdd0e4dde/LICENSE.txt#L60-L64 https://github.com/NVlabs/stylegan2/blob/cec605e0834de5404d5...) and Artbreeder is clearly commercial (https://artbreeder.com/pricing https://artbreeder.com/pricing) but is using StyleGAN (https://artbreeder.com/i?k=755521d9fe07456286a06f06 https://artbreeder.com/i?k=755521d9fe07456286a06f06), do you feel that Artbreeder is intentionally violating nVidia's license? For what it's worth, I fully support your project and I nearly launched something similar myself. But many people choose not to do projects such as yours due to licensing considerations, and I'm wondering in general how to reconcile the moral dilemma. On one hand, nVidia almost certainly doesn't care. On the other hand, they did say very clearly "thou shalt not use this model, or any derivative thereof, commercially." (It's tempting to believe the license only covers the code, not the model. But the same license appears in their download drive: https://drive.google.com/drive/folders/1QHc-yF5C3DChRwSdZKcx1w6K8JvSxQi7 https://drive.google.com/drive/folders/1QHc-yF5C3DChRwSdZKcx...)
- momokoko 6y agoCan we safely take your comment as legal advice in making business decisions about the use of the nVidia model? If so, you may want to consider the liability you are opening yourself up to. Since HN is an international site, it may be helpful to specify which geographic region you practice law out of as well. Thanks for the free legal advice! It definitely will save a few people the cost of paying for a legal interpretation of this situation. It is very generous of you.
- aspenmayer 6y agoHe’s not nvidia’s lawyer either so what does it matter either way? Please don’t do this here.
- gumby 6y agoFWIW I read sillysaurusx's comment as someone being friendly and pointing out a potential bug, not as snark or arrogance.
- programmarchy 6y agoHow did you make it so fast? Are the multiple combinations generated in parallel ahead of time once you breed images? Seems like the GPU power must be immense.
- sillysaurusx 6y agoThe 1024x1024 StyleGAN2 network can generate an image in around 200ms - 500ms. And Artbreeder may be powered by some beefy GPUs, so that might reduce the latency to something like 100ms. Throw in 100ms of server lag time, and it feels almost instantaneous. It's quite a technical achievement!
- joelS 6y agoNothing fancy, just running on AWS :) Specifically, https://aws.amazon.com/ec2/instance-types/g4/ https://aws.amazon.com/ec2/instance-types/g4/ - The uploading is actually the expensive part since it can take ~2 minutes per image.
- stevofolife 6y agoHow much are you paying to keep this up? Is it expensive to run GPU instances like this?
- battery_cowboy 6y agoI recall you had 'a font one' in the works, but I can't find anything about it now, is that still coming? The demo is really neat in the header image!
- joelS 6y agoI still have the model! I started working on the image -> vector conversion (which is more complicated than I expected) and then got distracted. I should finish that up, thanks for the reminder :)
- battery_cowboy 6y agoIt's such a cool idea, imagine a webfont that changes slightly each page load. I could make a document that slowly became more illegible after each new visitor, just for fun!
- aabbcc1241 6y agoI had a similar idea inspired by JWildfire but didn't got enough momentum to make it. You made it! And you've so much content on it already. Appreciate your effort. Spent half day having fun on your website :)
- joelS 6y agoThanks! Glad you liked it. JWildfire is super cool, I would love to add a fractal flame category to Artbreeder.
- netsec_burn 6y agoJoel, heads up the images that aren't supposed to be visible allow directory browsing. e.g. https://s3.amazonaws.com/artbreederpublic-shortlived/ https://s3.amazonaws.com/artbreederpublic-shortlived/ https://s3.amazonaws.com/artbreederpublic/ https://s3.amazonaws.com/artbreederpublic/ https://s3.amazonaws.com/ganbreederpublic/ https://s3.amazonaws.com/ganbreederpublic/ s3cli can navigate and download the images. Cool project!