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

Diffusion Bee: Stable Diffusion GUI App for M1 Mac

github.com

Post reply on HN
Print view View original

31–40 of 199 posts

Re: Diffusion Bee: Stable Diffusion GUI App for M1 Mac

#34

Post by josephg »

metadat wrote: post #24
simondotau wrote: post #21

Earlier quoted context omitted.

And the real questions come out. Seriously though, I imagine this is less a case of whether this specific implementation permits pornography, but whether any porn was included in the dataset it was trained on. No matter how good AI is, it only knows what it knows.

Based on the upvotes, inquiring minds _really want to know_!

It can make porn images, but I haven't been able to get consistently good results. There's probably ways to tweak the parameters to improve things, but I haven't figured that out.

It mostly understands what naked people look like, but the images I've generated involve a lot of accidental body horror. You get a lot of people with extra arms, weird eyes, or body parts in the wrong places. The fact that they explicitly removed porn from the training set comes through pretty clearly in the model.

I suspect it could be improved a lot with some specialized retraining. As far as I know, nobody has done that work yet.

Re: Diffusion Bee: Stable Diffusion GUI App for M1 Mac

#36

Post by josephg »

andreasley wrote: post #26
indigodaddy wrote: post #19

How long is the generation taking on say a 2021 MacBook M1 Pro?

55 seconds on a M1 Pro MacBook with 16GB RAM to generate a picture (with default settings, so 512x512, 25 steps). On a M1 Ultra, it takes 12 seconds with 64GB RAM (same settings). While computing, the Mac Studio is pulling 100 watts of power.

Based on the screenshot, it looks like this is set to default to 25 steps. I've been using 50 steps on my M1 macbook pro with 16gb of ram. It takes about 1m30s per image.

Re: Diffusion Bee: Stable Diffusion GUI App for M1 Mac

#38

Post by XorNot »

simondotau wrote: post #21
metadat wrote: post #18

Earlier quoted context omitted.

Does it work with pornographic or potentially pornographic prompts?

And the real questions come out. Seriously though, I imagine this is less a case of whether this specific implementation permits pornography, but whether any porn was included in the dataset it was trained on. No matter how good AI is, it only knows what it knows.

I mean I don't know why we have to talk around this point: yes, you can definitely generate pornography with SD. Be prepared to see a lot of weird things, because that super-realism really means it just drives right into the uncanny/body horror valley at full speed.

But the second point here is also wrong: the whole reason these models are interesting is because they can generate things they haven't seen before - the corpus of knowledge represents some type of abstract understanding of how words relate to things that theoretically does encode a little bit of the mechanisms behind it.

For example, it theoretically should be able to reconstruct human like poses it has never seen before provided it has examples of what humans look like and something which transposes to an approximate value - an obvious example in the context of the original question would be building photorealistic versions of a sketched concept (since somewhere in it's model is an axis which traces from "artistic depiction of a human" to "photograph of a human" in terms of style content).

Of course, most people aren't very good at drawing realistic human poses - it's a learned skill. But the magic of deep learning is really that eventually it doesn't need to be - we would hopefully be able to train a model which can be easily copied and distributed which represents the skill, and SD is a big step in that direction (whether it's a local maxima remains to be seen - it's dramatic, but is it versatile?)

Re: Diffusion Bee: Stable Diffusion GUI App for M1 Mac

#40

Post by KaoruAoiShiho »

fragmede wrote: post #9
keizo wrote: post #3

amazing, look forward to trying it out. has anyone done the same for windows?

Yes: https://nmkd.itch.io/t2i-gui No idea if it supports AMD GPUs though.

Do you have any info on perf compared to this: https://news.ycombinator.com/item?id=32805174
Post reply on HN