极简海报设计的风格LoRA,Minimalist poster design style LoRA, 触发词 trigger word : MinimalistPoster, 生成建议 Tips:：<ul><li>  权重推荐 weigh：0.6~0.8；</li><li>采样步数：30-50步数</li><li>  推荐采样器 Sampler：DPM++ SDE Karras / DPM++ 2M SDE Karras /Euler a</li><li>建议分辨率 resolution：512x768</li><li>  适合搭配大模型 Model：revAnimated etc.</li><li>  如果你想要英文文本，可以输入englishtext,</li></ul>  If you want English text, you can use "englishtext",<ul><li>  可以选择你想要的颜色主题，例如：bluetheme,</li></ul> You can choose the color theme you want, for example: "bluetheme",<ul><li>  用silhouett关键词以生成剪影,</li></ul> Use "silhouette" to generate silhouettes,<ul><li>  使用portrait,backlight提示词生成人物轮廓背光效果，</li></ul> Use "portrait", "backlight" to generate character outline backlight effects.<ul><li>  还可以尝试使用: monochrome,</li></ul> and also can use "monochrome", 有意思的是当你输入两个物体时，可能会将两个不同的物体创意性的结合在一起，使用silhouett,提示词时也可能将剪影进行融合。 What’s interesting is that when you input two objects, you may creatively combine two different objects together. When using "silhouette" you may also merge the silhouettes.Enjoy it.v2: 这一版综合效果非常优秀，更易使用，与模型的兼容性也更好。 使用anything-v5-PrtRE做底模时，画面更简洁。 另外这版本的物体融合效果也更明显。This version has excellent overall effects, is easier to use, and has better compatibility with models. When using anything-v5-PrtRE as the base model, the picture is simpler. In addition, the object fusion effect of this version is also more obvious.v1: 这一版的训练其实是过拟合的，适合口味较重的同学，而且对于剪影的混合效果比较好。This version of training is actually over-fitting, suitable for heavier tastes, and has a better blending effect on silhouettes.

极简海报 | Minimalist Poster

这一版综合效果非常优秀，更易使用，与模型的兼容性也更好。 使用anything-v5-PrtRE做底模时，画面更简洁。 另外这版本的物体融合效果也更明显。This version has excellent overall effects, is easier to use, and has better compatibility with models. When using anything-v5-PrtRE as the base model, the picture is simpler. In addition, the object fusion effect of this version is also more obvious.

MinimalistPoster_v2.safetensors

v2.0

v1.0

Minimalist Poster

The style of minimalist poster design, LoRA,Base model: anime AnyLoRA, (v2 version uses base model anything-v5-PrtRE)Training resolution: 512x768Trigger words:MinimalistPoster, Suggestions for generation: &nbsp; &nbsp;Recommended weight: 0.6~0.8; Sampling steps: 30-50 steps &nbsp; &nbsp;Recommended samplers: DPM++ SDE Karras / DPM++ 2M SDE Karras / DPM++ 2M Karras/ Euler/Euler a &nbsp; &nbsp;Suitable for large models: revAnimated, etc. &nbsp; &nbsp;If you want English text, you can input english text, &nbsp; &nbsp;You can choose the color theme you want, for example: blue theme, &nbsp; &nbsp;Use silhouette, keywords to generate silhouettes, &nbsp; &nbsp;Use portrait and backlight keywords to generate character silhouette backlight effects, &nbsp; &nbsp;You can also try using: monochrome, &nbsp; &nbsp;Interestingly, when you input two objects, they may creatively combine two different objects together, and when using silhouette keywords, they may also blend with each other.The first formal training of LoRA, the model was rushed to be released, this version of the model may not be mature enough, looking forward to more creative attempts from everyone and feedback, I will continue to iterate and update. Look forward to everyone's feedback.Enjoy it. V2This version has an excellent overall effect, easier to use, and better compatibility with models.When using anything-v5-PrtRE as the base model, the picture is cleaner.Additionally, the fusion effect of items in this version is more prominent.

Files with SHA256: 4e81222d6ed0421da392cdb16c03f894c1b70f37e7b5482b9accabc371a9d3e9