previous got yeeted cause spammers so behave : >>101432110>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
we're back
Will StabilityAI be able to fix Stable Diffusion 3?also first for (tr)ani is literal human garbage
>mfw Resource news07/16/2024>DataDream: Few-shot Guided Dataset Generationhttps://github.com/ExplainableML/DataDream>UltraPixel: Advancing Ultra-High-Resolution Image Synthesis to New Peakshttps://jingjingrenabc.github.io/ultrapixel>AccDiffusion: An Accurate Method for Higher-Resolution Image Generationhttps://github.com/lzhxmu/AccDiffusion>US Senate introduces bill to setup legal framework for ethical AI developmenthttps://www.techspot.com/news/103820-us-senate-introduces-ground-breaking-bill-setup-legal.html>KineTy: Kinetic Typography Diffusion Modelhttps://seonmip.github.io/kinety>DiffStega: Towards Universal Training-Free Coverless Image Steganography with Diffusion Modelshttps://github.com/evtricks/DiffStega>Arbitrary-Scale Video Super-Resolution with Structural and Textural Priorshttps://github.com/shangwei5/ST-AVSR07/15/2024>Google's Gemini AI scans Google Drive PDFs without permissionhttps://www.tomshardware.com/tech-industry/artificial-intelligence/gemini-ai-caught-scanning-google-drive-hosted-pdf-files-without-permission-user-complains-feature-cant-be-disabled>ComfyUI-ppm: Fixed AttentionCouple/NegPip, more CFG++ samplers, etchttps://github.com/pamparamm/ComfyUI-ppm07/14/2024>ControlNet-Union-SDXL updated, adding Inpaint and Tilehttps://github.com/xinsir6/ControlNetPlus/tree/main>ComfyBlog July: Versioned Releases, Beta UI, New Modelshttps://blog.comfy.org/new-models-versioned-release-beta-ui-new-models>danbooru2023-florence2-captionhttps://huggingface.co/datasets/KBlueLeaf/danbooru2023-florence2-caption>Nvidia RTX 50 family TDPs leaked by Seasonichttps://www.tomshardware.com/pc-components/gpus/nvidia-rtx-50-graphics-card-family-tdps-leaked-by-seasonic>GLSL Nodes for ComfyUIhttps://github.com/patriciogonzalezvivo/comfyui_glslnodes07/13/2024>Kolors + Controlnethttps://github.com/MinusZoneAI/ComfyUI-Kolors-MZ>ToonCrafter_with_SketchGuidancehttps://github.com/mattyamonaca/ToonCrafter_with_SketchGuidance
why did ldg get nuked?
>mfw Research news07/16/2024>VGBench: Evaluating Large Language Models on Vector Graphics Understanding and Generationhttps://arxiv.org/abs/2407.10972>InVi: Object Insertion In Videos Using Off-the-Shelf Diffusion Modelshttps://arxiv.org/abs/2407.10958>IDOL: Unified Dual-Modal Latent Diffusion for Human-Centric Joint Video-Depth Generationhttps://yhzhai.github.io/idol>OPa-Ma: Text Guided Mamba for 360-degree Image Out-paintinghttps://arxiv.org/abs/2407.10923>Benchmarking VLMs for Cultural Understandinghttps://arxiv.org/abs/2407.10920>Optical Diffusion Models for Image Generationhttps://arxiv.org/abs/2407.10897>MoE-DiffIR: Task-customized Diffusion Priors for Universal Compressed Image Restorationhttps://renyulin-f.github.io/MoE-DiffIR.github.io>When Synthetic Traces Hide Real Content: Analysis of Stable Diffusion Image Launderinghttps://arxiv.org/abs/2407.10736>Addressing Image Hallucination in T2I Generation through Factual Image Retrievalhttps://arxiv.org/abs/2407.10683>Evaluating Model Bias Requires Characterizing Mistakeshttps://arxiv.org/abs/2407.10633>WildVidFit: Video Virtual Try-On in the Wild via Image-Based Controlled Diffusion Modelshttps://arxiv.org/abs/2407.10625>InsertDiffusion: Identity Preserving Visualization of Objects through Training-Free Diffusion Architecturehttps://arxiv.org/abs/2407.10592>A Survey of Defenses against AI-generated Visual Media: Detection, Disruption, and Authenticationhttps://arxiv.org/abs/2407.10575>Local Action-Guided Motion Diffusion Model for Text-to-Motion Generationhttps://jpthu17.github.io/GuidedMotion-project/>How and where does CLIP process negation?https://arxiv.org/abs/2407.10488>Masked Generative Video-to-Audio Transformers with Enhanced Synchronicityhttps://arxiv.org/abs/2407.10387>Sora and V-JEPA Have Not Learned The Complete Real World Model - A Philosophical Analysis of Video AIs Through the Theory of Productive Imaginationhttps://arxiv.org/abs/2407.10311
>mfw MORE Research news>Noise Calibration: Plug-and-play Content-Preserving Video Enhancement using Pre-trained Video Diffusion Modelshttps://arxiv.org/abs/2407.10285>Beyond Prompt Learning: Continual Adapter for Efficient Rehearsal-Free Continual Learninghttps://arxiv.org/abs/2407.10281>Disrupting Diffusion-based Inpainters with Semantic Digressionhttps://arxiv.org/abs/2407.10277>Low-Rank Interconnected Adaptation Across Layershttps://arxiv.org/abs/2407.09946>LeRF: Learning Resampling Function for Adaptive and Efficient Image Interpolationhttps://arxiv.org/abs/2407.09935>VividDreamer: Invariant Score Distillation For Hyper-Realistic Text-to-3D Generationhttps://arxiv.org/abs/2407.09822>Explanation is All You Need in Distillation: Mitigating Bias and Shortcut Learninghttps://arxiv.org/abs/2407.09788>Layout-and-Retouch: A Dual-stage Framework for Improving Diversity in Personalized Image Generationhttps://arxiv.org/abs/2407.09779>TemporalStory: Enhancing Consistency in Story Visualization using Spatial-Temporal Attentionhttps://arxiv.org/abs/2407.09774>Diminishing Stereotype Bias in Image Generation Model using Reinforcemenlent Learning Feedbackhttps://arxiv.org/abs/2407.09551>How Stable is Stable Diffusion under Recursive InPainting (RIP)?https://arxiv.org/abs/2407.09549>MATE: Meet At The Embedding -- Connecting Images with Long Textshttps://arxiv.org/abs/2407.09541>BitNet b1.58 Reloaded: State-of-the-art Performance Also on Smaller Networkshttps://arxiv.org/abs/2407.09527>Putting GPT-4o to the Sword: A Comprehensive Evaluation of Language, Vision, Speech, and Multimodal Proficiencyhttps://arxiv.org/abs/2407.09519>Self-Consuming Generative Models with Curated Data Provably Optimize Human Preferenceshttps://arxiv.org/abs/2407.09499>Augmented Neural Fine-Tuning for Efficient Backdoor Purificationhttps://arxiv.org/abs/2407.10052>Image captioning in different languageshttps://arxiv.org/abs/2407.09495
>>101433215>ldgno idea, but /sdg got nuked to
What would happen if someone would stop spamming these links?
>>101433243all made by the banned poster
>>101433215>>101433243the guy who makes ldg went on a spam campaign here and jannies nuked his whole history
why is there a discord in OP?
>>101433243there is nothing strange in the archive for the ldg thread, i guess whoever mass reported sdg also reported ldg for some reason.
>>101433273My sweet summer child...
This is an illustration I made in the meanwhile. Just photoshop and lot of arrogance.
All threads that had /sdg/ in the op got deleted, it's clear this general is the problem
>>101433286i don't keep up with thread lore anymore, i wanted to talk about the new cascade model https://jingjingrenabc.github.io/ultrapixel/does anyone know how much vram it takes to inference in comfy? is it any good?
Is it summer fags?
>>101433307well done!
>>101433313Cascade was always around 20 GB. The problem here is the fact if you need to ask if it is any good you are not a synthographer but a troll.
>>101433273>>101433312>>101432232It's just the doxxcord spamming again
>Rate exceededrudei'll just get new nodes lmao gg
>>101433435You are too passive aggressive to arrange any opposition.
my kingdom for a normal looking bike
>>101433497sweet
>>101433497that is fun
>>101433463>arrange any oppositionwhat does this mean
Wait, people actually go to LDG?
>>101433556It is a double take on claimed passivity.
>>101433585Proper British would be "proposed" or "called".
>>101433585idk what you're on about desu
>>101433609>>101433556I answered this post.
>>101433581why shouldnt they? there are other image gens than sdg and dall-e .. so the general is legitimate, but as I seen it it moves pretty slow
>>101433581Only serious imggen discussions happen on /ldg/, here people use avatars and say a lot of off topic sutff
>>101433546>>101433554thankgyo
what do you think debo went to do?
>>101433665ur mom lul
>>101433676thread schizo
>>101433688Why are you not yet banned?
>>101433625sorry but still idk what you're trying to say
>>101433692Go on with your life then, that's your own personal problem.You are one of the posters on these threads who actually cross time zones, you are probably an American and think why does this guy speak different English - it's because he's from EU. American English is not everything, if you ever went to school US is not the center of this planet either. Faggots like you make me sick.
>>101433690why would i get banned? im always nice
>>101433728No you are not, you are pretty much passive piece of shit.
>>101433739better than actively messing up the thread i guess senpai
>>101433720i'm british, i just don't know what you're trying to say, "to arrange any opposition" doesn't really make sense, sorry anon
>>101433728I got banned for laughing at Trump being shot
>>101433759You are not from Britain.
>>101433758It's a war against empty words
>>101433779are you ok? you could stop at any moment you know?
>>101433759Of course you are. English is such a mythical language it takes few guys to decipher some words.But if I read some stuff in Finnish I can instantly call foreign borrowed words and forms, but let some words be because it's how they bend.
>>101433790I just wiped my ass with your opinions and wishes.
>mfw Research news07/16/2024>GayBench: Evaluating Large Language Models on Vector Graphics Understanding and Generationhttps://arxiv.org/abs/2407.10972>InVi: Object Insertion In Videos Using Off-the-Shelf Diffusion Modelshttps://arxiv.org/abs/2407.10958>IDOL: Unified Dual-Modal Latent Diffusion for Human-Centric Joint Video-Depth Generationhttps://yhzhai.github.io/idol>OPa-Ma: Text Guided Mamba for 360-degree Image Out-paintinghttps://arxiv.org/abs/2407.10923>Benchmarking VLMs for Cultural Understandinghttps://arxiv.org/abs/2407.10920>Optical Diffusion Models for Image Generationhttps://arxiv.org/abs/2407.10897>MoE-DiffIR: Task-customized Diffusion Priors for Universal Compressed Image Restorationhttps://renyulin-f.github.io/MoE-DiffIR.github.io>When Synthetic Traces Hide Real Content: Analysis of Stable Diffusion Image Launderinghttps://arxiv.org/abs/2407.10736>Addressing Image Hallucination in T2I Generation through Factual Image Retrievalhttps://arxiv.org/abs/2407.10683>Evaluating Model Bias Requires Characterizing Mistakeshttps://arxiv.org/abs/2407.10633>WildVidFit: Video Virtual Try-On in the Wild via Image-Based Controlled Diffusion Modelshttps://arxiv.org/abs/2407.10625>InsertDiffusion: Identity Preserving Visualization of Objects through Training-Free Diffusion Architecturehttps://arxiv.org/abs/2407.10592>A Survey of Defenses against AI-generated Visual Media: Detection, Disruption, and Authenticationhttps://arxiv.org/abs/2407.10575>Local Action-Guided Motion Diffusion Model for Text-to-Motion Generationhttps://jpthu17.github.io/GuidedMotion-project/>How and where does CLIP process negation?https://arxiv.org/abs/2407.10488>Masked Generative Video-to-Audio Transformers with Enhanced Synchronicityhttps://arxiv.org/abs/2407.10387>Sora and V-LIGMA Have Not Learned The Complete Real World Model - A Philosophical Analysis of Video AIs Through the Theory of Productive Imaginationhttps://arxiv.org/abs/2407.10311
>>101433824this smells like ban evasion now .. wtf fuck is up with you shit head?
>mfw MORE Research news>Cumsock Calibration: Plug-and-play Content-Preserving Video Enhancement using Pre-trained Video Diffusion https://arxiv.org/abs/2407.10285>Beyond Prompt Learning: Continual Adapter for Efficient Rehearsal-Free Continual Learninghttps://arxiv.org/abs/2407.10281>Disrupting Diffusion-based Inpainters with Semantic Digressionhttps://arxiv.org/abs/2407.10277>Low-Rank Interconnected Adaptation Across Layershttps://arxiv.org/abs/2407.09946>LigmaeRF: Learning Resampling Function for Adaptive and Efficient Image Interpolationhttps://arxiv.org/abs/2407.09935>VividDreamer: Invariant Score Distillation For Hyper-Realistic Text-to-3D Generationhttps://arxiv.org/abs/2407.09822>Explanation is All You Need in Distillation: Mitigating Bias and Shortcut Learninghttps://arxiv.org/abs/2407.09788>Layout-and-Retouch: A Dual-stage Framework for Improving Diversity in Personalized Image Generationhttps://arxiv.org/abs/2407.09779>TemporalStory: Enhancing Consistency in Story Visualization using Spatial-Temporal Attentionhttps://arxiv.org/abs/2407.09774>Diminishing Stereotype Bias in Image Generation Model using Reinforcemenlent Learning Feedbackhttps://arxiv.org/abs/2407.09551>How Stable is Stable Diffusion under Recursive InPainting (RIP)?https://arxiv.org/abs/2407.09549>MATE: Meet At The Embedding -- Connecting Images with Long Textshttps://arxiv.org/abs/2407.09541>BitNet b1.58 Reloaded: State-of-the-art Performance Also on Smaller Networkshttps://arxiv.org/abs/2407.09527>Putting GPT-4o to the Sword: A Comprehensive Evaluation of Language, Vision, Speech, and Multimodal Proficiencyhttps://arxiv.org/abs/2407.09519>Self-Consuming Generative Models with Curated Data Provably Optimize Human Preferenceshttps://arxiv.org/abs/2407.09499>Bugmented Waifu Neural Fine-Tuning for Efficient Backdoor Purificationhttps://arxiv.org/abs/2407.10052>Niggerniggermage captioning in different languageshttps://arxiv.org/abs/2407.09495
I am Schizo AnonAsk me Anything
@101433831d*bo
>>101433833Nobody cares faggot
>>101433904I only waited for your post.This is typical US post. English is still an international language.Not your special escape from the Queen of England. I did not expect you or anyone to be that careful. You are just busy with other things anyway.
>>101433938Cope faggot
>>101433949This one really hurt. At least you didn't try to correct my phrasing.
>>101433897Are you in the Discord?
>>101433985I am. my name is "Joaquim", I am an art director.
>>101433994Fuck off
>>101434001Seems like you are afraid of the talent.
>>101434016fix the hand Fritz
Nobody cares, just post booba
>>101434024I can fix anything you want. I only post images here for my fun.I enjoy your anger when it comes down to call me Fritz, it is really an honour of sorts my friend.
i unironically wish trani would leave forever
>>101434057I know you do debo
>>101434053okay d*bo ;^)
Any reason to pick Automatic1111 over SD.Next? The latter seems way more advanced. Looking to get back into SD after a break and always used Automatic1111.
>>101434073no im schizoanon
>>101434091Use CummiesUI
48mjust temporary block, increased the random delay instead of getting new nodeslots of posters playing up today, if they were my cat i'd hold them like a baby and shush them
shush shush, it's ok
>>101434091Also, is it all SDXL these days? I used to get decent results with SD2.1, but if SDXL is better I'll go to that.
>>101434131I hate you...
>>101434131
is that a display of power?last banwave schizo acted like he's gone now everything containing /sdg/ was nuked from orbit and he continues to be here?
>>101434146i know i know, shush now, you'll be ok>>101434151glistening
>>101434091they are very similar by now, a1111 just werks tho (once it werks) .. sd.next crashed on me multiple times when I last tried it, Id say both are very valid still, sd.next is in more active development tho and supports plenty of different diffusion models .. automatic commits to both .. so idk what the future holds for a1111
>>101434178Narcissism is a symptom of psychopathy.
>>101434141sdxl finetunes like juggernaut xl or pony far surpass sd15 sd21 by now in their respective target groups (juggernaut, realvisxl for general purpose, pony for 2d porn)
>>101434178
>>101434210>finetunesI assume you mean checkpoints by that? And thanks for the answer btw.>>101434179You too anon
>>101434267yw, same anon .. and yea checkpoints, finetuned on base model, just go to civitai.com and sort for most downloaded you will find some excellent ones.. be sure to read the discription tho, some are very specialized and or need weird quality prompts like the pony models .. you wont get far with a pony model when you want to make nice backgrounds instead of hyper realistic pussy
I feel like I am stuck... it hurts! I need to generate pederastic images and yet claim I am from California!
>>101434419Get a vpn
>>101434429What for?
>>101434429I am using VPN but for Finland->United States, these are my connections.
>>101434461I mean I am naked.
We need some new drama in these threads
>>101434474disgusting, put on some clothes
>>101434501In Finland and Sweden, when we are at home were are literally naked.
>>101434516won't that like horribly stain your chair and couch in the long term?what I like about scandinavia tho is sauna .. wish I had one
>>101434494Do you really?
>>101434474I always liked Finns.
>>101434516More Pinkeye in Sweden
>>101434535Yes comfy, we do. We miss your days of shilling.
>>101434091auto1111 support all the extensions, SD.Next and Forge don't
>>101434541Swedish are the box headsFinnish are the gods.
>>101434091there's no Auto1111 with forge optimizations though
>pw banned for avatarfagging (again)>deb* in shambles>julien still a piece of shit>koff redeemed>rainy girl laughing hystericallyschizo anon won
lads, when generating with an SDXL model, what base resolution should you use? I've read 832 x 1216 is good but when doing hires fix at 2x it lags my pc since it runs out of memory. also how are you supposed to inpaint an image with such high resolution??
>>101434516can confirm am eurotrash in california and am naked rn
>>101434588how do we know it's rainy girl or not, I don't know how he types
>>101434592On SDXL I use 1024 as a base, but it'll go up safely on mine with 8gb to like 2048x2048.XL doesn't seem to be as sensitive to aspect ratio fuckery and weird element uh, mutations based on dimensions like 1.5* was
>>101434596This is what god made you and I can bless you! This why they call me bad company until the day I die.
Guys, what are the requirements that I need in my PC to run Stable Diffusion? ( I wanna run it fast)Im searching this on the internet but every source says something different.(Sorry for bad english)
>>101434588i didnt cause any of this desu
>>101434615gpu-ram-cpu specs and we will see.
>>101434620yes you did julien
>>101434557people forget how masterful comfy was at baiting back in his prime
>>101434615VRAM, Nvidia, xformers installed in whatever setup you have.The rest matters little.The vulkan build for KoboldCPP and the like works as well as Cuda for LLM, but I don't know if there's a vulkan compute build of diffusion somewhere, if there were, those big 24gb meaty, cheap AMD cards would be really viable.
>polar bears and comic booksreally cool dude posts recently ty>M2N2P>Man you need to post your penisWhat?>Error: You seem to have mistyped the CAPTCHA. Please try again.What?>Verify you are human.>Typing so fast Linux Mint's compositor can't keep up with my ML typewriter functioning in the background, it told me this.>YHTMWhat the hell yo.
>>101434588can you shut the fuck up for a single thread d***?
>>101434566There are no "forge optimizations" there is only "a1111 with comfy code".
>>101434603hmm so is that 2048x2048 without hires fix? what's the point of hires fix then?
>>101434644oh god not this shit againyou're the schizo arent you?
>comfy STILL thinks A1111 stole comfy codeHerewegoagain
>>101434648I don't know, honestly. It seems the XL models I use (pony based) don't need it, but I remember how important it was on 1.5's.
>>101434644no one takes what you say seriouslyyou and your bf will forever be lolcows in this community
>they had to steal codemale_bear_guffaw.wav>VHHDVirtual DM: Hee hee!
Thank you.
>>101434661>>101434697>>101434702https://old.reddit.com/r/StableDiffusion/comments/1e3nslq/reforge_webui_updates_and_in_need_of_a_bit_of/>So, when I was doing some edits in the code, I did notice that most of the backend (about 80%) of Forge is yep, ComfyUI. They share the same files, code and such.
You think we'll be able to jailbreak OpenAI's Sora? Or will that be impossible? It seems really cool but they're whole "extensive testing to ensure safety and lack of offensive/nsfw/hateful/misinformative/dangerous content" makes it basically useless and trash.
>>101434668>>101434644comfy is little ego man who has a hard time accepting that people still prefer to use what he considers to be an "inferior" product made by a smallbrainit's okay comfy, us autists still love you
>>101434719Sorry for posting unedited image.This is from photoshop.
>>101434764This is what it means for a soldier to meditate.
>inb4 shitty 4 keyframe """animations""" reposted again
>>101434747That is unbelievable, you should go after them and get forge taken down
>>101434747screech in their repo again cumfart?
>>101434749you'd never have the hardware to run it
stop playing up!!
>>101434749>With cinematic shots, the ideas of ‘tracking’, ‘panning’, ’tilting’ or ‘pushing in’ are all not terms or concepts captured by metadata. As much as object permanency is critical for shot production, so is being able to describe a shot, which Patrick noted was not initially in SORA. “Nine different people will have nine different ideas of how to describe a shot on a film set. And the (OpenAI) researchers, before they approached artists to play with the tool, hadn’t really been thinking like filmmakers.” Shy Kids knew that their access was very early, but “the initial version about camera angles was kind of random.” Whether or not SORA was actually going to register the prompt request or understand it was unknown as the researchers had just been focused on image generation. Shy Kids were almost shocked by how much the OpenAI was surprised by this request.https://www.fxguide.com/fxfeatured/actually-using-sora/garbage
>>101434870I would do disgusting things to her
>>101434803Forge is already dead. Unfortunately I can't take the credit for killing it because a1111 is the one that killed it by telling the author to fuck off and refusing to merge any of the forge "improvements".
>>101434920Dad..thank you for replying. I love you dad.
>>101434890I bet you would, but she is virtual
>>101434920based auto
>>101434879sora is a very powerful world simulator but runway is the primier toolkit. runway brings in camera control, lighting control, masked inpainting, etc. when there's a marriage of the two sides into a full solution, videogen will finally be usable as a production tool
>>101434920Sorry dad... I forgot my Token.
>>101434920manic phase again?
>forge is ded so.. any alternatives? forge is the only one that works fast on my potato.
>>101434879i'm surprised sora doesn't work like that desu. i suppose they don't have the data. i have data like that for individual shots from tv&film, tagged by people in the tv&film industry, but not for video
>>101434936picrel>>101434971I remember the day it came out and directors pointed out it's useless if there is zero control over when things happen and the camera controls
https://www.youtube.com/watch?v=wo_h_FNlXbo>>101434947reforge, I guess. someone forked forge and is trying to merge in all the upstream a1111 stuff plus [more] comfy stuff. I can't link the updates in the news cuz he doesn't version/tag and just dumps all his updates in bloated reddit posts
>>101434979https://www.youtube.com/watch?v=BJBiRTX5-MgMariya.
>>101434947Forge was just the a1111 interface on top of ComfyUI so you can just use ComfyUI directly instead or another interface like SwamUI: https://github.com/mcmonkeyprojects/SwarmUI
>>101434978desu i can get data for it. ill get some clips from tv&film ready
>>101435013ask them to do a girl lying in grass
>>101434562https://files.catbox.moe/s2bbil.jpgreminded me of this
>>101434920you mean like SAI told pony to fuck off even though it's still the best model currently to date? then you release steaming turd SD3? fuck off from this general forever no one believes your lies anymore you schizo
>>101435050>it's still the best model currently to date?what a joke.. its the best at coom, its not the best overall
>>101435050the fuck is wrong with you debo?
>>101435079>>101435084better than your model that can't even lie on the grass properly
>>101434998So are Swarm and Comfy just gonna exist as separate UIs, even though mcmonkey left stability to join comfyorg? I would've imagined that there would be a merger into one unified interface that shares a collective vision.
>>101435095my model? you are obviously mistakenly think you only talk with one person here, I have no affiliation with SAI or any other model maker
rofl... schizo really deep in his own head now, popcorn time
>>101435050>best model currently to dategod no it's just the best at porn atm. same shit as AOM. the model is ruined by the dumbass decisions purplesmart did for moral reasons but also the stupid score tagging. he would be supporting all the dumbass shit safety does too. look at how well the model initiative is going btw kek
>>101435119i hope your depression will get worse soondont forget the flip faggot
>>101435109then fuck off nobody and let the big boys talk
>debo wishing death on aniunbelievable to see
>>101435102people use comfy over swarm for a reason
>>101435134says the smallest fry in the pan
So, some psycho can spam these threads for weeks.But when I mention "nigger" in a some post which is part of language play, I am going to get banned. Without remorse.Let everyone else run like a fucking hog.
>>101435119you were kind of away from sdg for a while when the pony drama was happening. I get that you had thoughts on the model and the author's approach but was there a reason you came out so aggro against him? seemed like there was bad blood there or something
>>101435161he is ban evading.. what can ya do
>>101435163debo, he's just an asshole, end of story
>>101435172who's ban evading
>>101435172It is not about ban evasion it is about the action.This guy has been going on for months.
>>101435119It's the best at anime and semi-realistic. I don't make porn with it. It would be fair to argue that merges like autismmix are better than the base model and are actually the best right now. But SAI supports ethical moral model stuff too.
>>101435177it takes one to know one
>>101435163>was there a reason you came out so aggro against him?all you have to do is have one conversation with him to know he just paid his way through it. he is pretty clueless on the nuances of training>>101435193>It's the best at anime and semi-realisticthat would be NAI for anime. it has untainted artist tags at least>semi-realisticit all looks the same to me most of the time I don't really know what's supposed to look good if it isn't my cup of tea
>>101435161mods are in it with the sai employees, they know everyone's IP and abuse it. you think your anonymity matters on a basket weaving website when hundreds of millions of dollars are being invested and passed around and on the line
>>101435224>lineYeah but most of the shit are paid shills on /v/ and /pol/ and Hiroshima sells everything forward now that AI is so damn popular.
>>101435224>>101435237take your meds
>>101435250??? Go back.
I don't know about you guys but I love Julien
>>101435217>all you have to do is have one conversation with him to know he just paid his way through it. he is pretty clueless on the nuances of trainingthe most classic comfy cope "b..b..but i'm so smart" doesn't matter how smart you are, it matters that other people are way more effective at doing the things that matter in this community, so many people use pony and pony derivatives for a reason
>>101435217pedo
heh, molotov cocktail
>>101435247kek "An TITI fa"
is that really the discord raid tier shit again of cumfart and his unsuccessful loser boyfriend julien? lmao
https://huggingface.co/xinsir/controlnet-union-sdxl-1.0/tree/mainall in one SDXL controlnet, actually works...pretty neat
https://files.catbox.moe/lb2099.jpgThis is for the kuffar.
>>101435281do you think anyone would even care about pony if NAI leaked? that's how you know it's hardly a SOTA model in any regard
>>101435344Anon...jesus christ
>>101435337canny test, first image:
>>101435344giwtwm
>>101435351No, it's Buddha we pray here.
>>101435361result: (prompt more important in controlnet)neat, all I did was change the character prompt (ganyu, from lumine).
Kuffar.https://files.catbox.moe/lpt7c7.jpg
>>101435119how hard is it to keep a notepad file handy, takes five seconds, meanwhile sd3 is the worst model ever released and can't be redeemed by any tags ever, sad
>>101435347100% yes, it's way more versatile for one thing but i honestly haven't seen anything particularly impressive from NAI at all
>>101435401>>101435344you won't do ani
>>101435432He is my colleague. Or inspiration.
>>101435432>"I won't do anyone" How would you describe this to someone who is from England? I won't deck you in?What is it?
>>101435463ani is a person bruv
>>101435407>how hard is it to keep a notepad file handywhy should this be a thing when I can just prompt an artist with any other model? I don't have anything to do with the prior model release btw. I only touched the new ones.>>101435410>i honestly haven't seen anything particularly impressive from NAI at allthere won't be too much difference other than the shitty things about pony aren't present. downside is no furry smut to improve genitalia or niche fetishes. if we could train loras or finetune NAI that wouldn't be an issue. I have to destroy the pony unet with loras in order to get something I want and even then the random ass style bleed that pops up usually ruins it>>101435474how did you know!?!??
>>101435463could not give a fuck about britbongs>>101435450Colleague? hmm....
>>101435474Is this an American way of saying something in English?I won't do Ani.What the hell?
>>101435483Remember, 20% of the population are speaking two languages and only 5% of the white population have never studied a second language.You are not such a 'dominant' 'white' population now aren't you?
using pony btw since there isn't any other choice fuuuuuck. nobody (even me) wants to drop 10k for a better trained xl smut model because we are feeling out what model is the new de facto best to train with model.
>>101435378and one more:
>>101435172who?
>Julius Cesar taking a selfie
>>101435481>why keep a notepad of pony tags when I can just keep a notepad of artist tags???
>>101435575>he can't remember artist names
why would SAI kneecap their model (sd3) by censoring it, and making the anatomy worse?there is a very good reason pony based models are good at making figures. cause the model fundamentally knows how to make all aspects of a figure.
damn, I think I have to give it to sd3 over pony
>>101435581pony was trained on a cuck model so if anything what you said doesn't even matter
what cellphone brand is Cesar using? it certainly aint an iPhone>>101435581laws, public outcry, general American bigotry, fear of being branded as cp peddler
>>101435485shut up bruv. little neek gunna get chefed up
>>101435485>I won't doIs this really an expression you don't understand? It means refuse to perform a task or action. In this context, you won't create an image of ani similar to the other one you created.
>>101435599if you train a model on everything, from tame to explicit stuff, you have a wide dataset so figures can be depicted with more accuracy than sd3 which is a neutered model.
>>101435607It is funny that Americans always think it is about lowest nominator.Trying to explain slowly. As if ... as if...In this context, I wouldn't even listen to your whining.
>>101435614sdxl could barely do breasts (just like sd3). I am not defending the ablation but sdxl had the same treatment. Every other open source model doesn't have smut either so you are in the same boat no matter what
>>101435607>I won't doIt is insulting. Maybe you have learned it in the dorms or something.
>>101435432>I won't do AniIt means I wont (do an action) to AniIn this case draw, generate, etc.It can also mean I won't "do" (having sex) with said person. But in this case the first one is the more appropriate from context.
lol the troll got baited so hard in explaining his troll msg .. I can't
>>101435629I used to teach ESL>>101435639What?
>>101435651I'm 40+ and have lived in the UK. I still think you are a cunt.Don't try to forward your action to me or I will cut your throat.
https://huggingface.co/xinsir/controlnet-union-sdxl-1.0neat, it works for canny, depth, openpose, AND sketch, all in one (SDXL). even works for pony based models. just rename it to (whatever).safetensors and it works for all of them.just set "prompt more important" in controlnet and it works every time.
>>101435665I think they key here is the fact I have NI number and lived in UK for over a decade and you don't have a shit.
>>101435677
>>101435666That was my first reply to the topic and I explained the syntax properly and without being rude. Maybe you are the one that is a massive cunt, so go ahead and fook off m8. Enjoy your shithole bitch
>>101435665Did they call you from that ESL number? Who is your NI number?
>What do you mean you "lost" an alien body?
yhyh b bout 45 mins
>>101435692one more: canny on a rei image = saber ayanamithis is cool because some SDXL controlnets didnt work with pony based models or with certain SDXL models.
>>101435730original:
>>101435695Thing is - I am more British than these white pastry guys here. Regardless of my language.This is why I love Britain, my second home.
>>101435695I'm going to deck you in.
>>101435730and one more, this with an asuka source image:
>>101435743>>101435757So don't come on American website and start shitting yourself and calling names when people give you honest answers to questions you ask. Also I know you don't have any knives to cut me with fool, you need a loisense for that.
>>101435763>le americanhttps://www.youtube.com/watch?v=uogMDqVv_Tc
>>101435696I worked at a uni. I don't know what an NI number is.
dall-e 3 through bing is proper desu. microsoft won
>>101435778It is the social security number in UK.
>>101435781can't do porn so garbage
>>101435781If I can't run it locally without the internet.... forgetaboutit
>>101435792who cares>>101435793when do you not have the internet though
>>101435802...there may be this time in the near future, what if!!
>>101435666>I'm 40+gross
>>101435677>neat, it works for canny, depth, openpose, AND sketch, all in one (SDXL)does that mean it has a larger memory footprint than the standalone models?
>>101435884Think of the beans
>>101435884Also 40+,, need knowledge to gen works of substance Anon.
>>101435884After your parents have paid your 'university', what is next?
Night Danger I feel sorry... Darkness has the power.
>>101435677>https://huggingface.co/xinsir/controlnet-union-sdxl-1.0wdym "just rename it" like i need to name it depth.safetensors for it to work as a depth net (in auto?)
sd3 background + pony inpainted foregroundnice but fucks shadows
have to describe it instead of just asking for juicy couture but it gets the font pretty close sometimes. this one's a knockoff she's had off the market though>>101435848i mean if there's no internet at all you've probably got more important things to do than generating images like staying alive
>>101435915Seething at anons here for answering all his ESL questions, apparently.
>>101435991Night....I am in the hands of Satan!
>>101435989you think losing internet is a matter of life and death? that's the most zoomie thing I've ever heard
>>101435986Nice but yeah, the inpainting is too obvious.
>>101435989That's a good point anon, kek
>>101435991https://www.youtube.com/watch?v=E3zBl1R6TWAThis song is made by a real world psychopath.
>>101436023Greenback my darling!
>>101435901I don't know all the technical details, but I think it's just all of the controlnet types in 1 file but loaded separately depending which you pickI see control type HARD EDGE, and control type OPENPOSE when using 2, some SDXL canny or depth models by *themselves* are 2.3 gig, this model with all of them is also 2.3 gig.
the crown is actually a juicy couture symbol. sure with local i could train a lora on my juicy dataset but this is just accessible to anyone for free with no effort. neat desu>>101436011no mate. the entire hypothetical scenario where there's no internet so you can't access bing is probably a matter of life or death>>101436017even if you had time there's a chance your local ui wouldn't work anyway if it's using any huggingface libraries because they connect every time by default, and you probably won't know how to set offline mode without the documentation that you can't access
https://civitai.com/models/583247?modelVersionId=650684
>>101435973just call it ControlnetUnionSDXL.safetensorsthen when you pick canny, depth, openpose, or sketch, you can just pick that as the model.
>>101436070I have noticed my ComfyUI needing to update to work, again..
>>101436070Dogs are let loose when war is waged.
>>101436070now bing is the pillar keeping people alive?keep going, I wanna see where this leads
I am a warrior and this is my song. I make devil himself afraid.
https://www.youtube.com/watch?v=Sam4lq2WHos
>>101436103what?>>101436109shush now darling its ok
>>101436150I know they hate me.But if you think you are good, please listen.
>>101436151You don't obviously know anything.
>>101436077weird artifacting
>>101436181yeah I guess, there are some more examples on the model page and i'll make more
Thank you! It is time to go down
I hope pw ended up doing some fun stuff today
>>101436439I think it's really interesting at how convincing it is with text, it doesn't look like inane gibberish but earnest words in a unique font or from an unknown language.
>>101436457I'm always fascinated by the AI language
spam general
>>101436525hes purging
>>101436474AI Finnish
>>101436631>>101436631>>101436631New
