<h1 id="hello-this-is-ether-blu-mix-ragtu1he4">🔹☁️ Hello, this is Ether Blu Mix ☁️🔹</h1>Ether Blu Mix is a stylized anime model specialized in creating character illustrations with vibrant colors, expressive backgrounds and a painterly feel. Based on Silicon29 and Counterfeit 3, I consider this my interpretation of a pastel-mix style of model.✨ Please Share Your Cool Creations Below! ✨<h2 id="*-version-7-update:-p4hiq6i3j">* Version 7 Update:</h2>Ether Blu 7, while being largely a weights focused change, looks to further enhance the lighting focused, slightly niji feel of EBM 6 as well as its painterly origins.Might be a little overcooked. Some prompts and generations, particularly those with heavier quality related keywords and negative prompts, are more likely to experience this. Please use the recommended negative prompts to help alleviate it.VAE is baked, same one I've always preferred.Thanks everyone, have fun. I'll be looking forward to your feedback!<h3 id="-vivk5jkf8"></h3><h3 id="*-huggingface-archive-48zjt7zvk"><a target="_blank" rel="ugc" href="https://huggingface.co/gamerdan69/EtherMix">* Huggingface Archive</a></h3><h3 id="-3aoj10e1s"></h3><h2 id="generation-recommendations-1uy2kb66b">☄️ Generation Recommendations</h2>Use whatever sampler, steps, CFG, clip-skip you prefer.Sample images were generated with<ul><li>Sampler: DPM++ 2M Karras</li><li>Steps: 20 - 28</li><li>CFG: 7 - 11</li><li>Clip Skip: 1</li><li>ENSD: 31337</li></ul>Increasing CFG can lead to more vibrant results.Negative Prompt Example:"NSFW, (worst quality, low quality:1.3)"I recommend this simple negative prompt if you wish to achieve a similar style to the preview images. You can further alter positive and negative prompts based on the type of style you'd like to achieve.Style focused negative prompt examples:* Photorealism feel:<ul><li>"sketch, drawing, anime, cartoon, illustration, NSFW, (worst quality, low quality:1.3)"</li></ul>* Flat Anime/Cartoon:<ul><li>"realistic, photorealistic, photograph, 3D, NSFW, (worst quality, low quality:1.3)"</li></ul><h3 id="recommended-vae:-885yegsew">Recommended VAE:</h3>The VAE is baked:<a target="_blank" rel="ugc" href="https://huggingface.co/hakurei/waifu-diffusion-v1-4/blob/main/vae/kl-f8-anime2.ckpt">Waifu Diffusion VAE kl-f8-anime2</a>Strongly recommended as the model was tuned with this VAE's color and saturation in mind.Sample images were further upscaled using AUTO1111 Hi-res fix. Please utilize it if you wish to achieve similar results.<ul><li>Latent (nearest-exact)</li><li>Upscale by: 2</li><li>Denoising strength: 0.55</li></ul>Are your results changing too much after Hi-res fix? Please try <a target="_blank" rel="ugc" href="https://openmodeldb.info/models/4x-Remacri">4x-foolhardy-Remarci.</a>(Also recommended for simple style illustrations, anime, cartoon, etc.)<ul><li>4x-foolhardy-Remarci</li><li>Upscale by: 2</li><li>Denoising strength: 0.45 - 0.55</li></ul><h2 id="-p6l412297"></h2><h3 id="prompting-advice-wn746qbr4">🖊️ Prompting Advice</h3>The model was mixed and tested using these prefixes and will give you the most consistent style results. Of course you're free to prompt however you'd like as booru-style models tend to be quite flexible in prompting.If you look at the prompts included with the sample images they demonstrate that even with a very loose and vague style of prompting you can achieve fun, high quality images.Experiment and give the AI some freedom to interpret your ideas!General guidelines:<ul><li>Prompt in the style of anime/booru-style models, "masterpiece, best quality, ..... "<ul><li>More anime style images</li></ul></li><li>Prompt in the style of photography/non-anime models, "professional photograph of ..... "<ul><li>More realistic images, probably higher style variety</li></ul></li></ul><h2 id="-yzz9wbwzl"></h2><h2 id="model-merging-info-v7-5r1wi9baj">📘 Model Merging Info - v7</h2>I have no training or real knowledge of machine learning. I merged the model using the information provided on <a target="_blank" rel="ugc" href="https://huggingface.co/WarriorMama777/OrangeMixs">WarriorMama777's AOM model page</a> as well as the info from <a target="_blank" rel="ugc" href="https://rentry.org/BlockMergeExplained">these</a> <a target="_blank" rel="ugc" href="https://rentry.org/Merge_Block_Weight_-china-_v1_Beta">Rentrys.</a> I very arbitrarily mixed the UBM weights of each model, modifying the values with trial and error until I was visually satisfied with it.I utilized the <a target="_blank" rel="ugc" href="https://github.com/hako-mikan/sd-webui-supermerger">SuperMerger</a> extension for AUTO1111 by hako-mikan.<h3 id="please-refer-to-the-'about-this-version'-for-more-detailed-information.-xticxffkg">Please refer to the 'About This Version' for more detailed information.</h3><h3 id="-i9ebjofa5"></h3><h2 id="credits-8kc79aht5">💙 Credits</h2>Thank you to WarriorMama777 for providing <a target="_blank" rel="ugc" href="https://civarchive.com/models/9942/abyssorangemix3-aom3">AOM</a> and other various mixes as well as detailing your workflow and inspiring me to try mixing my own models.Thank you to Luna for sharing your process on creating <a target="_blank" rel="ugc" href="https://civarchive.com/models/3850/kenshi">Kenshi</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/40199/aurora">Aurora</a>. It deepened my understanding of model making as well as inspired me to mix LORA's.Thank you to hako-mikan for creating the <a target="_blank" rel="ugc" href="https://github.com/hako-mikan/sd-webui-supermerger">SuperMerger</a> extension allowing a quicker workflow for merging.Thank you to gsdf, rqdwdw for <a target="_blank" rel="ugc" href="https://civarchive.com/models/4468/counterfeit-v25">Counterfeit.</a>Thank you to Xerxemi and Xynon for <a target="_blank" rel="ugc" href="https://huggingface.co/Xynon/SD-Silicon">SD-Silicon.</a>Thank you to YozoRaAru for <a target="_blank" rel="ugc" href="https://civarchive.com/models/12262/night-sky-yozora-style-model">Night Sky YOZORA</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/21200/color-box-model">Color Box.</a>Thank you to DynamicWang for <a target="_blank" rel="ugc" href="https://civarchive.com/models/84476/awpainting">AWPainting.</a>Thank you to Jemnite for <a target="_blank" rel="ugc" href="https://civarchive.com/models/89145/lunar-radiance-mix">Lunar Radiance.</a>Thank you to CTD for <a target="_blank" rel="ugc" href="https://huggingface.co/closertodeath/ctdlora">Dpep4, Xlimo and lightingv2.</a>Thank you to dolphinz for <a target="_blank" rel="ugc" href="https://huggingface.co/dolphinz/stlora/blob/main/loras/sttabi_v1.4-04.safetensors">Sttabi.</a>Thank you to SatyamSSJ10 for <a target="_blank" rel="ugc" href="https://huggingface.co/SatyamSSJ10/Satyam_Public_LORA/blob/main/Jordan_3.safetensors">Jordan 3.</a>Thank you to ashen-sensored for <a target="_blank" rel="ugc" href="https://huggingface.co/ashen-sensored/lora-isolation-collection">breakA lighting isolation.</a>Thank you to Terry9993 for <a target="_blank" rel="ugc" href="https://civarchive.com/models/95007/niji-default-style-niji">Niji Default style.</a>Thank you to those at the <a target="_blank" rel="ugc" href="https://huggingface.co/LMFResearchSociety">LMFResearch Society</a> for their archive of SD LORA's and checkpoints.Thank you to AUTO1111 for creating the web UI.Thank you to StabilityAI for starting everything with Stable Diffusion.Thank you to CivitAI and Huggingface for providing places where open source AI can flourish.Thank you to the entire SD community for continuing to openly share and create.

Ether Blu Mix

* Version 5 Update:Ether Blu Mix 5 looks to further enhance the vibrant, painterly look with my first attempt at LORA mixing. There's also a lot more <a target="_blank" rel="ugc" href="https://civitai.com/models/4468/counterfeit-v30">Counterfeit 3</a> allowing for more compositional variety.EBM 4 did a lot of things right; as a base it created detailed, vibrant images reminiscent of a classic anime style thanks to <a target="_blank" rel="ugc" href="https://civitai.com/models/12262/night-sky-yozora-style-model">YOZORA</a>. Though, personally, it seemed too biased at depicting characters at a somewhat close distance, typically creating waist-high portraits. Adding more weight to Counterfeit 3 should hopefully add more compositional variety, while slightly reducing YOZORA and <a target="_blank" rel="ugc" href="https://civitai.com/models/21200/color-box-model">Color Box's</a> weights to maintain EBM 4's style. <a target="_blank" rel="ugc" href="https://huggingface.co/nuigurumi/basil_mix">Basil Mix</a> was added as a supplementary realistic model in tandem with <a target="_blank" rel="ugc" href="https://civitai.com/models/4550/ayonimix">AyoniMix</a>. It's a personal favorite that plays nice with anime (booru tag) style models.EBM 5 is my first attempt at LORA mixing. I was initially hesitant at merging LORA's to checkpoints as they might introduce unwanted bias, since LORA's tend to be trained on small datasets. After plenty of testing, as well as learning from Luna's documentation of <a target="_blank" rel="ugc" href="https://civitai.com/models/3850/kenshi">Kenshi</a> and <a target="_blank" rel="ugc" href="https://civitai.com/models/40199/aurora">Aurora</a>, I applied a light mix of LORA's to help enhance the vibrant, painterly style, while also improving background expression.Your prompts might react differently as the base of the model is no longer just AOM3, so please experiment and have fun!📘 Model Merging Info - v5I have no training or real knowledge of machine learning. I merged the model using the information provided on <a target="_blank" rel="ugc" href="https://huggingface.co/WarriorMama777/OrangeMixs">WarriorMama777's AOM model page.</a> I very arbitrarily used the UBM weights given to merge and create AOM3 and its derivatives and modified the values with trial and error until I was visually happy with it.I utilized the <a target="_blank" rel="ugc" href="https://github.com/hako-mikan/sd-webui-supermerger">SuperMerger</a> extension for AUTO1111 by hako-mikan.▼ Merging<ul><li>MODEL A: AOM3.safetensors</li><li>MODEL B: Counterfeit-V3.0_fp16.safetensors</li></ul>Add Sum merge - Don't Use MBW<ul><li>Alpha: 0.5<ul><li>Model: AOM3-Counter3Sum-05</li></ul></li></ul>▼<ul><li>MODEL A: AOM3-Counter3Sum-05</li><li>MODEL B: nightSkyYOZORAStyle_yozoraV1PurnedFp16</li></ul>Add Sum merge - Don't Use MBW<ul><li>Alpha: 0.4<ul><li>Model: AOM3-Counter3Sum-05-YozoraSum-04</li></ul></li></ul>▼<ul><li>MODEL A: AOM3-Counter3Sum-05-YozoraSum-04</li><li>MODEL B: colorBoxModel_colorBOX.safetensors</li><li>MODEL C: AOM3-Counter3Sum-05</li></ul>Add Difference merge - Don't Use MBW<ul><li>Alpha: 0.2<ul><li>Model: AOM3-Counter3Sum-05-YozoraSum-04-ColorBoxDiff-02</li></ul></li></ul>▼<ul><li>MODEL A: AOM3-Counter3Sum-05-YozoraSum-04-ColorBoxDiff-02</li><li>MODEL B: pastelmix-better-vae-fp16.safetensors</li><li>MODEL C: Basil_mix_fixed.safetensors</li></ul>Weight Sum merge - Use MBW<ul><li>Alpha: 10,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0.2,0.1,0.2,0.2,0.2,0.2,0.2,0.1,0.1,0.1,0.2,0.2,0.1,0.2,0.2,0.2</li><li>Beta: 10,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0.2,0.1,0.2,0.2,0.2,0.2,0.2,0.1,0.1,0.1,0.2,0.2,0.1,0.2,0.2,0.2<ul><li>Model: AOM3-Counter3Sum-05-YozoraSum-04-ColorBoxDiff-02-PastelSumS1-02-BasilSumS1-02</li></ul></li></ul>▼<ul><li>MODEL A: AOM3-Counter3Sum-05-YozoraSum-04-ColorBoxDiff-02-PastelSumS1-02-BasilSumS1-02</li><li>MODEL B: ayonimix_V5.safetensors</li></ul>Weight Sum merge - Use MBW<ul><li>Alpha: 10,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0.2,0.1,0.2,0.2,0.2,0.2,0.2,0.1,0.1,0.1,0.2,0.2,0.1,0.2,0.2,0.2<ul><li>Final Model: EtherBluMix-5</li></ul></li></ul>▼ LORA Mix<ul><li>xlimo768: 0.25</li><li>dpep4-768-000004: 0.3</li><li>sttabi_v1.4-04: 0.05</li><li>Jordan_3: 0.2</li><li>lightingv2: 0.4</li></ul>Models Used:<pre><code>*AOM3.safetensors
D124FC18F0232D7F0A2A70358CDB1288AF9E1EE8596200F50F0936BE59514F6D

*Counterfeit-V3.0_fp16.safetensors
CBFBA64E662370F59D4AA2AA69BF16749FCE93846CCCE20506AEE5DF01169859

*pastelmix-better-vae-fp16.safetensors
D01A68AE76F97506363F387F5F28BB564AD9E20924844FD5945E600B72D39E79

*colorBoxModel_colorBOX.safetensors
93A20525F51798309AF22AC6E8285E61CF563A88F1EE00AA9F544CA939E63181

*nightSkyYOZORAStyle_yozoraV1PurnedFp16
E7BF829CFFD9EA034DA3B59FA13421F52AA4E4ED74FCBBEEFBF9CFB2F97C329E

*ayonimix_V5.safetensors
B46B04432B070FD88D69161B62C43EDCD4340DFC27FFB2E5FE7EDA64950C7E34

*Basil_mix_fixed.safetensors
0FF127093F5BE455057742C40CEF578407B6933F240EE8DC5ED0F3061196FB38
</code></pre>

Files with SHA256: cba54287a2f850eb8c55595921237a5353c8e06f98bd2a9d419c078acf67dc4b