Previous /sdg/ thread : >>101063127>SD3 info & downloadhttps://rentry.org/sdg-link#sd3https://education.civitai.com/quickstart-guide-to-stable-diffusion-3https://aitracker.art/viewtopic.php?t=57>Beginner UI local installFooocus: https://github.com/lllyasviel/fooocusEasyDiffusion: https://easydiffusion.github.ioMetastable: https://metastable.studio>Local installAutomatic1111: https://github.com/automatic1111/stable-diffusion-webuiComfyUI: https://github.com/comfyanonymous/ComfyUIAMD 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>Auto1111 forksSD.Next: https://github.com/vladmandic/automaticAnapnoe UX: https://github.com/anapnoe/stable-diffusion-webui-ux>Run cloud hosted instancehttps://rentry.org/sdg-link#run-cloud-hosted-instance>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://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
The pixart chads will release a more powerful model (that will be a 16ch VAE)
>mfw Resource news06/20/2024>Comfy Org launches comfy.orghttps://www.comfy.org/>ComfyUI-UniAnimatehttps://github.com/AIFSH/ComfyUI-UniAnimate>SD-Latent-Interposer Adds SD3 Supporthttps://github.com/city96/SD-Latent-Interposer06/19/2024>Florence2 in ComfyUIhttps://github.com/kijai/ComfyUI-Florence2>Ilya Sutskever, OpenAI’s former chief scientist, launches new AI companyhttps://techcrunch.com/2024/06/19/ilya-sutskever-openais-former-chief-scientist-launches-new-ai-company>ComfyUI_frontend: ComfyUI modernizedhttps://github.com/huchenlei/ComfyUI_frontend>RunwayML Gen-3 Alphahttps://runwayml.com/blog/introducing-gen-3-alpha/>ComfyUI-LuminaWrapperhttps://github.com/kijai/ComfyUI-LuminaWrapper>TroL: Traversal of Layers for Large Language and Vision Modelshttps://github.com/ByungKwanLee/TroL>diff-sampler: open-source toolbox for fast sampling of diffusion modelshttps://github.com/zju-pi/diff-sampler>Diffusers Adds≈ SD3 ControlNet and Multi-ControlNet Supporthttps://github.com/huggingface/diffusers/pull/8566>Better & Faster Large Language Models via Multi-token Predictionhttps://huggingface.co/facebook/multi-token-prediction06/18/2024>Release of training script of PCM-LoRA with Stable Diffusion 3.https://github.com/G-U-N/Phased-Consistency-Model#news>Autoregressive Model Beats Diffusion: Llama for Scalable Image Generationhttps://peizesun.github.io/llamagen/>Flash Diffusion: FlashSD3https://huggingface.co/jasperai/flash-sd3>SD3-Controlnet-Tile https://huggingface.co/InstantX/SD3-Controlnet-Tile>Sharing new research, models, and datasets from Meta FAIRhttps://ai.meta.com/blog/meta-fair-research-new-releases/>The next chapter for ComfyUIhttps://blog.comfyui.ca/comfyui/update/2024/06/18/Next-Chapter.html>DeepFuze: Deep Learning Tool Integrating Lipsync, Face Swap, VidGen and morehttps://github.com/SamKhoze/ComfyUI-DeepFuze
Cursed thread
>mfw Research news06/20/2024>Knowledge Fusion By Evolving Weights of Language Modelshttps://arxiv.org/abs/2406.12208>Unmasking the Veil: An Investigation into Concept Ablation for Privacy and Copyright Protection in Imageshttps://arxiv.org/abs/2406.12592>SeTAR: Out-of-Distribution Detection with Selective Low-Rank Approximationhttps://arxiv.org/abs/2406.12629>Federated Learning with a Single Shared Imagehttps://arxiv.org/abs/2406.12658>Instruct 4D-to-4D: Editing 4D Scenes as Pseudo-3D Scenes Using 2D Diffusionhttps://arxiv.org/abs/2406.09402>Updating CLIP to Prefer Descriptions Over Captionshttps://arxiv.org/abs/2406.09458>Between Randomness and Arbitrariness: Some Lessons for Reliable Machine Learning at Scalehttps://arxiv.org/abs/2406.09548>Grounding Image Matching in 3D with MASt3Rhttps://arxiv.org/abs/2406.09756>Bayesian Conditioned Diffusion Models for Inverse Problemshttps://arxiv.org/abs/2406.09768>ALGM: Adaptive Local-then-Global Token Merging for Efficient Semantic Segmentation with Plain Vision Transformershttps://arxiv.org/abs/2406.0993606/19/2024>Synergizing Foundation Models and Federated Learning: A Surveyhttps://arxiv.org/abs/2406.12844>LayerMerge: Neural Network Depth Compression through Layer Pruning and Merginghttps://arxiv.org/abs/2406.12837>VIA: A Spatiotemporal Video Adaptation Framework for Global and Local Video Editinghttps://arxiv.org/abs/2406.12831>Neural Approximate Mirror Maps for Constrained Diffusion Modelshttps://arxiv.org/abs/2406.12816>AITTI: Learning Adaptive Inclusive Token for Text-to-Image Generationhttps://itsmag11.github.io/AITTI>Extracting Training Data from Unconditional Diffusion Modelshttps://arxiv.org/abs/2406.12752>AGLA: Mitigating Object Hallucinations in Large Vision-Language Models with Assembly of Global and Local Attentionhttps://arxiv.org/abs/2406.12718>SUPER: Selfie Undistortion and Head Pose Editing with Identity Preservationhttps://arxiv.org/abs/2406.12700
!!! MESSAGE FOR COMFY !!!>the latest comfy master branch and the latest https://github.com/comfyanonymous/ComfyUI_TensorRT master branch>using workflows from https://github.com/comfyanonymous/ComfyUI_TensorRT/pull/30 to generate the images>picrel
>>101068679What are your gpu specs? Are you trying to use SD3?If so I think it is somehow botched in terms of that it tries to use tensor cores and if they are not available it doesn't what to do. Implementation issue because it defaults to Nvidia speed optimizations.Whatever I suggest don't waste your time or electricity with SD3 at all.
Are your Diffusers sufficiently Stable?
>>101068741>>101068762I like
>>101068719>What are your gpu specs?rtx 3060 12gb>Are you trying to use SD3?yes>Whatever I suggest don't waste your time or electricity with SD3 at all.i agree that its shit but it has the best prompt understanding out of all open models, i dont generate humans/animals that much. never used SD for anything professional, fun only. sd3 is fun
I got a spare spear
>>101068796She be the Green Witch of Greenwich
>>101068798>best prompt understanding out of all open modelsWhen you are unable to generate an image, any image for that matter - the so called prompt understanding doesn't matter.Why do you think you need to be a failed novelist in order to generated images anyway?
>>101068883Testing accurate depictions of fire.
>>101068914
>>101068605
>>101068857>When you are unable to generate an image, any image for that matter - the so called prompt understanding doesn't matter.i agree but without tensorrt it works>Why do you think you need to be a failed novelist in order to generated images anyway?you dont, but you do if you want the best prompt understanding. atleast from what i know
It's hard to get a skirt lift with the flap thing>>101068480These are my favorite Japanese ones so farhttps://suno.com/song/68c61b21-aa6b-4031-a1df-328d4392b0fchttps://suno.com/song/b5a4dc4b-8956-46df-bc57-c1c4bb394815
>>101068961I don't know I think you are trolling or otherwise you'd be more than happy with your pre-existing gens or pursuing for them instead of posting the same jpg across different threads, multiple times.
>>101068961Besides, the way you type is correlating with the psychopath already.Nice try, faggot.
in recent news: civitai becoming pretty much useless for genning rip poorfags
>>101068996i am happy with the gens, i am not happy with the inference time it takes to make them, i am very happy with the speedup tensorrt gives to other models (SDXL goes down from 13s to 7s)>>101069022lolwut
>>101069049Blaablaablaa. Get flushed down the toilet, will you.
>>101069064
With this texture I am hoping to achieve a contemporary absolution, something which we all know and dare to face in our daily lives.
>>101069187too high brow for sdg
>>101069187absolution of what?
>>101069217I only speak English, not American English.
>>101069236>chips are fries>fries are chipsunbelievable
>>101069235Now make her purple
>>101069217Absolution as in Christian faith, or even in regular Buddish beliefs you mongrel.
me on the left>>101068944
>>101069236/b/ullshit
>>101069253This is the most abstract paradox of it all.>>101069271Here is one American who has never traveled outside of his own state in his life. I do not envy you, my friend.
>>101069254I've already paid tribute to PW anon's girl last year, not interested in more.I will post more Green Witch of Greenwich.
Morning anons
>>101069298Green Witch of Greenwich - this is my home and it is called England!
>>101069298Why is she from Greenwich?
>>101069324morning
>"Aww, is the swamp making you uncomfortable? Don't be a sissy, Anon. Let us continue our walk, and you'll be just fine."
>>101069326She is from Greenwich, CT, for Reasons(TM)
>>101069114>anons remembering my old gens better than me
>>101069326>>101069394My American friends, time starts from there.
>>101069394Is Greenwich that swampy? I know it's hella expensive and where the hedge fund guys are.
>>101069394>>101069326Naval Time.
>>101069410She's not always in Greenwich.
>>101069324 morning
>>101069324gm>>101069263
>>101069324morming
>>101069324monring
>>101069324moorning
>>101069324Morning
>>101069324marning
Dalle3 local fucking when
>>101069587not anytime soon
Qocka is the worst retard
>>101069587we would need some insider to leak the source code, unlikely to ever happen
>>101069587FUTA-DIT-10B is coming soon
>>101069677It's more their terabytes of training data that interests me. I like being able to put in one character prompt with some vague flavor words and get like 80 different artistic styles.
>>101069681it's been announced it will only be an API model
>>101069636
>>101069677it's okay SAI has ｢REDACTED｣ who is notorious for bragging about leaking secret NDA tech
>>101069587Anon, i don't think that's possible.Knowing openAI the only way to get that would be to pay them millions and also own a computer that costs millions to run it.And I have a feeling that's why no employee nor hacker has tried to leak it
>>101067937sorry had to take care of some things before I could build. it's working fine. what are your build params? It's working fine for me
>>101069462
>>101069724It is more about the database than the "compute". Everyone can rent some renderfarm but sorting out billions of images via database queries is bit more problematic.If I had the money, I would establish a company.But I do not have.
>>101069587If you learn to prompt SD3M it's not that far ahead on adherence. It's harder to prompt right to get results, and you will have to sort through more crap but you have as an advantage the capability of photorealism, and you can be as politacally incorrect as you can imagine.
>>101069045this reads like a moba game patch>Increased cooldown on PonyXL ultimate from 50 to 55
>>101069755Local will never be on par with SaaS and it would be great if anons would come to terms with this reality. It's an arms race that we will always lose.
>>101069236american english is closer to ancestral english than british english
>>101069763>sorry had to take care of some things before I could build. its ok>what are your build params?picrel
>>101069496nice one debo
>>101069394That's a different kind of JAP than stable diffusion is usually known for.
>>101069755Best I can give you is tree fiddy>>101069789If it involves things thing like this >>101066987I will never be good at image gen ; ;
>>101069324evening
>>101069841It is not because US is a giant field of yockels pretending to be patriotic. Real English is ancient and allows somewhat different expressions. American English won't allow anything because you are automatically an 'outsider'. I can tell you a story about a restaurant owner from Britain who had perfectly normal British accent but when he went to Kentucky people kept asking him "what is your accent I can't understand you".
was trying to get a "man resting on a girls chest" lmao
The stars look very different today.
>>101069957>american accents can still be understood>some british ones are indecipherable Why is this?
>>101068654>>101069187great pictures
>>101070050I don't know, maybe travel more or refrain from posting.
>>101069957>who had perfectly normal British accentwas from Liverpool wasn't he
>>101070063
>>101070067Nah, just a normal British accent. Americans are so full of themselves they think >a) learning English (American English) is somewhat magical because 99% of them will never learn any foreign languages>b) Americans think some English expressions are offensive because some Californian surfer didn't read some books
>>101069187album cover af
>>101069729see >>101070014
>>101070153>security checks>but it's from chinaman so it must be okay
>>101070067Also: Naval English means any person on board speaking it are speaking English.
>>101070153So we're back? let's fucking gooooooooo
>>101069187I like this style
>>101070171moving the goalpost?
>>101070201wasn't the other anon, just pointing it out
>>101068614Can I use my cpu for stable diffusion?
>>101070208his claim was that it's been announced it will only be an API model
>>101070261always enjoyed these
>>101070032Looks good to me>>101070038MF Doom
>>101069957>American English won't allow anything because you are automatically an 'outsider'.Blame certain individuals, not the dialect. This applies even more to less educated individuals, no matter the country or race. Some personality types are "vigilantes"/"guards" that will be less tolerant of people expressing themselves more fluidly/liberally, and these can be found everywhere on Earth, even Britain. They will come across as being too "stiff" with what is allowed as "valid expression" in communication. This would be common sense to you if you spent more time outside.Go chill with some tea and crumpets then come back.
>>101070222in theory yes, if you got enough ram and your cpu is not a decade old
>>101070345Do you have a tutorial for it?
>>101070304
>>101069892Just experiment with prompts and basic settings
>>101070304>MF DoomRIP. also, ALLCAPS
>>101070115>a) learning English (American English) is somewhat magical because 99% of them will never learn any foreign languagesthat's an unfortunate part of the culture, and English being the international language (thanks to the British Empire and USA) leaves Americans spoiled - foreigners typically come in speaking some English to fit in (except Latino illegals), a lot of foreigners sharing the various social media sites speak some English, even if broken, and a lot of other kinds of popular sites are likely to be majority English speakers, even if they have a bunch of foreigners, the foreigners, again, speak enough English to more or less communicate
>>101069783
>>101069857trying to juggle things but it's fine on my end. did you try rebuilding the engine? is there a trace?>>101069892>I will never be good at image gen ; ;nah, you gucci! (get a burner phone and join the discord already)
>>101070507>international languagelingua franca
>>101070401>too angry to notice how awfully drafty it is down there
>>101070063neat
>>101070553Just imagine asking Fran's 6'5" dad's daughter out on a date past 9PM
>>101070380the easiest way should be easydiffusion, linked in op >>101068614https://easydiffusion.github.io/but keep in mind making a picture in low resolution might take about 10 minutes
>>101070717works on comfy tootried once, lost patience waiting before step1 was completed
>>101070717>easydiffusionthanks anon, going to try that one
>>101070538are engines distributable
>>101070672>Fran's 6'5" dadThat's a good idea for a future gen
>>101070793no it's per checkpoint and gpu. since ai'm using a 4090 it won't work on the 3060
I miss debo
>>101070538>trying to juggle things but it's fine on my end. could you please upload your engine and vae to huggingface or anywhere you prefer>did you try rebuilding the engine?i've tried re-building the engine many times and just rebuilt it, still getting the messy pictures
>>101070849same
>>101070748true, but I tested it on an old computer, installed and easy worked out of the box, detected no gpu and started with cpu without and need to adjust anything so this might be a good start for absolute beginners
>>101070672
>>101070849>>101070872>samefagleave, shoo
>pore
Semi realistic waifu
>Generation will be right backlmao
>>101070895women am I right
>>101070868>>101070538forgot to include logs from terminal (trace?)https://pastebin.com/K5zES8mk
not what you think>>101070517
>>101070970haha ya... "women"
Has anyone actually monetized their stable diffusion outputs?
>>101071028Ethereal farts and you can't convince me otherwise
>>101071028she guffing
i wonder if this will work with 10 files
>>101070913that post timing is pretty unfortunate, ngl lmao. not samefag though, this guy says "i miss debo" every day even though I'm here
>>101071087Are you ever coming back to the discord
>>101071087Please return to the discord
>>101071087thread schizo
>>101071087make a matrix room instead
>>101070222Yes you can, there is >https://github.com/rupeshs/fastsdcpuLast time I did it with i7-6700 and with plain Comfy over a year ago it took 5 minutes to generate 512 variant.15 minutes to upscale it to 1.5x or something.
>>101071152>>101071163doubt it>>101071177I still dk what matrix is. I'm pretty ok with just hanging out in sdg, personally
>>101070326Of course, I have been very priviledged with the fact I have only been dealing with international companies and regardless of my colleagues, they have always been very inclusive of course out of politeness.
>>101071087You got banned for that exact same post and posted the ban with your op. It's in the pastebin
>>101071205matrix.orgapp.element.ioelement.ioOSS alternative to d*scord
>>101070326They need to tolerate my slightly wrong emails etc.
>>101071163>>101071152They kicked him out
>>101071205>>101071241We will let him rejoin, it was a mistake
i guess technically it's a "leopard frog">>101071050>>101071066ok it is what you think>>101071028
>>101071048It always crosses my mind but I'm not sure, it just doesn't feel right, i feel like i can better than the people who already do, my lack of real talent shouldn't be rewarded
>>101071048possible, but a tough market outside of patreon lewds from the looks of it
>>101070717it seems to be working, 6 minutes for an image 512x512 which is better than not being able to, I can add more models or do inpainting and stuff like that, right?
>>101071281
>>101069677What source code? Get 100 million images, caption them with Llava 1.6, and train a 10B model with it. You're making it sound more complicated than it actually is.
>>101071241did he actually get banned, what a sperg
>>101071308It is the database what is needed.
>>101071325Their database will be shit.
>>101071308>muh LLMDOUBTIs that really as good as using well-educated humans with art experience?
>>101071347I was not talking about an exact database but about the fact database is the hardest thing and takes the most computing power in terms of disk space etc.Arranging some GPU to go over it is just a minor detail.
>>101068614Local processing on PCs should be banned. Everyone should be forced to use cloud storage and mainframes.This is the only way to prevent creation of illegal images. Lobotomize those who complain and turn them into McDonalds workers. Especially the GATE tards.
>>101071374A well educated human with art experience with an AI is better than one without an AI. You can keep digging holes with your bare hands though while they use a shovel.
>>101070817probably works on other sm89. ideally it would be configurable, like why can't i just compile sm80 and be compatible with all ampere, or sm86 for consumer>>101071374Amazon should put their warehouses wagies to work captioning, provably better than openai's kenyans
>>101071308Maybe it works with 15M images on a 5B model too. Or some other thing like that. We aren't facing established facts.
>>101071390It really isn't that hard to get 100 million images and buy a couple 14 TB HDDs. The images is probably the least difficult part of all this.
>>101071438This is why SD2 and SD3 were disasters.Your mentality is just a dead end.
>>101071438From where are you getting those 100 million images in the first place? Twitter? 4chan? How are you going to go through them?
>>101071463SD2 was a disaster because they refused to train on anything with flesh tone.SD3 was a disaster because they tried to reward the model away from generating anything with flesh tone.
>>101071483Haha, at least take care of your reddit spaces my friend. Maybe you have such a database, billions of images classified and ready to be submitted to GPU farm.I do not.
>>101071479Obviously you don't naively download a bunch of images, at least not to start. To start you would download every public domain piece of art ever made. Then extract images from every public domain movie. Then images from mass media publications. Then eventually you start some spiders on the rest of the internet.
Why are you generating illegal images instead of working at McDonalds? Don't you need money to live? You wouldn't sell those images would you, that would be a felony....? Just be honest and tell us what you're generating with that nice, expensive PC! Come on now, the government will figure out the truth sooner than later. Just give up those hard drives and those processors so this won't be harder than it needs to be.
still no SD3 CN support in comfy
>>101071520>illegal imagesUnless you're genning CP or making deepfakes for malicious purposes
>>101071514To be honest I didn't read your post.
>>101071507Interesting move of goal post using the exaggerated "billions" assertion. You don't need billions of images, but nice try OpenAI employee.
>>101071552At least you learned how to capitalize your sentences properly. You are still the same troll with no substance whatsoever.
>>101071544To be honest I think you're poor and useless to society and that's why you act this way. I'm sure throwing a tantrum really makes a difference.
>>101071576Why are you wasting your energy writing these posts?
>>101071590Likewise?
>>101071594I can keep doing this forever. One day, you will be classified as a psychopath and they will deny your dole.
>>101071612What kind of schizo rambling is this?
>>101071612>>101071626Kiss already
>>101070868>Error Code 1: Cuda Runtime (out of memory) probably this lol HF upload is taking forever, thanks for being patient>>101070919very cute
>>101071626>>101071624Look, Fran is here.
>>101071612>I can keep doing this forever>deny your dole.is this what they call projecting?
>>101071582Really, really nice.
>>101071661dole = benefit paid by the government to the unemployed.
>>101071649thank you for bothering to upload it anon!!!
>>101070403I noticed another thing about SD3M, it gives me a lot of Asian women, even when prompted. It also has a thing for bangs.
>>101071626You nigfeonbrainedtrainedlanguidtranedupbadgenesbadleanstradmeans Lord Jesus Christ moron lauren forum boring company Felon Musk
>>101071677nta but he clearly understood what it meant otherwise he wouldn't have been able to cast the aspersion of projection
>>101071710Oh you must be quite a writer just to deliver a quip like that, aren't you?
>>101071697>It also has a thing for bangs.don't we all
>>101071710>he clearly understoodI'm so sorry for you because you have never learned any foreign language whatsoever and think English language is something mythical. English and Spanish are the easiest languages on this fucking planet.Sure, maybe I won't be a PhD but I can sure as hell to finish up my daily office tasks.
>>101071738oh, you're arguing just for sake of arguing with anons. I see.
>>101071778No you don't see. You are just a pretentious little faggot.
>>101071305
>>101071778It's probably the person that was arguing about the vtuber Jordan Patterson yesterday, just ignore
>Mismatched elements: 8 / 5242880 (0.0%)the numbers Mason, what do they mean?!
DEbo MEET the DEboHe's the _____ _______ _____
>glowing panties
>>101071874Nice work fran-anon!
Imagine posting images on an imageboard
>>101071845world's best prompter
>>101071980Work on your writing, you are the same troll who lives here.Enjoy the attention psychopath.
>>101071890franon*
>>101071845>He's thecornerstone of sdg
>debo is crying
>>101072020troonyou*
>>101072020Still not posting your own gens.
Where do I find realistic horse cock??
>>101072066no
>>101072066At the horse store
>open SDG>"debo">"politics">corpo worship>doomposting >"thread celebrities">dreamshaper>girlcock>close thread
>>101072097Never bother to use caps.
>>101072105y use caps if smol do trick?
>>101072117no dole for you especially no heroin.
>>101072117Aight, this one is the last one. I hope your family dies in misery and I think they already did, my friend.
>>101071477>>101071669Can you share your workflow? My comfyUI gens are so ass, IDK what I'm doing wrong but I'd love to make sorYXT277mething like this
>>101072127why do you always post old people. would be good if she was 16
>>101072170I love you too :*
>>101072188I love you too
>>101072183You'll understand someday lad
>>101072188I will cut your throat.
>>101072230kinky
>>101072183>20s/30s counts as old
>>101072105He used caps with "SDG" tho
>>101072240
>>101072066in your mom
>>101072260I think it is the same vermin with phone account. Just try it.
>>101072171>Can you share your workflowi load from stable diffusion gens and change prompts
>>101072295oh shit
>>101072355nice
>>101072342i meant a111 gens
debo is a
>>101072379>>101072361Same poster.>>101072364Phone poster.
I'm not giving you a (You) just know you're wrong. >>101072388To a point, you could argue he's mentally raping a certain person who's been obsessed with him for over a year now.
>>101069807The advantage of Local is being able to dictate the content you want.SaaS has terms of service which can contradict the user's vision regardless of the quality. To get the most out of local, you need a monster machine, regardless of the optimizations local comes up with down the line. Not necessarily a Threadripper with an A100 but still a pretty penny on the GPU since expecting to move mountains with 4GB of VRAM is a fool's errand.
>>101072412What compels you to explete all this nonsense?
>>101072412he could also just join (recommit) the discord and stop being a faggot
>>101072393ye
>>101072435Do you only just ask questions?
imagine samefagging yourself, talking in the third person, anything of that sort
>>101072421I feel so afraid about dictating anything.>>101072441You have false sense of bringing thread together because you are psychopath and only way people like will stop is with the 9mm bullet.
>>101072436You know he's going to construe this as "they were begging me to rejoin the losers"
>>101072454Not really, maybe for American, but for the others this is just excrement.
>>101072470can't abuse him if he doesn't rejoin. I extended the olive branch.
>>101072454Maybe it is me who actually needs the license.
What's the lore on debo getting banished from the discord? anon wants to know
>>101072469you are gay
>>101072469You're getting hung up on the word.The point is not having to change or water down your prompts+ideas because a service like Dall-E keeps blocking it.
>>101072511Don't you have anything else to do in your life? I will leave this thread any moment now.I posted all the things I had on my heart.You are one of those garbage people who just sits there like a leech.Enjoy your leeching.
>>101072536I think you are an underage piece of shit.
Bye- for now,Don't let the psychopath eat you!
>>101072592but this is my fetish
>announcing your leave from the threadfaggot
>>101072567Says the anon acting like a child.
I don't understand what's going on in this thread
>>101072592Good bye anon, see you later.
>>101072647discordtroon shitflinging, nothing new
>>101072647Me neither, i just post gens.
When using Pony models, should i bother with Refiner?Ive never used it and i dont know how, is it worth playing around with it?
>>101072780Pretty much no one uses the refiner anymore.
>>101072780regardless of what others say I will always do a 2nd pass/hires fix on any model I use
>>101072780I have never used refiner for anything but i guess it's worth a shot if you think it can improve your gen
>>101072824cute
>>101071681https://huggingface.co/Fizzledorf/somerandomtrtengineandsd3vae/tree/mainI know it's not going to work but here you go
https://files.catbox.moe/6ciuny.jpg
"blonde woman laying in the grass"pajeet coded model
https://files.catbox.moe/8eolor.jpg
>>101072875>jpgnah anon, you have to share this masterpiece
>>101072904I don't share it across nogens, I would if I was with other artists but because this thread is 90% of text and the rest of the images are made by me and by couple of other anons, I won't do it.
>>101072917thank you for not providing the means to generate this slop
>>101072904You are the same psycho anyway. Learn how to use>>101072926 capital letters please.
>thread is just a couple of anons spamming and shit talking Huh
>>101072933boo
>>101072926You are the reason why no one posts here.
>>101072945Laura Kinney Special (this one is old though)https://files.catbox.moe/8eolor.jpgNow you psychopath can repeat and repost your shit little texts again and again.
>>101072937Welcome to /sdg/
>>101072858>im on linux>[06/20/2024-19:46:53] [TRT] [E] IRuntime::deserializeCudaEngine: Error Code 1: Serialization (Serialization assertion plan->header.pad == expectedPlatformTag failed.Platform specific tag mismatch detected. TensorRT plan files are only supported on the same OS/Arch platform they were created on.)oh shit.. sorry for wasting your time anon, thank you still
>>101072937You are the reason why, dear psychopath.
Did I miss anything important /sdg/
>>101072989Maybe go to /ldg/ instead, chang?
>>101073004Yes. The usual suspects are fighting again.
>Mismatched elements: 76 / 65536 (0.1%)>Mismatched elements: 5 / 524288 (0.0%)so annoying having to increase the tolerance for such a small number of elements
>>101073004the nogens are really, really mad.or in other words, no, nothing important.
>>101073018Just like you, nogen.
>>101073018>>101073032Sad times. What are they fighting over this time?
Here comes the sun.
>>101073033I've got no dog in this fight
>>101068785>>101068682Love your work.
>>101073027me on the left
>>101073040It's only about useless shit with no importance but yet someone has been doing it for almost a year by now.
>>101073051If you don't, do not reply then. Or if you do, be constructive.
>>101073093faggot
>>101073103At least capitalize your effort.
Migu>>101073100>>101073100>>101073100
>>101072900>>101072825Those are some nice Super Quokka Bros, quokkAnon! If you happen to catch the next thread then there's been an update to the paste at the usual spot. Thanks for playing!
>>101073197Thx anon, I'll be sure to check it out
1.5 is kino still
>>101073441that's 1.5?
>>101073458yea
>>101073458yeah, one of the biggest issues about 1.5 is its composition but now you can easily work it out with the many extensions there are available, with image quality you can always upscale which also has many alternatives, the one big thing about 1.5 is its prompt adherence, I saw some posts of someone using using 1.5 with t5 encoder, that showed a lot of promise
>>101073571with the t5 encoder?????
>>101073588it was on reddit
>>101072858i tried running linux headless, it built the engine file without OOMing but it still doesnt work>https://pastebin.com/Mx0XQrCGvery weird.
>>101071402Great, all the more reason to hire a bunch of them to do all that dataset prep. :)
