迷你世界SDXL-miniature world SDXL

详情

模型描述

介绍

迷你微缩世界的SDXL版本回来了!由于基础模型数量较多,SDXL的复杂度仍需进一步开发。但这是一个通用性极强的模型,只需简单的提示词,就能生成令人惊艳的微缩场景图片,无论是大型建筑还是小型室内场景都能创作。得益于SDXL的高精度,细节得到了极大提升。欢迎各位尝试不同模型、体验效果,并在评论区分享你的作品!

The SDXL version of the miniature world is back! Due to the number of base models, the complexity of SDXL still needs further development. However, it is still a highly versatile model. With simple prompts, you can obtain stunning miniaturized scene images. It can create everything from large buildings to small indoor scenes. Thanks to the precision of SDXL, the details have been greatly enhanced. Everyone is welcome to try various models, experience them, and post your images in the comment section!

使用方法

如果你使用的是WebUI,请尽量使用1.6版本,并适当开启refiner,建议使用:

采样器:Euler A

分辨率:1024*1024

步数:30-40

如需提升清晰度,可开启高清修复,使用低去噪强度修复。

LoRA权重:0.7-0.9 仅支持提示词格式!

建议正面提示词格式:你可以替换我[ ]中的粗体内容

(8k, RAW photo, best quality, masterpiece:1.3), lora:yourloraname:0.75, realistic, miniature, isometric, landscape, mini\(ttp\), (8k, RAW photo, best quality, masterpiece:1.2), [你的主要物体], isometric, miniature, landscape, [你希望添加的其他细节]

负面提示词:请适当调整

(worst quality:1.8), (low quality:1.8), (normal quality:1.8), (EasyNegative:1.8), lowres, blurry, hand, signature, normal quality,

If you're using webui, please try to use version 1.6 and turn on the refiner appropriately. It's recommended to use:

sampler: Euler A

resolution: 1024*1024

step: 30-40

If needed, you can turn on HD restoration, and use low denoising repair.

lora weight: 0.7-0.9 Only supports prompt format!

Suggested prompt format: You can replace the content in my [ ] in bold:

(8k, RAW photo, best quality, masterpiece:1.3), <lora:yourloraname:0.75>, realistic, miniature, isometric, landscape, mini\(ttp\), (8k, RAW photo, best quality, masterpiece:1.2), [your major object], isometric, miniature, landscape, [other detail you want to add]

Negative prompts: Please adjust as necessary:

(worst quality:1.8), (low quality:1.8), (normal quality:1.8), (EasyNegative:1.8), lowres, blurry, hand, signature, normal quality,

祝大家使用愉快,有问题告诉我

此模型生成的图像

未找到图像。