- 版本 [version]
- 下载 167
- 文件大小 35755
- 文件计数 1
- 创建日期 2026年8月5日
- 最后更新 2026年9月21日
face swap V2
这是一份开箱即用的换脸换装工作流(SDXL、IPAdapter、ControlNet),拖入 ComfyUI 即可运行,全图 39 节点 / 44 连线。核心流程 CLIPTextEncode ×2、PreviewImage ×2、LoadImage ×2,输出图像,参数已调好,替换输入图即可复现效果。主模型 diffusion_pytorch_model.safetensors、sdxl_vae_fp16fix.safetensors、ip-adapter.bin。下载后直接导入即可使用,建议显存 8GB 起。
- 主模型:diffusion_pytorch_model.safetensors、sdxl_vae_fp16fix.safetensors、ip-adapter.bin
- 核心节点:CLIPTextEncode ×2、PreviewImage ×2、LoadImage ×2
- 输出形式:输出图像
- 规模:39 节点 / 44 连线
- 建议显存:8GB 起
- 适用场景:人物换脸、虚拟试衣、商品换装与背景替换,电商模特图的主力玩法。
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": 572,
"last_link_id": 1054,
"nodes": [
{
"id": 7,
"type": "ControlNetLoader",
"pos": [
3110,
420
],
"size": [
451.42926025390625,
62.14959716796875
],
"flags": [],
"order": 0,
"mode": 0,
"outputs": [
{
"name": "CONTROL_NET",
"type": "CONTROL_NET",
"links": [
6
],
"shape": 3,
"label": "ControlNet"
}
],
"properties": {
"Node name for S&R": "ControlNetLoader"
},
"widgets_values": [
"diffusion_pytorch_model.safetensors"
]
},
{
"id": 13,
"type": "CLIPTextEncode",
"pos": [
3170,
630
],
"size": [
393.1996154785156,
101.7002944946289
],
"flags": [],
"order": 20,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 300,
"label": "CLIP"
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"links": [
12
],
"shape": 3,
"label": "条件",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"1girl,face-to-face, nbest quality, masterpiece,"
]
},
{
"id": 296,
"type": "PreviewImage",
"pos": [
6036,
457
],
"size": [
210,
246
],
"flags": [],
"order": 37,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 497,
"label": "图像"
}
],
"properties": {
"Node name for S&R": "PreviewImage"
}
},
{
"id": 295,
"type": "VAEDecode",
"pos": [
5779,
461
],
"size": [
210,
46
],
"flags": [],
"order": 36,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 491,
"label": "Latent"
},
{
"name": "vae",
"type": "VAE",
"link": 492,
"label": "VAE"
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
497
],
"shape": 3,
"label": "图像",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "VAEDecode"
}
},
{
… (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。



