• 版本 [version]
• 下载 68
• 文件大小 6690
• 文件计数 1
• 创建日期 2026年4月16日
• 最后更新 2026年9月21日

局部重绘

局部重绘 是一套面向局部重绘场景的 ComfyUI 工作流，全图共 11 个节点、15 条连线。链路以 CLIPTextEncode ×2、KSampler ×1、InpaintModelConditioning ×1 为主干，输出图像，适合需要稳定批量出图的同学。主模型 SD1.5majicmixRealistic_v7.safetensors、SD1.5vaeFtMse840000Ema_v10.safetensors。下载后直接导入即可使用，建议显存 4GB 起。

• 主模型：SD1.5majicmixRealistic_v7.safetensors、SD1.5vaeFtMse840000Ema_v10.safetensors
• 核心节点：CLIPTextEncode ×2、KSampler ×1、InpaintModelConditioning ×1
• 输出形式：输出图像
• 规模：11 节点 / 15 连线
• 建议显存：4GB 起
• 适用场景：蒙版局部重绘、主体替换、材质替换与瑕疵修复。

工作流节点图

图较大可横向滚动；点击图片查看高清矢量原图（浏览器中可自由缩放）。

资源头图

使用说明

下载工作流文件后在 ComfyUI 中直接拖入画布（或菜单 Load 导入）即可运行；缺失的自定义节点用 ComfyUI Manager 一键安装，模型按文件名放入对应的 models 目录。

源码预览

{
    "last_node_id": 44,
    "last_link_id": 76,
    "nodes": [
        {
            "id": 26,
            "type": "KSampler",
            "pos": [
                40,
                1830
            ],
            "size": [
                315,
                262
            ],
            "flags": [],
            "order": 8,
            "mode": 0,
            "inputs": [
                {
                    "name": "model",
                    "type": "MODEL",
                    "link": 73,
                    "label": "model",
                    "localized_name": "模型"
                },
                {
                    "name": "positive",
                    "type": "CONDITIONING",
                    "link": 51,
                    "label": "positive",
                    "localized_name": "正面条件"
                },
                {
                    "name": "negative",
                    "type": "CONDITIONING",
                    "link": 52,
                    "label": "negative",
                    "localized_name": "负面条件"
                },
                {
                    "name": "latent_image",
                    "type": "LATENT",
                    "link": 75,
                    "slot_index": 3,
                    "label": "latent_image",
                    "localized_name": "Latent图像"
                }
            ],
            "outputs": [
                {
                    "name": "LATENT",
                    "type": "LATENT",
                    "links": [
                        55
                    ],
                    "slot_index": 0,
                    "label": "LATENT",
                    "localized_name": "Latent"
                }
            ],
            "properties": {
                "Node name for S&R": "KSampler"
            },
            "widgets_values": [
                677064267227010,
                "randomize",
                20,
                7.5,
                "dpmpp_2m",
                "karras",
                0.65
            ]
        },
        {
            "id": 25,
            "type": "InpaintModelConditioning",
            "pos": [
                -200,
                2540
            ],
            "size": [
                216.59999084472656,
                138
            ],
            "flags": [],
            "order": 6,
            "mode": 0,
            "inputs": [
                {
                    "name": "positive",
                    "type": "CONDITIONING",
                    "link": 46,
                    "label": "positive",
                    "localized_name": "正面条件"
                },
                {
                    "name": "negative",
                    "type": "CONDITIONING",
                    "link": 47,
                    "label": "negative",
                    "localized_name": "负面条件"
                },
                {
                    "name": "vae",
                    "type": "VAE",
                    "link": 69,
                    "label": "vae",
                    "localized_name": "vae"
                },
                {
                    "name": "pixels",
                    "type": "IMAGE",
                    "link": 50,
                    "slot_index": 3,
                    "label": "pixels",
                    "localized_name": "像素"
                },
                {
                    "name": "mask",
                    "type": "MASK",
                    "link": 72,
                    "label": "mask",
                    "localized_name": "遮罩"
                }
            ],
            "outputs": [
                {
                    "name": "positive",
                    "type": "CONDITIONING",
                    "links": [
                        51
            
… (完整内容请下载文件查看)

代码较长，以上仅前 3800 字符 —— 点击下方按钮下载完整 JSON。