1. Training Size: 512x768

2. Weight: 0.7-0.8

3. Main Model: Orange v7

4. The image taken from the side will look better

5. Adding the keyword "(fair skin)" will increase the weight, otherwise the skin may appear somewhat dark

6. Personally, I feel that this Lora is a failure, but drawing multiple cards may still produce good images

7. The generated case study images provide the ideal effect for the development purpose of this Lora