先看说明
鹅群：
墨幽炼丹阁-太清阁：858495398
墨幽炼丹阁-上清阁：764481772
这是一个墨幽系列模型的专用负嵌入（embeddings），使用墨幽系列模型负面词只需要这一个负嵌入就可以稳定出图
下载后若文件名不一致请将文件名更名为：dedicated_to_artificial_humans
文件放入路径：/sd文件夹/embeddings
在负面tag中输入：(dedicated_to_artificial_humans:1.2)
墨幽专用负嵌入适用于：
鹅群：
墨幽炼丹阁-太清阁：858495398
墨幽炼丹阁-上清阁：764481772
This is a specialized negative embeddings for the Moyou series model. Using negative words in the Moyou series model only requires this negative embeddings to stabilize the plot
If the file name does not match after downloading, please rename it to: dedicated_ To_ Artistic_ Humans
File placement path:/sd folder/embeddings
Enter in the negative tag: (dedicated_to_artificial_humans: 1.2)
Moyou dedicated negative embedding is suitable for:
鹅群：
墨幽炼丹阁-太清阁：858495398
墨幽炼丹阁-上清阁：764481772
これは墨幽系列モデルの専用負埋め込み（embeddings）であり、墨幽系列モデルの負の語を使用するにはこの負の埋め込みだけで安定した描画が可能である
ダウンロード後にファイル名が一致しない場合は、ファイル名を：dedicated toartificial_humans
ファイルの配置パス：/sdフォルダ/embeddings
負のタグに次のように入力します。（dedicated to artificial _ humans：1.2）
墨幽専用負埋込用途：
Description
FAQ
Comments (6)
We don't know what it is, we don't know what it is for, there is no explanation, i don't read japanese nor chinese.. Good job !
Well, I know how to read Chinese and knows this is a negative embed, but I had no idea what it does.
So I put this into the postivie prompts with weight at 1.5 (OP recommends 1.2 as negative), and it seems to be some kind of stylistic thing, as the results are fingers that looks like maize (or the reverse).
Doesn't seem to work as well as bad_hands when it comes to fingers, compared with easy-negative and epic-realism on a photorealism model, it appears to generate individuals with a more "artificial" look with a hint of digital drawing.
@cliang96844 这是一个墨幽系列的专用负嵌入，只需要只一个负嵌入加上正面的1girl，moyou就可以稳定出图，这就是它的作用
The today released model is same type release with zero info and input as this embedding is.
Just skip it.
Seems to be spam.
我看的懂中文 我也不知道這是什麼
也沒有說明 也沒清晰的預覽圖
似乎是
專用於此製作者"墨幽"系列 checkpoint 的 負嵌入 embedding
但這個負嵌入是為了獲得什麼 沒說明
DeepL to EN:
Negative embedding for this author's "墨幽" series Model.
But what this negative embedding is for, it's not specified.
应该结合很多负面进行2次训练的综合，让负面提示词简洁。