Why 5? Because I merged so many times. This is the last version. I merged real pony,<a target="_blank" rel="ugc" href="https://civarchive.com/models/365041/real-pony?modelVersionId=515456">real pony</a>、<a target="_blank" rel="ugc" href="https://civarchive.com/models/257749/pony-diffusion-v6-xl?modelVersionId=290640">Pony Diffusion V6 XL</a>、<a target="_blank" rel="ugc" href="https://civarchive.com/models/133005/juggernaut-xl?modelVersionId=456194">Juggernaut XL</a>、<a target="_blank" rel="ugc" href="https://civarchive.com/models/43977/leosams-helloworld-xl?modelVersionId=570138">LEOSAM's HelloWorld XL</a>、<a target="_blank" rel="ugc" href="https://civarchive.com/models/147720/colossus-project-xl-sfwandnsfw?modelVersionId=592868">Colossus Project XL (SFW&amp;NSFW)</a>, and iterated and merged dozens of times. There are also some LORA, and some lora I trained myself. I can't remember clearly. I used various fusion methods. Anyway, this is it in the end.OK, for now, you can try to present all kinds of NSFW you want, and basically all of them can be achieved. The original version can present characters, and basically all of them can be presented, but this part will be more 3D. Of course, you can also improve the realism through keywords such as "Photorealistic", "realistic" or "realism", "photo taken with a canon Eos R5," etc.The example pictures have enlarged fixes using "Hires. fix" and face fixes using "ADetailer".I use webui.Sometimes you can also not use negative and score.There are still many imperfections. You can also try to use DPM++ 3M SDE plus Exponential, CFG 3.5-4.5 to generate a more delicate picture and see if you can reduce the errors.Wish you all a happy time.________________________________________________________________________________________________________为什么是5，因为大概融合了那么多那么多次。这是最后一个版本。融合了大概有<a rel="ugc" href="https://civarchive.com/models/365041/real-pony?modelVersionId=515456">real pony</a>、<a rel="ugc" href="https://civarchive.com/models/257749/pony-diffusion-v6-xl?modelVersionId=290640">Pony Diffusion V6 XL</a>、<a rel="ugc" href="https://civarchive.com/models/133005/juggernaut-xl?modelVersionId=456194">Juggernaut XL</a>、<a rel="ugc" href="https://civarchive.com/models/43977/leosams-helloworld-xl?modelVersionId=570138">LEOSAM's HelloWorld XL</a>、<a rel="ugc" href="https://civarchive.com/models/147720/colossus-project-xl-sfwandnsfw?modelVersionId=592868">Colossus Project XL (SFW&amp;NSFW)</a>，迭代融合大概几十次吧。还有一些LORA，还有一些我自己训练的lora.记不清楚了。用了各种融合的方式。总之最后就是这样了。OK，目前对于呈现你想要的各种NSFW,可以尽情尝试，基本都可以实现。原版可呈现角色，基本都可以呈现，但是这部分会较为偏3D感。当然，你也可以通过如“Photorealistic”、“realistic”或“realism”, “photo taken with a canon Eos R5,”此类关键词来提高真实度。示例图有使用“Hires. fix”放大修复，和使用“ADetailer”修复面部。我使用的是webui。有时也可以不使用负面和score。还是有很多不完美的地方，你也可以尝试用DPM++ 3M SDE加Exponential,CFG 3.5-4.5这样的方式生成更细腻的画面，看看能不能减少错误。祝各位愉快。

Files with SHA256: 433b26281d8d92f4dcfe379ddabaafa2c0a38f5ff42f61cd9756eabc40c945d9