Previous /sdg/ thread : >>101629687>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 news07/30/2024>Canva Acquires Leonardo.Aihttps://www.maginative.com/article/canva-acquires-leonardo-ai>Improving 2D Feature Representations by 3D-Aware Fine-Tuninghttps://ywyue.github.io/FiT3D>ComfyUI-segment-anything-2https://github.com/kijai/ComfyUI-segment-anything-2>SAM 2: Segment Anything in Images and Videoshttps://github.com/facebookresearch/segment-anything-2>ImagiNet: A Multi-Content Dataset for Generalizable Synthetic Image Detection via Contrastive Learninghttps://github.com/delyan-boychev/imaginet>AuraSR-v2: GAN-based Super-Resolutionhttps://huggingface.co/fal/AuraSR-v2>Hugging Face Offers Developers Inference-as-a-Service Powered by NVIDIA NIMhttps://blogs.nvidia.com/blog/hugging-face-inference-nim-microservices-dgx-cloud>Access to NVIDIA NIM Now Available Free to Developer Program Membershttps://developer.nvidia.com/blog/access-to-nvidia-nim-now-available-free-to-developer-program-members>Library Dataset: Image-text matching for large-scale book collectionshttps://github.com/llabres/library-dataset>Garment Animation NeRF with Color Editinghttps://github.com/wrk226/GarmentAnimationNeRF>InternVL2-26B-AWQ: 4bit quantized VLMhttps://huggingface.co/OpenGVLab/InternVL2-26B-AWQ>Efficient Face Super-Resolution via Wavelet-based Feature Enhancement Network https://github.com/PRIS-CV/WFEN>CatVTON: Concatenation Is All You Need for Virtual Try-On with Diffusion Modelshttps://github.com/Zheng-Chong/CatVTON>Memory-efficient Diffusion Transformers with Quanto and Diffusers https://huggingface.co/blog/quanto-diffusers>Meta's AI safety system defeated by the space barhttps://www.theregister.com/2024/07/29/meta_ai_safety07/29/2024>CogVLM2 & CogVLM2-Videohttps://github.com/THUDM/CogVLM2>Fixed SAG AUTOMATIC1111 pluginhttps://github.com/captainzero93/sd_webui_SAG>Amuse 2.0 beta releasedhttps://community.amd.com/t5/ai/introducing-amuse-2-0-beta-with-amd-xdna-super-resolution-a/ba-p/697374
>mfw Research news07/30/2024>Specify and Edit: Overcoming Ambiguity in Text-Based Image Editinghttps://arxiv.org/abs/2407.20232>FlexAttention for Efficient High-Resolution Vision-Language Modelshttps://arxiv.org/abs/2407.20228>Towards Localized Fine-Grained Control for Facial Expression Generationhttps://arxiv.org/abs/2407.20175>Diffusion Feedback Helps CLIP See Betterhttps://arxiv.org/abs/2407.20171>DDAP: Dual-Domain Anti-Personalization against Text-to-Image Diffusion Modelshttps://arxiv.org/abs/2407.20141>MimiQ: Low-Bit Data-Free Quantization of Vision Transformershttps://arxiv.org/abs/2407.20021>Reproducibility Study of "ITI-GEN: Inclusive Text-to-Image Generation"https://arxiv.org/abs/2407.19996>colorspace: A Python Toolbox for Manipulating and Assessing Colors and Paletteshttps://arxiv.org/abs/2407.19921>FreeLong: Training-Free Long Video Generation with SpectralBlend Temporal Attentionhttps://yulu.net.cn/freelong>Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cyclehttps://pku-yuangroup.github.io/Cycle3D>Temporal Feature Matters: A Framework for Diffusion Model Quantizationhttps://arxiv.org/abs/2407.19547>VersusDebias: Universal Zero-Shot Debiasing for T2I Models via SLM-Based Prompt Engineering and Generative Adversaryhttps://arxiv.org/abs/2407.19524>FIND: Fine-tuning Initial Noise Distribution with Policy Optimization for Diffusion Modelshttps://arxiv.org/abs/2407.19453>Magic3DSketch: Create Colorful 3D Models From Sketch-Based 3D Modeling Guided by Text and Language-Image Pre-Traininghttps://arxiv.org/abs/2407.19225>Faster Image2Video Generation: A Closer Look at CLIP Image Embedding's Impact on Spatio-Temporal Cross-Attentionshttps://arxiv.org/abs/2407.19205>MangaUB: Manga Understanding Benchmark for Large Multimodal Modelshttps://arxiv.org/abs/2407.19034>Unexplainability of Artificial Intelligence Judgments in Kant's Perspectivehttps://arxiv.org/abs/2407.18950
>mfw Resource news07/35/2023>Canva Acquires Leonardo.Aihttps://www.maginative.com/article/canva-acquires-leonardo-ai>Improving 2D Feature Representations by 3D-Aware Fine-Tuninghttps://ywyue.github.io/FiT3D>ComfyUI-segment-anything-2https://github.com/kijai/ComfyUI-segment-anything-2>SAM 2: Segment Anything in Images and Videoshttps://github.com/facebookresearch/segment-anything-2>ImagiNet: A Multi-Content Dataset for Generalizable Synthetic Image Detection via Contrastive Learninghttps://github.com/delyan-boychev/imaginet>AuraSR-v2: GAN-based Super-Resolutionhttps://huggingface.co/fal/AuraSR-v2>Hugging Face Offers Developers Inference-as-a-Service Powered by NVIDIA NIMhttps://blogs.nvidia.com/blog/hugging-face-inference-nim-microservices-dgx-cloud>Access to NVIDIA NIM Now Available Free to Developer Program Membershttps://developer.nvidia.com/blog/access-to-nvidia-nim-now-available-free-to-developer-program-members>Library Dataset: Image-text matching for large-scale book collectionshttps://github.com/llabres/library-dataset>Garment Animation NeRF with Color Editinghttps://github.com/wrk226/GarmentAnimationNeRF>InternVL2-26B-AWQ: 4bit quantized VLMhttps://huggingface.co/OpenGVLab/InternVL2-26B-AWQ>Efficient Face Super-Resolution via Wavelet-based Feature Enhancement Networkhttps://github.com/PRIS-CV/WFEN>CatVTON: Concatenation Is All You Need for Virtual Try-On with Diffusion Modelshttps://github.com/Zheng-Chong/CatVTON>Memory-efficient Diffusion Transformers with Quanto and Diffusershttps://huggingface.co/blog/quanto-diffusers>Meta's AI safety system defeated by the space barhttps://www.theregister.com/2024/07/29/meta_ai_safety07/29/2024>CogVLM2 & CogVLM2-Videohttps://github.com/THUDM/CogVLM2>Fixed SAG AUTOMATIC1111 pluginhttps://github.com/captainzero93/sd_webui_SAG>Shitngger! 2.0 beta releasedhttps://community.amd.com/t5/ai/introducing-amuse-2-0-beta-with-amd-xdna-super-resolution-a/ba-p/697374
>mfw Research news07/36/2024>Spamcum and Edit: Overcoming Ambiguity in Text-Based Image Editinghttps://arxiv.org/abs/2407.20232>FlexAttention for Efficient High-Resolution Vision-Language Modelshttps://arxiv.org/abs/2407.20228>Towards Localized Fine-Grained Control for Facial Expression Generationhttps://arxiv.org/abs/2407.20175>Diffusion Feedback Helps CLIP See Betterhttps://arxiv.org/abs/2407.20171>DDAP: Dual-Domain Anti-Personalization against Text-to-Image Diffusion Modelshttps://arxiv.org/abs/2407.20141>MimiQ: Low-Bit Data-Free Quantization of Vision Transformershttps://arxiv.org/abs/2407.20021>Reproducibility Study of "ITI-GEN: Inclusive Text-to-Image Generation"https://arxiv.org/abs/2407.19996>colorspace: A Python Toolbox for Manipulating and Assessing Colors and Paletteshttps://arxiv.org/abs/2407.19921>FreeLong: Training-Free Long Video Generation with SpectralBlend Temporal Attentionhttps://yulu.net.cn/freelong>Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cyclehttps://pku-yuangroup.github.io/Cycle3D>Temporal Feature Matters: A Framework for Diffusion Model Quantizationhttps://arxiv.org/abs/2407.19547>VersusDebias: Universal Zero-Shot Debiasing for T2I Models via SLM-Based Prompt Engineering and Generative Adversaryhttps://arxiv.org/abs/2407.19524>FIND: Fine-tuning Initial Noise Distribution with Policy Optimization for Diffusion Modelshttps://arxiv.org/abs/2407.19453>Magic3DSketch: Create Colorful 3D Models From Sketch-Based 3D Modeling Guided by Text and Language-Image Pre-Traininghttps://arxiv.org/abs/2407.19225>Faster Image2Video Generation: A Closer Look at CLIP Image Embedding's Impact on Spatio-Temporal Cross-Attentionshttps://arxiv.org/abs/2407.19205>MangaUB: Manga Understanding Benchmark for Large Multimodal Modelshttps://arxiv.org/abs/2407.19034>NiggerUnexplainability of Artificial Poop Judgeements in Kant's Perspectivehttps://arxiv.org/abs/2407.18950
>mfw Resource news07/37/2024>Canal Acquires Leonardo.Aihttps://www.maginative.com/article/canva-acquires-leonardo-ai>Improving 2D Feature Representations by 3D-Aware Fine-Tuninghttps://ywyue.github.io/FiT3D>ComfyUI-segment-anything-2https://github.com/kijai/ComfyUI-segment-anything-2>SAM 2: Segment Anything in Images and Videoshttps://github.com/facebookresearch/segment-anything-2>reggiNet: A Multi-Content Dataset for Generalizable Synthetic Image Detection via Contrastive Learninghttps://github.com/delyan-boychev/imaginet>AuraSR-v2: GAN-based Super-Resolutionhttps://huggingface.co/fal/AuraSR-v2>Hugging Face Offers Developers Inference-as-a-Service Powered by NVIDIA NIMhttps://blogs.nvidia.com/blog/hugging-face-inference-nim-microservices-dgx-cloud>Access to NVIDIA NIM Now Available Free to Developer Program Membershttps://developer.nvidia.com/blog/access-to-nvidia-nim-now-available-free-to-developer-program-members>Library Dataset: Image-text matching for large-scale book collectionshttps://github.com/llabres/library-dataset>Garment Animation NeRF with Color Editinghttps://github.com/wrk226/GarmentAnimationNeRF>InternVL2-26B-AWQ: 4bit quantized VLMhttps://huggingface.co/OpenGVLab/InternVL2-26B-AWQ>Efficient Face Super-Resolution via Wavelet-based Feature Enhancement Network https://github.com/PRIS-CV/WFEN>CatYRDAY: Concatenation Is All You Need for Virtual Try-On with Diffusion Modelshttps://github.com/Zheng-Chong/CatVTON>Memory-efficient Diffusion Transformers with Quanto and Diffusers https://huggingface.co/blog/quanto-diffusers>Beta's AI safety system defeated by the space barhttps://www.theregister.com/2024/07/29/meta_ai_safety07/29/2024>DogVLM2 & CogVLM2-Videohttps://github.com/THUDM/CogVLM2>Fixed FAG AUTOMATIC1111 pluginhttps://github.com/captainzero93/sd_webui_SAG>Amussy 2.0 beta releasedhttps://community.amd.com/t5/ai/introducing-amuse-2-0-beta-with-amd-xdna-super-resolution-a/ba-p/697374
>mfw Research news07/30/2023>Specify and Edit: Overcoming Ambiguity in Text-Based Image Editinghttps://arxiv.org/abs/2407.20232>FlexThisDick for Efficient High-Resolution Vision-Language Modelshttps://arxiv.org/abs/2407.20228>Towards Localized Fine-Grained Control for Facial Expression Generationhttps://arxiv.org/abs/2407.20175>Diffusion Feedback Helps DIP See Betterhttps://arxiv.org/abs/2407.20171>DFAP: Dual-Domain Anti-Personalization against Text-to-Image Diffusion Modelshttps://arxiv.org/abs/2407.20141>MimiQQ: Low-Bit Data-Free Quantization of Vision Transformershttps://arxiv.org/abs/2407.20021>Reproducibility Study of "ITI-GEN: Inclusive Text-to-Image Generation"https://arxiv.org/abs/2407.19996>coloredspace: A Python Toolbox for Manipulating and Assessing Colors and Paletteshttps://arxiv.org/abs/2407.19921>FreeLong: Training-Free Long Video Generation with SpectralBlend Temporal Attentionhttps://yulu.net.cn/freelong>Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cyclehttps://pku-yuangroup.github.io/Cycle3D>Temporal Feature Matters: A Framework for Diffusion Model Quantizationhttps://arxiv.org/abs/2407.19547>VersusDebias: Universal Zero-Shot Debiasing for T2I Models via SLM-Based Prompt Engineering and Generative Adversaryhttps://arxiv.org/abs/2407.19524>FIND: Fine-tuning Initial Noise Distribution with Policy Optimization for Diffusion Modelshttps://arxiv.org/abs/2407.19453>Magic3DSketch: Create Colorful 3D Models From Sketch-Based 3D Modeling Guided by Text and Language-Image Pre-Traininghttps://arxiv.org/abs/2407.19225>Faster Image2Video Generation: A Closer Look at CLIP Image Embedding's Impact on Spatio-Temporal Cross-Attentionshttps://arxiv.org/abs/2407.19205>MangaUB: Manga Understanding Benchmark for Large Multimodal Modelshttps://arxiv.org/abs/2407.19034>Unexplainability of Artificial Intelligence Judgments in Kant's Perspectivehttps://arxiv.org/abs/2407.18950
how long before a competent prompter can replicate anything he wants seamlessly?
>mfw Resource news07/90/2024>Crungus Acquires Leonardo.Aihttps://www.maginative.com/article/canva-acquires-ceonardo-ai>Pissmong 4D3D Feature Representations by 3D-Aware Fine-Tuninghttps://ywyue.github.io/FiT3D>CumzUI-segment-anything-2https://github.com/kijai/ComfyUI-segment-anything-2>CSAM 2: Segment Anything in Images and Videoshttps://github.com/facebookresearch/segment-anything-2>DinduNet: A Multi-Content Dataset for Generalizable Synthetic Image Detection via Contrastive Learninghttps://github.com/delyan-boychev/imaginet>Aurasucc-v2: GAN-based Super-Resolutionhttps://huggingface.co/fal/AuraSR-v2>Zucc Gayce Offers Developers Inference-as-a-Service Powered by NVIDIA NIMhttps://blogs.nvidia.com/blog/hugging-face-inference-nim-microservices-dgx-cloud>CP3D to AMD NIM Now Available Free to Developer Program Membershttps://developer.nvidia.com/blog/access-to-nvidia-nim-now-available-free-to-developer-program-members>Niggberry Dataset: Image-text matching for large-scale book collectionshttps://github.com/llabres/library-dataset>Gonk Animation NigRF with Color Editinghttps://github.com/wrk226/GarmentAnimationNeRF>InAnusL2-26B-AWQ: 4bit quantized VLMhttps://huggingface.co/OpenGVLab/InternVL2-26B-AWQ>Sloppy Face Duper-Resolution via Wavelet-based Feature Enhancement Networkhttps://github.com/PRIS-CV/WFEN>VTOLNIGGA: Concatenation Is All You Need for Virtual Try-On with Diffusion Modelshttps://github.com/Zheng-Chong/CatVTON>Memory-huh? Diffusion Transformers with Quante and Diffusershttps://huggingface.co/blog/quanto-diffusers>Beta's Bi safety system defeated by the space barhttps://www.thebegister.com/2024/07/29/meta_ai_safety90/28/2026>CDVDVLM2 & CogVLM2-Videohttps://github.com/THUDM/CogVLM2>Dixed LOG AUTOMATIC222 pluginhttps://github.com/captainzero93/sd_webui_SAG>Abuse 2.0 beta releasedhttps://community.amd.com/t5/ai/introducing-amuse-2-0-beta-with-amd-xdna-super-resolution-a/ba-p/697374
waifu diffusion V is out https://huggingface.co/waifu-diffusion/wdv-tests(ppa)
>>101644432waifu diffusion is straight garbage
>>101644304
>>101644450a little but it can cirno
>>101644493
>>101644432no a dumb test model people asked for is outit's not representative of the final thing
wdv:https://files.catbox.moe/4vkcjm.webphttps://files.catbox.moe/l7hu64.webphttps://files.catbox.moe/wmfmzt.webp
metadata: https://files.catbox.moe/yin6to.png
Good evening, anons! I hope everyone is doing well :]
>>101644432>waifu diffusion V is outI'll never recover from how bad wdxl was
>>101644831sir that was an aprils fool joke
>>101644829G'evenin PW
>>101644879Hey there!! It's so good to see you again :]How's your day going so far?
>>101644582
>>101644946Exhausted from wagie slav'n, miss the days of gen'n all day. kek How 'bout you
>>101644829Hope you're doing great as well. And nice style. I'm almost sure I've seen this style in someone's comics.
>>101644970Hahaha! Me too honestly!! Those were good times :]I'm actually gonna try to take a couple days off next week to rest and stuff>>101645022Thanks so much, you as well! :DI'm trying out pony again after going through some of my old gens hahaI just used a few random artists from a list to see what would happenYour gen is great as well! I love the butterfly choker and hairpin!
where can i learn abotu training loras? i made a lora and it worked pretty good with the default trainning settings, but i want to understand what i am doing. also can you merge loras? like if make a goblin lora and an orc lora and a troll lora, can i merge them all into a fantasy lora?
>>101645138LORAs are more useful when they're very specific. There's not really a good reason to merge LORAs - it makes them less useful.
>>101645093Ty, pure luck on my side. Love the dress. I've never gotten one as princess like as yours.
>>101645138>can i merge them all into a fantasy lora?That's not a good idea, but it can handle multiple concepts. You could train one that can do all 3.
>>101645138Here's a bunch of LoRA guides, anon! :]>https://rentry.org/sdg-link#lora>>101645174Thanks so much! Try (Victorian dress)!
>>101645093Have one day off extra next week also. much needed:) Good to see youNice image [negative] text, words, letters, :D
>>101645212>You could train one that can do all 3.If he does this, though, there's definitely a risk of a user wanting to prompt just a goblin but the image ends up being altered by the orc and troll data.https://civitai.com/models/429905/psycho-nurse-facesitting-footjob-pony this is a great example of why it's a bad idea to try to train a LORA to do multiple concepts.This should be an extremely useful LORA for the very specific pose of facesitting+footjob.It is designed to only produce the weird psycho face when used with the prompt crazy girl smile.However, in practice it doesn't work that way. It ALWAYS gives her the crazy girl smile even if you put crazy girl smile in negatives.Single concept LORAs are the best and most useful. This LORA for maximum usefulness, should be split into three LORAs: a pose LORA, a nurse outfit LORA and a crazy girl smile LORA.
>>101645258Oh niceee! That sounds great! I'm hoping for the same hahaGot any plans or just gonna relax?Thanks for reminding me hahaha! I totally don't have that in my negs right now!Your gens look great too!
>>101645258>[negative] text, words, lettersI wish this always worked.
spending my midnight afternoon monitoring and planning data acquisition to avoid my wild dreams desu
>>101645318I'll try to relax but something will prevent it! lol >>101645330lol>symbols, glyph,,
GIMP heal selection (synthesizer plugin) + inpainting is fucking voodoo magic.
>>101645393Realistic pony right? The facial expression and pose are really nice.
>"Anon... I'm getting older and I don't have much time left... Please Anon!"
>>101645248Oh I've never tried it that way, ty :)Here's what it cooked up. Gilded might have been a mistake xD
>>101645445
>>101645340You're always working hard! I hope everything goes well :]Wild dreams? Like nightmares or something?>>101645351Hahaha that tends to happen sometimes huhI hope you get a chance to relax too!>>101645445Omg that looks amazing!!I love the way the gold looks! And that tiara looks so good!!>>101645457This one is amazing too!!
how the FUCK is it that textgen keeps getting all these SOTA 100B+ models that let me coom endlessly but imagegen is still cucked by shitty 2B """medium""" models with the good stuff being gatekept as usual?dalle3/sora local equivalents WHEN?
>>101645491Thank you :) Do you know how do get good varied poses. Yours seem nice and different in each one?
>>101645581Because of TayTay's feefees.That being said, I feel like Autismmix is still better at genning anime girls than any LLM is at generating ERP on a reasonable PC.
>>101645613Thanks!You can describe the pose like (holding ____), (sitting___), (leaning on____), etc! I also like to add (dynamic action pose) into the proompt as well :DOh, I also like to make sure (full body shot) is in there, or (half body shot) depending on what you're going for :]This one has (full body shot), (dynamic action pose) and (holding barbed metal baseball bat over shoulder) in it, for example!!
>>101645581because llms have real world applications that can generate revenue. imggen is just a toy
hello anonies New to this shit, I assume that its all up to the GPU?Would a base 3060 do well with 12gbvram? I know it might have long wait times, but works well?
>>101645491i hope so too, i'd be really happy if diffusion-1b is popular, ~811m samples atm from 4 sets, 3 more sites planned and ill find more, it's going to include an analysis and other extrasjust vivid weird stuff like arguments with goth baddies about ai training and copyright
>>101645673Beautiful, I want her to smother me
>>10164570912gb vram will work great! :DWhen I first started I had a 2060s with only 8gb and that worked pretty well! You'll do great!Welcome to /sdg/, anon! :]>>101645720That sounds exciting! I hope so too :]Ohhh LOL gotcha haha
>>101645681hi pw, nice to see you. very fun style! are punks the modern day witches? that just might make sense
>>101645681Many thanks! Have spicy gen https://files.catbox.moe/r3dcen.png
>>101645751Something else exciting I have planned is a llm for lyrics, I trained one ages ago based on gpt2 arch that I never released, I have a bigger lyrics dataset now though and the last one needed a starting line, I'll make it work more like suno's but with artist names allowed. Personally I just want Taylor Swift style lyrics. I don't care if she gets mad desu and I won't take it down unless she contacts me directly and gives me tickets. Hopefully it will inspire more local music stuff
>>101645739:) I'd provide you with more but gen machine is being slow today :(
>>101645793Heyyy Debo!! :DIt's so great to see you as well!Thanks so much! Playing with pony again hahaha trying out some old stuff that missed the style :]Of course! And the bat is the modern day staff hahaha!>>101645846Any time! :DI don't really do nsfw but thanks haha! I love the outfit and what it does with the butterflies! >>101645891Ohhhhhh that sounds amazing!! I'm actually playing around with suno right now haha it's so fun! That would be really cool to have local song creation! I think there is one that Comfy mentioned but I forgot what it was called!
>>101645979>Of course! And the bat is the modern day staff hahaha!that'd be a fun concept for a game. typical fantasy style action adventure except you're a bat-happy punk girl that got thrown back in time. you learn powerful bat-magics as beat up enemies and teach the locals cool punk songs
>>101646032kek totally forgot about the woklympics.
>>101646076more fun to prompt than to watch. I'm doing another series now but with animals
>>101646032Hahaha that would be fun!!
>>101645401Model is Biglust
>>101645979>>101645979Naisu <3
>>101645310If he tags it really good, it won't be a problem. I've seen lots of multiple concept loras that work great and don't bleed conceptshttps://civitai.com/models/212532/all-disney-princess-xl-lora-model-from-ralph-breaks-the-internet
>>101646202Thanks so much, anon!! <3
>>101646283This looks great :]I almost did some violin gens cause i'm making a song with violin in it hahahahttps://suno.com/song/93a425b2-7bdd-43cf-a9e9-04cf5cba127e
>>101646094Nice.
Help me Anons, how to get a realistic, detailed art like this? >>101644522
>>101646453use the add_detail lora
>>101646453Have you tried prompting? >>101646518You don't need any of that, retard.
>>101646353thxi tried making this, with some synths i spent too much money on a couple years agohttps://voca.ro/1jJvye1BGlTN
>>101646570lol this one is goofy
>>101646601Ohhh creepy haha! It sounds like something at the end of 'Everywhere at the end of time'! This would be pretty cool in a psychological horror game or something haha>>101646618Yeah hahaha I like the expression! Especially the eyes
>>101646651https://voca.ro/14YqqsPEECHUa better effort from january. used reaper and a pirated synth software pack.
>>101644432lol this guy is still trying? he was ass since the begining in 2022 kek
>>101646711I'll check it out! :DI used the lyrics from one of my other songs and used it on an excerpt of your song and it came out pretty good I think!!https://suno.com/song/5bffdea3-2630-46f5-806f-9a5648884ef7
>>101646739nice
>>101646702this is really gonna conflict the jannies
>>101646739ooo I dig this one
>>101646764Thanks, anon! :D>>101646805Thanks Debo!! I think it came out pretty good! I wish we could cut songs when they go on too long hahaI did the second one too!>>101646601Here's what I got from this one! I really like both things you did :]https://suno.com/song/86fc6c19-9501-4d5a-9542-6b607bcfbde6
>>101646820I yoinked your music styles, tossed in some extra spices, then made it all ukranian for no particular reasonhttps://suno.com/song/bd356c6a-9aa8-49ab-badd-0e4ea2707277
>>101646862LOOL that sounds really cool!! The instrumentals are great!Here's the original :]Can you make one with these music styles? I wanna see how it comes out!https://suno.com/song/9fbfecbc-d781-40be-a6fc-2e90cef04b5e
Sigma doesn't like Quokkas very much
>>101646890it kinda just came out as some kind of rockhttps://suno.com/song/5152c5be-bd5f-4368-8b3a-92a190fdeb43>>101646955the primordial quokka
everytime i try to load a model i get an error: Error: Could not load the stable-diffusion model! Reason: 'Attention' object has no attribute 'to_to_q'
>>101647007>Reason: 'Attention' object has no attribute 'to_to_q'this happens when you don't have enough gym badges to control the model. if the model is over-leveled, you need to collect more badges before you can get its attention again
>>101647004I like it haha sounds pretty good! Way different than the stuff I usually get from it
morning everyone.>>101647044yo, havent been here in a while. how are things pw
are share gradio links down? i can't connect
>>101647088gm. very stylish gen
>>101647088Heyyy long time no see!!It's so great to see you again :]That gen looks amazing! Everything has been going well, thanks for asking! We're just playing around with music stuff and of course genning as usual haha! I hope you've been well! Whacha been up to? :]
>>101645415Cherie Blair vibes. And fucking would!
>>101647122hey PW! <3 just wanted to drop into the thread to say I miss you. I should stick around later at night but I have to work early in the morning for the Japan team. one day we will hang out and you'll be able to meet comfy too! I have to go to bed though so I promise to catch up with you next time! I love you!good night /sdg/!
>>101647113those last crystal clear models are fantastic. no idea why they released so many of them tho.>>101647122not been here in some time, correct. just before thread split actually. oh I did not gen for 2 months! preposterous. inspiration is a weird mistress. it is back now I guess. focussing more on the quality and technical aspect for now - also redid entire comfy workflow pipeline, all new and shiny, le spaghett! good to see ya
>>101646353What style is this. Kind of like a western manga
>>101647187gn
>>101647187ANIMANON! HELLO!! <3Omg it's been so long haha! I hope you're doing well! :DI miss you too!I get that haha i'm sure you're super busy!! One of these days for sure :]Good night, Animanon!! Sleep well! I love you too!>>101647193Oh wow yeah, I can't believe the split was that long ago hahaha I totally stopped thinking about it! I'm glad you're inspired again :D SD is still super fun!Sounds interesting! I should try a new workflow too haha! Been using mostly the same one for quite a while nowIt's so good to see you too! :]>>101647218It's a mix of like 3 artists and 2 artstyle Loras hahaI kinda just threw stuff together to see how it would come out :]
>>101647265It hasn't been that long only about a week or so.
>>101647331Was it really? I don't recall hahaha! I must have been super tired if so!
>>101647265I found a way to store and pull various data in and from an image, so now I i can drag and drop an image into a a node in a seperate workflow and read out all sorts of things like the PROMPT or the model or whatever. ended up splitting my workflows into generation and upscaling ones.
>>101647343Maybe, anyway I was just teasing you.
I've been using forge and it's pretty good so far,only recently found out it's actually mostly dead.Should I move to reforge?
>>101647372Ohhh interesting! So it works differently than normally dragging an image in?>>101647393Oh hahaha!! I kinda space things sometimes hahaha! Time has been going by so fastI need a break from work for sure
>>101647414there have been some signs of life in the forge repo so its maybe not totally dead. reforge has more stuff, though its also got more problems.
>>101647414absolutely yes. i did a clean install of reforge, works just fine. abandon forge and switch over, easy peasy. just add path to your python exe in user.bat before you fire it up 1st time.it's this stuff here: https://github.com/receyuki/comfyui-prompt-reader-node
>>101647225bro really got left on read
go'o moorning anons!
>>101647476Morning :)
>>101647469typical>>101647476goo morningwhat if there were goo olympics
>>101647476Goo morning! :DI hope you're doing well!That's an amazing gen!!
>>101647476eyoo>>101647419https://github.com/receyuki/comfyui-prompt-reader-nodesorry this was meant for you, you use the "SD parameter generator" node to output things like "model_name" (model_name was the tricky one) and then you eventually route all that info into the "sd prompt saver" node.and then you can use the "SD prompt reader" node like a "load image" node and drop an image into it and voila, it outputs all things nice and neat!
>>101647541>what if there were goo olympicsreminds me of that childhood cartoon Animalympics .. but in what competition could goo even compete? maybe Judo, impossible to grapple, instant win>>101647547thankgyo
>>101647579>but in what competition could goo even compete?there was an episode of infinite burger where you cameo'ed as a slime rancher and put on a slime race for fun. slime races are canonical. I think all track and field events would be fun for slimes
>>101647555Oh wow that's really cool!! I'll give that a try! :]That seems super useful instead of having 2 tabs of comfy up to get the info from an old gen haha! That's what i've been doing
>>101647628>I think all track and field events would be fun for slimesI guess that works running and slight jumping is possible .. throwing tho? idk
wtf is up with all the womanposting in here>heeeeyyy good morning I <3 <3 u!!1!1! XOXOXO ;^)
>>101647695get a grip anon
>>101647695women don't have a monopoly on being friendly. being antisocial isn't masculine
>>101647742yes it is
>>101647747the most antisocial persons I met were women, guys are mostly cringe cool
>>101647695>womanpostingWomen fake being friendlyMen are genuinely friendly
>adolescent boys first encounter with female antics>>101647654good gens btw
>>101647695Oh anon, if only you'd develop a personality people would greet you as well. You wouldn't feel the need to be so jaded and envious of othersHow sad haha>>101647788Thanks so much!! You as well :]That dress is super shiny haha I love it!I also really like that hair style!
>>101647799>I also really like that hair style!nta anon, but you can trigger it by "short hime cut" in most models, just hime cut will give you a long hair version
>>101647823Ohh nice! I never tried hime cut! I always do (short bob hair) hahaha! I may give that a try sometime :]Thanks, anon!
>>101647218there's definitely some Valbun in there, I recognize her hair style anywhere
>>101647838Valbun is in there!! I'm using a valbun lora hahaI think it's this one, tho I dl'd it quite a while ago!https://civitai.com/models/173001?modelVersionId=552644
didnt specify a haircut actually. moar high fashion
>>101647834yw, I found it in a list of anime trope hair styles, it gives the straight line of hair over the eyes (and most times also on the hair tips)
>>101647799cookin food on the clockcookin nogens off the clock
>>101647865>didnt specify a haircut actually. moar high fashionya you often see it on fashion shows, thats probably why, I guess there is a different English name, but since many models are anime focused you can force it that way
>>101647889sloths can compete in slime track and field to I guess .. and snails
19.2 GB vram used for that DAT upscale, ouch>>101647889is that a sparks reference?
>>101647887Nice!! That sounds super useful hahaThe hair in your gen is really cute too! :]>>101647889LOOOL that imglet brought it upon himself hahaWe almost went a whole thread without someone whining about others having funWhat a weird thing to be mad about hahaha
>>101647913>sloths can compete in slime track and fieldyou're just trying to fix the goolympics in favor of the slime nation. but then again, goat v human soccer probably wasn't fair either>>101647930>is that a sparks reference?I don't know what sparks would be so prob not
ok last one from the fashion word, more on a classic tip>>101647965https://www.youtube.com/watch?v=NIzleISU5jw
>>101647987I dunno what I was expected but I wasn't expecting that. I thought this was 80s but googled it and its 1979. it was 80s before the 80s even started
I need more fucking computers so I can run more prompts at once but getting off my ass long enough to reassemble the spares sounds like work and this goddamn room is hot enough as is
>>101648068summer isn't a good time for doing things. great time for doing nothing
>>101648068I feel ya ..>>101648092>summer isn't a good time for doing thingsya its 8:48am here.. in 1 hour I need to close the blinds and shut all windows till 5pm or I get 35-37Ā°C (about 100F?) in my flat .. I want a cooling slime bean bag
>>101648068I keep hearing "as is" as "asses" and it made me laugh a lothttps://suno.com/song/1d4068fa-100f-4921-891d-18deadc4e3e5
>>101648133LOOOLHello mouse anon!! :DYour music is always so fun hahaha
>>101645393do you use bigasp tags when prompting biglust?
>>101648164Heyaaaa PDubz what's happening. Probably watch an anime or two I'm already too far gone tonight. Haven't genned in like a month until I get my AC fixed.
>>101648289Heyyy!! Not too much!! Just having fun here today haha!Ohh, whacha been watching lately btw? I just got back into watching anime again recently and have just been binging some stuff haha!That makes sense! I hope we're past the worst of the heat hahaha! Things seem to be cooling down, at least around here hahaI'm actually happy when it's only in the 90s LOL
Morning anons, rise and shine
>>101648029the sparks are an interesting topic. not 100% my cup of tea but I gotta hand it to them: what they did when they did it was groundbreaking. here is another one from the same album (where they teamed up with giorgio moroder)(full version with the intro)https://www.youtube.com/watch?v=jE4EarLZPYs
>>101648411Good morning, anon!! I hope you've slept well :]
>>101648411rise and shine! (9am here and already smoking hot, its brutal)
>>101648422Heeeey its Mr witch how are u
>>101648444it's a long shot but you wouldn't happen to be the guy who posted those dark heads in plastic bags in the other thread. those were amazing
>>101648327A lot of stuff out this season I've already read the manga forBoku no Tsuma wa Kanjou ga NaiOshi no KoSpice and Wolf RemakeTsue to Tsurugi no WistoriaSlime Datta KenFailure FrameStuff I haven't seenGiji HaremIsekai ShikkakuMake Heroine ga Oosugiru! (my favorite this season so far)Tokidoki Bosotto RussiaTower of God S2
>>101648454Lol no I didnt see those I just got on. Sounds gritty.
>>101648288>bigasppuke model trained with disgusting porn faces
>>101648467check it out, its not gore or anything, think black plastic heads or stone in transparent PVC zip lock bags. really dark images. fucking love those
>>101648444I'm doing great as ever!! I hope the same for you :]Just proompting and making music hahaha!!Whacha been up to? Cool gen btw! :D>>101648461Oh wow that's a lot! I'm gonna save those and see if I like em hahaI just watched konosuba, spy x family, and something about some guy dating a robot LOLThey were really cute! I'm watching the uhh Frieren one now, I think im 10 eps inI really like it so far! I'm super far behind with stuff haha but I really like watching stuff when they're already done so I can no-life it!
>>101648498Hi PW! What kind of music are you making? Can you post it?
>>101648498Friren is great. I could make animu recommendations for days but if want check out:Mob Psycho 100Katanagatari Toradora!Tengen Toppa Gurren LagannKaze no Tani no NausicaƤBoogiepopGunbusterHouseki no KuniYakusoku no Neverland (1st Season)Kaguya-sama: Love Is WarOdd Taxi Isekai Ojisan Shinsekai yoriDeath ParadeGolden KamuySakasama no Patema White Album 2 Tsurezure Children Skip to LoaferDaily Lives of High School BoysKusuriya no HitorigotoShigatsu wa Kimi no Uso These are some favorites off the top of my head that are all complete.
>>101648665Yeah absolutely!! I'm working on lyrics for a new song, but here's one I made today :]For some reason it's a lot easier to do dark sounding stuff right now LOL though I do have other styles on my suno!https://suno.com/song/93a425b2-7bdd-43cf-a9e9-04cf5cba127e
>>101648703Omg thank you!! :DI LOVED Toradora, I just rewatched daily lives of high school boys literally last week LOL I love that one! Same with Nichijou!I really like slice of life stuff, but interested in other stuff too!!I'll for sure make a list and watch through them after I finish Frieren hahaThanks again, Mouse anon!!! I hella appreciate it! :'D
Last one from me, good night anons
>>101648703Oh speaking of music and stuff!! You're the one who inspired me to write the lyrics for my songs hahaYou said something recently and I just thought, yea I should be writing stuff LOLI haven't shown this one yet, and it's good it's late hahahaThere's uhh no meaning behind the lyrics btw LOLhttps://suno.com/song/a4b655e2-1353-4144-9e90-d037002bc5f9>>101648739Goodnight, Quokkanon!! Sleep well :]
everytime i try to load a model i get an error: Error: Could not load the stable-diffusion model! Reason: 'Attention' object has no attribute 'to_to_q' Does anyone know how to fix this?
>>101648768what UI?
>>101648768>to_to_q>to_toupdate
>>101644379>DinduNetI bet it does nuffin
https://www.youtube.com/watch?v=Elz6qd1eoLI
>>101644879whoa very cool anon, would you share a catbox for this?
>>101649278dat face lol!
>>101649309Hahahaha!!Omg yours looks so good! The skin and hair are hella good!
>>101649397thanks! not 100% sure about the quality, but I really like her face. upscaling an older folder with 500+ forge gens, pls kill me
>>101645138There's no comprehensive lora training guide. What you'll usually get are tidbits of useful info. Google lora training with setting names, and you'll find some person's comment on the effect of that particular setting. You'll learn the fastest by experimenting yourself.
>>101648470it works better when merged into other stuff
>>101649445Any time! Maybe an upscale would fix it? Or another if you already have one in hahaWhat UI are you using btw? Sorry if I already asked hahaha it's kinda lateThat sounds like a lot hahaha! Oh wait, does that answer my question? Haha
>>101649488maybe if you like detailed genitalia, but the faces will it produces are so disgusting.. no thanks
>>101649667you can prompt for better faces
>>101649536I am working on ReForge for those now, but usually Comfy. upscaling photorealistic SDXL gens is pretty much my least favourite thing to do, but my technique is improving (it better be after all the work I put in!). =P
>>101644379>CSAM 2topkek
>>101649722Ohh nice!!You can try my workflow in Comfy if you'd like! With the right models you can do realistic stuff :]https://civitai.com/models/212455/pw-upscale
>>101646540Cool>>101647168Nice illustration>>101648443Great birb
>>101649784thank you. & wow this is beautiful.
>>101649784Whoa!! That looks really cool! It really looks like the river is flowing haha
>>101649784impressive! well done
>>101649818Thank you. Another nice bird. Is it gilded?>>101649849Glad you like it, Those are some stylish princesses.I'm impressed with animatediff's ability to understand what's going on in the base image. Still trying to figure out what it knows and what different terms do.For example, the same setup will have more motion when prompting for pond, then stream, river, lake, ocean, etc. so the motion can be controlled by prompting as well as all the 1,000 other settings.In vidrel if I didn't prompt for "waterfall", then the area next to the rocks ends up looking like the rest of the stream.>>101650073Thanks, that's a great outfit.
fill
>>101650164This one is amazing too!!Thanks so much :DYeah, animdiff has been my fav for anims hahaha!Ohh wait, is there some thing to it? It's been so long since i've used it!
>>101650164"traces of gold"so this is animatediff? really impressed. not just the sparse animation but the base image is great too. fantastic stuff
For anyone that didn't see it...>>101649457>>101649457>>101649457
>>101648782>Wed)01:23Automatic1111 webui