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

DALL·E now available in beta

openai.com

Post reply on HN
Print view View original

561–570 of 579 posts

Re: DALL·E now available in beta

#561

Post by ryanmercer »

softwaredoug wrote: post #398

Surprised by the lack of comments on the ethics of DALL-E being trained on artists content whereas copilot threads are chock full of devs up in arms over models trained on open source code. Isn’t it the same thing?

"DALL-E being trained on artists content"

Well, I can't go ask Caravaggio or Gentileschi to paint my query since they've been dead hundreds of years. But being able to to feed a query containing much more modern concepts in and get a baroque style painting in that specific style is wonderful.

Plus what has already been said about a lot of art being an imitation/derivation of previous works.

Re: DALL·E now available in beta

#562

Post by latexr »

jaggs wrote: post #321
berberous wrote: post #20

Earlier quoted context omitted.

I think they are reacting to competition. MidJourney is amazing, was easier to get into, gives you commercial rights, and frankly I found more fun to use and even better output in most instances.

nightcafe.studio is also free and good. Very good.

Gave it a try. After each image (all disappointing) I dumbed down the prompt, finally ending in “dog”. Didn’t even handle that.

Re: DALL·E now available in beta

#563

Post by jaggs »

latexr wrote: post #562
jaggs wrote: post #321

Earlier quoted context omitted.

nightcafe.studio is also free and good. Very good.

Gave it a try. After each image (all disappointing) I dumbed down the prompt, finally ending in “dog”. Didn’t even handle that.

I guess it depends on what you like/enjoy? It's not good at photorealistic, but it comes up with some pretty entertaining (and pretty?) 'arty' type stuff. I go on regularly just to play around for fun.

Re: DALL·E now available in beta

#564

Post by smusamashah »

andybak wrote: post #498
smusamashah wrote: post #286

Earlier quoted context omitted.

They won't. DALL-E images are mostly not as high quality. The high quality stuff which everyone has been sharing is result of lots of cherry picking.

It's a lot better than you are claiming. Mind if I ask if you have access personally?

Yes I have. And I realized it as soon as I started experimenting that mind blowing results are mostly cherry picking.

It's very good at generating art style images. These kind of images are mostly amazing most of the times. But the Photorealistic images only work with cherry picking.

Re: DALL·E now available in beta

#565

Post by andybak »

smusamashah wrote: post #564
andybak wrote: post #498

Earlier quoted context omitted.

It's a lot better than you are claiming. Mind if I ask if you have access personally?

Yes I have. And I realized it as soon as I started experimenting that mind blowing results are mostly cherry picking. It's very good at generating art style images. These kind of images are mostly amazing most of the times. But the Photorealistic images only work with cherry picking.

> And I realized it as soon as I started experimenting that mind blowing results are mostly cherry picking

Me and you must have very different definitions of "cherry picking". For prompts that fall within it's scope (i.e not something unusually complex or obscure) I get usable results probably 90% of the time.

Can you give me some examples of prompts that you tried where you found good results difficult to obtain?

Re: DALL·E now available in beta

#566

Post by yunwal »

williamcotton wrote: post #440

I would like to point out that there is not as much of an uproar on this forum over DALL-E utilizing other people’s photographs, illustrations, et al, as there is around Copilot utilizing other people’s code. Why do you think that is?

One reason is that copilot could be easily prompted to "create" line for line reproductions of people's code. AFAICT, you can't do this with DALLE (even if you, for example, try to input the caption of an image directly).

Re: DALL·E now available in beta

#567

Post by tough »

jnovek wrote: post #316
peteforde wrote: post #228

I have been having a blast with DALL-E, spending about an hour a day trying out wild combinations and cracking my friends up. I cannot imagine getting bored of it; it's like getting bored with visual stimulus, or art in general. In fact, I've been glad to have a 50/day limit, because it helps me contain my hyperfocus instincts. The information about new pricing is, to me as someone just enjoying making crazy imagines…

My heart sank when I saw the pricing model. I’ve been creating generative art since 2016 and I’ve been anxiously waiting for my invite. I wont be able to afford to generate the volume of images it takes to get good ones at this price point. I can afford $20/mo for something like this but I just can’t swing $200 to $300 it realistically takes to get interesting art out of these CLIP-centric models. Heck, the initial 5…

They have a form for discounted plans if you can't afford it, might be worth a try too

Re: DALL·E now available in beta

#568

Post by MrBra »

l33tman wrote: post #510
antiterra wrote: post #475

Earlier quoted context omitted.

> 6-9 means a joykilling minute+ wait Excet we’re already waiting 90+ seconds for dall-e mini.

min-dalle with MEGA model params takes about 20 seconds to generate 9 images on a RTX3080 if you run it locally, including the param loading time (which is about 8 of those seconds)

How much VRAM will minDALL-E plus the MEGA model take?

Re: DALL·E now available in beta

#569

Post by MrBra »

minimaxir wrote: post #216
neya wrote: post #186

I'm curious to know - does the community have any open source alternatives to DALL.E? For an initiative named OpenAI, keeping their source code and models closed behind a license is bullshit in my opinion.

The best open-source-but-actually-can-be-run-on-simple-infra analogous to DALL-E 2 is min-dalle: https://github.com/kuprel/min-dalle

How is minDALL-E better than DALL-E mini?

I can run DALL-E mini with the MEGA model with 12 GB of VRAM. What are the requirements for minDALL-E in terms of VRAM?

Re: DALL·E now available in beta

#570

Post by MrBra »

kuprel wrote: post #164
zikduruqe wrote: post #147

Earlier quoted context omitted.

Don't spend money. Use https://www.craiyon.com

This one is faster, I ported it https://replicate.com/kuprel/min-dalle

I can run DALL-E mini with the MEGA model with 12 GB of VRAM. What are the requirements for minDALL-E and MEGA, in terms of VRAM?
Post reply on HN