Different from the first generation, although this time did not succeed, the background and foreground characters can be seen to have some sense of separation, this sense of separation will be generated when there are fewer tags, once the positive description of the tag becomes more and more, the foreground and background will begin to merge. Next I will continue to study how to completely separate the foreground and background.
跟第1代不太一样，这次虽然也没有成功，不过背景和前景人物能看出有一些分离感，这种分离感在tag少的情况下会产生，一旦正面描述的tag变多，前景和背景就开始融合。接下来我会继续研究怎样让前景和背景完全画风分离。
Different from the first generation, although this time did not succeed, the background and foreground characters can be seen to have some sense of separation, this sense of separation will be generated when there are fewer tags, once the positive description of the tag becomes more and more, the foreground and background will begin to merge. Next I will continue to study how to completely separate the foreground and background.
示例图形使用的tag：
1girl,
photo background,
Negative prompt: (bad and mutated hands), (worst quality), (low quality), (normal quality),
Steps: 30, Sampler: Euler a, CFG scale: 15, Size: 512x640, Model hash: df333d88d5, Model: 23-3-21-fish mix1.1-Experimental-half, Denoising strength: 0.5, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires steps: 10, Hires upscaler: R-ESRGAN 4x+