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

221–230 of 519 posts

Re: Stable Diffusion 2.0

#221

Post by nextlevelwizard »

LASR wrote: post #27

I am a solo dev working on a creative content creation app to leverage the latest developments in AI. Demoing even the v1 of stable diffusion to the non-technical general users blows them away completely. Now that v2 is here, it’s clear we’re not able to keep pace in developing products to take advantage of it. The general public still is blown away by autosuggest in mobile OS keyboards. Very few really know how far…

>Demoing even the v1 of stable diffusion to the non-technical general users blows them away completely.

What do the results have to do with "non-technical" people? I am blown away every time I run stable diffusion of the images I get out from it.

Re: Stable Diffusion 2.0

#222

Post by kennyloginz »

xnyan wrote: post #66
dougmwne wrote: post #62

Earlier quoted context omitted.

Are you kidding? Many times corporate decisions are being made effectively at random. Thinking that the average company operates with a 999 batting average is a total fantasy.

When our c suite decides on an ad campaign and tells our artists to draw normal humans, those people have 3 legs or upside down teeth exactly 0% of the time. Humans have many many limitations, but with every model I’ve tested there’s a set of errors that would virtually never be made by any human.

As an outsider, this rings true to me. I still don’t see any reduction of hours involved in producing professional level works. Generating YouTube thumbnails, sure.

Re: Stable Diffusion 2.0

#223

Post by govg »

wmwragg wrote: post #217
ftufek wrote: post #81

Earlier quoted context omitted.

To put things in perspective, the dataset it's trained on is ~240TB and Stability has over ~4000 Nvidia A100 (which is much faster than a 1080ti). Without those ingredients, you're highly unlikely to get a model that's worth using (it'll produce mostly useless outputs). That argument also makes little sense when you consider that the model is a couple gigabytes itself, it can't memorize 240TB of data, so it "learned"…

Well that would be ~4000 people each with an Nvidia A100 equivalent, or more with less, this would be an open effort after all. Something similar to folding@home could be used. Obviously the software for that would need to be written, but I don't think the idea is unlikely. The power of the commons shouldn't be underestimated.

It's not super clear whether the training task can be scaled in a manner similar to protein folding. It's a bit trickier to optimise ML workflows across computation nodes because you need more real time aggregation and decision making (for the algorithms).

Re: Stable Diffusion 2.0

#224

Post by Hamuko »

fastball wrote: post #76
kmeisthax wrote: post #45

Earlier quoted context omitted.

The main reason why Stable Diffusion is worried about NSFW is that people will use it to generate disgusting amounts of CSAM. If LAION-5B or OpenAI's CLIP have ever seen CSAM - and given how these datasets are literally just scraped off the Internet, they have - then they're technically distributing it. Imagine the "AI is just copying bits of other people's art" argument, except instead of statutory damages of up to…

Is artificially generated CSAM that doesn't actually involve children in its production not an improvement over the status quo?

"Artificially-generated CSAM" is a misnomer, since it involves no actual sexual abuse. It's "simulated child pornography", a category that would include for example paintings.

Re: Stable Diffusion 2.0

#226

Post by sussmannbaka »

fragmede wrote: post #190
xnyan wrote: post #66

Earlier quoted context omitted.

When our c suite decides on an ad campaign and tells our artists to draw normal humans, those people have 3 legs or upside down teeth exactly 0% of the time. Humans have many many limitations, but with every model I’ve tested there’s a set of errors that would virtually never be made by any human.

I agree. Cars break down and crash, they'll never replace horses.

the cars we're talking about here have a random amount of wheels and sometimes morph into cosmic horrors mid-ride.

Re: Stable Diffusion 2.0

#227

Post by pygy_ »

wokwokwok wrote: post #175
ftufek wrote: post #81

Earlier quoted context omitted.

To put things in perspective, the dataset it's trained on is ~240TB and Stability has over ~4000 Nvidia A100 (which is much faster than a 1080ti). Without those ingredients, you're highly unlikely to get a model that's worth using (it'll produce mostly useless outputs). That argument also makes little sense when you consider that the model is a couple gigabytes itself, it can't memorize 240TB of data, so it "learned"…

Quite right, but… > That argument also makes little sense when you consider that the model is a couple gigabytes itself, it can't memorize 240TB of data, so it "learned". The matter is really very nuanced and trivialising it that way is unhelpful. If I recompress 240TB as super low quality jpgs and manage to zip them up as single file that is significantly smaller than 240TB (because you can), does the fact they are…

> Can’t you just engineer the prompting better so that it generates “by Greg Rutkowski“ images without being trained on actual images by Greg?

You couldn't teach a human to do that without them having seen Greg's art. There are elements of stroke, palette, lightning and composition that can't be fully captured by natural language (short of encoding a ML model, which defeats the point).

Re: Stable Diffusion 2.0

#228

Post by bowsamic »

TaylorAlexander wrote: post #154
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…

Ah I am glad to see someone else talking about using public domain images! Honestly it baffles me that in all this discussion, I rarely see people discussing how to do this with appropriately licensed images. There are some pretty large datasets out there of public images, and doing so might even help encourage more people to contribute to open datasets. Also if the big ML companies HAD to use open images, they would…

No one is ever going to stop using all the available images until there is a law against it. Why would they?

Re: Stable Diffusion 2.0

#229

Post by jetbooster »

acidburnNSA wrote: post #156

Awesome. I'm installing on Ubuntu 22.04 right now. Ran into a few errors with the default instructions related to CUDA version mismatches with my nvidia driver. Now I'm trying without conda at all. Made a venv. I upgraded to the latest that Ubuntu provides and then downloaded and installed the appropriate CUDA from [1]. That got me farther. Then ran into the fact that the xformers binaries I had in my earlier attempt…

Thanks for reminding me why I shouldn't go to my computer right now and try getting this working with my 2070!

Re: Stable Diffusion 2.0

#230

Post by kennyloginz »

butMebbe wrote: post #46
malermeister wrote: post #37

Earlier quoted context omitted.

What kind of stuff does your app do that blows people away?

Wraps an ML model that blows people away in opinionated UX

That’s a little vague, so forgive me if I’m assuming too far.

You are making changes to a products UX based on graphical inference?

I could see a decent business supporting the logic problems a UX designed from AI graphics would introduce ;)

Post reply on HN