Previous /sdg/ thread : >>101555885>Beginner UI local installEasyDiffusion: https://easydiffusion.github.ioFooocus: https://github.com/lllyasviel/fooocusMetastable: https://metastable.studio>Local installAutomatic1111: https://github.com/automatic1111/stable-diffusion-webuiComfyUI: https://github.com/comfyanonymous/ComfyUISD.Next: https://github.com/vladmandic/automaticAMD GPU: https://rentry.org/sdg-link#amd-gpuIntel GPU: https://rentry.org/sdg-link#intel-gpu>Use a VAE if your images look washed outhttps://rentry.org/sdvae>Run cloud hosted instancehttps://rentry.org/sdg-link#run-cloud-hosted-instance>SD3 info & downloadhttps://rentry.org/sdg-link#sd3>Try online without registrationsd3: https://huggingface.co/spaces/stabilityai/stable-diffusion-3-mediumtxt2img: https://www.mage.spaceimg2img: https://huggingface.co/spaces/huggingface/diffuse-the-rest>Models, LoRAs & upscalinghttps://civitai.comhttps://huggingface.cohttps://aitracker.arthttps://openmodeldb.info>Animationhttps://rentry.org/AnimAnonhttps://rentry.org/AnimAnon-AnimDiffhttps://rentry.org/AnimAnon-Deforum >Index of guides and other toolshttps://rentry.org/sdg-linkhttps://rentry.org/rentrysd>View and submit GPU performance datahttps://vladmandic.github.io/sd-extension-system-info/pages/benchmark.htmlhttps://docs.getgrist.com/3mjouqRSdkBY/sdperformance>Share image prompt info4chan removes prompt info from images, share them with the following guide/site...https://rentry.org/hdgcbhttps://catbox.moe>Discord6wUwtcJsr2>Related boards>>>/h/hdg>>>/e/edg>>>/d/ddg>>>/b/degen>>>/vt/vtai>>>/aco/sdg>>>/trash/sdg
>mfw Resource news07/24/2024>SV4D: Dynamic 3D Content Generation with Multi-Frame and Multi-View Consistencyhttps://sv4d.github.io>Open-Sora-Plan Report v1.2.0https://github.com/PKU-YuanGroup/Open-Sora-Plan/blob/main/docs/Report-v1.2.0.md>Official global launch of Kling AI's International Version 1.0https://klingai.com>INF-LLaVA: Dual-perspective Perception for High-Resolution Multimodal Large Language Modelhttps://github.com/WeihuangLin/INF-LLaVA>FoRA: Low-Rank Adaptation Model beyond Multimodal Siamese Networkhttps://github.com/zyszxhy/FoRA07/23/2024>Mammoth - An Extendible (General) Continual Learning Framework for Pytorchhttps://github.com/aimagelab/mammoth>Differentiable Convex Polyhedra Optimization from Multi-view Images https://github.com/kimren227/DiffConvex>Cinemo: Consistent and Controllable Image Animation with Motion Diffusion Modelshttps://maxin-cn.github.io/cinemo_project>Automatic Generation of Fashion Images using Prompting in Generative Machine Learning Modelshttps://github.com/georgiarg/AutoFashion07/22/2024>Intel AI Playground beta Official Launchhttps://game.intel.com/us/stories/introducing-ai-playground/>SD3 Unbanned: Community Decision on Its Future at Civitaihttps://civitai.com/articles/6221/sd3-unbanned-community-decision-on-its-future-at-civitai>GenTron: Diffusion Transformers for Image and Video Generationhttps://github.com/lavinal712/GenTron>ComfyUI-Lumina-Next-SFT-DiffusersWrapperhttps://github.com/Excidos/ComfyUI-Lumina-Next-SFT-DiffusersWrapper>PhotoMaker V2 Released with improved ID fidelityhttps://github.com/TencentARC/PhotoMaker#-new-featuresupdates>T2V-CompBench: A Comprehensive Benchmark for Compositional Text-to-video Generationhttps://t2v-compbench.github.io/>ComfyUI_FL-Trainer: Train Loras in ComfyUIhttps://github.com/filliptm/ComfyUI_FL-Trainer/tree/main>Video Diffusion Alignment via Reward Gradienthttps://github.com/mihirp1998/VADER
>mfw Research news07/24/2024>PartGLEE: A Foundation Model for Recognizing and Parsing Any Objectshttps://provencestar.github.io/PartGLEE-Vision>MovieDreamer: Hierarchical Generation for Coherent Long Visual Sequenceshttps://aim-uofa.github.io/MovieDreamer>DreamVTON: Customizing 3D Virtual Try-on with Personalized Diffusion Modelshttps://arxiv.org/abs/2407.16511>Harmonizing Visual Text Comprehension and Generationhttps://arxiv.org/abs/2407.16364>DreamDissector: Learning Disentangled Text-to-3D Generation from 2D Diffusion Priorshttps://chester256.github.io/dreamdissector>Outfit Anyone: Ultra-high quality virtual try-on for Any Clothing and Any Personhttps://humanaigc.github.io/outfit-anyone>Channel-Partitioned Windowed Attention And Frequency Learning for Single Image Super-Resolutionhttps://arxiv.org/abs/2407.16232>CLII: Visual-Text Inpainting via Cross-Modal Predictive Interactionhttps://arxiv.org/abs/2407.16204>MxT: Mamba x Transformer for Image Inpaintinghttps://arxiv.org/abs/2407.16126>Diffusion Prior-Based Amortized Variational Inference for Noisy Inverse Problemshttps://arxiv.org/abs/2407.16125>Fréchet Video Motion Distance: A Metric for Evaluating Motion Consistency in Videoshttps://arxiv.org/abs/2407.16124>Test-Time Low Rank Adaptation via Confidence Maximization for Zero-Shot Generalization of Vision-Language Modelshttps://arxiv.org/abs/2407.15913>CatVTON: Concatenation Is All You Need for Virtual Try-On with Diffusion Modelshttps://arxiv.org/abs/2407.15886>Adversarial Attacks and Defenses on Text-to-Image Diffusion Models: A Surveyhttps://arxiv.org/abs/2407.15861>Image Classification using Fuzzy Pooling in Convolutional Kolmogorov-Arnold Networkshttps://arxiv.org/abs/2407.16268>Is 3D Convolution with 5D Tensors Really Necessary for Video Analysis?https://arxiv.org/abs/2407.16514>HDRSplat: Gaussian Splatting for High Dynamic Range 3D Scene Reconstruction from Raw Imageshttps://arxiv.org/abs/2407.16503
>>101562401appreciate the baking as always debo
>>101562516i, too, prefer to pretend that which is, will always be. it's lindy. mmhmmm
Haven't been here in ages.whats new, fags?
despite what you may think, i didn't start posting bad gens. alas, I AM touched your sense of taste and made it only seem that way! it happens, there's no accounting for God, i'm afraid! ¯\_(ツ)_/¯
>>101562625THE CIA MADE MY IMAGE NOT ATTACH! I AM VERY ANGRY!!!!
you will never die, anon, tee hee. if you do, i'll be here waiting for you!
>>101562656thing is, if you don't, it gets a lot worse.
i'll
honestly i should go do literally anything else, let this shithole die as Now trickles across the pacific ocean towards Chickity China... time was, californiggers, at the least, contributed to this shitole.>>101562721>>101562724just as suddenly as the previous realization occurred to him, the opposite suddenly became real. he didn't notice until hours later. why would he?
could you prompt female pov, female hand casting spell?
>>101562382Hahaha yea! Time continually flows in it's monotony! I enjoy it though!It did! You have great gens, anon! :D>>101562376Cuuuute!!>>101562392Hahaha! A dangerous idea! But that shouldn't break the rules, i'd imagine!>>101562459That was not my intention whatsoever, anon!Sorry if you think it came off that way!
Mods really do be getting buttmad when you call out their 1girl posts huh
on the one hand, if everything was me, things would be a lot less complicated than they are. unfortunately, "we exist", and that makes everything a lot more complicated! where is your God now? ah right, he's here with me, inline for this exclusive club, Planet Earth. limited time only! oopse you already missed it! maybe next time!!>>101562746HALT! identify yourself!
>>101562766no mod
>>101562766Naaaah it's just mass-reporting imglets bothering them hahaOtherwise I wouldn't be here :]A slap on the wrist for the things I do!>>101562772Hello! I am but a simple green cat! :]No purple witches here hahaha!
>>101562743>female povevery 1girl post in here is a 'female pov'
>>101562804there is a transparent tentacle lora that came out, I want to try water tentacles playing with Azula with an extended arm implied to be Katara using bending
>>101562804I don't know if I've actually seen a gen from pov of the 1girl that was generated
>>101562857the 1girl is always holding the camera
>>101562880but then it would look like a fps
>>101562857If you look up "POV" on tiktok and you'll see thousands of women misunderstanding what point of view means and confusing it with "mfw." I was joking about that.
>>101562902I don't look at tiktok so I had comprehension failure
https://files.catbox.moe/xay6pd.jpgMistakes were made.
>>101562993I would assume that's SD3 but it can't do boobs
>>1015629982dnPony_V10 at 2048 instead of 1024. I don't know really what triggers that as sometimes it'll scale just fine.
goo night anons
>>101563045Goodnight
if I just want to make a basic SDXL aspect ratio image and upscale it by 2x,should I hiresfix straight from text2img or upscale it through img2img? To me it seems img2img gives me way less problems and can't tell which hiresfix upscaler works best
>>101563009It's the size. Generating above 1024 will cause a lot of body horror.
>>101563064in a1111 you can also send it to "extras." Doing that is less memory intensive than hires fix and you won't get out of memory errors. I just use hires fix though
what are some good male models to use in MF lewds?
>>101563181very nice
>>101563227ty
https://suno.com/song/0de6620f-5d21-4a62-a53a-000b3dea3898>M8PSK8
https://suno.com/song/1a3a170c-b386-4c1c-a07c-2379eb2cea60>MHMWSHNo, you hush. I pay in blood.
>ban evading purple schizo alt avatar
https://suno.com/song/e77e9e2a-9380-4607-9449-43f6b4fcbb09>0V8HPNo V* no HP this thread dead.
For the life of me I can't get it to generate an actual tub full of maple syrup. At first it kept filling it with water, then when I neg prompted water it put a tub of maple syrup jars in the tub
>>101563543Imagine caring
>>101563570Try Liquid + ColorLike "Dark Amber Liquid"If I try "Maple Syrup" it just fills it with water but it'll follow the "<COLOR> Liquid" prompt
>>101563791No luck on translucency of any kind though.
120m and 5m
>>101563791>>101563804But if I add say, a romanian cat girl to the bath tub it stops following the color prompt for the liquid
would buying one of these laptop cooling pads noticeably help my mobile 3060?
>Generate me a GANimal>Breaks the rules four times overOh, Gemini...>DJMNYHahahahaha.https://suno.com/song/7f836fd1-e852-4985-a9fa-1139acf44bd3>DSPSRI don't need a DSP, sir.
>>101564062lmao
i'm sure if anyone else were to release the exact same datasets they'd be getting a lot more recognition ;-;
>>101564554what did you release?
>>101564554Well, if people just knew about your magnificient datasets maybe you would get more recognition in the first place.
https://suno.com/song/eb53496b-075a-40e8-afa8-65c4afaf8824
>>101562401Complete newfag here. I planned on trying the first three programs (EasyDiffusion, Fooocus, Metastable) but before I did I wanted to see what limitations each one posed (if any). Do all programs limit violence/sexual content/copyrighted content or are some better than others at allowing you the most freedom?
>>101565741The UI doesn't limit the content of the images you can make, that will be down to the checkpoint/model you use with it.
>6k of 41k captioned truly is suffering
>>101565741If you're after pornhttps://civitai.com/models/520661Will serve you well with any UII like EasyDiffusion because python is a greasy whore and I ain't fixing it
stage 1, database of actors for the personal details, work, pictures etc, then extract embeddings from the pictures for facial recognition
>>101565930I don't know what any of that means but goodluck.
Must I really wageslave another month to get dual 4090s
>>101566055or one commission for diaper furry scat
>>101565939i don't need luck. i need more nodes and less bytes and that american apparel girl in just tights
i think trani is shit (as a human)
gm anons
What tags do I use to get the fisheye effect? This gen somehow got it without having the "fisheye" tag and I'd like to generate some more with but but having that tag in there isn't working.
>>101566292good morningi hate trani too
only two more weeks until the 30m coke fund is(n't) secured
>>101566476do you have an ani shrine in your house or something? beginning to believe this is tsundere behavior
>>101566535i executed a ritual once against traniunfortunatly im not good at that so it removed his dad from existence instead of him
Anyone aware of a working steganography extension for webui? All the solutions I saw needed a separate tool.I was able to write one but Python Pillow mangles the image.
>>101566292Good morning good morning! And also good morning /sdg/! Hope the GENs go well today.
mornin
>niggerjak moment
>>101566955good morningi hope you are feeling well
>>101566055>dual 4090syou can't chain them. what's the point?
I'm listening through the backlog of suno tracks>>101566292gm>>101566904gm>>101566955gm
gm ser today we are dropping two zeroes
133m and 6m37k actorshow many will it take to fill the void
>>101564062Its more cost efficient to simply elevate the back corners of your laptop on two books or something, less the hot air flow out rather than be absorbed by the table surface
I'm having a great day, how about you guys?
>>101567285fair, didnt wake up feeling especially miserable, like most of this week
>>101567344Sounds like an improvement. Good stuff. You been sick or something?
>>101567372sick in the head, nothing new, and struggling with food/money, but my mood alternating between feeling o.k. and constantly fantasizing about suicide is normal for me.
>>101565409yoink, mine nowhttps://suno.com/song/7d074682-35e2-4bba-85da-829c1b5f4b85
How come some people manage to use multiple loras in one gen and get decent results?If I use anything beyond just one, I start getting artifcats, at 3 or above I get very severe melting/burning. Is there something Im missing?
>>101567448>3 or abovethere's your problem right there. Try stick to 1.0 or less.
>>101567448depends on the model and the loras, and the strength of the loras, usually have to lower the strengths if you use more of them
>>101567400Wow, I hope things get better for you anon
>>101565409https://suno.com/song/4e8ba8ea-9654-4ce7-bfc5-6c073b01a3a7Oh anon. Good morning. I cut your song to pieces and served it.>HWSAKXYep.
I wonder if any tech firms have trained models on chip data/blueprints and seeing what they come up withno cloudflare i'm no longer human you fucking daft cunts
>>101567665>no cloudflare i'm no longer human you fucking daft cuntsBUILT DIFFERENT
>ipadapter dev released yet another schizo nodekinda wish this guy focused on improving what already has instead of "inventing" so many new nodes
Morning anons
>>101567665https://medium.com/intuitionmachine/the-alien-look-of-deep-learning-generative-design-5c5f871f7d10I also remember reading a story about an AI-designed antenna which was bizarre looking too. I'm fascinated by the prospect of AI-driven engineering design because AI can do fundamentally unintuitive things a human mind wouldn't think of. I also think a lot about what will happen once there's an infinite-context reasoning LLM agent that will be able to consume all research papers and output undiscovered breakthroughs. there has to be countless discoveries waiting in the vast sea of scientific data and research projects that were printed and reviewed but never cross-referenced. maybe we already discovered room temperature super-conductors but never realized it because its a needle in the massive haystack of data
>>101568168gm
>>101568289hot off the press>McGAN: Generating Manufacturable Designs by Embedding Manufacturing Rules intoConditional Generative Adversarial Networkhttps://arxiv.org/pdf/2407.16943
i miss schizo anon
>>101568567you are the schizo anon. we've figured that out already
>>101568667he banks on there being enough newfags to not know but in reality most of the userbase has been here for at least two years
Hello guys, newfag here.
>>101568819hello newfren
>>101567440try this onehttps://suno.com/song/71665827-215b-46db-96d6-234b177bafbb
https://youtu.be/vIMdDa4d-EI?si=xnRTeMPDNLLDTK4O
>>101568990goes pretty hardhttps://suno.com/song/f62c0178-73bd-410e-bc61-39ee001e15f5alternative take that gives the lyrics more spacehttps://suno.com/song/42f0c8d0-13de-4d88-9886-3cc44037e905
vegOni
>>101568168moring/aferoon
Apparently this is kitty porn.I don't see what the big deal is.
>>101569709cute, model?
>>1015697412dnpony_v10It's pretty damn flexible
Local is finally starting to catch up to NAI v3.Give it a bit more time and we'll get there.
>>101569908Which models are you using?I'd love to get away from NAI, seriously, but...
>>101570390>k-on X touhou>classic flat 2d anime stylea man of culture I see
I'm just testing out stupid shit right now.But yes, I absolutely love Touhou (mostly for the music) and good old Anime.It hurts to admit that it's old.
>all the good posters have leftit's a damn fucking shame
And we all (save for one) know why Damn shame
>>101570223Neta v2https://civitai.com/models/410737/neta-art-xlIt's not quite NAI v3 level but it's the closest thing right now imoIt's about the same as NAI v3 for SFW stuff, way worse for NSFW unfortunately.
>>101570549If it's the same for SFW, then it's perfect for me since that is 99% of what I gen.But it's hard to believe because, holy shit, NAI 3 is pretty good at it. And it knows so many characters, shows and whatnot...
>>101570571Just try it out, it's free lolIf you don't like it, stick with NAI instead.
>>101570515no, he knows. he doenst want to admit it
>>101570580I will. Thanks, anon.
>>101568560>McGOON: Generating Latent Designs by Embedding Tensor Seed Elf Rules into Conditional Generative Adversarial Network
>>101567665I have a hunch that Activision has used some form of machine learning to design their Call of Duty maps throughout this year. Just the layout. How do I know? They all have sort of same features but of course this is speculation unless someone from the inside tells more. They seem somewhat different than previously too. Sure it all needs lot of manual work to ensure the map is playable afterwards and testing and so on.
>>101570719Yes, I'm redeeming the card; I'm redeeming the cards, okay!
>>101570716I remember Anarchy Online had a random mission layout system that was pretty decent back in early 2k. I can't remember exactly how it was done but that it was pretty varied and reliable.
>>101566955late ohayonice aesthetic
>>101570787ty
>>101570892Very disturbing, cool. In my opinion it's pretty difficult to generate genuinely upsetting images but this one has certain elements.
>>101570892very debo-coded
>>101570992the best time /sdg/ ever had was when you were banned
>>101571004he lacks the ability to self reflect
>>101571084Do you still have the one of her sitting spread in the sundress? I forgot to save it.
Okay I can't find any info on this by searching and maybe I'm just retarded but I need help. I'm using easy diffusion and I want to use wildcards or something like ( this OR that ) for dynamic prompts but I can not figure out what the syntax for that would be. The only syntax they have on their github wiki is basic shit like increasing the weight of a keyword. I saw there's like a dynamic prompt extension for A111 or whatever but I don't know if/how I could use that on ED
>>101571049my spirit animal
>>101571123I think it's like:Red | Blue | Orange | HorsecockAnd it'll start a queue for each variantFrom what I remember.
>>101571142nop, that doesn't do them individually, it does pic related for some reason
>>101571213I feel like she should be a grungy hacker with cables snaking all around her to be more on-brand
>>101571247
>>101571109100 folders, 10,695 files in my /output, I probably have it lol but which one?
she keeps the most valuable data on a sbc hidden in her hair
>>101571279now we're talking. might need (cables:2.0) for the kind of data you work with though
>>101571333Tropical setting, blue and red sundress, sitting, spread, knees apart, she's sweaty, think she has a flower in her hair. I guess it was about the 17th
"the government uses your name to track you" so she doesn't have one. i called her Uma once because she dyed her hair black after we watched Pulp Fiction, big mistake, the next day i'm deceased, load of hassle to sort it out
blog tiemi noticed for one lora that i'm training was taking a ridiculously long time per step, about 22s/iterationmade no sense to me because im using the same exact settings as the last lora i trained which trained at ~7s/it all the training images were saved as pngso i saved them as jpg and the training is [as of right now] training at 5.6s/it the more you know!/blog
>d***
>>101571794Yeah PNG is...slow as fuck for some reason.For everything
I'm going to bed. Good night and have fun genning.
>>101571923Goodnight
do you know that feeling when waking up and simple hoping tranis depression got a lot worse?
>>101571956the feeling of being a loser?
>>101571934Why don't you attach an image to your Good morning and Goodnight replies?
>julien
>>101571956too well actually
>be /sdg/>love koff>hate aniunironically why?
>>101571976Very little I gen could be posted here
>>101572070koff is a guy with genuine problems which he should work on, he needs help, a real friend and a hugtrani is just literal human garbage end of story
>>101572070Regulars are mentally ill ain't they
>>101572075Care to post an example? You allude to having at least some.
>>101572116Paranoid delusions of self-importance
>>101572070sulky waifu > dancing waifu
>https://civitai.com/content/rules/real-peopleWhat's wrong with these fucking people
>>101572070koff at least makes progresstrani just reposts the same shit for months "but now i applied tooncrafter im making anime realsies!!11"
>>101572186are you dumb?
>>101572070no one hates any except one guy who has a super obsessive parasocial relationship with him
>
https://x.com/HelloCivitai/status/1816594588523724916Bit of collab between different devs
>singular schizo theory
>>101572272You have the biggest parasocial relationship with multiple anons who rarely interact with you anymore kek
>>101572321no catjak I have the biggest social parasite stop constructing other people's bullshit into fake storiesI'm not writing this captcha on the library wifi after being banned repeatedly.
meds
Crying again?
Don't make him start cry bullying like the last dozen times.
video/film set is going to be a lot of work desu. the plan is top 250 imdb, split into scenes/shots then annotations for each like shot type, camera angle, characters, objects, dialogue and audio descriptions
>>101572939>split into scenes/shotshow will transitions be detected?>shot type, camera angle, characters, objects, dialogue and audio descriptionssynthetic captioning?are there any open source vlms that even do that well?
Good think AI generated images don't need supplemental oxygen.
>>101573277Nice.
>>101573203>how will transitions be detected?ssim works well>synthetic captioning?for characters i'm building a database of actors, extract embeddings, then face detection model on frames from the shot, embeds from that and compare to the database. for shot type/camera angle i'll experiment with some vlms but i doubt they'll be any good, i have already have a dataset of frames with annotations that might help to finetune something. dialogue from subtitles probably. i'd like as much as possible to be automated, manual annotations where neededas a related thing i also thought of extracting frames every N seconds from tv and film then captioning+masking the characters
How do you guys upscale in comfyui? Latent upscaling looks way worse than realesrgan does in easy diffusion
>>101573295ty
>>101573348Using a separate workflow, SD3 with an Ultimate SD upscale node.
>>101573601I'm not sure why but upscaling with comfy is 10X slower than on WebUI. At least for me.
After awhile my prompts just turn into word soup like a misconfigured LLM install
>>101573644It's the one thing I only miss about a1111, the ultimate sd upscaling was amazing.
151m and 7m, might be about it for the second, first has a lot more to go200k actors36/250 films to download, i'm going to reencode them so they're all the same
>>101573277hi humu anon, haven't seen you for a few
how owl feelthe new 4chan ad is even worse than the prev
>>101573810hey debo
>>101573821hi debo
Adaptive optimizers don't really adapt that well. :(
>>101573810Hey, yeah. I'm around. I've been working on implementing a steganographic script. I'm not a python dev so it's quirky. Plus there doesn't seem to be much support for SD devs.
1girl looking for mentor
>>101573821>>101573831hi debos>>101573904idk if this is interesting to you at all but it popped up recentlyDiffStega: Towards Universal Training-Free Coverless Image Steganography with Diffusion Modelshttps://github.com/evtricks/DiffStega
>>101573982>Training-Free CoverlessInteresting. I've seen a few but no implementation for SD webui.I've been able to do it with PNG format, quite trivial, but jpg was harder. I'm using this https://github.com/daniellerch/stegolab/blob/master/J-UNIWARD/j-uniward.py which works well except that PIL used in Python, mangles the image internally. So I'm a dead end for now.
desu you could use generative AI like this for mega encryption that'd never be bruteforcedyou'd have to go deterministic obviously but...
>>101574054No need for all that inefficiency. There are already encryption methods that can't be bruteforced.
>>101574154I know, but it'd be a fun experiment anyways
>>101574172No it wouldn't. Stop being oppressive.
>>101574221can you do a wise koi-girl sage?
>>101563791that got some alright results, thank yahttps://files.catbox.moe/r5clkz.png>>101563900for me when I said 'bathing in a tub of maple syrup' it worked pretty alrighthttps://files.catbox.moe/dlaf30.png
>>101574314I tried>a monster girl at a lake, monster hands, koi pattern, pale skin with orange spots and black spots, very long beautiful hair, fin ears, gentle smile
>>101574631a bit off the mark but very cool aesthetic regardless
7900xtx or 4080 super?
>>101574657Yeah, not sure how to actually get a koi pattern
>>101574788I'd consider that a koi pattern. maybe adding (fish scales) would help?
Hey AItards, normally I avoid this thread but I heard some boomer ranting about 'AI stealing our jobs' and thought about a question you all might be able to answer;Has there been any succesful attempts to replace keyframe inbetween animators with AI? Hobbyist efforts are interesting, but I'm more curious about professional solutions. I know the artists they have filling in between key frames are practically slaves anyway, so nobody cares about them.
How to fuck do I quantize a sd model. I cant fucking figure out yow to use sd.cpp.
>>101574832even if its never happened yet, wouldn't it be inevitable regardless?>>101574852open the model in notepad and delete every 7th character
>>101574893>7thStop trolling, it's the 12th.
>>101562401New Update to Lora:X-Men Evolution Style SDXL V2:https://civitai.com/models/578582v2 changes: Much better faces and cleaner outputs.Has character prompts now (if you want to force a certain character. Or have their influence)
>>101574948>>101574948>>101574948