Oct 28 Update - cloud version

for Users outside Mainland China,visit this link for one-click cloud version:

https://www.runninghub.ai/#/post/1846020690389385217/?utm_source=kol02-RH009

Flux Consistent Realistic Character Pro: Dress on Realistic Clothes (1st edition)

This workflow is a Pro version of Flux Consistent Realistic Characters

https://openart.ai/workflows/coonhound_gracious_99/flux-consistent-realistic-character_pulid_faceswap-version-_pulid/kNyovSvGSnpYFA7VHBmU

What can this do:

Put the real clothes in the real world on the virtual model defined by you, including the front, side, back, specific posture, and the scene you specify

How to use:

Unlike the dressing model based on geometric swarping such as CatVTON, our path needs to be trained with LoRA

Step 1: Take multi-angle clothing photos and train LoRA based on the HunyuanDiT checkpoint (customization required)

Step 2: Load LoRA and corresponding tags in the workflow

Step 3: Customize your model features and scenes, generate female selection 1, male selection 2, and enlarge 1.5 times by default

Step 4: Click Run. Remember to select one of the four in the image chooser node

Soon, you will have your own supermodel for your clothing without real persons

Applicable Clothes:

Pure colors, regular color matching, fewer patterns, fewer texts, large and clear patterns clothes, highly recurrence

Not applicable Clothes:

Irregular color matching, multiple lines of text, many and dense patterns (such as Hanfu), irregular recurrence

Enjoy!

Cloud version available (including 4 sets of clothing LoRA):

https://runninghub.cn/#/post/1843491534351335425/?utm_source=kol01-RH009

Author contact

WeChat: xdtr6688

mail：[email protected]

------------------------------------------------------------------------------------------

FLUX多角度一致性写实人物Pro: 穿上真实的衣服（第一版）

此工作流是多角度一致性写实人物的进阶版

https://openart.ai/workflows/coonhound_gracious_99/flux-consistent-realistic-character_pulid_faceswap-version-_pulid/kNyovSvGSnpYFA7VHBmU

功能：

将物理世界中真实存在的服装，穿着到由你定义的虚拟模特身上，包括正面、侧面、背面、特定姿势，以及你指定的场景

使用方法：

不同于CatVTON等基于几何变形方式的穿衣模型，我们的路径需要通过LoRA训练

Step 1：拍摄多角度服装照片，基于混元模型训练LoRA（需定制）

Step 2：在工作流中加载LoRA以及对应tag

Step 3：自定义你的模特特征和场景，生成女性选择1，男性选择2，默认放大1.5倍

Step 4：点击运行。注意别忘了在image chooser节点四选一

很快，无需真人模特，你的服装有了专属的代言人

适用对象：

纯色，有规律的颜色搭配，较少花纹，较少文字，图案大且清晰的服装，复现程度较高

不适用对象：

无规律的颜色搭配，多行文字，花纹多且密集（如汉服），复现程度无规律

enjoy！

云端版本（包含4套服装LoRA）：

https://runninghub.cn/#/post/1843491534351335425/?utm_source=kol01-RH009

作者微信：xdtr6688