The large model generates relatively simple images, but with a wide variety of materials. It can depict individuals, groups, people with animals, and various scenes, making it ideal for use with different versions of Lora.

Recommended Parameters:

Sampling Method: Euler a  or DPM++ 2M Karras

Iterations: The displayed image uses 20

CFG :   5-7

Image Size: Recommended 512*768

(Suggested to enable resolution repair, with 20 iterations and a repaint range of 0.3)

Welcome to provide feedback and follow for future updates.