Live data from Hacker News

DALL·E: Introducing Outpainting

openai.com

221–230 of 302 posts

Re: DALL·E: Introducing Outpainting

#221
post #110

Earlier quoted context omitted.

> I can only guess either safety from misuse I still don't understand what this would mean. Where are all of the terrible things that were supposed to happen, now that Stable Diffusion is available? We've been able to create completely photorealistic fiction for decades now. See any movie with CGI for an example of whole worlds, and people, that don't exist. The bar has gradually been lowering (see the amazing CGI th…

Stable Diffusion allows anyone to make kiddie porn with a half-second of curiosity/effort. Maybe you didn't know about that, maybe you think it's NBD, but in any case, that is the tire fire which aspiring AI majors want to avoid.

I can't quite decide if that's good or bad.

On one side there's no problem to like CP if it's AI generated and no real kids are harmed.

On the other side it may desensitize the public to this type of content.

Re: DALL·E: Introducing Outpainting

#222

Earlier quoted context omitted.

Doesn't make sense to me why OpenAI has kept DALL-E closed source for so long. I can only guess either safety from misuse or leveraging it for money. At this rate though, Stable Diffusion is going to dwarf it

I don't follow this stuff very closely - is there any open-source model for text generation that outclasses GPT-3? Stable Diffusion has been released for barely a week and already seems like the clear winner. It doesn't seem like any of the open (actually open) text models have made as much of a splash. Of course maybe it's just because text is less visually impressive than images.

There are some open models as good as initial GPT-3 (which wasn't hard), but whatever they did to create InstructGPT hasn't been reproduced as far as I know, and it's the first one to really seem magical.

Re: DALL·E: Introducing Outpainting

#223

Earlier quoted context omitted.

It's sort of both. OpenAI, being an outgrowth of the AI doomerist community, does have a bunch of people who really do think the technology is too dangerous to be given to the masses. This happens to mesh perfectly with the other group of people at OpenAI who want to make tons of revenue. It's a harmonious alignment for everyone! Except, y'know, us.

Content creators, like artists, also happen to hate filters. They do not want to have San Francisco VC culture induced political correctness imposed on their work. This helps Stable Diffuse to quickly gain popularity.

That's not "VC political correctness", it's "you can't use credit cards or be hosted in many countries if your online service produces porn".

Re: DALL·E: Introducing Outpainting

#224
post #213

Earlier quoted context omitted.

> be an explosion of believable fake items I think that's a good thing. A slow creep is dangerous, because people may not notice, and would be tricked. An explosion of insanity makes everyone realize they can't believe everything they see on the internet.

It will be all of the above so it doesn’t really matter since it’s all going to be indistinguishable from the real thing.

Which has been the case for the last few decades. Nobody should be trusting anything they see on the internet.

Re: DALL·E: Introducing Outpainting

#225

Earlier quoted context omitted.

>Doesn't make sense to me why OpenAI has kept DALL-E closed source for so long. >leveraging it for money

It was a long gap between DALL-E 1 and 2, a whole year. In that time they just sat on it, didn't release anything. Such a bummer. My theory is that they wanted to hype everyone up even more for the grand commercial release. Funny thing is that people didn't stand still and invented diffusion and other CLIP guided image synthesis methods, and DALL-E 2 copied the method, completely changing from the first architecture.…

The fact that any of this stuff works is so surprising that none of it could possibly have been planned ahead on anyone's part. StableDiffusion has some real novel research of its own in there; the methods OpenAI and Google used couldn't produce a model that trains or runs as fast as SD.

Re: DALL·E: Introducing Outpainting

#226
post #84

Meanwhile someone has already built a photoshop plugin for Stable Diffusion that you can use today to do basically the _exact_ same thing: https://old.reddit.com/r/StableDiffusion/comments/wyduk1/sho...

But aren’t the results from stable diffusion not nearly as good as DALLE2?

You don't have to drive SD with a text prompt; if you take the model weights you can feed anything you want in there and generate an image out of it.

Since DALLEmini and DALLE2 are more "creative" (since they use a better text transformer) you can use them to generate the input and SD to refine it for more fine detail.

https://twitter.com/hardmaru/status/1559861001163788289

Re: DALL·E: Introducing Outpainting

#227

Earlier quoted context omitted.

I find Stable Diffusion better overall, but it has downsides. Stable Diffusion tends to be more creative than DALL-E, but does a lousy job of following directions, especially complex ones. DALL-E is good if I know what I want specifically. I can think of ways to fix Stable Diffusion since it's open-source. I think I could bridge the gaps as I see them in about a weekend of hacking. I'm not sure when I'll get that wee…

Setting a high cfg parameter, like 13, drastically helps with the prompt following. That said, for me, I agree that dalle does much better pencil sketches.

Raising the cfg ("classifier-free guidance") scale is essential for following the prompt, but if you raise it too high the image gets weird and saturated.

According to Google's Imagen paper this is literally because the pixels get multiplied by the cfg scale and start clipping; they have a technique called dynamic thresholding that replaces it. Not sure if SD uses this, but I saw Emad hinting they were training an Imagen model…

Re: DALL·E: Introducing Outpainting

#228

Earlier quoted context omitted.

One of SDs big limitations (understanding from what I had read about it) is positional prompts. dall-e seems to understand x on top of Y, but simple diffusion does not.

img2img drawing should take care of that

I'm experimenting with the base model right now, I'm going to be very excited to try that one out too.

Re: DALL·E: Introducing Outpainting

#229

Feels like a race to the bottom. More features, lower cost, every week. No idea where it’ll level out, but I like it. Just bought some more Dalle credits today because it’s so much fun. This is a revolution in ‘art technology’ it’s like Steve Job’s bicycle for the mind. Best I could do a month ago was a stick figure in MS Paint, but now..

> Best I could do a month ago was a stick figure in MS Paint That is still the best you can do... which happens to be about the best I can do! Just like my introduction to the computer at a young age has atrophied my handwriting quality.

Just switch to a PalmPilot.

Re: DALL·E: Introducing Outpainting

#230
post #189

Earlier quoted context omitted.

Our brain is apparently just a 4 gb large arrangement of electrical weights.

Not to be pedantic, but we have on the order of 100B neurons, and afaik each of them can be connected to thousands of other neurons. I assume we probably have a ways to go before we're encoding the amount of information a brain can comprehend.

Also, your brain knows how to look up things in external storage, but ML models have to keep everything in their weights even if it's not a good fit.
Post reply on HN