MM_PortraitView

这是个工具系的LORA，可最大限度地生成标准正面肖像照片，我通常用于检测MIX模型是否受到污染，以及Lora的正面镜头泛化能力。(放心使用，不对自身面部产生影响)

将文件portraitView.safetensors放置在stable-diffusion-webui\models\lora中，然后输入提示词(请注意修改对应LORA文件名)

<lora:portraitView:1>,(portraitView,portrait,solo,from front,eye contact,looking at viewer,simple background),

\\无人物类Lora加载情况下，如果只能生成侧脸，通常说明这个MIX模型是一个已经被训练样本有限的人物类LORA污染过的模型。\\

\\加载人物Lora情况下，如果只有侧脸，说明你的训练集严重缺少正面样本，可以通过增加正面样本数量来提高正面生图成功率。

This is a tool of LORA that maximizes the generation of standard frontal portrait photos, which I commonly use to detect contamination of MIX models, as well as Lora's frontal lens generalization capabilities.

Put the portraitView.safetensors into stable-diffusion-webui\models\lora and enter the prompt

<lora:portraitView:1>,(portraitView,portrait,solo,from front,eye contact,looking at viewer,simple background),

If only side faces can be generated without character class Lora loading, it usually means that the MIX model is a model that has been contaminated by a character class LORA with a limited training set.

In the case of loading the character Lora, if there is only a side face, it means that your training set is seriously lacking positive samples, you can increase the success rate of positive images by increasing the number of positive samples.