Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Stable Diffusion 2.0

stability.ai

Post reply on HN
Print view View original

411–420 of 519 posts

Re: Stable Diffusion 2.0

#411

Post by sebzim4500 »

akie wrote: post #394
CuriouslyC wrote: post #387

Earlier quoted context omitted.

Removing NSFW content is fine, people who care about that can work around it easily. Removing celebrities and commercial artists was a mistake though and I expect this will need to be really impressive in other ways or people aren't going to bother using it.

It's remarkable, this sense of entitlement people have. You literally have a computer program here that can make photorealistic imagery of almost ANYTHING you ask it to, which was impossible even half a year ago, and here you are complaining that people won't use it unless it incorporates all of the protected imagery of famous artists and celebrities. Amazing.

If version 2 is worse than version then obviously a lot of people will use version 1. That doesn't make them entitled.

Re: Stable Diffusion 2.0

#412

Post by Xelynega »

bonoboTP wrote: post #324
LunaSea wrote: post #320

Earlier quoted context omitted.

There are already laws against it but enforcement is laking, as always.

If you can't process/digest copyrighted content with algorithms/machine learning then Google Search (the whole thing, not just Image Search) is dead. So no, it's not at all clear where the legal lines are drawn. There have been no court cases yet, regarding the training of ML models. People are trying to draw analogies from other types of cases, but this has not been tried in court yet. And then the answer will likel…

> If you can't process/digest copyrighted content with algorithms/machine learning then Google Search (the whole thing, not just Image Search) is dead.

Not if Google honors the robots.txt like they say they do. Hosting content with a robots.txt saying "index me please" is essentially an implicit contract with Google for full access to your content in return for showing up in their search results.

Hosting an image/code repository with a very specific license attached and then having that licensed ignored by someone who repackages that content and redistributes it is not the same as sites explicitly telling Google to index their content.

A much closer comparison IMO would be someone compressing a massive library of copyrighted content and then redistributing it and arguing it's legal because "the content has been processed and can't be recovered without a specific setup". I don't think we'd need prior court cases to argue that would most likely be illegal, so I don't see how machine learning models differ.

Re: Stable Diffusion 2.0

#413

Post by kmeisthax »

juliendorra wrote: post #375
ajuc wrote: post #325

Earlier quoted context omitted.

> It would be interesting to see if this technology can erode the copyright concept a bit Copyright law (especially in US) only ever changes in the direction that suits corporations. So - no. What I expect instead is artists being sued by a big tech company for copyright violations because that big tech company used the artist Public Domain image for training their copyrighted AI and as a result it created a copyrigh…

My bet is that big corporations won’t risk suing anyone over a supposed copyright on generated images,as there is a good chance that a court ends up stating that all AI generated images are in fact public domain (no author, not from the original intent and idea of a human) You can already see the quite strange and toned down language they use on their sites. (And for some the revealing reversal from we licence to you…

So, the US Copyright Office will already refuse to issue a copyright for text-prompt-generated AI art, at least if you try a stunt like naming the artist to be the AI program itself.

However, even if an image is not copyrightable, it can still infringe copyright. For example, mechanical reproductions of images are not copyrightable in the US[0] - which is why you even can have public domain imagery on the web. However, if I scan a copyrighted image into my computer, that doesn't launder the copyright away, and I can still be sued for having that image on my website.

Likewise, if I ask an AI to give me someone else's copyrighted work[1], it will happily regurgitate its training set and do that, and that's infringement. This is separate from the question of training the AI itself; even if that is fair use[2], that does nothing for the people using the AI because fair use is not transitive. If I, say, take every YouTube video essay and review on a particular movie and just clip out and re-edit all the movie clips in those reviews, that doesn't make my re-edit fair use. You cannot "reach through" a fair use to infringe copyright.

[0] In Europe there's a concept of neighboring rights, where instead of issuing you a full copyright you get 20 years of ownership instead. This is intended for things like databases and the like. This also applies to images; copyright over there distinguishes between artistic photography (full copyright) and other kinds of photography (20 years neighboring right only). This is also why Wikimedia Commons has a hilarious amount of Italian photos from the 80s in a special PD-Italy category.

[1] Which is not too difficult to do

[2] My current guess is that it is fair use, because the AI can generate novel works if you give it novel input.

Re: Stable Diffusion 2.0

#414

Post by Xelynega »

smeagull wrote: post #342
LunaSea wrote: post #320

Earlier quoted context omitted.

There are already laws against it but enforcement is laking, as always.

Can't wait for law enforcement to start arresting people for looking at images.

As long as they aren't repackaging and redistributing them, why would looking at them be illegal?

Re: Stable Diffusion 2.0

#415

Post by dmix »

machina_ex_deus wrote: post #399
in3d wrote: post #363

In addition to removing NSFW images from the training set, this 2.0 release apparently also removed commercial artist styles and celebrities [1]. While it should be possible to fine tune this model to create them anyway using DreamBooth or a similar approach, they clearly went for the safe route after taking some heat. 1. https://twitter.com/emostaque/status/1595731407095140352?s=4...

I predicted back when they started backpedaling that there's a chance that sd1.4 or 1.5 will be the best available model to the general public, for a very long duration, because the backlash will force them to self-castrate themselves. You can see nobody likes this new model in any of the stable diffusion communities. It's a big flop and for a good reason. The reason it was so successful in the first place was becaus…

These models will always work best with open datasets and open platforms for this reason.

Social media/"AI ethics" pressure groups will eventually come from these organizations (see Meta's recent debacle with Galactica). Being an unknown org without these pressures was a big reason Stable Diffusion got so popular in the first place.

Re: Stable Diffusion 2.0

#416

Post by educaysean »

nbzso wrote: post #204

The crimes against the creative people are getting better and better. What a time to live, when your entire career burns to dust just because. I hope AI gets these programmers jobs soon. Then we all can go to the woods and have a good life, finally.

Thanks for your premature concern, but we'll be fine. Despite how it may appear to a layperson such as yourself, the value of human creativity is in no way diminished by the release of this tool or others like it.

Re: Stable Diffusion 2.0

#418

Post by Xelynega »

etienne618 wrote: post #367
TeMPOraL wrote: post #258

Earlier quoted context omitted.

> Specifically, Wikimedia Commons images in the PD-Art-100 category, because the images will be public domain in the US and the labels CC-BY-SA. Doesn't the "BY" part of the license mean you have to provide attribution along with your models' output[0]? I feel you'll have the equivalent of Github Copilot problem: it might be prohibitive to correctly attribute each output, and listing the entire dataset in attribution…

I feel quite strongly that there is a large difference between Stable Diffusion and Copilot: with the size of the training set vs the number of parameters, it should be very difficult if not impossible for Stable Diffusion to memorize and, by extension, copy paste to produce its outputs. Copilot is trained on text and outputs text. Coding is also inherently more difficult for an AI model to do. I expect it will memor…

Let me guess that you write more code than visual art?

Isnt it a bit anthropomorphic to compare the two algorithms by "how a human believes they work" instead of "what they're actually doing different to the inputs to create the outputs"?

These are algorithms and we can look at how they work, so it feels like a cop-out to not do that.

Re: Stable Diffusion 2.0

#419

Post by kmeisthax »

TeMPOraL wrote: post #258
kmeisthax wrote: post #58

Is there a good explanation of how to train this from scratch with a custom dataset[0]? I've been looking around the documentation on Huggingface, but all I could find was either how to train unconditional U-Nets[1], or how to use the pretrained Stable Diffusion model to process image prompts (which I already know how to do). Writing a training loop for CLIP manually wound up with me banging against all sorts of stra…

> Specifically, Wikimedia Commons images in the PD-Art-100 category, because the images will be public domain in the US and the labels CC-BY-SA. Doesn't the "BY" part of the license mean you have to provide attribution along with your models' output[0]? I feel you'll have the equivalent of Github Copilot problem: it might be prohibitive to correctly attribute each output, and listing the entire dataset in attribution…

If I was generating image labels I absolutely would need to worry about that. However, since we're only generating images alone, we don't need to worry about bits of the labels getting into the output images.

The attribution requirement would absolutely apply to the model weights themselves, and if I ever get this thing to train at all I plan to have a script that extracts attribution data from the Wikimedia Commons dataset and puts it in the model file. This is cumbersome, but possible. A copyright maximalist might also argue that the prompts you put into the model - or at least ones you've specifically engineered for the particular language the labels use - are derivative works of the original label set and need to be attributed, too. However, that's only a problem for people who want to share text prompts, and the labels themselves probably only have thin copyright[0].

Also, there's a particular feature of art generators that makes the attribution problem potentially tractable: CLIP itself was originally designed to do image classification. Guiding an image diffuser is just a cool hack. This means that we actually have a content ID system baked into our image generator! If you have a list of what images were fed into the CLIP trainer and their image-side outputs[1], then you can feed a generated image back into CLIP and compare the distance in the output space to the original training set and list out the closest examples there.

[0] A US copyright doctrine in which courts have argued that collections of uncopyrightable elements can become copyrightable, but the resulting protection is said to be "thin".

[1] CLIP uses a "dual headed" model architecture, in which both an image and text classifier are co-trained to output data into the same output parameter space. This is what makes art generators work, and it can even do things like "zero-shot classification" where you ask it to classify things it was never trained on.

Re: Stable Diffusion 2.0

#420

Post by troad »

machina_ex_deus wrote: post #399
in3d wrote: post #363

In addition to removing NSFW images from the training set, this 2.0 release apparently also removed commercial artist styles and celebrities [1]. While it should be possible to fine tune this model to create them anyway using DreamBooth or a similar approach, they clearly went for the safe route after taking some heat. 1. https://twitter.com/emostaque/status/1595731407095140352?s=4...

I predicted back when they started backpedaling that there's a chance that sd1.4 or 1.5 will be the best available model to the general public, for a very long duration, because the backlash will force them to self-castrate themselves. You can see nobody likes this new model in any of the stable diffusion communities. It's a big flop and for a good reason. The reason it was so successful in the first place was becaus…

As someone completely unfamiliar with SD but interested in playing around with it in the future, what exactly should I download, to have a fully local instance of 1.4 or 1.5?
Post reply on HN