<h2 id="if-you-like-the-model-please-share-your-images-so-i-can-make-more-loras-and-models">If you like the model please share your images so i can make more Loras and models</h2>I recommend using hirefix res: i normally use 1.35-1.5, 20 hire steps, 0.35-0.5 denoising strengthIt is not necessary but it improves the hands and eyes, (which is very important to me so...)<h3 id="my-models-are-now-in-seaart-if-you-make-images-on-that-plataform-please-share-them-here-too">MY MODELS ARE NOW IN <a target="_blank" rel="ugc" href="https://www.seaart.ai/es/models/detail/0e9836712a7fc989db9607bd73d0e6af">SEAART</a>, IF YOU MAKE IMAGES ON THAT PLATAFORM PLEASE SHARE THEM HERE TOO</h3><h2 id="updates:">UPDATES:</h2><h1 id="v8.5">v8.5</h1>ngl, this version is here because I merge a model with my lora and then with DanLiuWei V4 (one of the best versions of the model, lets be real) I like how the skin and lighting of this version looks, and im thinking in using this version as base training model for next versions of the lora<h1 id="v8.0">v8.0</h1>update base lora trying to add more characters<h1 id="v7.0">v7.0</h1>update base lora and merged with quality some quality loras<h1 id="v6.5">v6.5</h1>now the base lora is trained with Wai-illustrious-sdxl v15<h1 id="v6.0">v6.0</h1>updated database and i used another base model to try how it goes... then Wai-illustrious v15 went out... now i slightly regret it :(<h1 id="v5.05.5">v5.0/5.5</h1>A HUGE update to the base lora I used and i trained with a version of One Obsession with DanLiuWei v4.0, no more extra Loras, just mine :3for v5.5 I made some merges between 5.0, a version with a slightly updated base lora and DanIllustrious "other version", I think this one has a better anatomy and it sacrifices some shading but use the one you prefer :D<h1 id="v4.0">v4.0</h1>Like v3.0, updated lora trained with janku and more lora merges to make the shadows better<h1 id="v3.0">v3.0</h1>updated base lora and a merge of some other loras<h1 id="v2.0">v2.0</h1>I updated the lora i originally use so its a little more flexible, but is still not perfect, i suggest that if you use short hair, add to your negative prompt inner-colored hair if you dont want it to show upI also merged more loras into the model<h2 id="to-generate-images-i-recommend:">To generate images I recommend:</h2><ul><li>20-25 steps</li><li>1024x1024</li><li>4.5-5 CFG</li><li>ADETAILER</li></ul>

DanLiuWei Mix

updated base lora merged with Janku

danliuweiMix_v30.safetensors

v3.0

v8.9

v8.5

v8.0

v7.0

v6.5

v6.0

v5.5

v5.0

v4.0

v2.0

v1.0

Files with SHA256: e6d07fc05ce9ba23e64c9d71fc66a75d6d565a40ac671aad6993280d3eeec168