Trigger: chairtied,
General: chairtied, tied to chair, sitting, bound, bondage, bdsm, rope, arms behind back, bound legs,
Description
FAQ
Comments (14)
Chair ties are among my favorites. I really like this LoRA but I have a suggestion. Maybe it's feasible, maybe not:
There's a time and a place for the arms to go over the top of the chair and behind, through an opening in the chair and behind, or for the upper arms to stay in front and to the character's sides. This wouldn't solve everything, but there should be a way to train the data on "arms behind chair" or "arms through chair," to make it easier to generate thhe desired position. I might have had a little luck using "arms behind chair" in the negative prompt, just because Illustrious can sometimes understand positioning prompts.
I understand what you mean and I confess I didn't really consider that. In this case though I got the training images/txt files from a friend so I only trained the Lora.
I might try gathering/tagging images myself for a more flexible chair tie Lora including having the arms tied to the chairs arms as well but for now I have no plans train a new version.
Also if you like bondage themed images/Lora might I interest you in a bondage themed Ai Discord server I'm in? We have plenty of people who makes images and Lora for bondage, ties, gags, poses and more. No worries if you aren't interested of course.
Interested? Sure am. But I'm an alter ego, so I may not be joining until I create a separate Discord account.
@gagologist Well here's a link if you ever make an alt account, we always welcome fellow bondage fans. https://discord.gg/GKmeeu6u
Looks interesting. I'm curious.
What are the main points this LoRA adds, over just using base Illustrious?
I'll test it later myself, but just thought to ask first in case a compact answer happens to exist.
One issue I have with base Illustrious with chairties is that it likes to draw characters tied on a chair, not tied to a chair.
Particularly, having the character's legs tied to the chair legs, or having her legs together and bent so that her feet are under the seat, seem to be out-of-distribution for Illustrious. Also, in base Illustrious, it's difficult to pose her arms to behind the backrest.
It's more to help get a chair tie pose consistantly but if you can get a good chair tie on base IL then you might not need this Lora.
@Lancelot69lol Thanks for the info!
The main issues I've had with base IL with this topic are getting the character tied to a chair, as mentioned, and the general lack of understanding for perspective in SDXL (chair legs not making geometric sense, etc.).
I'll give the LoRA a whirl when I find time, to compare how it fares against base IL.
@Mathemagic Sorry for the late reply and I hope the Lora helps.
@Lancelot69lol ...holy cow.
Tested this last night. Let's just say... strongly recommended.
Images and a review to follow later when I get around to it.
@Mathemagic I'm super happy to hear that! Also if you like bondage Ai images I am in a Discord server where we make/share both images and Lora related to bondage. If you happen to have any interest in that let me know and I'll send you an invite but no worries if not.
@Lancelot69lol I do like bondage AI images, but I don't use Discord much. So maybe later.
It does sound like a good idea to have a sharing platform, given the hostility toward GenAI in much of the internet. I'll grant the points about a deluge of low-effort AI slop, and human artists not wanting to be imitated by a machine. But as for images that look as good as hand-drawn ones - ones where the creator actually did some inpainting to enhance detail and fix the AI's mistakes - I don't get it. Isn't GenAI just another new tool for creatives to do their thing? That's actually why I post my images on CivitAI - this place is built around GenAI, and fellow enthusiasts tend to gather here.
But it's difficult to get an audience for your images, and even more difficult to find nice images to look at. I suppose some LoRA pages work as aggregators, but then, Illustrious-XL requires much fewer LoRAs than previous model generations did (which, on balance, is a good thing).
This reminds me, another LoRA creator @SimianHose said on their page that everyone wants to make AI art, but no one wants to look at it. I find this a curious phenomenon. Maybe it's that every creative wants to fulfill their own artistic vision, and GenAI brings the time cost low enough to make it feasible as the primary way of getting new images that align with one's own tastes?
But I can also see that the site UI is missing some useful features. Particularly, there is no way to track new image posts by creators you follow (or even, by specific creators, and/or by specific tags).
As for just searching, the tags tend to be rather generic. I don't want to search for all BDSM images if I want to see clothed chairties, specifically. Some things I consider important don't even have their own tag, and many that I'm indifferent to, do have a tag.
Anyway, I'm getting offtopic. I'll return with a set of images in the near future. :)
@Mathemagic I view Ai as a tool one that artists can use to greater effect than those who lack drawing skills. You make lots of good arguments as well and I agree. I like looking at high quality Ai images but do find myself sometimes bored of it and look for an artists image.
Searching for Ai images can be tough as like you said tags are often just to generic. I can totally see the point that @SimianHose made as well. I've wanted to draw things for a long time but lack both the skill and time to learn. Ai is a super easy way to fulfill my desire to "create" art without the years of learning I'd have to do.
@Lancelot69lol I meant creatives in a general sense, including also e.g. composers and programmers. GenAI has potential to augment visual artists, but it also levels the playing field so that one can now substitute other types of technical skill in place of drawing skill, thus opening up visual art for non-visual creatives. Even just having persistence and an eye for detail goes a long way now.
At the same time, I understand where visual artists are coming from. I don't want AI to write my code for me, either. It's just not good enough just yet. But very similarly, vibe coding is a thing now, and that changes the dynamics.
I agree that GenAI is easier than drawing. It has certainly democratized some aspects of the level of illustration that used to be possible only for professional CG artists.
But I wouldn't say it's super easy to get satisfying results, at least yet, with the SDXL architecture (including Illustrious). A technical element remains, and some images still take an hour or more to create (even though that is very fast compared to comparable manual work). In complexity, I'd say that GenAI as a hobby is on par with DSLR photography. And it's still not easily controllable. Creating consistent OCs with GenAI is hard.
Personally, I've almost exclusively switched to GenAI. I rarely bother with human-created images any more. My tastes are far enough from mainstream that searching for anything interesting generally takes longer than just creating custom art with GenAI, inpainting time included. I often want some specific combination of multiple elements, so the real options are either GenAI or drawing. This is what I meant by creatives perhaps preferring to fulfill their own artistic vision rather than looking at art made by others.
But it is still interesting to see what others create with this tech, as well as the workflows they use to arrive at their results.
I think Anderson was right: the 21th century will be the century of mass customization. Right now, the S-curve is just barely lifting off. I'm no futurist, but if I were to predict, I'd say things are going to get pretty wild.
@Mathemagic I couldn't agree more, honestly you're on point with this topic.
I really feel you with just making your own images being easier then searching as well. Hopefully people from all different fields will use Ai to enhance their own skills. The creativity will be explosive.