Recommended Lora strength : 0.8 ～ 1.0
The generation of wide picture size might be difficult.
In this case, I recommend to use the particular words such as split view, multiple views.
(see my sample pic)
v1.0 : The face in the original image has been included in the training.
v2.0 : Since the model file are only trained the breasts, the face depends on the base model.
Description
Re-training with the improved material pictures.