<h2 id="click-here-to-join-the-discord-and-share-your-images-love"><a target="_blank" rel="ugc" href="https://discord.gg/8NYk7Gfm5M">Click here to join the discord and share your images</a> ♥</h2><h2 id="new:-my-wildcards-+-visualizer-can-be-found-here!-(wildcard-files-coming-soon-love)"><a target="_blank" rel="ugc" href="http://wildcards.corbin.wtf">NEW: My wildcards + visualizer can be found here! (wildcard files coming soon ♥)</a></h2>♥ ａ　ｍｏｄｅｌ　ｆｏｒ　ｇｏｏｎｅｒｓ ♥Mugen is a stylized anime/realistic model for SD 1.5, PDXL and Illustrious.All sample images were generated using ADetailer on the face and hands (no ADetailer on the hands in the Illustrious images).Please check the About this version section for more model specific info on my older models.Mugen IL:Like most Illustrious models, using "masterpiece, best quality, high quality" in your positive prompt is the best practice.Basic positive and negative prompts: pos: 1girl, huge breasts, cleavage, breast suppress, masterpiece, best quality, high qualityneg: worst quality, low qualityAll sample images for the Illustrious model were generated without inpainting and using DPM++ 2M Karras at 4 CFG - using this sampler at higher CFGs will cause burn-in and other artifacting. Euler a Karras at 7 CFG is my usual choice for sampler/scheduler (variable but can go higher than DPM++ 2M Karras). Mugen XL:For Mugen XL v1-3.1, prompting isn't much different across versions, but experimenting is always recommended.Try adding "3d, blender, " to your negative prompt near the beginning (preferably after your score tags) to avoid that style if you don't want it. XL v1 and v2 heavily default to that style of image. While it's not as apparent in XL v3.1 with proper style tags in your positive prompt (or LORAs), it's still frequent enough to be worth prompting out.Mugen 1.5:For Mugen v1-3, generally just use the guidelines given in v3.♥

Mugen, Infinity

New merge aiming to retain the main 3d/blender style of Mugen, Infinity XL v3.1 on the Illustrious base model. All model testing/sample image creation was done without LoRAs. Other artstyles are achievable, however, during testing most generations that tried to change the artstyle/artist would only lightly effect the final result with the artist tag used, not fully transforming the overall artstyle of the image.Like most Illustrious models, using "masterpiece, best quality, high quality" in your positive prompt is the best practice.Basic positive and negative prompts: pos: 1girl, huge breasts, cleavage, breast suppress, masterpiece, best quality, high qualityneg: worst quality, low qualityAll sample images for the Illustrious model were generated without inpainting and using DPM++ 2M Karras at 4 CFG - using this sampler at higher CFGs will cause burn-in and other artifacting. Euler a Karras at 7 CFG is my usual choice for sampler/scheduler (variable but can go higher than DPM++ 2M Karras).

Files with SHA256: d81becab766d573e1524649da3fa70c0c2788c9fee5eadf097885e34be89f22a