Running Stable Diffusion in 260MB of RAM • Hacker News Board

Re: Running Stable Diffusion in 260MB of RAM

#51

Post by stavros » Fri, Jul 21, 2023, 1:15 AM UTC

isoprophlex wrote: post #12

Incredible! If only there was some cheap hackable eink frame, you could make a fully self contained artwork from eink panel + rpi that's (slowly) continuously updating itself..!

You can use mine:

https://www.stavros.io/posts/making-the-timeframe/

You just put an image on some HTTP server and it shows it.

Re: Running Stable Diffusion in 260MB of RAM

#52

Post by yjftsjthsd-h » Fri, Jul 21, 2023, 1:31 AM UTC

hadlock wrote: post #37

boredemployee wrote: post #15

Earlier quoted context omitted.

Do you have a google colab link?

There is no shortage of google collab stable diffusion tutorials on the web

Which is why asking for one high quality starting point is such a useful question.

Re: Running Stable Diffusion in 260MB of RAM

#53

Post by NooneAtAll3 » Fri, Jul 21, 2023, 1:51 AM UTC

isoprophlex wrote: post #12

Incredible! If only there was some cheap hackable eink frame, you could make a fully self contained artwork from eink panel + rpi that's (slowly) continuously updating itself..!

2 years ago people were already hacking updateable photo frames out of them

https://www.youtube.com/watch?v=YawP9RjPcJA

Re: Running Stable Diffusion in 260MB of RAM

#54

Post by jojobas » Fri, Jul 21, 2023, 2:15 AM UTC

isoprophlex wrote: post #12

Incredible! If only there was some cheap hackable eink frame, you could make a fully self contained artwork from eink panel + rpi that's (slowly) continuously updating itself..!

Listen to the speech in the room, based on hot topics generate a set of pictures for tomorrow.

Re: Running Stable Diffusion in 260MB of RAM

#55

Post by mirekrusin » Fri, Jul 21, 2023, 2:34 AM UTC

johnklos wrote: post #28

In 260 megs of RAM?!? I'm going to try this on my Amiga! Check back in a few months for my results...

I remember rendering few ray traced balls over night on Amiga, good times.

Re: Running Stable Diffusion in 260MB of RAM

#56

Post by jfoster » Fri, Jul 21, 2023, 3:07 AM UTC

3 hours for 8 bit. I wonder what it would be if going further. Greyscale? Black & white?

Re: Running Stable Diffusion in 260MB of RAM

#57

Post by nicollegah » Fri, Jul 21, 2023, 4:52 AM UTC

Kuinox wrote: post #27

nicollegah wrote: post #25

Wait are these inference times real? 1 second on a Raspi? Do I get this right? This is faster than on my GPU. What's going on here?

I think it's the inference time per iteration.

ahh thanks

Re: Running Stable Diffusion in 260MB of RAM

#58

Post by MorbidCuriosity » Fri, Jul 21, 2023, 5:10 AM UTC

Minor49er wrote: post #38

zirgs wrote: post #30

"It runs Stable Diffusion" is the new "It runs Doom".

Now I'm wondering: could a monkey hitting random keys on a keyboard for an infinite amount of time eventually come up with the right prompts to get GPT-4 to produce code that compiles to a faithful reproduction of Doom?

Yes. Infinity is weird.

Re: Running Stable Diffusion in 260MB of RAM

#59

Post by janAkali » Fri, Jul 21, 2023, 5:54 AM UTC

Minor49er wrote: post #38

zirgs wrote: post #30

"It runs Stable Diffusion" is the new "It runs Doom".

Now I'm wondering: could a monkey hitting random keys on a keyboard for an infinite amount of time eventually come up with the right prompts to get GPT-4 to produce code that compiles to a faithful reproduction of Doom?

Now I'm wondering: if there were two monkeys hitting random keys on a keyboard for an infinite amount of time, one in the gpt-4 prompt and another straight typing 0s and 1s who would produce Doom code faster?

Re: Running Stable Diffusion in 260MB of RAM

#60

Post by jlokier » Fri, Jul 21, 2023, 5:58 AM UTC

MorbidCuriosity wrote: post #58

Minor49er wrote: post #38

Earlier quoted context omitted.

Now I'm wondering: could a monkey hitting random keys on a keyboard for an infinite amount of time eventually come up with the right prompts to get GPT-4 to produce code that compiles to a faithful reproduction of Doom?

Yes. Infinity is weird.

No, because GPT-4 has finite memory, its context length, and its random number generator for output selection is probably pseudo-random with finite memory.

If the random number generator is pseudo-random, this makes GPT-4 a deterministic finite-state machine, and the output sequence does not necessarily contain all possible subsequences no matter how many times the monkey types a new random key. Put differently, some output subsequences may be inaccessible no matter which keys are input. Same if the random number generator is truly random but its value cannot select among all possible output tokens, only a subset provided by the GPT at each step.