Dreamy and Beautiful Learning Style

底模选择： 太郎FLUX-写实风格大模型 底模训练

解决问题： 质感屏蔽掉了用Flux-dev-fp8的人物AI油质皮肤的感觉！

要保持该风格的完美发挥，提示词中应尽量减少干扰lora发挥的其他色调描述词

此模型可用在创意设计，经典美术艺术。叠加lora人物场景便面有一层光晕附着，制作卡牌，人像壁纸等等创意设计上面。

在线生图参数设置

使用「在线生图」功能时，请参考以下参数配置来进行设置。

推荐权重： 0.8~1

触发词：abcdefg style

采样推荐：Euler

迭代步数：20-30

CFG : 在线webui建议保持默认3.5，comfyUI建议设置为1

推荐尺寸：1024x1536， 1024分辨率以上 效果更好

高分辨率修复：4x_NMKD-Siax_200k/步数20/重绘幅度0.3~0.5

Flux对文字的理解能力很强，建议提示词以小作文形式去写，可以借助市面上的AI语言模型去对自己的提示词扩充润色再丢进去。

描述句式示例：供参照

abcdefg style, . this is a digital illustration featuring a serene, ethereal woman with a surreal, dreamlike quality. the artwork is executed in a smooth, hyper-realistic digital painting style with soft, blended colors and subtle shading. the subject is a young woman with pale skin and closed eyes, exuding a sense of calm and introspection. her hair is voluminous and fluffy, resembling a cloud, with a gradient of white to pale pink hues, adding to the ethereal ambiance. she wears an off-the-shoulder, long-sleeved white dress with puffed sleeves, which contrasts with her hair, creating a striking visual effect. her hands are delicately placed over her chest, and her nails are manicured with a glossy finish. the background is a gradient from deep purple at the bottom to a lighter pink at the top, enhancing the dreamy atmosphere. sparkling stars are scattered around her head, adding a magical touch. the overall composition and color palette evoke a sense both of tranqu

abcdefg风格。这是一幅数字插图，描绘了一个宁静、空灵的女人，具有超现实、梦幻般的品质。这件艺术品以流畅、超现实的数字绘画风格进行，具有柔和、混合的色彩和微妙的阴影。拍摄对象是一位皮肤苍白、闭着眼睛的年轻女子，流露出一种平静和内省的感觉。她的头发蓬松蓬松，像一朵云，有白色到淡粉色的渐变，增添了空灵的氛围。她穿着一件露肩长袖蓬松袖白色连衣裙，与头发形成鲜明对比，营造出引人注目的视觉效果。她的双手小心翼翼地放在胸前，指甲修得很光滑。背景是从底部的深紫色到顶部的浅粉色的渐变，增强了梦幻般的氛围。闪闪发光的星星散落在她的头上，增添了一种神奇的感觉。整体构图和调色板都给人一种曲曲的感觉

To maintain the perfect expression of this style, other tone descriptors that interfere with Lora's performance should be minimized as much as possible in the prompt words

This model can be used in creative design and classic art. Overlaying Lora character scenes creates a layer of halo attachment, making creative designs such as cards, portrait wallpapers, and more.

Online image parameter settings

When using the "Online Image Generation" function, please refer to the following parameter configurations for settings.

Recommended weight: 0.8~1

Trigger word: abcdefg style

Sampling recommendation: Euler

Iteration steps: 20-30

CFG: For online webui, it is recommended to keep the default 3.5, and for comfyUI, it is recommended to set it to 1

欢迎大家尝试和返图，愿大家玩的愉快！