tbf, it may have been implicitly trained, designed only to add this style to whatever you prompt

@Grim0r is it?

@Grim0r If it is, then it can be explicitly stated in the details. My initial tests didn't yield good results in V1 so I dug through the couple display images and inferred what that keyword was. Why go through the effort of making something potentially so good, then not spending a few seconds to write up something that helps people use your resource better?

@Kaladae Exactly. It's not supposed to be a puzzle, lol. The programs are hard enough to figure out. I just got A1111 working the other day, then spent days screwing around with the dependencies, just to end up resetting it back to default because it turns out that the people that make the Xformers(memory optimizer for A1111) are way behind on their shit.

you can extract the tags or the keywords this or any other lora trained on if you drop it here https://lorainfo.tools/

@Romlo I believe some of the civit tools for A1111 can also show lists of trained keywords used and their frequency. I have over 30K resources downloaded, and the ones that don't have immediate access to specific details, I tend to ignore, and eventually delete. An extra 30-45 seconds even across that much (I'm pathetic and have no life obviously) means a lot of extra time, and a lot less probability to actually use something. But that's probably a great tool for others as well if they're willing to invest the extra steps.

@Romlo Thx. Will have to check it out.

@Kaladae 30k.. Jesus... Thought I was bad for downloading 2-300 gigs in models/loras over the first couple weeks. I'm off A1111 a while ago though. Since then, I tried SD Next, and been on comfyui for the past few days. comfy is best, least inherently restrictive, and very customizable. Just takes extra work to set it up. But if you insist on a more easy-mode approach, I still recommend SDN over A1111, it's a big upgrade. though, so far nothing beats the PNG info tab of A1111, it worked perfect for pulling metadata from images and sending it all to the right places. That aside, most things are better in SDN, including the dependency shit. Still needed an older version of python I think.. 3.12.3 iirc.. But newer than A1111, and everything else worked better.

@Lazman Lol yeah I literally have about 24TB of stuff downloaded. I've actually been slowing removing some of the ones I definitively don't use anymore. Mostly SD1.5 and SDXL that isn't pony. A1111 used to be completely unusable for me, it just ran like liquid dog feces. Fortunately with the last update from Forge rolling into it, it's helped a LOT also since I found to not use VAE full in the options, A1111 is my go to now. I do a lot of inpainting and tweaking (usually) after the initial generation though so the layout is more intuitive for me, everything is right there and I don't have to have different layouts for anything I'm trying to do. I may try SDN, but so far A1111 has everything I desire to use, and since the VAE things, it runs a lot better too. I get my little display pictures that I can change to remind me which does what, and I don't have an impossibly long navigation menu that requires me to scroll for hours with just names using Comfy, not actually ever remembering anything by name like I used to. Granted, the amount of resources it has to load slows startup down by a considerable amount, but using things I want to is worth waiting a little extra over just using the exact same thing I know works most of the time. I actually have to upgrade my server storage soon now rofl.

@Kaladae Yea, NGL, 24tb in resources for this, is just retarded. I mean, tell me that you're at least doing this as a career, or running a big project of some kind. But if you're just downloading with the pokemon style 'gotta catch em all' mentality of getting every lora/model you see, just to make nudie images of your favorite characters, then ya really gotta rethink things. Cuz a lot of things can be recreated with much fewer resources.

For one, most models are just basically oversized loras. They took a base pony/sd/sdxl model, slapped a style on it and gave it a name. In many cases, the exact same things can be produced with the base model, plus maybe an extra lora with a similar style. knowing this, I've become pretty picky about which new models I choose to download. This also helps when it comes to sorting them out.

I agree that comfy has a crappy system for going through models/loras. I either gotta memorize the name of what I want, or scroll a lot to find it (especially in my loras, which I do have a lot of). But if you're gonna insist on using a restricty thing like A1111 anyways, you could try comfyui with the workload/tab manager thing, I imagine someone may have made an extension for better sorting the models/loras. In-fact, I'm certain I seen one that looks similar to SDN, in which it has image thumbnails for each of your models. Best one I liked for loras was easydiffusion. Sadly it's abandonware at this point. But I loved the lora thing for that, especially when I got the extension for saving profiles. But, alas, pretty much every software is better in nearly every other respect.

But if you must use a full on 100% user friendly GUI, SDN is the way to go. RIP to the PNG info tab from A1111, but if you didn't use that, or don't care so much about it, nearly everything is better in SDN. Especially the setup, since it doesn't rely on an outdated version of python(iirc..). Well, I know comfy uses the newest version of python, I forget if SDN used the newest, or just newer than A1111. In any sense, I remember it having an easier setup.

in any case, it's an aweful waste to have that much resources, and be using a dated software to utilize them. Much of a pain as it may be to set up, once you set it up and save your workflows, you can reuse them at any point, and do many more things in one workflow, as opposed to doing them in several steps in another webui.

Side note; I just reread the OP. Turns out I was wrong. If you go to a seach engine, Brave, Google, w.e, and type in 'lorainfo' like that with no spaces, a site should pop up, top of the results that allows you to drag/drop loras into a box, and it will give you a full list of tags. The jury's still out on whether editing those tag weights and re-saving the lora will actually affect the outcome of image generations made with it, but at the very least, it's helpful to know what the lora is geared towards, and what keywords to use for it.

@Lazman Nah, just a hobby. I just go full out on the hobbies I enjoy, that and having a few spare servers laying around, it gave me a chance to play with them and finally learn the software side of things. When I lose interest enough or find something I enjoy better, they'll probably be completely wiped and recommissioned. That just happened to be 24TB. I have been clearing out some things and reducing clutter, it's just taking a while. I also do hold to the notion that just because something is newer doesn't mean it's better. I do play with a lot of the ones I've acquired, some more than others. My tastes and styles can change frequently, so I like having options. Pony/XL tends to be a little heavier so that's fewer actual resources per same storage requirements as SD1.5, which I have been cleaning out. It's a time waster to help relax and take my mind off of chronic pain, so consider it retarded or not, the goal isn't always to just do crap the same way a thousand different times. It's not to me, and that's all that matters :)

I do use the crap out of the PNGs for A1111, so I won't be looking to lose that. I still have the last good Forge version installed that runs much better, but lama cleaner doesn't work, so I've been defaulting to A1111. I do still go back to OG Forge though because it works, and it still works well, even if a little dated. I don't mind if things get a little complex, I just want to be able to quickly and easily make stuff. Like I said this is a hobby for a purpose, doesn't have to be good or amazing or anything, I just need it to relax and distract me, like youtube, except I get a lot more engagement here. I just want to be able to jump in and accomplish what I have floating around in my mind. At least, as best I can explain it to the AI anyways.

And then the gremlins kick in and run amuck and I have to make it better and inpaint and tweak and "what if" and I just find A1111/Forge best for that. Lol.