Lynne Ingram, a Somerset beekeeper and the chair of the Honey Authenticity Network UK, said: “The market is being flooded by cheap, imported adulterated honey and it is undermining the business of genuine honey producers. The public are being misinformed, because they are buying what they think is genuine honey.”

The UK is one of the biggest importers of cheap Chinese honey, which is known to be targeted by fraudsters. Honey importers say supply chains and provenance are carefully audited, but there has been no consensus on how technical tests should be applied, or which are most reliable.

A fun bit of perspective that I like to mention in discussions about this. Roll back a bit over a century:

Scientific American, November 2, 1907

Artificial Honey

Prof. Herzfeld, of Germany, recently brought out some interesting points regarding the manufacture of artificial honey in Europe. It is noticed that when we bring about the inversion of refined sugar in an almost complete manner and under well determined conditions, this sugar solidifies in the same way as natural honey after standing for a long time, and it ran be easily redissolved by heating. Owing to the increased production of artificial honey, the bee cultivators have been agitating the question so as to protect themselves, and it is proposed to secure legislation to this effect, one point being to oblige the manufacturers to add some kind of product which will indicate the artificial product. On the other hand, it is found that the addition of inverted sugar to natural honey tends to improve its quality and especially to render it more easiIy digested. Seeing that sugar is about the only alimentary matter which is produced in an absolutely pure state, its addition to honey cannot be strictly considered as an adulteration. Bees often take products from flowers which have a bad taste; and the chemist Keller found that honey coming from the chestnut tree sometimes has a disagreeable flavor. From wheat flowers we find a honey which has a taste resembling bitter almonds, and honey from asparagus flowers is most unpalatable. Honey taken from the colza plant is of an oily nature, and that taken from onions has the taste of the latter. In such cases, the honey is much improved by the addition of inverted sugar. Prof. Herzfeld gives a practical method for preparing this form of sugar. We take 1 kilogramme (2.2 pounds) of high-quality refined sugar in a clean enamelware vessel, and add 300 cubic centimeters (10 fluid ounces) of water and 1.1 grammes (17 grains) tartaric acid. This is heated at 110 deg. C. over an open fire, stirring all the while, and is kept at this heat until the liquid takes on a fine golden yellow color, such operation lasting for about three quarters of an hour. By this very simpIe process we can easily produce artificial honey. Numerous extracts are now on the market for giving the aroma of honey, but none of them will replace the natural honey. However, if we take the artificial product made as above and add to it a natural honey having a strong aroma, such as that which is produced from heath, we can obtain an excellent semi-honey.

You’re probably better off asking on !imageai@sh.itjust.works.

If you want to try there, I can throw some ideas out.

Asklemmy, despite the oft-confusing community name that generates a lot of these sort of things, isn’t really intended as a general “ask any question” community, but for “thought-provoking” questions. Their Rule 5 excludes stuff like this.

The mods tend to delete stuff like this; I’ve had a few questions that I’ve spent time answering and then had the post deleted with the answers, which is kinda frustrating if you’ve put effort into an answer.

If you ask there, I’d suggest indicating which system you used to generate the image, as it’ll affect the answer.

That’s a pretty good AK. The models I’ve used have pretty consistently done poorly with firearms, though I’ve not tried specialized models.

EDIT: And military hardware. I was thinking about trying an F-22 Flux LoRA that I saw on civitai. Vanilla Flux can’t do much by way of accurate military ships, vehicles, or aircraft.

I wish people had a perverse love of main battleships

You ever seen Drachinifel’s channel on YouTube? He’s into the gun-era stuff, which overlapped a lot with the British Empire’s era.

Ever played the Rule the Waves video game series?

Because only an idiot would move there.

Ehhhh…

I mean, you can make houses that can handle hurricanes and flooding. They will cost more and may place some aesthetic constraints on the house, but it’s like fires or earthquakes or anything else; you can build to deal with it.

For flooding, you can elevate a house.

https://experience.arcgis.com/experience/203f772571cb48b1b8b50fdcc3272e2c

It looks like you can even modify an existing house, which I think is a little bonkers, but maybe it’s cheaper:

https://www.fema.gov/sites/default/files/documents/fema_elevating-your-house-chapter-5.pdf

You can also build buildings that will stand up to debris blown by high winds; heavier support, possibly concrete or steel frames.

I agree with the sentiment regarding, say, people who just keep rebuilding the same beach house after it gets wiped out by storms, but it is possible to have houses that can take rougher conditions. Just costs more.

EDIT: That being said, I would not personally want to move to Florida due to the humid and toasty summers and mosquitoes. Not my cup of tea.

But everyone’s got their own preferences.

Oh, cool, there is actually a !credibledefense@sh.itjust.works.

Not much traffic, but…

seven stages of anything

Five stages?

https://en.wikipedia.org/wiki/Five_stages_of_grief

The only way you can do that is if Congress signs off on it.

Every other state has an incentive not to permit that, because then that state gets two senators of its own.

Congress has only ever permitted a state to split a single time – West Virginia from Virginia, during the American Civil War, where West Virginia was willing to side with the Union, and contained some militarily-important rail and water infrastructure.

Texas also negotiated the right to have the ability to split into five states if it wanted down the line at the time it joined, but I recall reading that it was considered to no longer be an exerciseable option after the American Civil War.

EDIT:

https://en.wikipedia.org/wiki/Admission_to_the_Union

Article IV, Section 3, Clause 1:

New States may be admitted by the Congress into this Union; but no new State shall be formed or erected within the Jurisdiction of any other State; nor any State be formed by the Junction of two or more States, or Parts of States, without the Consent of the Legislatures of the States concerned as well as of the Congress.[4]

EDIT2: Correction; Kentucky was also split from Virginia and Maine from Massachusetts. The Kentucky split happened before the US Constitution was ratified. Maine was part of the Missouri Compromise, to keep slave and free states in balance when Missouri joined as a slave state.

https://www.amazon.com/s?k=vanilla+bar&ref=nb_ss_gw

There are various vanilla-flavored bars.

https://www.military-ranks.org/decorations

https://en.wikipedia.org/wiki/List_of_naval_battles

Most recent three entries as an example:

• August 21, 2024: 21 August - In Red Sea, the Greek flagged tanker Sounion was attacked and exchanged gunfire with two Houthi manned fast attack craft. The Sounion was then hit by three projectiles and set afire before being abandoned by its crew.

https://en.wikipedia.org/wiki/Attacks_on_the_Sounion

Two Houthi fast attack craft engaged in a firefight with the Sounion armed guards before three projectiles struck the tanker.

Sounds like that may have been small arms, if armed guards were involved. There was a subsequent boarding action.

• 15 July, 2024 - In the Red Sea, the Panamaninan flagged tanker Bentley-I is unsuccessfully attacked by a Houthi sea drone before engaging in a gun battle with two Houthi manned fast attack craft, which are driven off.[29]

I assume that the sea drone attempted to ram. The attack craft were using guns.

• 5 March, 2024 – Sinking of the Sergey Kotov – Russian Bykov-class patrol ship, RFS Sergey Kotov is attacked and sunk by Ukrainian sea drones in the Kerch Strait.[28]

Probably ramming.

We may, surprisingly-enough, be at more of a knife-fight range than we have for a long time in naval combat, though I suppose one could argue that maybe USVs should be thought of as more like long-range torpedoes or something than watercraft.

6900

kagis

If that’s 16GB, that should be more than fine for SDXL.

So, I haven’t done much with the base Stable Diffusion XL model. I could totally believe that it has very little Sailor Moon training data. But I am confident that there are models out there that do know about Sailor Moon. In fact, I’ll bet that there are LoRAs – like, little “add-on” models that add “knowledge” to a checkpoint model on Civitai specifically for generating Sailor Moon images.

Looks like I don’t have vanilla SDXL even installed at the moment to test.

downloads vanilla

Here’s what I get from vanilla SDXL for “Sailor Moon, anime”. Yeah, doesn’t look great, probably isn’t trained on Sailor Moon:

Sailor Moon, anime
Steps: 20, Sampler: DPM++ 2M, Schedule type: Karras, CFG scale: 7, Seed: 2, Size: 1024x1024, Model hash: 31e35c80fc, Model: sd_xl_base_1.0, Token merging ratio: 0.5, Version: v1.9.4-169-ga30b19dd

searches civitai

Yeah. There are. Doing a model search just for SDXL-based LoRA models:

https://civitai.com/search/models?baseModel=SDXL 1.0&modelType=LORA&sortBy=models_v9&query=sailor moon

75 results for ‘sailor moon’

goes to investigate

Trying out Animagine, which appears to be a checkpoint model aimed at anime derived from SDXL, with a Sailor Moon LoRA that targets that to add Sailor Moon training.

I guess you were going for an angelic Sailor Moon? Or angelic money, not sure there…doing an angelic Sailor Moon:

Doing a batch of 20 and grabbing my personal favorite:

(masterpiece, best quality, very aesthetic, ultra detailed), intricate details, 4k, aausagi, long hair, double bun, twintails, hair ornament, parted bangs, tiara, earrings, blue eyes, heart choker, blue sailor collar, red bow, white shirt, see-through, elbow gloves, white gloves, multicolored skirt, white skirt, lora:sailor_moon_animaginexl_v1:0.9, standing, angel, halo

Negative prompt: (worst quality, low quality, very displeasing, lowres), (interlocked fingers, badly drawn hands and fingers, anatomically incorrect hands), blurry, watermark, 3d

Steps: 30, Sampler: Euler a, Schedule type: Automatic, CFG scale: 7, Seed: 11, Size: 1024x1024, Model hash: 1449e5b0b9, Model: animagineXLV31_v30, Token merging ratio: 0.5, Lora hashes: “sailor_moon_animaginexl_v1: e658577df088”, Version: v1.9.4-169-ga30b19dd

Time taken: 8 min. 18.0 sec. (to do a batch of 20, and with 30 steps, whereas I typically use 20…used 30 because the LoRA example image did, don’t wanna go experiment a bunch).

I grabbed some of those prompt terms from the example images for the Sailor Moon LoRA on Civitai. Haven’t really tried experimenting with what works well. I dunno what’s up with those skirt colors, but it looks like the “multicolored skirt, white skirt” does it – maybe there are various uniforms that Sailor Moon wears in different series or something, since it looks like this LoRA knows about them and can use specific ones, as they have those different skirts and different prompt terms on the example images.

I just dropped the Animagine model in the models/Stable-diffusion directory in Automatic1111, and the Sailor Moon Tsukino Usagi LoRA in the models/Lora directory, chose the checkpoint model, included that <lora:sailor_moon_animaginexl_v1:0.9> prompt term to make the render use that LoRA and some trigger terms.

That’s 1024x1024. Then doing a 4x upscale to a 16GB 4096x4096 PNG using SwinIR_4x in img2img using the SD Ultimate Upscale script (which does a tiled upscale, so memory shouldn’t be an issue):

The above should be doable with an Automatic1111 install and your hardware and the above models.

EDIT: On Civitai, when you view example images, you can click the little “i” in a circle on the image to view what settings they used to create them.

EDIT2: It looks like the same guy that made a LoRA for Sailor Moon also did LoRAs for the other characters in her series, so if you want, like, training on Sailor Mars or whatever, looks like you could grab that and also add knowledge about her in.

Well, for me, the selling points are:

• Versus earlier versions of USB, it’s reversible. This isn’t a game changer, I guess, but it’s definitely nice to not have to fiddle plugs around all the time.

• I don’t know if it’s the only form of USB that does USB PD – I’d guess not – but in practice, it seems to be pretty strongly associated with USB PD. Having USB PD isn’t essential, but it makes charging larger devices, like laptops, a lot more practical. I can lug around a power station that doesn’t need to have an embedded inverter.

I still feel that it’s kind of physically small and weak compared to USB A. That’s an okay tradeoff for small portable devices that don’t have the space for larger connectors, but I’m kinda not enthralled about it on desktop. I worry more about bending connectors (and I have bent them before).

So for me, I’d say that it’s definitely nice, but not really in a game changing sense. I could do the things it can do in somewhat-worse ways prior to USB-C.

Thanks for trying it out! Both the inpainting and outpainting – the expansion – worked better than I’d expected, though I dunno if that’s exactly what M0oP0o’s after.

Lots of odd artifacting, slow creation time and yes it had some issues with sailormoon.

It probably isn’t worth the effort for most things, but one option might also be – and I’m not saying that this will work well, but a thought – using both. That is, if Bing Image Creator can generate images with content that you want but gets some details wrong and can’t do inpainting, but Midjourney can do inpainting, it might be possible to take a Bing-generated image that’s 90% of what you want and then inpaint the particular detail at issue using Midjourney. The inpainting will use the surrounding image as an input, so it should tend to try to generate similar image.

I’d guess that the problem is that an image generated with one model probably isn’t going to be terribly stable in another model – like, it probably won’t converge on exactly the same thing – but it might be that surrounding content is enough to hint it to do the right thing, if there’s enough of that context.

I mean, that’s basically – for a limited case – how AI upscaling works. It gets an image that the model didn’t generate, and then it tries to generate a new image, albeit with only slight “pressure” to modify rather than retain the existing image.

It might produce total garbage, too, but might be worth an experiment.

What I’d probably try to do if I were doing this locally is to feed my starting image into the thing to generate prompt terms that my local model can use to generate a similar-looking image, and include those when doing inpainting, since those prompt terms will be adapted to trying to create a reasonably-similar image using the different model. On Automatic1111, there’s an extension called Clip Interrogator that can do this (“image to text”).

Searching online, it looks like Midjourney has similar functionality, the /describe command.

https://docs.midjourney.com/docs/describe

It’s not magic – I mean, end of the day, the model can only do what it’s been trained on – but I’ve found that to be helpful locally, since I’d bet that Bing and Midjourney expect different prompt terms for a given image.

Oh I also tried local generation (forgot the name) and wooooow is my local PC bad at pictures (clearly can’t be my lack of ability it setting it up).

Hmm. Well, that I’ve done. Like, was the problem that it was slow? I can believe it, but just as a sanity check, if you run on a CPU, pretty much everything is mind-bogglingly slow. Do you know if you were running it on a GPU, and if so, how much VRAM it has? And what you were using (like, Stable Diffusion 1.5, Stable Diffusion XL, Flux, etc?)

@M0oP0o@mander.xyz, as far as I can tell, you always use Bing Image Creator.

And as far as I can tell, @Thelsim@sh.itjust.works always uses Midjourney.

I don’t use either. But as far as I know, neither service currently charges for generation of images. I don’t know if there’s some sort of different rate-limit that favors one over the other, or another reason to use Bing (perhaps Midjourney’s model is intentionally not trained on Sailor Moon?), but I do believe that Midjourney can do a few things that Bing doesn’t.

One of those is inpainting. Inpainting, for those who haven’t used it, lets one start with an existing image, create a mask that specifies that only part of the image should be regenerated, and then regenerate that part of the image using a specified prompt (which might differ from the prompt used to generate the image as a whole). I know that Thelsim’s used this feature before with Midjourney, because she once used it to update an image with some sort of poison witch image with hands over a green glowing pot, so I’m pretty sure that it’s available to Midjourney general users.

I know that you recently expressed frustration with Bing’s Image Creator’s current functionality, wanted more.

Inpainting’s time-consuming, but it can let a lot of images be rescued, rather than having to just re-reroll the whole image. Have you tried using Midjourney? Was there anything there that you found made it not acceptable?

it’s just ComfyUI behaving this way. Nothing is loaded, no LLMs

Oh, okay, then I’m probably wrong on VRAM, then. On my system, it needs to actually run the nodes before the VRAM gets allocated. Sorry! I thought I had it…

Just realized that I was dyslexic with the ordering directive, which worked well with Flux apparently not understanding it, since I wound up with the desired ordering; I’d intended to write “left to right”.

Not yet! One thing that AI generated images right now are not so good at is maintaining a consistent portrayal of a character from image to image, which is something you want for illustrating a story.

You might be able to do something like that with a 3d modeler to pose characters, generate a wireframe, and then feed that wireframe into ControlNet. Or if you have a huge corpus of existing images of a particular character portrayed in a particular way, you could maybe create new images with them in new situations. But without that, it’s hard to go from a text description to many images portrayed in a consistent way. For one image, it works, and for some things, that’s fine. But you’d have a hard time doing, say, a graphic novel that way.

I suspect that doing something like that is going to require having models that are actually working with 3D internal representations of the world, rather than 2D, at a bare minimum.

it starts flipping frames between the nodes and a different set of nodes.

Yeah, I don’t know what would cause that. I use it in Firefox.

Maybe try opening it in Chromium or a private window to disable addons (if you have your Firefox install set up not to run addons in private windows?)

I’m still suspicious of resource consumption, either RAM or VRAM. I don’t see another reason that you’d suddenly smack into problems when running ComfyUI.

I’m currently running ComfyUI and Firefox and some relatively-light other stuff, and I’m at 23GB RAM used (by processes, not disk caching), so I wouldn’t expect that you’d be running into trouble on memory unless you’ve got some other hefty stuff going on. I run it on a 128GB RAM, 128GB paging NVMe machine, so I’ve got headroom, but I don’t think that you’d need more than what you’re running if you’re generating stuff on the order of what I am.

goes investigating

Hmm. Currently all of my video memory (24GB) is being used, but I’m assuming that that’s because Wayland is caching data or something there. I’m pretty sure that I remember having a lot of free VRAM at some point, though maybe that was in X.

considers

Let me kill off ComfyUI and see how much that frees up. Operating on the assumption that nothing immediately re-grabs the memory, that’d presumably give a ballpark for VRAM consumption.

tries

Hmm. That went down to 1GB for non-ComfyUI stuff like Wayland, so ComfyUI was eating all of that.

I don’t know. Maybe it caches something.

experiments further

About 17GB (this number and others not excluding the 1GB for other stuff) while running, down to 15GB after the pass is complete. That was for a 1280x720 image, and I was loading the SwinIR upscaler; while not used, it might be resident in VRAM.

goes to set up a workflow without the upscaler to generate a 512x512 image

Hmm. 21GB while running. I’d guess that ComfyUI might be doing something to try to make use of all free VRAM, like, do more parallel processing.

Lemme try with a Stable Diffusion-based model (Realmixxl) instead of the Flux-based Newreality.

tries

About 10GB. Hmm.

kagis

https://old.reddit.com/r/comfyui/comments/1adhqgy/how_to_run_comfyui_with_mid_vram/

It sounds like ComfyUI also supports the --midvram and --lowvram flags, but that it’s supposed to automatically select something reasonable based on your system. I dunno, haven’t played with that myself.

tries --lowvram

I peak at about 14GB for ComfyUI at 512x512, was 13GB for most of generation.

tries 1280x720

Up to 15.7GB, down to 13.9GB after generation. No upscaling, just Newreality.

Hmm. So, based on that testing, I wouldn’t be incredibly surprised if you might be exhausting your VRAM if you’re running Flux on a GPU with 12GB. I’m guessing that it might be running dry on cards below 16GB (keeping in mind that it looks like other stuff is consuming about 1GB for me). I don’t think I have a way to simulate running the card with less VRAM than it physically has to see what happens.

Keep in mind that I have no idea what kind of memory management is going on here. It could be that pytorch purges stuff if it’s running low and doesn’t actually need that much, so these numbers are too conservative. Or it could be that you really do need that much.

Here’s a workflow (it generates a landscape painting, something I did a while back) using a Stable Diffusion XL-based model, Realmixxl (note: model and webpage includes NSFW content), which ran with what looked like maximum VRAM usage of about 10GB on my system using the attached workflow prompt/settings. You don’t have to use Realmixxl, if you have another model, should be able to just choose that other one. But maybe try running it, see if those problems go away? Because if that works without issues, that’d make me suspicious that you’re running dry on VRAM.
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EDIT: Keep in mind that I’m not an expert on resource consumption on this, haven’t read material about what requirements are, and there may be good material out there covering it. This is my ad-hoc, five-minutes-or-so-of testing; my own solution was mostly to just throw hardware at the problem, so I haven’t spent a lot of time optimizing workflows for VRAM consumption.

EDIT2: Some of the systems (Automatic1111 I know, dunno about ComfyUI) are also capable, IIRC, of running at reduced precision, which can reduce VRAM usage on some GPUs (though it will affect the output slightly, won’t perfectly reproduce a workflow), so I’m not saying that the numbers I give are hard lower limits; might be possible to configure a system to operate with less VRAM in some other ways. Like I said, I haven’t spent a lot of time trying to drive down ComfyUI VRAM usage.