Please consider buying my two albums Sisters of the Wicked and Danger Ahead, my book Dracula and also please check out my YouTube channel. Civitai won't support creators, so any purchase is a real help. Thanks.
DRACULA https://www.amazon.com/dp/B0FM4DXYPL
Sisters of the Wicked https://amazon.com/music/player/albums/B0DSJY3ZHY?marketplaceId=ATVPDKIKX0DER&musicTerritory=US&ref=dm_sh_CdyeuxIBNBDOjZ1hpB1PPJxT0
Danger Ahead https://www.amazon.com/music/player/albums/B0DB2DV354
Music In The Machine https://www.youtube.com/@musicinthemachine/videos
Acorn Is Boning XL V3
V3 came about after I had been away from SDXL for probably close to a year. I was working with a company trying to make realistic characters for use in online chat avatars that were NSFW. They were using Lustify. In working with it, I found it to have some good qualities, but not very good prompt adherence. I decided to go back to my original Acorn Is Boning XL as the base for a new model. V1 didn't look quite as realistic as V2, but it had better prompt adherence, so I used block merging to graft on the better visual look of Lustify onto the better composition and prompt adherence of Acorn Is Boning XL. I made a few versions for testing and found two of them to be very good and fairly similar. Here I posted version 3.0 and posted v3.1 for online generation over at Tensor.Art. https://tensor.art/models/860244001473710239/Acorn-Is-Boning-XL3-3.1
You can try it out over there if you like and compare.
Acorn Is Boning XL V2
https://www.youtube.com/@musicinthemachine/videos
I would appreciate a follow and any comments or any posted results of images. Thanks.
For V2 I was able to improve the photorealism as well as the consistency of better hands.
Like with my primary model, Acorn Is Spinning XL V4, I changed my default generation resolution to 1024 X 1328. I think this change was part of the reason for the hand improvement as I get improvement in my previous models at that resolution as well.
I get tired of all the same old porn poses so in this initial gallery, I have included a mix of regular poses along with some img2img stuff and some fantasy themed porn. If you are trying to reproduce my images and can't, it's probably one of the img2img pics. Enjoy!!
My go to sampler is still DPM++ 2M SDE Heun Exponential, typically either 30 or 60 steps with CFGs from 2 to 7, more often 2 to 3.5.
Acorn Is Boning XL
This is the first SDXL version of my Acorn Is Boning model. If you enjoy it, please follow me. I enjoy working on models and hope to keep doing it. Getting in the Creator's Program is based on followers so far so I need all I can get. It is much appreciated.
https://www.youtube.com/@musicinthemachine/videos
I use some of my models for videos on my YouTube page if you want to check it out.
As with my previous 1.5 versions, I made a randomizer prompt that takes advantage of dynamic prompts. You can set the seed to randomize and run off batches. Go back to tweak those than stand out to you. Enjoy.
highest quality photograph of beautiful naked woman {(vaginal sex:1.5) with man, erection | (from above, BLOWJOB, big cock) | (manually masturbating:1.3) with (blue glass dildo insertion:1.3) | (kneeling), (receives cum in open mouth from man's cock:1.4) | (squatting, peeing piss:1.4) | (licking penis, cum:1.4) | (giving deep throat blowjob:1.4) | (having sex with another woman:1.4), (manual dildo inserted:1.5) | (POV BLOWJOB:1.4) | (having anal sex:1.4) | (kneeling masturbation, touching own pussy:1.4)} {in the bathroom | in hot tub | in the kitchen | by the pool | in bedroom | on rooftop in the city}, {ginger | blonde | brunette} {long | medium | short} hair, {petite | skinny | chubby} body, {(large breasts:1.3) | (medium breasts:1.3) | (small breasts:1.3) | (huge breasts:1.3)}, (intricately detailed skin), (perfect lips), 18 years old, high contrast
Negative: (watermark:1.5), (anime, eye contact, misaligned:1.4), (missing limbs:1.5), (blurry), (deformed iris), (deformed legs) (extra feet, extra hands), (lipstick), (two dicks)
DPM++ 2M SDE Heun Exponential is my preferred sampler but many work just fine.
I prefer to use 40 steps or more, but you can get good results at lower steps as well, at least down to 20.
CFGs from 3 to 9 all work pretty well.
Description
FAQ
Comments (36)
Thank you for your continued work on this model, and thank you for the model (especially version 1). Without testing version 2 at all, but just using the same prompt, I like the realism in version 1 of AcornIsBoning, better.
I am glad you like it. New versions are always tough. Some things seem better and others worse almost always. Even going back to my 1.5 models, there are still things I like better about Acorn Is Spinning Photo V2 over some of the later ones, even though the later ones are definitely better overall in many ways. Boning XL V1 was definitely lightning in a bottle and came out better than I was even expecting. Some things are definitely improved with V2, but as always, there may be some things it doesn't do as well too. All feedback is welcome though once you play with V2 a bit. Of course V1 will always be here if you end up still liking it better.
I'm gonna give it to you straight. I've been working with V2 for the last couple hours. There's a marked improvement in skin texture quality, but I am having enough issues that I'm starting to feel like it's a side-grade to V1 more than an upgrade. Understand, everything i say comes from a place of love, I am your #1 fan. Consistency of anatomy is worse than V1. Hands, feet, genitals. I chose Acorn is boning for my merged model primarily because of it's amazing prompt adherence and very consistent hands. Prompt adherence is less flexible in V2. If you ask for a pose that generally has some variation between iterations, you only get it one way with V2. That is a bit of an exaggeration, but it's close to the truth. Even my example image I uploaded, when using the term kneeling, there are a very limited number of very specific kneeling poses that appear. 1024x1328 consistently results in elongated bodies for me. It certainly ups the pixel count, but it comes at too great a cost. Lastly, loras barely work. I tried at least a dozen loras and the model is very resistant to them. The lora's I had success with were custom loras that i trained explicitly on acorn is boning. I used one of them with one-piece swimsuits on my example image. My merged model is also a nightmare with loras. I caused that by merging the model with too may loras. Maybe that's the cause here? Overall, this is a great effort, with some clear signs of improvement, but it was frustrating to get results that with V1 are more easily obtained than with any other model on civitai.
That is interesting. I wonder if it is changed based on the software or systems used. I get better and more consistent hands and no elongation on the bodies with V2. For a long time I used 1024 X 1280 and when I switched to 1024 X 1328 I immediately got vastly improved results both on Boing and Acorn Is Spinning. I use ForgeUI for all of my generating and definitely get different results if I try to generate the same pic using Comfy or some other software. With past models and especially back in the 1.5 days it didn't make as much difference from one software to another. It's always hard to put out a new version though because there is almost always some things that seem better and usually some degree of trade off with other things that don't seem as good. I will keep experimenting though of course. I suppose I really should test models on multiple different software but I have just found ForgeUI to be vastly better than others so I tend to only use Forge. Of course I do this mostly as a hobby and because I am a tech guy. To this point I don't get any income or anything for my models, so I guess I don't have a lot of incentive to test on Comfy or others. I know a lot of people love Comfy, but to me it is just a frustration to deal with it. To me the UI should all be there without having to build it yourself. :)
I haven't tried Loras with the new model yet. I didn't merge any Loras into this version either. I use block merging when working on new models and usually am only merging in the image texture most of the time. I will admit, my methods are normally focused purely on improving image quality over other factors like prompt adherence. Since I first started working with models, my goal has always been photo realism. My pet peeve has long been those people who name their models "realism" this or that and then all the output looks like Pixar. My goal with a V2 of Boning was to get the look closer to matching my main Acorn Is Spinning XL models in image quality. I love Boing XL V1, but it has always bothered me that the photo quality is noticeably less than my Acorn Is Spinning XL models. V2 definitely is a step closer but I might have to experiment more to see I can sustain the image quality without affecting the prompt adherence.
@Seeker70 I sorta took off on a rant, but I didn't intend it as a complaint. The reason I am a big fan and appreciate your work is we have very similar goals. Photorealism is my priority, over prompt adherence or flexibility, and your work is about the best there is. I was using comfy. That clearly accounts for the issue with image dimensions. I have both and am happy to use forge with V2. To compensate for my rant, I'll train a couple of my lora datasets against V2 to produce superior quality loras for the model and upload them for everyone. I'll upload something over the weekend. And lastly, I'm finishing up my V2 merged model (which includes your work, so, thanks) and I got an idea from LeoSam that turned out to be very interesting. He merged the SPO lora into his latest model and felt it made a big difference. I tried it on my merge, and I think it goes a long way towards better hand and feet accuracy, as well as color balance and other photographic characteristics. Try adding it to one of your merge projects and see what you think. https://civitai.com/models/510261?modelVersionId=567119
@gurilagardnr I will take a look. I'll admit I am not much of an expert at merging Loras into the model. Before I did Acorn Is Boning XL, I did try merging NsfwPovAllInOneLoraSdxl into my Acorn Is Spinning XL V2. I had some help from TripleHeadedMonkey on how to set up Comfy to merge in Loras. It did work, but it wasn't good enough at the time for me use it as my first SDXL Acorn Is Boning model. I shelved that for a while and eventually when I looked at a few of the newer porn models I decided on onlyfornsfw118_v20 and did about 10 different merges with Acorn Is Spinning XL V2 and that eventually ended up being Acorn Is Boning XL. I mostly use Super Merger and Merge Block Weighted extensions in Forge to do my merges. Unfortunately the last update on Forge seems to have broken something and my Super Merger tab is gone. I posted on github about it but haven't seen any fix yet. Luckily, I have 2 separate installs of Forge on my laptop. One updates and one I have to manually update. The old one is still working, so I will use that one without updating until they fix the new one.
What do you normally use to merge Loras into models?
@Seeker70 Forge is under major construction right now, definitely don't update your working copy. I have used both kohya-ss and comfyui to merge loras. I'll make you a straight-forward workflow that you can just load into comfy and use. It's not hard. Just select your models, adjust your ratio, and click go. I'll upload something and link it here later today.
@gurilagardnr That would be cool. Thanks.
@Seeker70 well, i've been fighting with the checkpoint-lora merge comfy workflow for an hour, something isn't workin, so, there will be a delay. Sorry. Worked Saturday. I dunno, might have to reinstall. Not gonna send you something I can't confirm is working. stand by.
@gurilagardnr No problem. Anytime you have it is fine.
@Seeker70 It was a bug, they fixed it in the overnight update. Actually, it's ridiculously simple. Open up comfyui. Double-click on the screen, search and load these three nodes: Checkpoint Load, Lora Load, Checkpoint Save. That's it. Connect them together (in that order) and you can merge lora's into a model.
@gurilagardnr I'll take a look at it. My Comfy has been acting odd. I was trying to get Flux working on it and haven't been able to yet. I mess with it again but I may not have time today. Have a AD&D game tonight that goes until the late hours. :)
@gurilagardnr yeah, I faced the same thing yesterday with "checkpointsave" node. Had to open an issue on github, and after that comfy fixed it in like 15 minutes. https://github.com/comfyanonymous/ComfyUI/issues/4281
@coyotte yup, that was the error. Thanks for saving my sanity and tapping comfy on the shoulder about it.
Just wanted to thank you for v2, very subtle changes and, but in some cases incredibly good with the right prompt. I use v1 and v2 and hope that one day there will be a flux version :D
Thanks. I am glad you like them. I think V2 has considerable better photo qualities, but V1 might be better for prompt adherence. Since I started working with models, my goal has been to push for the bleeding edge of photorealism over everything else, so I think V2 is a good step in that direction even if it may lack a bit on some other areas. Luckily, anyone can use either or both. As for Flux, we'll see. I generated my first flux image about a half hour ago. I don't know yet what the differences will be in trying to work with flux for merges and how long til there will be anything to merge with. I will keep an eye on it. https://civitai.com/images/23861104
once upon a time there was a model...
months ago i downloaded a really good nsfw model named aphroacronxl_v10. it was a good model and i liked it a lot. one day i went to see if there were any updates to it and it was gone. my heart was broken. but i still had my copy...so i used it heavily.
one day for shits and grins i searched the internet for my favorite model. what could have happened to it. after much searching, i finally found it hidden on another site. still the original version. much to my surprise, the 'author/poster' said it was a merge of afroditexl_20Bkdvae and acornIsBoningXL_v10. so i scurried back to my 'safe place' civitai. i then searched for these two models. YAY, again to my surprise these two models still exist. what's better is they both have been updated. so now i'm downloading them both.
i'm a happy little camper. once downloaded i will merge them to try and recreate the model i was so fond of. then, all will be well with the universe.
and we all lived happily ever after...
the end
I really liked aphroditexl when it was first out. I had my main model Acorn Is Spinning XL and I don't remember if XL V2 was out yet. There were a number of porn type models out but none of them were super realistic looking. That's why I decided to make one with the base of my Acorn Is Spinning XL V2 photo model. I did still have the old aphroditexl model on my computer and it was in the mix for Acorn Is Boning XL. Much earlier, there was a model called Venus that was one of the first models to add in good nudes to SDXL. It was pretty raw and not really a very good checkpoint except for the raw image texture which I mixed into what was at the time Acorn Is Spinning XL. So anyway, eventually Acorn Is Boning XL had the base of my Acorn Is Spinning XL, Venus, AphroditeXL and onlyfornsfw118 and probably a little bit of some others. Unfortunately I am terrible at having a written recipe for any of my stuff. It's just a lot of mixing and testing.
@Seeker70 i may have to try acorn is spinning? i really like the nsfw content. it's nice having a model you have to work to get girls to close their legs... :) usually i have to (((prompt))) them to spread their legs.
@tedbiv Yeah, to me Acorn Is Spinning is my primary baby. Acorn Is Boning in the 1.5 version was an accident. I was working on a new Acorn Photo 1.5 model and just ended up with one version that did women and men together really well. That original one never had a huge following, but when I came out with Acorn Is Boning XL, that one really took off. Way more downloads than any single version of my Acorn Is Spinning models. I like both, but my goal as always been to be the ultimate photo realism model. I was trying to bring Boning closer to that as well will V2 and I think I did, even if it may not be quite as good with prompt adherence.
great model! are "selfie" type photos viable in this checkpoint? been using the weighted "selfie" keyword, but it seems to not result into a selfie type photo. any help?
What did you merge into Acorn is Boning XL 2?
To be honest, I can't remember exactly what went into it. I experimented a lot using Supermerger. I believe the main addition was blending in a bit of NatVis: Natural Vision in the image texture blocks while leaving the composition blocks as Boning XL V1.
@Seeker70 Thanks. Which are the texture blocks and which are the composition blocks when doing block merge?
@SubtleShader Output layers are more impactful for aesthetic.
Input layers are more impactful for composition.
Fine Details Layers: Minor, but important aesthetic polish. Sub surface scattering, shadow textures, eye lashes, nipple textures, etc
INPUT 00 -> OUTPUT 11
INPUT 01 -> OUTPUT 10
INPUT 02 -> OUTPUT 09
INPUT 03 -> OUTPUT 08
INPUT 04 -> OUTPUT 07
Major Details Layers: General Aesthetic and Flair, Photographic or Anime
INPUT 04 -> OUTPUT 07
INPUT 05 -> OUTPUT 06
INPUT 06 -> OUTPUT 05
INPUT 07 -> OUTPUT 04
INPUT 08 -> OUTPUT 03
INPUT 09 -> OUTPUT 02
Core Composition Layers: Pose, Outfit, Background, Base Elements of the Image
INPUT 10 -> OUTPUT 01
INPUT 11 -> OUTPUT 00
M00
https://civitai.com/articles/2370/model-merging-management-how-to-merge-stable-diffusion-models-to-fit-your-style
This is the guide I referenced when I was working on my SDXL models a lot. Tons of info, but a long read.
@Seeker70 Thanks. Sounds like you used the GRAD_V preset.
@SubtleShader I tried several. The Wrap16 is one of the major ones for putting the image quality onto the other models composition.
@Seeker70 Thanks will try! Which would you use the put the composition on other models image quality? Just reverse Wrap16? Or swap the models?
@SubtleShader @SubtleShader I think the way I did it was to put the porn model in A since the composition is the part that handles the things such as poses. I put my Acorn Is Spinning XL V2 in as the B model to try to transfer the photorealistic quality of it over to the porn model that was more cartoony than photo. It worked out pretty well. My Acorn Is Boning XL V1 still seems to be about the best porn model on the site. I was just looking through some of the other SDXL porn type models and a lot of them look pretty terrible.
@Seeker70 Thanks for the information, Seeker70! Unfortunatelly, did not have much luck with block merge. Something tells me it works better with anime models and less with photorealistic models. Anyway, I will release a new version of my Big Love model today. It includes Acorn Is Boning XLv2. I love the cute faces and spontanity that it adds.
@Seeker70 the link "https://civitai.com/articles/2370/model-merging-management-how-to-merge-stable-diffusion-models-to-fit-your-style" is 404 now. do you know if it has a new location? search didn't bring anything up
@kunde2 No, unfortunately I don't know.
@Seeker70 the guide is gone anybody save or copy it?
@HiProfileArt Here are the parts of the old guide that I had copied that I had found the most useful. I made it into an article. Hope it is helpful. https://civitai.com/articles/12503/super-merger-usage
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.