Type
LORA
Publish Time
2023-08-07
Base Model
SD 1.5
Trigger Words
batman
Copy
Version Introduction
v1.0 In the dreamthbooth training, light-colored clothing doesn't work, hair color is fixed, it lacks dramatic lighting effects, and side poses fail.
v1.1 It doesn't resemble a real person, the background isn't delicate enough, and it needs to activate contrlnet pose to match.
v1.2 It's closer to a real person's image, no need to activate contrlnet, and the background can be improved through pre-training materials and prompt word weighting.
