This is the model used when making lora.

https://civitai.com/models/232545/badbro31

Trained with 768,768

I confirmed that it works great in a 512*1024 resolution full body shot.

weight : 0.5~0.8

trigger : Basically, entering

'negligee'

is stable.

And depending on your taste,

'black dress', 'brown dress', 'red dress', 'see-through', 'blue dress', 'green dress', and 'pink dress' are useful.