这是做儿童插画的lora,训练图用了92张训练图,
底模用的是bestqualityPastelmix_bestqualityPastelmix,尺寸是512x768,描述词简单描述就好了,
例如女孩, 帽子, 花, 独奏, 草, 站立, 鸟, 脸红, 全身, 渐变背景, (插图:1.0), 杰作, 最佳质量...
我试了其他偏2次元风格的大模型也合适,做儿插场景也可以,权重调看着调就好,不过有时候会出现结构错误 o(一︿一+)o;
The following is machine translated
This is lora for children illustration, 92 training maps were used for the training maps.
The base model is bestqualityPastelmix_bestqualityPastelmix, the size is 512x768, and the descriptors are simply descriptive, such as girl, hat, flower, solo, grass, standing, bird, blush, full body.
For example girl, hat, flower, solo, grass, standing, bird, blush, full body, gradient background, (illustration: 1.0), masterpiece, best quality...
I'm not sure if this is a good idea, but I'm sure it's a good idea to try it out, so I'm not sure if it's a good idea.