作為一個土下座圖片愛好者我訓練了這個lora，基於noobreal checkpoint。這是我第一個lora。

As a fan of Dogeza images, I trained this LoRA based on the NoobReal checkpoint. This is my first LoRA.

如果想出寫實系的圖，建議配合noobreal使用

If you want to create realistic-style images, I recommend using it with NoobReal.

這個lora會將更好的還原土下座姿勢，大多會以俯拍視角呈現。存在著手部和腳部變形、腰部過長的問題，但可以靠大量出圖解決

This LoRA improves the reproduction of the Dogeza posture, mostly presented from a top-down perspective. However, issues such as hand and foot deformation and an excessively long torso may occur, but these can be resolved through generating a large number of outputs.

我建議配合以下prompt使用

I recommend using it with the following prompt:

naked dogeza,prostration,humiliation,defeat,seiza,back of head,(kneeling down on the ground:1.103),(face down:1.34),(clothes on floor:1.158),(removed clothes are on the ground:1.158),(folded clothes:1.216),all fours,(face on floor:1.103),(heads on floor:1.05),hands own hands together,(hands on floor:1.103),(kneeling down on the ground:1.103),(face on hands:1.05),(nude:1.276),breasts press, ,full body,<lora:Cdogeza:0.8>,(Cdogeza:1)

optional:

from above

我沒用試過在其它模組上的成效

I haven't tested its performance on other models.

關鍵詞是:Cdogeza

The keyword is: Cdogeza.

現在，坐好沙發，讓女孩們知道該向誰臣服吧

Now, just sit back on the chair and let the girls know who they should bow to