<h1 id="heading-1421">ReMaster+1.2.2</h1>ReMaster+ 1.2.2 is a merged model from 1.2.1 and <a target="_blank" rel="ugc" href="https://civarchive.com/models/212253/baxl-bartstylexl-or-blue-archive-flat-celluloid-style-fine-tune-on-animaginexl-v3-or">BAXL</a>thanks to 4BEraser (I actually used the BAXL with minus weight, LOL)<h2 id="heading-83">BAXL doesn't permit any commercial uses and ReMaster+1.2.2 follows BAXL's permission.</h2>(ReMater+1.2.1 follows Animagine XL 3.0's)Recipes are in my blog(<a target="_blank" rel="ugc" href="https://stablediffusionmaster.blogspot.com/2024/01/merged-model-remaster-recipe.html">Merged model ReMaster+ recipes(리마스터+ 레시피)</a>)PS. I recently used ComfyUI. A1111 makes noise sometimes for the base model of Animagine XL 3.0. So, some sample images are not reproduced in A1111<hr /><h1 id="heading-1421">ReMaster+1.2.1</h1>ReMaster+1.2.1 is a merge model based on ReMaster+1.2 and <a target="_blank" rel="ugc" href="https://civarchive.com/models/317792/chenkin-anime-hotbaby">chenkin-anime-hotbaby</a>.chenkin-anime-hotbaby is a finetuned model of Animagine XL 3.0. Thanks to Chenkin.Please note that 1.2.1 may generate NSFW images without any NSFW prompts. 1.2 may be suitable for SFW images.<ul><li>nsfw into negative prompts may help this problem but it might not be helpful. So, I recommend adding rating: general into positive prompts and nude, nsfw, sex into negative prompts.</li><li>Also, best quality and masterpiece can generate NSFW things even though there's no NSFW hint</li></ul><h1 id="heading-1421">ReMaster+</h1>I merged ReMaster1.0 with animagine-xl-3.0(<a target="_blank" rel="ugc" href="https://civarchive.com/models/260267/animagine-xl-v3">Animagine XL V3 - V3.0 | Stable Diffusion Checkpoint | Civitai</a>) (This is really good model to generate anime images. I think this is SOTA model for free)Thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/cagliostrolab">cagliostrolab (Cagliostro Research Lab) (</a><a target="_blank" rel="ugc" href="http://huggingface.co">huggingface.co</a><a target="_blank" rel="ugc" href="https://huggingface.co/cagliostrolab">)</a>Prompting tips:<ul><li>(<s>realistic:0.01) </s>(realistic:0.5) would generate good semi-realistic images. Please check the prompts of the sample images. (realistic:0.01 may cause artifacts)</li><li>realistic in negative prompts will generate less realistic images.</li><li>low CFG(less than 7) scale for realistic, high CFG(higher than 7) for animes</li></ul>Recipes: <a target="_blank" rel="ugc" href="https://stablediffusionmaster.blogspot.com/2024/01/merged-model-remaster-recipe.html">Merged model ReMaster+ recipes(리마스터+ 레시피) (</a><a target="_blank" rel="ugc" href="http://stablediffusionmaster.blogspot.com">stablediffusionmaster.blogspot.com</a><a target="_blank" rel="ugc" href="https://stablediffusionmaster.blogspot.com/2024/01/merged-model-remaster-recipe.html">)</a>Please note that the Licence in <a target="_blank" rel="ugc" href="https://civarchive.com/models/260267/animagine-xl-v3">Animagine XL V3 - V3.0 | Stable Diffusion Checkpoint | Civitai</a>This model follows the same license of AnimagineXL3<hr /><h1 id="heading-1422">V1.0</h1>I tuned Reproduction(<a target="_blank" rel="ugc" href="https://civarchive.com/models/118729/reproduction-or-sdxl">Reproduction | SDXL - 2v1.2 | Stable Diffusion Checkpoint | Civitai</a>) by merging custom LoRas.Thanks to <a target="_blank" rel="ugc" href="https://civarchive.com/user/Aikimi">Aikimi</a>"gothic, realistic, makeup, masterpiece, best quality" would make good images.Please see the prompts for sample images.

ReMasterXL

Changed the base model from animagineXL v3.0 to animagineXL v3.0-base

remasterxl_remasterV12.safetensors

Files with SHA256: c200fb1ff35fa2eec75d0e40726e33cf7ad661f4c2e36ec7ca733ffcb1c596d7