Why release this version?
为什么要发布这个版本？
Firstly, this is in response to fans' suggestions to enhance the "oily" effect on characters.
Secondly, it is to experiment with new "age adjustment methods" and "oily lighting effects" to verify the feasibility of their principles. Currently, the approach seems feasible, and we will continue to add datasets for training to achieve better results.
首先，这是应粉丝的建议，调高人物“油腻感”的效果。其次，进行新的“年龄调整方式”和“油光效果”的实验，验证其原理可行性，目前来看方案可行，之后会继续增加数据集进行训练来获取更好的效果。
Compared to version 0.5
与0.5版本的相比
Advantages
优势
1.Uses a new "oily lighting effect" method, making the "oily" effect on characters more pronounced.
1.使用新的“油光效果”方式，人物“油腻感”更加明显。
2.Due to the replacement of the age adjustment method, the age of generated male characters is more balanced, and elderly male characters will not appear. Young male characters can now be generated correctly. If you want to further increase the age, please add "milf, mature_female" to the prompt. This model has optimized these two prompts to increase age without significantly changing the image.
2.由于替换了年龄调整的方式，生成男性时年龄更加均衡，不会出现老龄男性。正太角色现在可以正确生成。如果想要继续提升年龄，请在提示词中同时添加“milf,mature_female”，本模型对这两个提示词进行了优化，可以在尽量不改变画面的情况下提升年龄。
3.Uses illuNoob as the base, resulting in better character recognition.
3.使用了illuNoob作为基底，角色识别更加良好。
4.Better compatibility with breast size prompts (small breasts, medium breasts, large breasts, huge breasts, gigantic breasts). Please see the comparison images for details.
4.对胸部大小提示词（small breasts,medium breasts,large breasts,huge breasts,gigantic breasts）有更好的兼容性，详情请查看对比图。
Disadvantages
劣势
1.Limb anatomy may be worse (depending on the situation).
1.肢体解刨可能更差(视情况而定)。
2.Lower compatibility with artist tags.
2.画师tag兼容性变低。
3.Requires enabling facial restoration.
3.需要开启面部修复。
Generation Parameters
生成参数
euler_ancestral_cfg_pp
cfg：1
832*1216/1024*1472
This model does not require quality words such as "masterpiece, best quality, newest, absurdres, very aesthetic," etc. Although it may seem counterintuitive, quality words do not improve your generated images (personal understanding).
本模型无需使用质量词如“masterpiece, best quality, newest, absurdres,very aesthetic”等等，虽然很反常识，但质量词并不会使您的生成图变得更好（个人理解）。
For the basic text-to-image workflow, please refer to the example images. You can save the example images and drag them into ComfyUI to view.
基础文生图工作流请查看作例图，可将作例图保存后拖入comfyui中查看
This is still an experimental version, mainly used to experiment with new methods. Of course, this version itself is not bad, as can be seen in the example images.
这仍然是一个实验版本，主要用来实验新的方式，当然这个版本本身并不差，这一点在作例图中可以查看。
Once again, thank you all for your downloads and likes. Due to the new promotion slot auction system, I cannot afford the "advertising fee buzz," so I need your likes and comments even more!
再一次感谢大家的下载和点赞，由于新的宣传位拍卖系统，个人无力承担其“广告费的buzz”，所以更加需要大家的点赞和评论！
.jpeg)