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

81–90 of 519 posts

Re: Stable Diffusion 2.0

#81

Post by ftufek »

kmeisthax wrote: post #74
echelon wrote: post #61

Earlier quoted context omitted.

> train this from scratch If you're talking about training from scratch and not fine tuning, that won't be cheap or easy to do. You need thousands upon thousands of dollars of GPU compute [1] and a gigantic data set. I trained something nowhere near the scale of Stable Diffusion on Lambda Labs, and my bill was $14,000. [1] Assuming you rent GPUs hourly, because buying the hardware outright will be prohibitively expen…

I have... ~11TBs of free disk space and a 1080ti. Obviously nowhere close to being able to crunch all of Wikimedia Commons, but I'm also not trying to beat Stability AI at their own game. I just want to move the arguments people have about art generators beyond "this is unethical copyright laundering" and "the model is taking reference just like a real human".

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".

But if you want to create custom versions of SD, you can always try out dreambooth: https://github.com/XavierXiao/Dreambooth-Stable-Diffusion, that one is actually feasible without spending millions of dollars on GPUs.

Re: Stable Diffusion 2.0

#82

Post by datalopers »

braingenious wrote: post #72
datalopers wrote: post #40

Earlier quoted context omitted.

They’ve ensured the only way to create CSAM is through old-fashioned child exploitation, meanwhile all perfectly humane art and photography is at risk of AI replacement. This is a huge missed opportunity to actually help society.

LMFAO What do you propose? The FBI releases a CSAM data set for devs to use for “training”? Would you be the one to create the model? Would you run a business that sells synthetic CSAM?

Without the changes they made to Stable Diffusion, it was already able to generate CP. That's why they restricted it from doing so. It did not have child pornography in the training set, but it did have plenty of normal adult nudity, adult pornography, and plenty of fully clothed children, and was able to extrapolate.

Anyway, one obvious application: FBI could run a darknet honeypot site selling AI-generated child porn. Eliminate the actual problem without endangering children.

Re: Stable Diffusion 2.0

#83

Post by tempthrowaway11 »

Gigachad wrote: post #11
argella wrote: post #6

Earlier quoted context omitted.

Bummer. AI porn is fun.

The future is probably models trained almost exclusively on porn.

They’re already out there, although they’re hard to find via Google - people are doing wild things like “merging” hentai models with models trained on real life porn to get realistic poses and lighting with impossible anatomy.

The scary thing is that you can then train it further with things like DreamBooth to start producing porn of celebrities… or, even more worrying, people you know.

Seriously folks, we are within a year or less of this being trivial. It’s already possible with a lot of work today.

Re: Stable Diffusion 2.0

#84

Post by Jack000 »

echelon wrote: post #61
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…

> train this from scratch If you're talking about training from scratch and not fine tuning, that won't be cheap or easy to do. You need thousands upon thousands of dollars of GPU compute [1] and a gigantic data set. I trained something nowhere near the scale of Stable Diffusion on Lambda Labs, and my bill was $14,000. [1] Assuming you rent GPUs hourly, because buying the hardware outright will be prohibitively expen…

Depends on the dataset. You can probably get decent results by restricting the modality of the images (faces, cars, bedrooms etc)

I trained from scratch with 4x3090 and while it’s not as good as SD it’s surprisingly better with hands.

Re: Stable Diffusion 2.0

#85

Post by adamdusty »

fastball wrote: post #75
echelon wrote: post #61

Earlier quoted context omitted.

> train this from scratch If you're talking about training from scratch and not fine tuning, that won't be cheap or easy to do. You need thousands upon thousands of dollars of GPU compute [1] and a gigantic data set. I trained something nowhere near the scale of Stable Diffusion on Lambda Labs, and my bill was $14,000. [1] Assuming you rent GPUs hourly, because buying the hardware outright will be prohibitively expen…

Quoted post unavailable.

"nowhere near"

Re: Stable Diffusion 2.0

#86

Post by dimaor »

88stacks wrote: post #60

Awesome, I’ve put stable diffusion on an api to train a model for anyone to use for free. I’m adding 2.0 to it as we speak! https://88stacks.com

How is this free? Is it possible to download the checkpoints?

I'm asking because I'm running SD locally but my GPU is not good enough to train new checkpoints and while I get the time to work on improve I wanted to use this API in order to generate some models for an illustration book I am working on.

Re: Stable Diffusion 2.0

#87

Post by charcircuit »

I dislike how they call their model open source even though there are restrictions on how you can use the model. The ability to use code however you want and not have to worry about if all the code you are using is compatible with your use case is a key part of open source.

Re: Stable Diffusion 2.0

#88

Post by snek_case »

braingenious wrote: post #72
datalopers wrote: post #40

Earlier quoted context omitted.

They’ve ensured the only way to create CSAM is through old-fashioned child exploitation, meanwhile all perfectly humane art and photography is at risk of AI replacement. This is a huge missed opportunity to actually help society.

LMFAO What do you propose? The FBI releases a CSAM data set for devs to use for “training”? Would you be the one to create the model? Would you run a business that sells synthetic CSAM?

Stable diffusion is able to draw images of bears wearing spacesuits and penguins playing golf. I don't think it actually needs that kind of input to generate it. It's clearly able to generalize outside of the training set. So... Seems it should be possible to generate that kind of data without people being harmed.

That being said, this is a question for sociologists/psychologists IMO. Would giving people with these kinds of tendencies that kind of material make them more or less likely to cause harm? Is there a way to answer that question without harming anybody?

In the mean time, stay away from 4chan.

Re: Stable Diffusion 2.0

#89

Post by myaccount9786 »

josephcsible wrote: post #23
cma wrote: post #18

Earlier quoted context omitted.

(Edit: it may have removed that wording now: https://github.com/Stability-AI/stablediffusion/commit/ca86d... ) They can force model upgrades too: > The New AI Model Licenses Have a Legal Loophole (OpenRAIL-M of Stable Diffusion) https://www.youtube.com/watch?v=W5M-dvzpzSQ

I don't understand why so many people call Stable Diffusion open source.

You can download the model weights and run them offline. At least, you could in v1.4. I assume this is still possible on v2.0?

Re: Stable Diffusion 2.0

#90

Post by bongobingo1 »

mberning wrote: post #43
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…

I don’t know anybody that is blown away by keyboard auto suggest. It’s wrong as often as it is right. Not saying it isn’t useful, but let’s not oversell it.

> Impressive isn't it!

>> Yeah! Don't they make a trillion dollars a year? How is it so crappy?

Post reply on HN