- Model introduction: SDXL model, which is a scene of lora;
- Creative overview: do not use desert, deep forest, deep sea three types of scene pictures for training, pictures of rocks, classrooms, corals, mountains, planets, warships, planes, stars and other trivial elements, I summarize as debris;
- Generate suggestions: If there is only one lora, the weight is on 1; If used with other LORAs, 0.4 is recommended when a high resolution repair is turned on;