The model has enhanced the text-to-image effect under simple prompt descriptions. By adding the trigger word "add_atmosphere" at the beginning, even if only a simple item name or picture content is input, the light and shadow atmosphere of the generated image will be significantly enhanced. This model is also applicable to image-to-image redrawing under the Flux.1 model, and can enhance the light and shadow atmosphere of realistic photos by reducing the redrawing intensity.

模型加强了在简单提示词描述下的文生图效果，在最开始处加上触发词add_atmosphere 即使只是输入一个简单的物品名称或画面内容，生成画面的光影氛围感也会明显增强。该模型同样适用于Flux.1模型下的图生图重绘，通过降低的重绘幅度实现对写实类照片的光影氛围增强。