这是 80 张图片的测试版本。我使用 AI-toolkit 修改的脚本训练了这个 LoRA 模型。
该模型的训练集基于与“Xiaohong Dou”相同的数据集,并添加了一些额外的数据,并使用 FLUX 进行训练。该模型的理念是针对服装和人体进行训练。然而,当训练集的图像数量超过一定数量时,可能会导致构图异常。因此,该模型偶尔可能会生成扭曲的手脚,但大多数情况下效果良好。
采样方法(Sampler):Euler simple
采样迭代步数:20
宽度:768,高度:1280

作品参照

a girl wearing a black belly dance costume, with intricate details and a graceful pose.
Her hair is pulled back in a bun and she has a serene expression on her face,
standing in a serene natural setting.
She is looking directly at the camera with a calm and composed expression. The cave's walls are rocky, and there are plants and foliage surrounding the water.
Steps: 20, CFG scale: 4, Sampler: Euler, Seed: 692958814229013

a girl in a black dress sitting on a bed, with a white cloth in the background. She is wearing a belly dance costume, with intricate details and a graceful pose. Her hair is pulled back in a bun and she has a serene expression on her face.

a girl in a blue dress standing in a serene natural setting. She is positioned in a cave-like environment with a water body in the background. The water appears calm, reflecting the surrounding greenery. She is wearing a belly dance costume, with intricate details and a graceful pose. She is looking directly at the camera with a calm and composed expression. The cave's walls are rocky, and there are plants and foliage surrounding the wat
Steps: 20, CFG scale: 3.5, Sampler: Euler, Seed: 1016525942000268
本内容及资源均来自网络,版权归原作者所有,仅供个人学习研究,请勿商用和侵犯他人肖像,如需商用请联系原作者。若原作者内容有侵权之处请速联系我们,我们将会在24小时内删除。