固定提示词可以用(masterpiece,best quality,beautiful and aesthetic:1.2), (high detailed skin:1.2), 8k uhd, dslr, soft lighting, high quality, film grain,随意些就好。 固定提示词可以用(masterpiece,best quality,beautiful and aesthetic:1.2), (high detailed skin:1.2), 8k uhd, dslr, soft lighting, high quality, film grain,随意些就好.
一般来说,只需要找一张图反推画面内容物,通过模型再生成就是一张时尚摄影了,例如示例图:1boy, male focus, black hair, jacket, white footwear, shoes, full body, hand in pocket, pants, shirt, solo, sitting, flower, stuffed toy, stuffed animal, teddy bear, long hair。
The specialized model of fashion photography is very, very unsuitable for everyday style, and the example image does not use any Lora model.
Fixed prompt words can be used (masterpiece, best quality, beautiful and aesthetic: 1.2), (high detailed skin: 1.2), 8k uhd, dslr, soft lighting, high quality, film grain, just be casual.
Generally speaking, only one image needs to be used to infer the content of the image, and through model regeneration, it becomes a fashion photography, such as example images: 1 boy, female focus, black hair, jacket, white footwear, shoes, full body, hand in pocket, pans, shirt, solo, sitting, flower, stuffed toy, stuffed animal, teddy bear, long hair.
It is recommended to cooperate with the Controlnet model for better results. Adjusting the color tone of the image through tile or shuffle models can effectively improve the frame.
Yuanjing still needs an adetailer, but there is a problem with the male face being repaired to look like a female.
In addition, the facial Lora trained based on chilloutmix is not easy to reproduce. If necessary, you can try using this model to retrain the face, but I did not try (lazy). Alternatively, use this model to generate images and then use your commonly used model for secondary modifications.
请先
!