Earlier quoted context omitted.
200% slower is 3x as long, yes.
I love making fun of people that don’t understand percentages… wait, wat?
Who is the intended butt of your joke here?
And explain to me why it isn't you?
Running Stable Diffusion XL 1.0 in 298MB of RAM
91–100 of 166 posts
Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#92Amazing feat, but of course takes forever to generate an image (in the Readme states 11 hours)
On a raspberry Pi. Zero 2.
Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#93Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#9411 hours remind me of doing raytracing on my Amiga 500 back in the day. It was definitely an overnight job for the "final" render.
Heh sometimes I am still doing that. modern bidirectional raytracers can do some interesting tricks. and I wanted to see caustics(the bright lines in pools). but caustics despite being bright are actually statistically rare. to get good caustics you have to unbound the render engine and just let it cook overnight.
And the end result, a single image of a mediocre scene by a poor artist with amazing caustics. I won't be quitting my day job.
Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#95[flagged]
Before you waste your time, this is a commercial product and you need to pay $30 to buy their model to run it.
Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#96Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#97Fascinating. The money quote: "OnnxStream can consume even 55x less memory than OnnxRuntime while being only 0.5-2x slower" The trade-off between (V)RAM use and inference time sounds like it could be advantageous in some scenarios, and not just when RAM is constrained like in the RPi case. I actually wonder if this weight unloading approach can be used to handle larger batch sizes in the same amount of RAM, in effect…
From my (albeit naive) reading, it doesn't appear that that they've reduced the amount of memory bandwidth required, simply the size of the working set required. Since inference is generally memory bandwidth bound once you reach the level of 'does this model even fit in the given system', I'd imagine that this technique wouldn't help much for greater throughpit via larger batch sizes. Just one instance is probably al…
That's true. But assuming the required memory bandwidth is not already maxed out by this, there might still be a narrow but workable "Goldilocks zone" for this technique to be useful.
Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#98Earlier quoted context omitted.
200% slower is 3x as long, yes.
I love making fun of people that don’t understand percentages… wait, wat?
117. 472% of grade school students are unable to readily convert between fractions and percentages.
38.157% of informally provided statistics are made up on the spot under the assumption nobody will actually check.
Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#9911 hours remind me of doing raytracing on my Amiga 500 back in the day. It was definitely an overnight job for the "final" render.
Same, (albeit a little later) with dodgy copy of 3DSMAX on a 386.
Re: Running Stable Diffusion XL 1.0 in 298MB of RAM
#100Earlier quoted context omitted.
This is more than a little melodramatic. https://frame.work/ and the https://mntre.com/ MNT Reform: Exist
If my country decides to ban the ownership of general purpose computers for individual persons, they would order the customs service to stop import of any computer hardware that enabled general purpose computing. Now I would not be able to have any computer shipped to me from outside my country, so I could no longer buy from either of those vendors you linked. Furthermore, it also would mean that I would not be able…
This is a slippery slope to the extreme.
What country outside of North Korea has banned the ownership of general purpose computers, or even considered/tried to?