とらいし６６６

Based on miaomiao v1.2 training

It will be effective for the upgraded version 1.3 as well as other anima base models.

Lora does not have trigger words, so you only need to adjust the weights to achieve the desired style.

During my training, I chose a Lora that could achieve sufficient coloring effects and could also restore the face shape and hairstyle even under high-intensity conditions. However, the effect might be weakened when loading multiple Loras.

If you want a certain degree of coloring, or if you wish to retain more of the style of other lora models

Recommendation strength: 0.3

If your character Lora has some overfitting issues, or if you want a more distinct skin style Recommendation strength: 0.5 - 0.7

If you need to completely restore the facial features and the distinctive hairstyle of that style of painting

Recommendation strength: 0.8 - 1.0

If you use the character Lora, in my test, I recommend that the weight of style Lora be slightly greater than that of the character Lora.

(At least for me, myself, it is like this)

基于miaomiao v1.2训练

对升级版本1.3和其他动漫类anima模型都可以生效

该lora没有触发词，调节权重即可获得风格化

Anima虽然有该画师的tag，但很明显还原度不够好

如果你想达到一定的着色效果，或者希望保留更多其他洛拉模型的风格

推荐强度：0.3

如果您的角色洛拉存在过拟合问题，或者您希望拥有更明显的皮肤风格

建议强度：0.5 - 0.7

如果你需要完全恢复那种绘画风格的面部特征和独特的发型

推荐强度：0.8 - 1.0

如果你使用角色Lora，在我的测试中，我建议风格lora的权重应该略高于角色Lora

（角色lora大部分自带一定程度的风格，所以你想要风格化效果明显就要尽量拉高风格lora的权重）