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）。