格温的Lora，它在2D和3D下的表现似乎都还可以。

它不需要提示词就可以工作，尽管如此，我还是建议加上提示词稳定它的权重，否则在提示词太多或者其他Lora的干扰下它的权重可能不够。

It does not need prompt words to work, but I recommend adding prompt words to stabilize its weight, otherwise it may not have enough weight if there are too many prompt words or other Lora interference.

tag：long hair, twintails,twin drills, blue hair, blue eyes,star-shaped pupils,

推荐权重/weight: 0.8~1

例图模型/Sample graph model：cetus-Mix-v3.5
采样方法（Sampler）：DPM++SDE Karras

放大算法（Hires.fix）：R-ESRGAN 4x+Anime6B (0.3)

不要把放大的重绘幅度设定得太高，否则生成的图像可能会变得模糊。

Do not set the zoom redraw too high, or the resulting image may become blurry.