资源合集：闲话AI官网 talk2ai.cn

本工作流旨在将真人照片转换为独特的漫画风格。

• 默认使用肖像大师自动生成写真照片，用户可自行替换为图像输入节点。
• 使用了WD14反推词对生成的写真反推，可以替换成普通的输入节点。
• 漫画风格使用了自训练的Lora模型：残照金 SDXL - Hoko Manga Style，可以在civitai下载（  https://civitai.com/models/647408/zanshou-kin-sdxl-hoko-manga-style  ）
• 大模型用到了SDXL 1.0 ArienMixXL 亚洲人像和Animagine XL V3.1，也都可以在civitai下载。
• 没有加入高清放大的节点，可以自行加入。

版本 1.1 更新说明

解决了在将工作流导入最新版 ComfyUI 时出现的验证错误。具体错误信息"Validation error: Array must contain at most 2 element(s) at 'nodes[25].pos'"已经得到修复。

This workflow is designed to transform real-life photographs into unique manga-style illustrations.

• By default, the workflow uses comfyui-portrait-master. Users can easily substitute this with image input node.
• You can replace the WD14 node yourself, which is used to backpropagate image keywords
• Custom Manga Style: Utilizes a specially trained Lora model: Zanshou-kin(残照金) SDXL Hoko Manga Style. This model is available for download on civitai ( https://civitai.com/models/647408/zanshou-kin-sdxl-hoko-manga-style ).
• Base Models: SDXL 1.0 ArienMixXL 亚洲人像 and Animagine XL V3.1, both accessible on civitai.
• While not included by default, users can easily integrate high-resolution upscaling nodes.

Version 1.1 Update Notes

Fixed the validation error that occurred when importing the workflow into the latest version of ComfyUI. The specific error message "Validation error: Array must contain at most 2 element(s) at 'nodes[25].pos'" has been resolved.