No special prompt words are required, but the more detailed the prompt words, the betterDescription: A Lora that can significantly increase light and shadow contrast, hair texture, and background details1. The weight is between 0.5 and 1. The higher the weight, the better the performance, but the relevance of the prompt words will be reduced2. Works better on poor images3. Can be used together with other Lora4. Use most of the 2D models, if the effect is poor, you should reduce the weight appropriatelyTrain with the Lion optimizer and noise-offsetset the block weights to increase generalizationps: A very confusing Lora,sometimes the effect is inexplicably good, and sometimes it is not obvious~~o(≧ω≦)o///////////////////////////////////////////////////不需要专门的提示词，但图片提示词请越详细越好描述：一个能显著增加光影对比，头发质感，背景细节的Lora1.权重在0.5到1之间，权重越高表现越好，但对提示词的还原度会降低2.对于较差的图像作用更好3.与其他Lora可以一起使用4.使用大部分2D模型，如果效果较差应该适当降低权重使用Lion优化器（optimizer）和noise-offset训练设置分层权重以增加泛化性ps:一个相当玄学的Lora，效果有时候莫名其妙的好。~~o(≧ω≦)o

Enhanced lighting details|光影细节增强|detailed__hikari|v1.0|Lora|2D|

trained 3 times with 3 different training set

Files with SHA256: 0b4e471fdc51ae2f1e7260b11ae8b8a88160748437b2af22119ab684d361dbfd