461–470 of 496 posts
Will I be able to generate an image of Xi as Pooh bear? Simple question. Just make it and post here the link to the result. Or will I receive a message that this prompt is not permitted and any attempt to get around the block will result in my account being suspended - like in Midjourney?
Adobe tends to be really old school and doesn’t have APIs. Or if they have an API it’s too expensive or crippled to be useful. That’s Adobes 1980s DNA…. everything is a client application. It’s a pity cause it looks like it would be fun to build with.
Earlier quoted context omitted.
I would consider both Stable Diffusion and MidJourney to be startups and both are better by far than the established companies like OpenAI, and there are dozens of LLMs soon catching up with GPT3/4. As you say, it will be a very interesting next 6-12 months..
Stable Diffusion is a thing you can run on your computer for free and train models for. It's from a research university. Is this one of those more expansive definitions of startup?
Earlier quoted context omitted.
I would consider both Stable Diffusion and MidJourney to be startups and both are better by far than the established companies like OpenAI, and there are dozens of LLMs soon catching up with GPT3/4. As you say, it will be a very interesting next 6-12 months..
Dall-E is still ahead of the competition for some tasks. Midjourney has a very polished look but it lacks the depth of understanding that Dall-E can manage. I regularly hit prompts that I need to jump into Dall-E for.
Earlier quoted context omitted.
I've looked a few times, but have not seen any research on assigning provenance to the weights used in a particular inference run. It's a super interesting space for a bunch of reasons. But the naive approach of having a table of how much each individual training item influenced every weight in the model seems impossibly big. For DALL-E 2's 6.5B parameters and 650m training items, that's 4.2 quadrillion associations.…
> And then you have to figure out which weights contributed the most to an output. Why do you have to, though? What do you hope to trace back to exactly?
Earlier quoted context omitted.
They also have the resources to build a huge training set, together with people who willingly upload their art and photos to them, which they can use to make the training set better than publicly available data.
Just to be really clear what we do and do not train on: Firefly was trained on Adobe Stock images, openly licensed content and public domain content, where copyright has expired. https://firefly.adobe.com/faq We do not train on user content. More info here: https://helpx.adobe.com/manage-account/using/machine-learnin...
If you go to https://stock.adobe.com/ and search for `-generative` you can see it filters out ~3M of them, but then go to the "undiscovered content" filter, and it appears to me like the vast majority of it is generative too.
Earlier quoted context omitted.
> thousands of different art styles Are you saying all those art styles are original? They are not. Artists influence artists. Sometimes the influence is strong, other times just a glimmer. Perhaps the burden of responsibility shouldn't be on Ai machine learning, but the human prompt author who chooses the instruction: "...in the style of Jo Blow". Who owns the ethical dilemma in that case? Who is being unfair to Jo…
I think this is a futile discussion. I think originality certainly exists, and I've seen it may times in the art world. I could be wrong, but you seem to imply originality is a made up concept? Do original thoughts even exist? Are our comments original or just regurgitation? Do you believe nothing ever created throughout history is original then? If you believe originality does exist, where and how do you draw the li…
The keyword you missed is "style". Your word, that I quoted.
As in, "the style of that artist". I wasn't referring to individual pieces of original art.
"Black and white street photography, gritty contrast, low wide angles, urban decay, weathered characters"... could be a style shared by many, that Ai learns.
Nobody owns a "style". A style is merely a convenient way to describe art. It belongs to the art critic more than the artist.
Many have made art in the style of say Andy Warhol - minimal flat colors, mass produced product imagery. Or surrealist styles, "like Dali", with dreamscapes, imaginary landscapes, twisted architecture and optical illusions. Things Ai can learn without being accused of stealing. Unless you're arguing a kind of "theft of humanity".
> Also, the approach of waving responsibility onto individuals is too naïve.
You want others to take responsibility for your actions?
Prompt: "image like Jon Doe's fantasy illustrations"... Prompting with intent. That's on you, not others to regulate your adult choices.
Before Ai, you could hire an artist to help you learn the style of another artist, living or dead. Maybe a drawing style you like. So you hire a skilled artist to point you in the right direction for acquiring a similar result with the pencil. Nothing ethically unsound there. Ai just skips the lessons and time it takes you to draw in the style of your favourite artist. Nothing has been stolen.
Earlier quoted context omitted.
is civitai.com literally just 90% japanese porn?
You can hide NSFW stuff, and apparently they’re going to implement a way to hide anime.