ComfyUI使用InstantID实现根据图片生成相似图片，类似换脸，AI摄影
ComfyUI-InstantID：https://github.com/ZHO-ZHO-ZHO/ComfyUI-InstantID
进入 ComfyUI\custom_nodes
git clone https://github.com/ZHO-ZHO-ZHO/ComfyUI-InstantID.git
cd ComfyUI-InstantID
pip install -r requirements.txt
需要下载的模型
下载 https://huggingface.co/InstantX/InstantID/tree/main/ControlNetModel 中的 config.json 和 diffusion_pytorch_model.safetensors ，将模型填入 ComfyUI/custom_nodes/ComfyUI-InstantID/checkpoints/controlnet
下载 https://huggingface.co/InstantX/InstantID/tree/main 中的 ip-adapter.bin ，将模型填入 ComfyUI/custom_nodes/ComfyUI-InstantID/checkpoints
下载 https://huggingface.co/DIAMONIK7777/antelopev2/tree/main 中的所有模型，将其放入 ComfyUI//custom_nodes/ComfyUI-InstantID/models/antelopev2 中
放置结果如下
如果报错本地没有openai/clip-vit-large-patch14或者CLIP-ViT-bigG-14-laion2B-39B-b160k
需要把它们下载到本地hanggingface的cache内，可以使用huggingface_hub下载，之前我写过 https://www.cnblogs.com/qcy-blog/p/18165717
据我测试，clip-vit-large-patch14需要全部下载，CLIP-ViT-bigG-14-laion2B-39B-b160k只需要下载一部分
地址 https://hf-mirror.com/laion/CLIP-ViT-bigG-14-laion2B-39B-b160k/commits/main
地址 https://hf-mirror.com/openai/clip-vit-large-patch14/tree/main
工作流 https://github.com/ZHO-ZHO-ZHO/ComfyUI-InstantID/blob/main/INSTANTID WORKFLOWS/V2.0 InstantID_locally_pose_ref【Zho】.json
注意这两个要改成绝对路径,就是上面放好的模型路径,基础模型得是SDXL,要用较低的CFG和步数，推荐CFG=3-5，Steps=20-40,这个越大越像
结果
或者使用这个工作流
{
"last_node_id": 33,
"last_link_id": 62,
"nodes": [
{
"id": 30,
"type": "CLIPTextEncode",
"pos": [
772,
221
],
"size": {
"0": 291.9967346191406,
"1": 128.62518310546875
},
"flags": {
"collapsed": false
},
"order": 9,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 59,
"label": "CLIP"
},
{
"name": "text",
"type": "STRING",
"link": 57,
"widget": {
"name": "text"
},
"label": "文本"
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"links": [
61
],
"shape": 3,
"slot_index": 0,
"label": "条件"
}
],
"properties": {
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"1girl,(8k, RAW photo, best quality, masterpiece:1.3),(realistic,photo-realistic:1.37),soft light,(extremely beautiful face:1.1),big eyes,an extremely delicate and beautiful girl,depth of field,blurry background,delicate,beautiful,beautiful face,beautiful eyes,beautiful girl,delicate face,delicate girl,8k wallpaper,(best quality:1.12),(ultra-detailed:1.12),(highres:1.12),hyper detailed,ultra-detailed,high resolution illustration,colorful,8k wallpaper,highres,Cinematic light,ray tracing,(8k, RAW photo, best quality, masterpiece, ultra highres, ultra detailed:1.2),(realistic, photo-realistic:),outdoors,forest,butterfly,green glowing plants,long_sleeves,(blond hair:1.2),(lolita:1.2),dress,lace,spirit,lace,ribbon,hair_ribbon,nature,Dreamy forest,cowboy_shot,flowers meadows,senlin,leaf,fairy,elf ears,"
]
},
{
"id": 31,
"type": "CLIPTextEncode",
"pos": [
770,
409
],
"size": {
"0": 286.3603515625,
"1": 112.35245513916016
},
"flags": {
"collapsed": false
},
"order": 10,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 60,
"label": "CLIP"
},
{
"name": "text",
"type": "STRING",
"link": 58,
"widget": {
"name": "text"
},
"label": "文本"
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"links": [
62
],
"shape": 3,
"slot_index": 0,
"label": "条件"
}
],
"properties": {
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry"
]
},
{
"id": 4,
"type": "CheckpointLoaderSimple",
"pos": [
100,
253
],
"size": {
"0": 480.5506286621094,
"1": 98
},
"flags": {},
"order": 0,
"mode": 0,
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"links": [
38
],
"slot_index": 0,
"label": "模型"
},
{
"name": "CLIP",
"type": "CLIP",
"links": [
59,
60
],
"slot_index": 1,
"label": "CLIP"
},
{
"name": "VAE",
"type": "VAE",
"links": [
3
],
"slot_index": 2,
"label": "VAE"
}
],
"properties": {
"Node name for S&R": "CheckpointLoaderSimple"
},
"widgets_values": [
"XL写实 realvisxlV40_v40LightningBakedvae.safetensors"
]
},
{
"id": 29,
"type": "LoadImage",
"pos": [
1149,
240
],
"size": {
"0": 293.4960021972656,
"1": 392.0699768066406
},
"flags": {},
"order": 1,
"mode": 0,
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
56
],
"shape": 3,
"label": "图像"
},
{
"name": "MASK",
"type": "MASK",
"links": null,
"shape": 3,
"label": "遮罩"
}
],
"properties": {
"Node name for S&R": "LoadImage"
},
"widgets_values": [
"未标222题-1.jpg",
"image"
]
},
{
"id": 14,
"type": "ControlNetLoader",
"pos": [
979,
84
],
"size": {
"0": 409.7744445800781,
"1": 82.64219665527344
},
"flags": {},
"order": 2,
"mode": 0,
"outputs": [
{
"name": "CONTROL_NET",
"type": "CONTROL_NET",
"links": [
37
],
"shape": 3,
"label": "ControlNet",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "ControlNetLoader"
},
"widgets_values": [
"diffusion_pytorch_model.safetensors"
]
},
{
"id": 15,
"type": "InstantIDFaceAnalysis",
"pos": [
1036,
-32
],
"size": {
"0": 315,
"1": 58
},
"flags": {},
"order": 3,
"mode": 0,
"outputs": [
{
"name": "FACEANALYSIS",
"type": "FACEANALYSIS",
"links": [
36
],
"shape": 3,
"label": "FACEANALYSIS",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "InstantIDFaceAnalysis"
},
"widgets_values": [
"CUDA"
]
},
{
"id": 13,
"type": "InstantIDModelLoader",
"pos": [
1041,
-141
],
"size": {
"0": 315,
"1": 58
},
"flags": {},
"order": 4,
"mode": 0,
"outputs": [
{
"name": "INSTANTID",
"type": "INSTANTID",
"links": [
35
],
"shape": 3,
"label": "INSTANTID",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "InstantIDModelLoader"
},
"widgets_values": [
"ip-adapter.bin"
]
},
{
"id": 19,
"type": "ApplyInstantIDAdvanced",
"pos": [
1484,
-46
],
"size": {
"0": 315,
"1": 338
},
"flags": {},
"order": 11,
"mode": 0,
"inputs": [
{
"name": "instantid",
"type": "INSTANTID",
"link": 35,
"label": "instantid"
},
{
"name": "insightface",
"type": "FACEANALYSIS",
"link": 36,
"label": "insightface"
},
{
"name": "control_net",
"type": "CONTROL_NET",
"link": 37,
"label": "control_net"
},
{
"name": "image",
"type": "IMAGE",
"link": 34,
"label": "image",
"slot_index": 3
},
{
"name": "model",
"type": "MODEL",
"link": 38,
"label": "model",
"slot_index": 4
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 61,
"label": "positive"
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 62,
"label": "negative"
},
{
"name": "image_kps",
"type": "IMAGE",
"link": 56,
"label": "image_kps",
"slot_index": 7
},
{
"name": "mask",
"type": "MASK",
"link": null,
"label": "mask"
}
],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"links": [
40
],
"shape": 3,
"label": "MODEL",
"slot_index": 0
},
{
"name": "positive",
"type": "CONDITIONING",
"links": [
41
],
"shape": 3,
"label": "positive",
"slot_index": 1
},
{
"name": "negative",
"type": "CONDITIONING",
"links": [
42
],
"shape": 3,
"label": "negative",
"slot_index": 2
}
],
"properties": {
"Node name for S&R": "ApplyInstantIDAdvanced"
},
"widgets_values": [
0.1,
0.85,
0,
1,
0,
"average"
]
},
{
"id": 24,
"type": "EmptyLatentImage",
"pos": [
1565,
354
],
"size": {
"0": 210,
"1": 106
},
"flags": {},
"order": 5,
"mode": 0,
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
54
],
"shape": 3,
"label": "Latent",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "EmptyLatentImage"
},
"widgets_values": [
1016,
1016,
1
]
},
{
"id": 1,
"type": "KSampler",
"pos": [
1849,
34
],
"size": {
"0": 267.6806945800781,
"1": 534.6439208984375
},
"flags": {},
"order": 12,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 40,
"label": "模型",
"slot_index": 0
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 41,
"label": "正面条件"
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 42,
"label": "负面条件"
},
{
"name": "latent_image",
"type": "LATENT",
"link": 54,
"label": "Latent",
"slot_index": 3
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
1
],
"shape": 3,
"label": "Latent",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "KSampler"
},
"widgets_values": [
898106851306632,
"randomize",
4,
1.5,
"dpmpp_sde",
"karras",
1
]
},
{
"id": 2,
"type": "VAEDecode",
"pos": [
2153,
74
],
"size": {
"0": 210,
"1": 46
},
"flags": {},
"order": 13,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 1,
"label": "Latent"
},
{
"name": "vae",
"type": "VAE",
"link": 3,
"label": "VAE",
"slot_index": 1
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
2
],
"shape": 3,
"label": "图像",
"slot_index": 0
}
],
"properties": {
"Node name for S&R": "VAEDecode"
}
},
{
"id": 3,
"type": "SaveImage",
"pos": [
1532,
723
],
"size": {
"0": 742.4053344726562,
"1": 825.8528442382812
},
"flags": {},
"order": 14,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 2,
"label": "图像"
}
],
"properties": {
"Node name for S&R": "SaveImage"
},
"widgets_values": [
"ComfyUI"
]
},
{
"id": 16,
"type": "LoadImage",
"pos": [
928,
798
],
"size": {
"0": 543.4951782226562,
"1": 643.7646484375
},
"flags": {},
"order": 6,
"mode": 0,
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
34
],
"shape": 3,
"label": "图像",
"slot_index": 0
},
{
"name": "MASK",
"type": "MASK",
"links": null,
"shape": 3,
"label": "遮罩"
}
],
"properties": {
"Node name for S&R": "LoadImage"
},
"widgets_values": [
"未2222标111题-1 (9).jpg",
"image"
]
},
{
"id": 33,
"type": "LoadImage",
"pos": [
2325,
715
],
"size": {
"0": 771.886962890625,
"1": 855.3436889648438
},
"flags": {},
"order": 7,
"mode": 0,
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": null,
"shape": 3,
"label": "图像"
},
{
"name": "MASK",
"type": "MASK",
"links": null,
"shape": 3,
"label": "遮罩"
}
],
"properties": {
"Node name for S&R": "LoadImage"
},
"widgets_values": [
"ComfyUI_01673_.png",
"image"
]
},
{
"id": 32,
"type": "SDXLPromptStyler",
"pos": [
225,
413
],
"size": {
"0": 458.0843811035156,
"1": 476.97576904296875
},
"flags": {},
"order": 8,
"mode": 0,
"outputs": [
{
"name": "text_positive",
"type": "STRING",
"links": [
57
],
"shape": 3,
"label": "正面条件",
"slot_index": 0
},
{
"name": "text_negative",
"type": "STRING",
"links": [
58
],
"shape": 3,
"label": "负面条件",
"slot_index": 1
}
],
"properties": {
"Node name for S&R": "SDXLPromptStyler"
},
"widgets_values": [
"looking_at_viewer,1woman,Asia,thin,portrait,Thirty-five years old",
"lowres, low quality, worst quality:1.2, text:1.2, watermark, painting, drawing, glitch, deformed, mutated, ugly, Letters, Freckles, Wrinkles, Pimples, illustration",
"game-cyberpunk game",
true,
true,
true
],
"color": "#233",
"bgcolor": "#355"
}
],
"links": [
[
1,
1,
0,
2,
0,
"LATENT"
],
[
2,
2,
0,
3,
0,
"IMAGE"
],
[
3,
4,
2,
2,
1,
"VAE"
],
[
34,
16,
0,
19,
3,
"IMAGE"
],
[
35,
13,
0,
19,
0,
"INSTANTID"
],
[
36,
15,
0,
19,
1,
"FACEANALYSIS"
],
[
37,
14,
0,
19,
2,
"CONTROL_NET"
],
[
38,
4,
0,
19,
4,
"MODEL"
],
[
40,
19,
0,
1,
0,
"MODEL"
],
[
41,
19,
1,
1,
1,
"CONDITIONING"
],
[
42,
19,
2,
1,
2,
"CONDITIONING"
],
[
54,
24,
0,
1,
3,
"LATENT"
],
[
56,
29,
0,
19,
7,
"IMAGE"
],
[
57,
32,
0,
30,
1,
"STRING"
],
[
58,
32,
1,
31,
1,
"STRING"
],
[
59,
4,
1,
30,
0,
"CLIP"
],
[
60,
4,
1,
31,
0,
"CLIP"
],
[
61,
30,
0,
19,
5,
"CONDITIONING"
],
[
62,
31,
0,
19,
6,
"CONDITIONING"
]
],
"groups": [],
"config": {},
"extra": {},
"version": 0.4
}
