<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/c0cde621-bd91-426d-b233-e372060b8027/width=525/c0cde621-bd91-426d-b233-e372060b8027.jpeg" /><a target="_blank" rel="ugc" href="https://civarchive.com/user/AshtakaOOf">Aha</a> said I lobotomized this model 💀<h2 id="description-565ea39lb">Description</h2>Lobotomized Mix is a highly capable base model built on <a target="_blank" rel="ugc" href="https://civarchive.com/models/833294/noobai-xl-nai-xl">NoobAI-vpred</a> with extremely high character fidelity and prompt responsiveness. It has a strong predilection for complex lighting effects and functions best with clear well defined prompt tags. Character lora responsiveness is good. Style lora responsiveness is varied, ranging from "well you can clearly see some sort of effect" and "this mimicks the style perfectly". It is generally capable prompting up to three characters interacting with any issue or need for regional conditioning. Prompting four characters is possible but attribute-blend becomes increasingly likely the more characters are prompted. Prompting more than four distinct characters is probably impossible on the SDXL architecture without aid from extensions (though this does not count for clones or twins or otherwise visually distinct characters).<h2 id="usage-requirements-1xejr7cc1">Usage Requirements</h2>Standard NoobAI quality tags apply. Check NoobAI documentation for details. In addition, the <code>very aesthetic</code> positive and <code>displeasing</code> negative tags have some minor effect. The recommended structure according to LAXHAR Labs for ordering prompts is: <code>artist tags, [your prompt here], quality tags</code>This is a V-pred ZSNR model and will not work with Automatic1111-webui. In order to get images out of vpred models you need to either switch the <a target="_blank" rel="ugc" href="https://github.com/AUTOMATIC1111/stable-diffusion-webui/tree/dev">dev branch</a> on Automatic1111 webui, <a target="_blank" rel="ugc" href="https://github.com/comfyanonymous/ComfyUI">ComfyUI </a>or <a target="_blank" rel="ugc" href="https://github.com/Panchovix/stable-diffusion-webui-reForge">reForge</a>. My personal recommendation is to switch to reForge.In order to properly replicate the example images, you will need to enable the following extensions on reForge, listed in order of importance."APG's now your CFG" and Guidance Limiter for reForge<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8afbddd3-dc47-4ba3-b270-c6095bc3b096/width=525/8afbddd3-dc47-4ba3-b270-c6095bc3b096.jpeg" />Perturbed-Attention Guidance<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6246283f-bd42-451c-aad5-abe36a65b397/width=525/6246283f-bd42-451c-aad5-abe36a65b397.jpeg" />and Mahiro CFG for reForge<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/a46ca7bd-4017-4437-8fdb-a157084f80ab/width=525/a46ca7bd-4017-4437-8fdb-a157084f80ab.jpeg" />Recommended VAE is the one by LAXHAR Lab member heziiiii. You can find it <a target="_blank" rel="ugc" href="https://huggingface.co/heziiiii/noob_vae_test">here</a>.<h2 id="license-dn7rlwnk6">License</h2>Same as NoobAI.<h1 id="q-and-a-sjus41qes">Q &amp; A</h1><ul><li>Q: What's the recipe? A: Version 1 is just a straight lora/lycoris merge. <a target="_blank" rel="ugc" href="https://civarchive.com/models/833294/noobai-xl-nai-xl">NoobAI-XL (vpred)</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/316075?modelVersionId=1270820">this LyCORIS</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/723360?modelVersionId=1204158">this LoRA</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/723360?modelVersionId=964220">this LoRA</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/915440?modelVersionId=1200214">this LyCORIS</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/856285?modelVersionId=1247042">this LoRA</a>, and <a target="_blank" rel="ugc" href="https://civarchive.com/models/502316/spo-locon?modelVersionId=558316">this LoCON</a>. Weights are 0.6, 0.5, 0.5, 0.25, 0.5, and 1. You can take them and make your own mix or whatever, I don't give a shit. Hiding recipes for merges is lame and for gigalosers. Version 1.5 was originally an attempt to convert <a target="_blank" rel="ugc" href="https://civarchive.com/models/138257?modelVersionId=1390019">MIX-GEM-XLBD1</a> to vpred, but I ended up using so many weights from LobotomizedMix that it ended up just being another version of lobotomized mix. The <a target="_blank" rel="ugc" href="https://github.com/ljleb/sd-mecha">sd-mecha</a> code for this comes from @illyaeater (thanks imi). Mostly it just SLERP merges the dissimilarity between LobotomizedMix_v1,MIX-GEM-XLBD1, and the parallel component between the two using SDXL1.0 as a base, at alpha 0.5 which is blurred Gaussian kernel to prevent sharp transitions and introduce artifacts, onto SDXL1.0. (For what a SLERP merge is, see the sd-mecha documentation). After that it's just a simple merge of replacing index 0 of every unet attention layer from the former with the latter. Code is <a target="_blank" rel="ugc" href="https://pixeldrain.com/u/1ipKuoLT">here</a>.</li><li>Q: Difference between versions? A: Check the version sidebar. Alternatively, X/Y comparison charts for:<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/posts/12723440">V1 vs V1.5</a></li></ul></li><li>Q: Does this model have innate artist tags? A: Yes, but they're not very strong. You either have to prompt them at high weights to overcome the base style or they're not readily apparent.</li><li>Q: Tips and or tricks? A: If you have a composition you like but you don't like the way the final generation comes out, experiment with changing the prompt and sampler during a hires fix.</li><li>Q: How to get good eyes and fingers? A: Hires fix. Hires fix always. Models trained with 4CH VAE innately suck at producing clean results off the bat unless you're going for a simple style so you need to run the second denoise pass to clean up noise, even if you don't upscale.</li><li>Q: Pony? A: No.</li><li>Q: Why are all the seeds 114514? A: <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/54dbe8e4-a3c4-4876-9ad4-145acf9e80d8/width=525/54dbe8e4-a3c4-4876-9ad4-145acf9e80d8.jpeg" /></li></ul>

Lobotomized Mix

Very minimal differences. Text encoder is the same, many of the weight blocks are also the same, slight adjustment to some blocks regarding lighting. Primarily LobotomizedMix_v10 was multi_domain_alignment merged with <a target="_blank" rel="ugc" href="https://civitai.com/models/138257?modelVersionId=1390019">MIX-GEM-XLBD1</a> and some other unet blocks remerged from v1. Compared to v1, the primary effect is that the color blue is a lot less dominant and the color yellow is much dominant. As a result, warm colors now take precedent over cold colors.Pros:<ul><li>Shine, luminescence, caustics, and other warm color dependent lighting elements now are slightly improved</li><li>Improved backgrounds, large anatomical features like limbs see improvement, fingers are sometimes better sometimes worse</li><li>Clothing and object interaction is better on average, may be worse in individual seeds</li><li>Default style greatly improved, also better at wide aspect ratios</li></ul>Cons:<ul><li>Shadows, darkness, black backgrounds, and other cold color depending lighting elements are worse</li><li>Artist keyword effectiveness heavily reduced, default style is harder to move away from</li><li>Seed variation reduced, randomizing seed now produces less dramatic effects</li></ul><a rel="ugc" href="https://huggingface.co/LMFResearchSociety/CheckpointArchive/blob/main/Jemnite/LobotomizedMix_v15.safetensors">Huggingface</a>.

Files with SHA256: 1074964b2c9fbe1000e89136b668f2f442a2e4777f093c55823b576c865a3547