This is a lora optimized specifically for contrast views. Although ZIT can originally draw multiple views, there are some problems:1. Each subgraph needs to be described in detail.2. Some descriptions for a single sub-picture may inevitably appear in all sub-pictures, such as exposing breasts.This lora only needs to simply describe the initial clothing, and then it can automatically generate clothing for other sub-pictures.The most stable number of subgraphs is 2, and 3 also performs well. When the number of subgraphs exceeds 3, it is very unstable (exposure may be out of order).这是一个专门针对对比视图进行优化的lora。虽然ZIT原本就可以绘制多视图，但存在一些问题： 1、需要详细描述每个子图。2、某些针对单个子图的描述可能不可避免地在所有子图中出现，比如露出胸部等。本lora只需要简单地描述初始服装，即可自动生成其他子图的服装。最稳定的子图数量是2个，3个也有不错的表现，子图数量3个以上时非常不稳定（暴露度可能会乱序）。

Files with SHA256: 13cef19ff2c0d1df1566d5f3a68a4051bee8757f5c07c092905a134a2704136f