Adult Cartoons?Last Thread: >>7008097>>> The one link to rule them all <<<rentry.org/sd-mashup^ almost everything you will need is linked here, for beginners and advanced users.Check that link before asking noob questions.>> Quick and easy Google Colab running the latest Web UI <<colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb^ Command-line arguments already set for the best experience, includes a VAE, literally idiot proof, specify a path to your models if you honk, just click the play button to run each cell, and click the gradio link in the final output.>> Washed out colors? Get a VAE <<The "one link to rule them all" above has some in the models link here rentry.org/sdmodels^ Place them in the VAE folder for webui, then you can select one in the settings tab.>> Getting empty black images? <<Add this to your command-line arguments: --no-half-vae>>> More useful info <<<github.com/AUTOMATIC1111/stable-diffusion-webui/wiki
>>7020609catbox for this? Do you have a LORA or some keyword for consistent faces or is this just the sampler doing sameface for everything?
>>7020579anyone have an idea of what model/merge this guy might be using?https://www.pixiv.net/en/artworks/104619957
>>7020732>an idea> Model hash: 4de704d8
>>7020736How can you find models by hash if you don't have it locally?
>>7020736oh does pixiv retain the sd metadata i never thought to check an image i assumed it got stripped
>>7020740Sometimes it will be in one of the images of a set.>>7020739Maybe google or get lucky in recognizing it.
>>7020741I tried Google before for some hash and got nothing, but for this one it says "basil mix" so that works. Civitai should definitely let you find by hash but doesn't seem to.
I swear hypernetworks were on this list before.
>>7020683https://files.catbox.moe/qleifo.png
>>7020762Hypernetworks are now loaded through prompt, no? You can use Prompt S/R
>>7020928>>7020929>>7020930>>7020931>>7020933plz no more of these pixelated works
i am once again asking for the model and catbox
>>7020929What is this bimbo model?
>>7020928Nice style. Did you pixelate with another tool or did you find the right prompts for this?
>>7020579Does anyone have an up to date guide on how to properly install dreambooth and LoRA extensions on Linux? I've been trying to get them to work for like 2 days and I I just realized that the guys that keep getting shillled here (as if they are not written like utter dog shit) are written with Windows machines in mind, which probably explains why even using the extensions install tab in the web UI doesn't install anything, even though I see the sd_dreambooth_extension folder in the extensions folder. Does anyone even have this shit working properly on linux? Paper space uses an Ubuntu vm so it's starting to seem like it's currently a lot more difficult if not impossible to train LoRAs on it because people love to assume everyone has access to a multi-thousand GPU on hand. if somehow someone else was able to get this running on Linux how did you do it and what am I fucking up? I've been going back and forth between literally every single guy linked here as well as YouTube and the training either fails for unknown reasons or the extension flat out does not show up on the web UI tabs.
Hi everyone, researchanon from /g/ here:I just released a new massive in-depth tutorial on advanced merging.It's still WIP, but should still contain some stuff that you probably didn't know:https://rentry.org/BlockMergeExplainedAny feedback is appreciated!
>>7021441it's a custom merge using Babes from civitai and a temogam dreambooth checkpoint>>7021462I used the pixelart extension from automaticwork really well
>>7021557Damn anon, cheers for this!
>>7020931Is there any chance you can upload your custom model. Checking the archives it looks like the temogam models were taken down. There’s a particular character I’m interested in generating a Lora of using your model. If not no worries as it can be kind of a drag to upload large files.
>>7021562Interesting thanks.
>>7021557can you do one on advanced prompting?
>>7021587thanks>>7021751I believe there are enough prompting guides out there already, so it's not a priority. I might at some point, but for now I'll focus on polishing and finishing up that merge guide
>>7021717Sure, here's the modelStill a work in progress thohttps://mega.nz/file/YoFDmIYT#rKRTQ5qHsHvFEzjz5z7XiV9OZ9UCzRJMZzUmztER9b0The model is pretty versatile, it knows and follows the SD artists and styles, while maintaining the bimbo look of the subjectYou'll need to use "temogam style" in your prompts to get the full bimbo look, and it requires a bit of wrangling with the prompt to get some nice resultsBeware, it can sometimes output some hideous fleshpiles.This model is a merge using Babes, Temogam Dreambooth, Dreamscape, SD1.5, Fellatrix Bimbo Dreambooth, Melkor Manchin Dreambooth, and god-knows what else
>>7021769Awesome, downloading now. I guess lastly, do you remember the prompts you used when making this picture? This is the specific style I’m going for, but it has been a couple of days so if you don’t remember it’s all good.
>>7021773(At least, I think this is one of yours)
>>7021773here's the catboxhttps://files.catbox.moe/2d9voj.jpegit's not the same but I used the same prompt/seed/etc. with the modelwhy?because I deleted the original model by mistake, and I had to remake it from merge history, but I think I did not use the correct merge ratio for a stepit's a bit different, but not muchpic related is the original, catbox is with the model
>>7021769Can it be merged with gape or anything gape for some nsfw stuff?I tried a merge now of regular babes with gape and well, I dont know what to say It might be good?
>>7021557good shit anon this is super informative and interesting, thank you for posting
>>7021769Finally fellatrix gross shit was useful for something eh.
>>7021793you can merge it, but be careful with the ratios cause it can happen to remove the bimbo stuff
>>7021778Fantastic, I really appreciate all your help. Getting the reference folder together and crossing my fingers at the moment. If my Judy Neutron model turns out good I’m gonna be one happy camper.
>>7021802I wonder what works. I already have a pretty decent anime nsfw one.But a large insertionish, belly bulge bimbo model would be really nice.I still think URPM and moistmix are the best modesl I tried yet in terms of putting something decent by accident.Although I have yet to try with real people etc.
Why does nobody ever talk about CLIP guided diffusion?https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/5902I've kept this PR in my local repo since it was released but nobody seems to be working on it or interested in it. I wish I were smart enough to improve it because as it stands its memory-hungry as fuck, but I've gotten all of my best results with it turned on.
>>7021769>"temogam style"What is that style?
>>7021812it's the magical dreambooth keyword that makes the bimbos appear on the screenit's a porn artist that draws mostly bimbos
>>7021816Oh an artist\style prompt?I see. I was getting more anime Look.
>>7021806
>>7021823we must go even higher res
>>7021803I’ve got a good feeling about this. Just seeing the quality of the images it’s generating off the images it’s training on look super good.(Dunno why the Lora drop down to convert to ckpt hasn’t appeared yet, but it may happen after. The first times I did it I sorta wigged out the Webui so it may have appeared early)
>tfw too stupid to into stable diffusion>no 4chan bros will make me a pic when i request it>just want to see lot latina chicks in american flag bikinis and stockings doing ahegao faces
>>7021916it's easy, the only thing holding you back should be hardware
>>7021921Pretty much this. Just find the most recent YouTube tutorial on installation and use and you should be good to go. It seems more intimidating than it actually is.Hell, now you can just download a zip with it all set up and read the features on the wiki.
>>7021769You don’t have a list of the prompts this model uses do you? Some of the images generated from the training look just incredible.
>>7021921>the only thing holding you back should be hardwareNope. Thanks to services like Colab, you can use it without a powerful machine. Stable Diffusion requires only a little bit of effort to use.
>>7022023
>>7022024
>>7022026How can I make a full body shot? I can't force it no matter what. If I try to force it, it's still a matter of luck. Maybe it's the model?
>>7022011>>7021921>>7021931okay while try and figure this shit out will someone PLEASE just make a latina wearing an american flag bikini and stockings? it will be motivation as i try to make my own
>>7022031combo of model, prompt, and dimensions. go 576x768 and put a framing prompt like "full length portrait" or "cowboy shot" in the front. also if you get a torso shot you like, send it to img2img and do outpainting mk2 with only "up" checked, the recommended settings, and an unchanged prompt.
>>7022067Best I can do.
>>7022179this is great brother. if you could do one where she isn't super THICC and just normal level sexy you'd be a godi'm trying to learn how to do it myself but my computer has an Intel Irix Xe card, I don't think I have the hardware to do it. I'm doing more research
>>7021916she's disappointed in you
>>7022192this is good but the face is busted. how are you making these? if i don't have the hardware is there a site I can do it through?
>>7022195im suing stable diffusion, my rig is weak though, only 1080.
>>7022199damn anon this is exactly what i am trying to make. how are you doing this? i'm new to this so i'm trying to make sure i have the hardware in the first place
>>7022179Can you upload it in catbox? Have this.
>>7022199using not suing kek>>7022202Stable diffusion, I followed the guide in sticky and i Have a GTS 1080SC so I'm limited
>>7022224i was just doing some research and watching some how-to videos and determined i don't have the hardware to do stable diffusion on my machine. i probably could but it would take hours rather than minutesi have been looking into upgrading my computer this year so maybe i can look forward to that down the lineone last request, can you do a latina in a flag bikini but in a cartoon/animated style?
>>7020766And for this, please.
>>7022238>>7020766something like this style?
>>7022216Sure!https://files.catbox.moe/q2c93c.png
Replaced URPM with PoV Skin Texture and kept grape1 and AOM2 https://civitai.com/models/4481/pov-skin-texture-dreamlike-r34
>>7022322this is that shit i like
Out of curiosity how long should training a Lora with around 7,000 steps take? Can’t tell if my settings are optimized or not.
>>7022352On Colab with Network Dim at 128 and Batch Size 2, it would take around 1h20m
>>7022364>>7022322p-post her in an American flag bikini and thigh highs please...
>>7022352> 7,000 stepsWhy?
>>7022364Yeah, I’m gonna definitely need to look into optimizing>>7022388I used the recommended settings for making a Lora based on a person, but it doesn’t look like the recommended were optimal. I also ran into an error so all my time waiting to see if my model would be good to go was wasted:TypeError: stat: path should be string, bytes, or os.Pathlike or integer, not listThe settings and everything were the same as the last Lora I made so I got no idea what happened. Maybe because my source checkpoint was a safetensor file instead of a ckpt? But that doesn’t seem like it would make that big of a difference.
>>7022388If you train a set of 200 images, with loop count of 5 at 7 epochs, it would take 7000 steps. I don't know about him, but I'm liking the results when the steps are between 5k and 7k.
>>7022440Yeah, my training is not going too good. I think I need to scrap everything and work a lot smaller till I get the basics down. Pretty disappointed in my model but oh well. Tomorrows a new day.
i have a 144 mb .safetensors file, is it an embedding, hypernetwork, or lora? and where do i put it?
>>7022550Lora. This is where you put it stable-diffusion-webui/models/Lora
This shit is surprisingly fun. I recently bought a 3080 with 12gb of vram, so i realised i could finally do AI shit, and so far i'm genuinely quite impressed with what i'm able to make so far. Sure sometimes it's so botched it's goofy, and other times it's just subtly wrong, but it still produces an image i could jerk to.Only question now is, for these images where the only messed up thing is say an erroneous arm/leg or messed up hands, is there any way to fix it? I've heard of inpaint but idk how to use it and stuff. any tutorials?
>>7021832more pixels
>>7022600thanks. it seems to put the girl into crazy overcontrast but a little inpaint can improve things
this one could plausibly have been made from scratch
ur tribe has been captured by busted amazons wat do
>>7021769Thank you for the experimental generosity Temoganon. Do I have to use a custom VAE with this or can it just work with the model itself?
>>7022195google collab
we are belle now.
>>7023405
>>7023405Do you actually find that attractive or are you just making fun of people who are into deformed looking human beings?
>>7023407You're funny
>>7023409Wish more girls thought so.
>>7023405>>7023406>>7023409love me some thick thighsgawdamn
>>7023113Model and/or hypernetwork used?
>>7022785>>7023405Elaine, you look like one of the JuJu women of BurmaPeterman.jpg
Is there a real-life model that obeys complex positions? I tried some from danbooru. I am struggling with UPM, HASDX and f222.>>7022116It's working now.. thanks for helping me, anon.
Curious question is this software hard on your graphics card or SSD.Outside the out of memory issues which there are work around for upscaling. It does not seem to be. But when its working it does seem to be quite heavy.
>>7023578Try LORAs. I haven't used them but people get nice results with them
>>7023569>>7023570>>7023571>>7023573>>7023576>>7023577absolutely not a fan on the bimbo style
I tried to train a lora and it gave me a few folders called 1000, 2000, 3000, etc. Each one has a sample grid, a lora_unet.pt, a lora_textencoder.pt, and a file called args_json. Where do i put these? Did i fuck up somwhere or have i not done all the steps?
>>7023571This one is nice, I for some reason like\enjoy bimbos when its not too exaggerated. either fit or curvy.Just enough of the bimbo lips and curves but not much bimbo fuck doll trash if that makes any sense. I guess its just the fitness angle that I enjoy maybe.
>>7023589>>7023594no offense to bimbo enjoyers.. just an opinion
>>7023589I don't enjoy your pimples
I feel like it was a long walk to get the slingshotkini hypernetwork to work,
>>7023600None taken, I share this place with futa faggots and worse.I like all types of women. Shit gets annoying and grating when it loses elegance.Lately I have been into slim long legged fair skinned women with nice big tits and perky ass, but smallish hips. Not fat.I cant seem to make her in SD even with some models made in 3D.I just want to make a nice Sorceress, Bayonettta type girl I guess but without the ay lmao proportions. I prefer statuesque western faces.
>>7023611theyre not pimples! they are pefectly sized boobies
>>7023621see, now this is a proper take
>>7023621I certainly have been trying but my prompts are trash, I have yet to figure this shit out. Although I have learned a lot.
>>7021916mfw when i literally followed the links and got it working on a gtx 960
>>7023571
>>7023571Prompt?
why when using anything v4 i get worst images than 3.0????
>>7021916I gotchu senpai. Also, like others said, check out colab. It's free, though time limited each day... Which is probably for the best, given how addicting this stuff is, haha.>>7023721v4 is fake, it's just some dude who made a merge and called it v4.
Has anyone manged to get something like the simpsons working, I tried the r34 model but it wasn't a good result.
>>7023721You aren't the only one who has noticed that. Rumor says AnyV4 was made in hurry and just merged AnyV3 with AOM2 and other models
>>7022238>>7023746here's an anime-ified one for ya too, free of charge
I literally just found out about this and will make my mission to use this for creating nudes of big tiddy girls I know IRL
>>7023626She makes me have feelings. ◉‿◉
:>=====8
>>7023868Okay, still a bit uncanny but a good one. Is it based on someone or something the program created?After learning how this shit works, I think its a bit of a reach to call these programs "AI", but I guess it sounds cool.Jesus fuck I can imagine a future going with this shit integraded with everything, movies, games, create animations, models form the get go. Modding will be insane.
>>7023872i dont remember wich model i used but it was a personal merge but it makes this
>>7023113Not as cute as the real one.
>>7023841not /aco/
someone said it would be possible to make a Gwen tennyson lora with some images so heres my stashtough i didn't put in images where shes featured with other girls like this one, hope thats okayhttps://easyupload.io/1poe6z
>>7023618I've been trying too.what model, configs, and prompts???
Any feedback?
>>7024110My feedback is keep doing whatever the hell it is you're doing anon
>>7024119Seriously, what should i improve?
Tried to install the thing, but im stuck on the python bit. I installed it and all but it doesn't work.
>>7024110Her front and left are melting with the rest of her hair, the earring is too high, no back hair, part of the neckpiece lacks depth, the weird highlight + black strap makes her right leg look like a sausage, her middle and ring fingers are missing the joints.
>>7024144Wat do you mean weird highlight? Are you refering to the colors?
>>7024149
>>7024162A fuck i didnt see that.
>>7024110I don't like the donut. It's distracting. Otherwise, great!
>>7024175>>7024168>>7024162>>7024144>4149▶>>7024162>>>7024144>>7024139>>7024119Thanks everyone, i was afraid her head was too big and the colors were fucked.
>>7024110>>7024177Long as it's not too out there, you can just hand wave it away by calling it your unique style
>>7023583What is being used to generate this image?
>>7021769Im new to this, how do you install this?
URPM looking damn fine
>>7024387doesn't producing hyper-realistic imagery devoid of any stylization kinda defeat the purpose of posting to /aco/ instead of /s/ or /hc/?
>>7024095Anon, it would save a lot of time if you prepared the images for training. You`ll need to crop her face in each image, resize it to 512x512, create a text file with the same name as the image and place the Danbooru tags describing her in each image. An example of how the tags should look:1girl, orange hair, green eyes, nose blush, smileYou can use this tool to help you with the tags: https://github.com/starik222/BooruDatasetTagManager/releases/tag/v1.4.1
>>7024234You can check the prompt: https://files.catbox.moe/5ydc3a.pngModel is mix of "PoV Skin Texture - Dreamlike r34" grape1 and AOM2 nsfw.Doesn't matter if you replace Pov Skin Texture with URPM
>>7020929>>7020930>>7020931Wait... it can already generate pixel art? It's fucking over.
>>7024577https://github.com/WuZongWei6/PixelizationThe tech has been here for a while.
>>7024409Those paddles are so disgusting that I wouldn't say that is hyper realistic. That is waaaay too weird fetish for me.
Ive got the whole thing set up locally, but Ive hit a wall with what model everyone one seems to be using? I get that many are using "merged" models (however that works) But I just have the default StableDiffusion as well as waifu diffusion and none of my shit is looking a tenth as good as any here
>>7024626go to civitai.com find something you like, click the "i" on images to get their prompts and you can start experimenting.
>>7024626Copy prompts from threads and civitai. Play with the prompts. At first I didn't want to use words like: "high resolution", "high sharpness" because I considered it cheating but if you want realistic or nice detail, you have to use generic words.
any Reiq's art style (from JigglyGirls) lora available ?
Anyone got any diffusions of gloryhole content?
Has anyone successfully created a lora that highly resembles a rl person? What I did was throw a handful of models I think are attractive into a lora, generate images using that lora about 2000 times, pick out about 100 that I think are exceptionally well drawn and trained a second level of loras using those images. And here are the results. While it does look nice, it looks too hand drawn and doesn't look like a photo anymore. I've seen images and gifs here that look like it could have been that person walking with a camera. How was that achieved? What kind of settings or training set was used?
>>7024572do you have a link to the models you used? also what configuration you used to combine them? i can't figure out how to get the lighting effect on the left that you managed, it looks beautiful
>>7024567>https://github.com/starik222/BooruDatasetTagManager/releases/tag/v1.4.1i have absolutely no idea what any of this means
>>7024567so on that site, i just randomly click on one of these numerous buttons and something happen that... should happen?or something?
>>7025189Click on BooruDatasetTagManager.v1.4.1.zip, extract the zip and run BooruDatasetTagManager.exe.When you open the program, go to File - Load Folder. After you load the folder, click the + symbol on the All tags side, add 1girl, and save all. This will generate a text file for all images. Then start tagging.
>>7025056You can find both from CivitAi by searching: "grape" and "urpm"https://civitai.com/models/2583/grapefruit-hentai-modelhttps://civitai.com/models/2661/uber-realistic-porn-merge-urpmAbyssOrangeMix2 you can find here: https://huggingface.co/WarriorMama777/OrangeMixs/tree/main/Models/AbyssOrangeMix2I merged with Merge Block Weight (https://github.com/bbc-mc/sdweb-merge-block-weighted-gui) and there isn't standardized way to share recipes. If you are a beginner, I recommend learning to prompt with one model and after that add little bit other models to your main model. You can recycle your old prompts to your new model. I use grape1 in my models because I know how to prompt with it though I don't like it's style. Lately I have been able to get rid of grape1.
>>7025272that explains a lot! thanks so much.
Whoever was posting vintage news presenter images in the previous thread I would like to say thanks, I couldn't post because site wide banned for the the innocuous of a reply. Do you have an art blog for more of your stuff?
>>7025429Sure,http://dontvaccinate.com/Do your research!
>>7025509>>7025429Sry, wrong url: nevervaccinate.com
flower fingers
>>7025429You're welcome anon, and nope
So for shits and giggles I used chatgpt to generate some prompts and it didn't do that bad of a job imo. Needs tweaking, sure, but when does using someone elses Prompt-Fu not need tweaking?
>>7023502underrated
>>7025779Damn you should start one, feel free to keep posting here in the meantime. ;-)
>>70244091. Like the other anon said, they already look disgusting. I want better than real, not a trash fire2. I looked, and neither had a SD thread upIf you guys know any better place to do real looking stuff, that's also not garbage, please tell
>>7025860>>7025861i really like the style
>>7025888It's the Decker loras I'm trying to make. link is a few threads ago.
>>7025272works like a charm! thanks again.
Bros, how does one prompt a tummy like this?With what kind of words?
working on dafne keen pics, this is an old prompt but i just to bump the thread.
>>7025925I don't, but do you have a catalog of images for this tummy type? I would love to train a model on this as it makes my monkey neurons activate
>>7025957Will always appreciate dafne keen content especially now she's 18
>>7025970i only began doing it because she turned 18,so prompting her this way is pefectly fine
>>7025925so i tried adding the words >fat, chubby, ((plumb)), wide hipsto my prompts, this were the results, while not exact i would say im getting closer.
>>7025994Amazing work my guy, don't stop there
>>7026014for now, i'll keep posting dafne,
>>7025994Do you have skindentation? It adds bulgy bits (indentations) of the skin/flesh which may help the effect
>>7026057ill try that, when i get back home
>>7025994"Deep skin", "pregnant", "stomach bulge" and "voluptuous" might work also. Add "obese" and "overweight" to the negative prompt if they look too unhealthy.>>7025916Nice work
What model is /asdg/ using? I've been dicking around with Berry's mix for hours and this is all I have to show for it.
>>7026075Top one looks like moebius.
>>7026261go to CivitAi, choose model that has pictures with prompts (the "i" icon on each pictures bottom corner), edit the prompt and see how keywords work.
>>7026042what model did you use for that?
>>7026261I've been trying my damndest to generate Holly Short from the Artemis Fowl books.
Merry christmas
>>7026517The eyes are crooked. Try the prompt "symmetrical eyes"
>>7026517mastahpiece
Expansions? Inflations?
>>7020579Any idea how this image was prompted and what model. I really like the lighting effect and the facial style.
while not what i wabted, i do like the way the words ""Deep skin", "skindentation", skin "stomach bulge" and "voluptuous" defined her shaped,i did put pregnant on negatives, that term just has too much power
>>7025925this is the closest i've ever been to that tummy
oh yeah!
there, now she is perfect
>>7026627I don't know anything about Stable Diffusion. Making models of specific characters are called a Lora right? Mind posting yours for Alex? This is really good.
>>7026871I'll need to change a few things first before sharing. Right now, I can't control which type of image you get. Sometimes you get clean images like >>7026627 or blurry screencap type like >>7026782 and this. Hopefully, by tomorrow I will have a better lora.
>>7026986Okay, thank you anon.
>>7026338it's the mellkor mancin dreambooth merged with >>7021769 at 0.3
>>7026993>>7026871Nah, I was wrong. It's better to have extra input before training a new one. Here's the link for the lora https://litter.catbox.moe/l30jtm.7zThere are 2 lora files alex.safetensors and alex-000007.safetensors. alex gives cleaner images but sometimes it deviates too much from the show's style. alex-000007 follows the show's style better but the images are too blurry sometimes. There are also 14 images that you can use with PNG Info.It was trained using Anything v4.5 but it should work with blends. The settings and negative prompts are the NAI standard. For the prompt, you don't need to use her name, just tags that describe her.1girl, solo, short hair, blue hair, brown eyes, dark-skinned female, dark skin, pink lips, smile, closed mouth, yellow bodysuit - This will generate her or close most of the time.Keep in mind that this model isn't flexible. I trained it to generate her but if you lower the lora strength enough. you can kinda get her in other styles.
>>7020579
Any anons got anything Eva Vlaardingerbroek inspired or know where or how?
no more AWSfagging for me, I finally got my RTX A2000 12 GB so I'm gonna proompt a lot more
who are some artists some of you guys would be interested in seeing a lora for?
>>7027325Olivia. Just ain't made it there yet, still working on this Decker one.
>>7027325I like old-timey comics. Is there a lora for artists like Hal Foster, Jack Kamen or Wally Wood?
>>7027327got a link to their art?>>7027334if you could find a good collection of their stuff i wouldn't mind taking a shot at it
>>7027339Can't find Kamen's work collected in one place, just a lot of EC comics.For Foster I can link you to a collection of Prince Valianthttps://getcomics.info/other-comics/prince-valiant-collection/as for Wallace Wood, there's this:https://getcomics.info/tag/the-life-and-legend-of-wallace-wood/
>>7027339Negative, thats why I aint made it yet. Gonna have to manually download from the net looks like. I got lucky and found a pack of decker pics.
>>7027364anon i don't even know who the "Olivia" is that i'm supposed to be searching for
Snakes are unironically great clothes.
any feedback???
>>7026029nice
>>7027382noit's perfect
Where is the goddamn Abby Shapiro model
>>7027096Bimbo creator, are you here?
>>7027487yes
>>7027496Are you created the SamDoesBimbos and now working on OmegaBimboMix?
>>7027455Make it anonBe the change you want to see
>>7027501YesOmegaBimboMix is just meme name I chose when I started a new bimbo model from scratch.It's based on a few dreambooth models of mine: Temogam, Iranon, WeakJ (weakJ is shit, because he has shit art but them lips tho); and a few other that anons posted in these threads: the pit bimbo, fellatrix, melkor mancin.Plus it has the old SamDoesBimbos in it.Right now I'm just merging old merges together, like a caveman banging rocks, and see what sticks.Kinda pleased with the current iteration, merging the melkor dreambooth again seemed to go in the direction I .
>>7027510The dreambooths that you used in SamDoesBimbos, ThePit Bimbo and Copeseethemald are you trained them yourself, or somebody else did?
>>7027519NoSomeone else didThe Pit Bimbo was posted in this general a few months agoThe Copeseethemald I got it during the entire SamDoesArts drama from the SD resource goldmine rentry: https://rentry.org/sdgoldmine#dreambooth-modelsI don't know if the links are still active, since the rentry hasn't been updated in a while
>>7027523Ok, I have them.What do you think about the Babes model on civitai?
>>7027455>no Abigal Shapiro model>no Katy Perry model>no AOC model>no Hayley Atwell modeletc.But we have Phil Collins...https://civitai.com/models/5427/phil-collins-80s
>>7027523Do you plan to publish OmegaBimboMix on civitai?
>>7027529I saw it, but I never used since I thought it was the model I madeLooks similar to what I made, so I don't think I'll use itThe bimbos look nice, as always>>7027538Probably notBut I'll post it here
Has anyone fapped to AI porn?
new thread when ready >>7027557
>>7027553nope;)
>>7027523Babes is SamDoesBimboes with added analog, dreamlike, redshift, and sdHeroBimbo(not sure if you used it in SamDoesBimboes).Do you address somehow freckles covering the whole image problem in OmegaBimboMix?If you publish OmegaBimboMix on civitai, more people will enjoy it, SamDoesBimboes and Babes are pretty popular models.I need to walk with my dog now.How can I summon you in the future?
>>7027565just post in these generals on /aco/I check them daily
>>7027571Ok, thanks!
>>7027565>Do you address somehow freckles covering the whole image problem in OmegaBimboMix?I don't have this issue, but it could stem from an Add Diff merge with too much weight. I noticed that it can happen,Right now I'm using the Merge Block WEighted extensions and it's pretty good>>If you publish OmegaBimboMix on civitai, more people will enjoy it, SamDoesBimboes and Babes are pretty popular models.I couldn't care lessSomeone will post it there on my behalf.I don't really care
>>7022735This is what I'm working on right now. I'm a fucking moron and I've still been able to follow some of this stuff.
fixed the double navel in a previous post
i'm looking at these images and one disadvantage of versatile models is that they make cartoonish/smooth nipples by default. never mind the weird prolapsing nipples sometimes.
>>7027370Olivia is what she is known professionally. and when you google "Olivia art" it comes back Olivia De Berardinis.
can anyone dump images they've made/have of dark-skinned/black women?
>>7020579Anyone else update recently and have their models change? Just updated today and now my PNG info sent to txt2img does not produce the same exact outcome, same seed, same sampler, same model, but the model hash changed. pic related.
>>7027504>>7027530Sigh... I started training
has there ever been an /asdg/ thread that has reached image limit?
>>7020928>>7020929>>7020930>>7020931>>7020933I have no instance and I must PROOMPT
>>7027530
>>7028577I hate how ghoulish this bitch looks even irl.
>>7027562I couldn't fap to that trash either
>>7028377what even is the limit? the thread sinks when it has 300+ text replies, no?
>>7028956The image limit is 300 images.
>>7028284yes. I updated with git pull a few days ago and I suspect it. it was after that I started getting color problems with vae and img2img too
>>7029241reddit claims the hash is read differently and the update happened two days ago without announcement. I think a lot of people are about to get their prompts ruined, kek
>>7029420How do you go back to previous? Is it on github and you just replace files or what?
>>7029844Before you do, I figured out that the old PNG info if sent to both img2img and txt2img still works and will produce the same outcome as before but there is some UI glitch that will show an empty model in the dropdown. If you just send info and dont touch any model settings you will get the same outcome.This video explains how to roll back an update in the first few minutes.https://www.youtube.com/watch?v=0fkGd9wIhrA
Is the face ok?
>>7020928>>7020929>>7020930>>7020931>>7020933Where's the PROOMPT. Where is it!? Give me the PROOMPT or ill crack your head open!
