領域指定生成とは？
デザインにおいて、レイアウトは重要な要素です。
単に「街の中で壁に寄りかかっている女性」というプロンプトで画像を生成するだけでは、魅力的な作品は作れません。
画面の右端にレンガの壁があり、そこに寄りかかっている女性、そのすぐ手前に街灯を配置し、画面左はオブジェクトを減らして余白を作り出す……。
好きな場所に好きなオブジェクトを置いた画像を生成するための技術が「領域指定」です。
プロンプトで位置を指示する
最もシンプルな方法は、プロンプトにそのまま位置関係を書く方法です。
{
"id": "18404b37-92b0-4d11-a39c-ae941838eb83",
"revision": 0,
"last_node_id": 45,
"last_link_id": 64,
"nodes": [
{
"id": 35,
"type": "FluxGuidance",
"pos": [
836,
190
],
"size": [
211.60000610351562,
58
],
"flags": {},
"order": 6,
"mode": 0,
"inputs": [
{
"name": "conditioning",
"type": "CONDITIONING",
"link": 56
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
57
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "FluxGuidance"
},
"widgets_values": [
3.5
],
"color": "#2a363b",
"bgcolor": "#3f5159"
},
{
"id": 38,
"type": "PreviewImage",
"pos": [
1568,
190
],
"size": [
430.8328552246094,
446.9476623535156
],
"flags": {},
"order": 9,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 58
}
],
"outputs": [],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "PreviewImage"
},
"widgets_values": []
},
{
"id": 8,
"type": "VAEDecode",
"pos": [
1408,
190
],
"size": [
140,
46
],
"flags": {},
"order": 8,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 52
},
{
"name": "vae",
"type": "VAE",
"link": 62
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"slot_index": 0,
"links": [
58
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "VAEDecode"
},
"widgets_values": []
},
{
"id": 43,
"type": "VAELoader",
"pos": [
1112.7188720703125,
507.60614013671875
],
"size": [
270,
58
],
"flags": {},
"order": 0,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "VAE",
"type": "VAE",
"links": [
62
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "VAELoader"
},
"widgets_values": [
"ae.safetensors"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 33,
"type": "CLIPTextEncode",
"pos": [
518,
378
],
"size": [
414.71820068359375,
108.47611236572266
],
"flags": {
"collapsed": true
},
"order": 5,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 60
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
55
]
}
],
"title": "CLIP Text Encode (Negative Prompt)",
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
""
]
},
{
"id": 27,
"type": "EmptySD3LatentImage",
"pos": [
722,
471
],
"size": [
315,
106
],
"flags": {},
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"slot_index": 0,
"links": [
51
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "EmptySD3LatentImage"
},
"widgets_values": [
1024,
1024,
1
]
},
{
"id": 42,
"type": "DualCLIPLoader",
"pos": [
185.0587921142578,
235.1116485595703
],
"size": [
270,
130
],
"flags": {},
"order": 2,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "CLIP",
"type": "CLIP",
"links": [
59,
60
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "DualCLIPLoader"
},
"widgets_values": [
"clip_l.safetensors",
"t5xxl_fp8_e4m3fn.safetensors",
"flux",
"default"
],
"color": "#432",
"bgcolor": "#653"
},
{
"id": 45,
"type": "UnetLoaderGGUF",
"pos": [
779.269287109375,
59.45874786376953
],
"size": [
270,
58
],
"flags": {},
"order": 3,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"links": [
64
]
}
],
"properties": {
"cnr_id": "ComfyUI-GGUF",
"ver": "b3ec875a68d94b758914fd48d30571d953bb7a54",
"Node name for S&R": "UnetLoaderGGUF"
},
"widgets_values": [
"FLUX_gguf\\flux1-dev-Q5_0.gguf"
]
},
{
"id": 6,
"type": "CLIPTextEncode",
"pos": [
507,
190
],
"size": [
301.84503173828125,
128.01304626464844
],
"flags": {},
"order": 4,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 59
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
56
]
}
],
"title": "CLIP Text Encode (Positive Prompt)",
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"A photo showing a banana on the left and an apple on the right placed on a desk\n"
]
},
{
"id": 31,
"type": "KSampler",
"pos": [
1070,
190
],
"size": [
315,
262
],
"flags": {},
"order": 7,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 64
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 57
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 55
},
{
"name": "latent_image",
"type": "LATENT",
"link": 51
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"slot_index": 0,
"links": [
52
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "KSampler"
},
"widgets_values": [
12345,
"fixed",
20,
1,
"euler",
"normal",
1
]
}
],
"links": [
[
51,
27,
0,
31,
3,
"LATENT"
],
[
52,
31,
0,
8,
0,
"LATENT"
],
[
55,
33,
0,
31,
2,
"CONDITIONING"
],
[
56,
6,
0,
35,
0,
"CONDITIONING"
],
[
57,
35,
0,
31,
1,
"CONDITIONING"
],
[
58,
8,
0,
38,
0,
"IMAGE"
],
[
59,
42,
0,
6,
0,
"CLIP"
],
[
60,
42,
0,
33,
0,
"CLIP"
],
[
62,
43,
0,
8,
1,
"VAE"
],
[
64,
45,
0,
31,
0,
"MODEL"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 0.9090909090909091,
"offset": [
-85.05879211425781,
40.54125213623047
]
},
"frontendVersion": "1.23.4",
"VHS_latentpreview": false,
"VHS_latentpreviewrate": 0,
"VHS_MetadataImage": true,
"VHS_KeepIntermediate": true
},
"version": 0.4
}
「左にバナナ、右にリンゴ」
Stable Diffusionのテキストエンコーダは位置関係をほとんど理解できなかったのですが、Flux以降のモデルでは、ある程度は位置関係を反映してくれるようになってきています。
それでも、複雑な構図になると破綻しやすく、厳密な領域指定というよりは、ゆるいレイアウトの希望を伝える手段です。
Inpaintingを繰り返す
一度画像を生成してから、Inpaintingを何度も繰り返す方法です。
{
"id": "18404b37-92b0-4d11-a39c-ae941838eb83",
"revision": 0,
"last_node_id": 47,
"last_link_id": 72,
"nodes": [
{
"id": 8,
"type": "VAEDecode",
"pos": [
1774.15185546875,
167.77081298828125
],
"size": [
140,
46
],
"flags": {},
"order": 9,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 52
},
{
"name": "vae",
"type": "VAE",
"link": 62
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"slot_index": 0,
"links": [
58
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "VAEDecode"
},
"widgets_values": []
},
{
"id": 42,
"type": "DualCLIPLoader",
"pos": [
185.0587921142578,
235.1116485595703
],
"size": [
270,
130
],
"flags": {},
"order": 0,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "CLIP",
"type": "CLIP",
"links": [
59,
60
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "DualCLIPLoader"
},
"widgets_values": [
"clip_l.safetensors",
"t5xxl_fp8_e4m3fn.safetensors",
"flux",
"default"
],
"color": "#432",
"bgcolor": "#653"
},
{
"id": 35,
"type": "FluxGuidance",
"pos": [
836,
190
],
"size": [
211.60000610351562,
58
],
"flags": {},
"order": 6,
"mode": 0,
"inputs": [
{
"name": "conditioning",
"type": "CONDITIONING",
"link": 56
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
65
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "FluxGuidance"
},
"widgets_values": [
3.5
],
"color": "#2a363b",
"bgcolor": "#3f5159"
},
{
"id": 33,
"type": "CLIPTextEncode",
"pos": [
518,
378
],
"size": [
414.71820068359375,
108.47611236572266
],
"flags": {
"collapsed": true
},
"order": 5,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 60
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
66
]
}
],
"title": "CLIP Text Encode (Negative Prompt)",
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
""
]
},
{
"id": 47,
"type": "InpaintModelConditioning",
"pos": [
1136.67138671875,
187.0927276611328
],
"size": [
270,
138
],
"flags": {},
"order": 7,
"mode": 0,
"inputs": [
{
"name": "positive",
"type": "CONDITIONING",
"link": 65
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 66
},
{
"name": "vae",
"type": "VAE",
"link": 72
},
{
"name": "pixels",
"type": "IMAGE",
"link": 67
},
{
"name": "mask",
"type": "MASK",
"link": 68
}
],
"outputs": [
{
"name": "positive",
"type": "CONDITIONING",
"links": [
69
]
},
{
"name": "negative",
"type": "CONDITIONING",
"links": [
70
]
},
{
"name": "latent",
"type": "LATENT",
"links": [
71
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.43",
"Node name for S&R": "InpaintModelConditioning"
},
"widgets_values": [
true
]
},
{
"id": 43,
"type": "VAELoader",
"pos": [
767.5374755859375,
432.8213806152344
],
"size": [
270,
58
],
"flags": {},
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "VAE",
"type": "VAE",
"links": [
62,
72
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "VAELoader"
},
"widgets_values": [
"ae.safetensors"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 38,
"type": "PreviewImage",
"pos": [
1934.15185546875,
167.77081298828125
],
"size": [
444.1662292480469,
578.058837890625
],
"flags": {},
"order": 10,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 58
}
],
"outputs": [],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "PreviewImage"
},
"widgets_values": []
},
{
"id": 45,
"type": "UnetLoaderGGUF",
"pos": [
1133.6611328125,
68.74612426757812
],
"size": [
270,
58
],
"flags": {},
"order": 2,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"links": [
64
]
}
],
"properties": {
"cnr_id": "ComfyUI-GGUF",
"ver": "b3ec875a68d94b758914fd48d30571d953bb7a54",
"Node name for S&R": "UnetLoaderGGUF"
},
"widgets_values": [
"FLUX_gguf\\flux1-fill-dev-Q4_K_S.gguf"
],
"color": "#323",
"bgcolor": "#535"
},
{
"id": 46,
"type": "LoadImage",
"pos": [
753.7998657226562,
559.425048828125
],
"size": [
296.30230712890625,
481.77783203125
],
"flags": {},
"order": 3,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
67
]
},
{
"name": "MASK",
"type": "MASK",
"links": [
68
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.43",
"Node name for S&R": "LoadImage"
},
"widgets_values": [
"clipspace/clipspace-mask-2027464.8999999762.png [input]",
"image"
]
},
{
"id": 6,
"type": "CLIPTextEncode",
"pos": [
507,
190
],
"size": [
301.84503173828125,
128.01304626464844
],
"flags": {},
"order": 4,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 59
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
56
]
}
],
"title": "CLIP Text Encode (Positive Prompt)",
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"the red cone on top of the cube"
]
},
{
"id": 31,
"type": "KSampler",
"pos": [
1436.15185546875,
167.77081298828125
],
"size": [
315,
262
],
"flags": {},
"order": 8,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 64
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 69
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 70
},
{
"name": "latent_image",
"type": "LATENT",
"link": 71
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"slot_index": 0,
"links": [
52
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.39",
"Node name for S&R": "KSampler"
},
"widgets_values": [
123,
"fixed",
20,
1,
"euler",
"normal",
1
]
}
],
"links": [
[
52,
31,
0,
8,
0,
"LATENT"
],
[
56,
6,
0,
35,
0,
"CONDITIONING"
],
[
58,
8,
0,
38,
0,
"IMAGE"
],
[
59,
42,
0,
6,
0,
"CLIP"
],
[
60,
42,
0,
33,
0,
"CLIP"
],
[
62,
43,
0,
8,
1,
"VAE"
],
[
64,
45,
0,
31,
0,
"MODEL"
],
[
65,
35,
0,
47,
0,
"CONDITIONING"
],
[
66,
33,
0,
47,
1,
"CONDITIONING"
],
[
67,
46,
0,
47,
3,
"IMAGE"
],
[
68,
46,
1,
47,
4,
"MASK"
],
[
69,
47,
0,
31,
1,
"CONDITIONING"
],
[
70,
47,
1,
31,
2,
"CONDITIONING"
],
[
71,
47,
2,
31,
3,
"LATENT"
],
[
72,
43,
0,
47,
2,
"VAE"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 0.620921323059155,
"offset": [
-85.05879211425781,
31.253875732421875
]
},
"frontendVersion": "1.23.4",
"VHS_latentpreview": false,
"VHS_latentpreviewrate": 0,
"VHS_MetadataImage": true,
"VHS_KeepIntermediate": true
},
"version": 0.4
}
-
- ベースとなる画像を生成する
-
- 編集したい領域にマスクをかけてInpaintingする
-
- 必要に応じて、別の領域にもマスクを変えてInpaintingする
あまりスマートではないと思うかもしれませんが、確実で安定した手法です。プロンプトが混ざらず、LoRAもマスク外には基本的に影響しません。各領域を完全に独立したステップとして扱えます。
弱点は、別々に生成するため対象同士の絡み合いができないこと。人物同士が握手している画像などは、目線が合わなかったり違和感が出やすいです。
Conditioning Set Area（Regional Prompting系）
画像の各位置に異なるテキスト条件を適用しようとする手法です。Cross-Attention層を利用して、領域ごとに別のプロンプトを使います。
{
"id": "e524c983-e762-4a7d-a5cb-d0f3a36bde28",
"revision": 0,
"last_node_id": 21,
"last_link_id": 27,
"nodes": [
{
"id": 15,
"type": "ConditioningCombine",
"pos": [
795.6497192382812,
-233.876220703125
],
"size": [
211.060546875,
46
],
"flags": {},
"order": 11,
"mode": 0,
"inputs": [
{
"name": "conditioning_1",
"type": "CONDITIONING",
"link": 18
},
{
"name": "conditioning_2",
"type": "CONDITIONING",
"link": 19
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"links": [
22
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "ConditioningCombine"
},
"widgets_values": [],
"color": "#432",
"bgcolor": "#653"
},
{
"id": 19,
"type": "ConditioningCombine",
"pos": [
1058.4512939453125,
60.54191207885742
],
"size": [
211.060546875,
46
],
"flags": {},
"order": 12,
"mode": 0,
"inputs": [
{
"name": "conditioning_1",
"type": "CONDITIONING",
"link": 22
},
{
"name": "conditioning_2",
"type": "CONDITIONING",
"link": 27
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"links": [
24
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "ConditioningCombine"
},
"widgets_values": [],
"color": "#432",
"bgcolor": "#653"
},
{
"id": 13,
"type": "CLIPTextEncode",
"pos": [
62.78541564941406,
45.40777587890625
],
"size": [
341.8740234375,
152.28765869140625
],
"flags": {},
"order": 7,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 15
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
14
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"RAW photo of an apple"
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 6,
"type": "CLIPTextEncode",
"pos": [
62.78541564941406,
-569.9358520507812
],
"size": [
341.8740234375,
152.28765869140625
],
"flags": {},
"order": 5,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 3
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
11
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"RAW photo of a banana"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 21,
"type": "LoadImageMask",
"pos": [
62.78541564941406,
262.41204833984375
],
"size": [
341.8740234375,
333.62274169921875
],
"flags": {},
"order": 0,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MASK",
"type": "MASK",
"links": [
26
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "LoadImageMask"
},
"widgets_values": [
"apple (2).png",
"red",
"image"
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 20,
"type": "LoadImageMask",
"pos": [
62.78541564941406,
-352.93157958984375
],
"size": [
341.8740234375,
333.62274169921875
],
"flags": {},
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MASK",
"type": "MASK",
"links": [
25
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "LoadImageMask"
},
"widgets_values": [
"banana (2).png",
"red",
"image"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 17,
"type": "CLIPTextEncode",
"pos": [
62.78541564941406,
660.7514038085938
],
"size": [
341.8740234375,
152.28765869140625
],
"flags": {
"collapsed": false
},
"order": 8,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 21
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
27
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"A beautiful analog-style photograph,wood table"
],
"color": "#223",
"bgcolor": "#335"
},
{
"id": 7,
"type": "CLIPTextEncode",
"pos": [
62.78541564941406,
877.7556762695312
],
"size": [
341.8740234375,
104.42913055419922
],
"flags": {},
"order": 6,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 5
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
6
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"text, watermark, low quality"
]
},
{
"id": 8,
"type": "VAEDecode",
"pos": [
1721.425537109375,
180.1886444091797
],
"size": [
177.93228149414062,
46
],
"flags": {},
"order": 14,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 7
},
{
"name": "vae",
"type": "VAE",
"link": 13
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"slot_index": 0,
"links": [
10
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "VAEDecode"
},
"widgets_values": []
},
{
"id": 5,
"type": "EmptyLatentImage",
"pos": [
1020.22705078125,
334.2445068359375
],
"size": [
315,
106
],
"flags": {},
"order": 2,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"slot_index": 0,
"links": [
2
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "EmptyLatentImage"
},
"widgets_values": [
768,
768,
1
]
},
{
"id": 14,
"type": "ConditioningSetMask",
"pos": [
458.8059997558594,
45.40777587890625
],
"size": [
270,
102
],
"flags": {},
"order": 10,
"mode": 0,
"inputs": [
{
"name": "conditioning",
"type": "CONDITIONING",
"link": 14
},
{
"name": "mask",
"type": "MASK",
"link": 26
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"links": [
19
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "ConditioningSetMask"
},
"widgets_values": [
1.5,
"default"
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 11,
"type": "ConditioningSetMask",
"pos": [
465.2082824707031,
-569.9358520507812
],
"size": [
270,
102
],
"flags": {},
"order": 9,
"mode": 0,
"inputs": [
{
"name": "conditioning",
"type": "CONDITIONING",
"link": 11
},
{
"name": "mask",
"type": "MASK",
"link": 25
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"links": [
18
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "ConditioningSetMask"
},
"widgets_values": [
1.5,
"default"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 4,
"type": "CheckpointLoaderSimple",
"pos": [
-419.08935546875,
178.1886444091797
],
"size": [
315,
98
],
"flags": {},
"order": 3,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"slot_index": 0,
"links": [
1
]
},
{
"name": "CLIP",
"type": "CLIP",
"slot_index": 1,
"links": [
3,
5,
15,
21
]
},
{
"name": "VAE",
"type": "VAE",
"slot_index": 2,
"links": []
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CheckpointLoaderSimple",
"models": [
{
"name": "v1-5-pruned-emaonly-fp16.safetensors",
"url": "https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/resolve/main/v1-5-pruned-emaonly-fp16.safetensors?download=true",
"directory": "checkpoints"
}
]
},
"widgets_values": [
"📷-v1.x\\real-dream-15.safetensors"
]
},
{
"id": 12,
"type": "VAELoader",
"pos": [
1723.3863525390625,
289.3052978515625
],
"size": [
270,
58
],
"flags": {
"collapsed": true
},
"order": 4,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "VAE",
"type": "VAE",
"links": [
13
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "VAELoader"
},
"widgets_values": [
"vae-ft-mse-840000-ema-pruned.safetensors"
]
},
{
"id": 3,
"type": "KSampler",
"pos": [
1375.4254150390625,
178.1886444091797
],
"size": [
315,
262
],
"flags": {},
"order": 13,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 1
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 24
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 6
},
{
"name": "latent_image",
"type": "LATENT",
"link": 2
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"slot_index": 0,
"links": [
7
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "KSampler"
},
"widgets_values": [
4444,
"fixed",
20,
8,
"dpmpp_2m",
"karras",
1
]
},
{
"id": 10,
"type": "PreviewImage",
"pos": [
1984.327880859375,
-58.26746368408203
],
"size": [
582.5722045898438,
626.802978515625
],
"flags": {},
"order": 15,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 10
}
],
"outputs": [],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "PreviewImage"
},
"widgets_values": []
}
],
"links": [
[
1,
4,
0,
3,
0,
"MODEL"
],
[
2,
5,
0,
3,
3,
"LATENT"
],
[
3,
4,
1,
6,
0,
"CLIP"
],
[
5,
4,
1,
7,
0,
"CLIP"
],
[
6,
7,
0,
3,
2,
"CONDITIONING"
],
[
7,
3,
0,
8,
0,
"LATENT"
],
[
10,
8,
0,
10,
0,
"IMAGE"
],
[
11,
6,
0,
11,
0,
"CONDITIONING"
],
[
13,
12,
0,
8,
1,
"VAE"
],
[
14,
13,
0,
14,
0,
"CONDITIONING"
],
[
15,
4,
1,
13,
0,
"CLIP"
],
[
18,
11,
0,
15,
0,
"CONDITIONING"
],
[
19,
14,
0,
15,
1,
"CONDITIONING"
],
[
21,
4,
1,
17,
0,
"CLIP"
],
[
22,
15,
0,
19,
0,
"CONDITIONING"
],
[
24,
19,
0,
3,
1,
"CONDITIONING"
],
[
25,
20,
0,
11,
1,
"MASK"
],
[
26,
21,
0,
14,
1,
"MASK"
],
[
27,
17,
0,
19,
1,
"CONDITIONING"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 1.030876132930519,
"offset": [
-181.07522170324756,
348.5181404697794
]
},
"frontendVersion": "1.24.1",
"VHS_latentpreview": false,
"VHS_latentpreviewrate": 0,
"VHS_MetadataImage": true,
"VHS_KeepIntermediate": true
},
"version": 0.4
}
理屈はきれいですが、実際には境界がにじんだり、きっちり切り替わらないことが多く、実用性は高くありません。
また、LoRAを領域指定することはできません。
Latent Composite（潜在空間での合成）
潜在空間の段階で画像を合成する方法です。
{
"id": "e524c983-e762-4a7d-a5cb-d0f3a36bde28",
"revision": 0,
"last_node_id": 47,
"last_link_id": 69,
"nodes": [
{
"id": 6,
"type": "CLIPTextEncode",
"pos": [
-74.14247131347656,
-228.59242248535156
],
"size": [
341.8740234375,
152.28765869140625
],
"flags": {},
"order": 9,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 3
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
37
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"RAW photo of a banana"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 47,
"type": "VAELoader",
"pos": [
1464.0460205078125,
-286.2972106933594
],
"size": [
270,
58
],
"flags": {
"collapsed": true
},
"order": 0,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "VAE",
"type": "VAE",
"links": [
66
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "VAELoader"
},
"widgets_values": [
"vae-ft-mse-840000-ema-pruned.safetensors"
]
},
{
"id": 12,
"type": "VAELoader",
"pos": [
2289.68603515625,
149.98648071289062
],
"size": [
270,
58
],
"flags": {
"collapsed": true
},
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "VAE",
"type": "VAE",
"links": [
13
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "VAELoader"
},
"widgets_values": [
"vae-ft-mse-840000-ema-pruned.safetensors"
]
},
{
"id": 43,
"type": "EmptyLatentImage",
"pos": [
-74.14247131347656,
3.208286762237549
],
"size": [
341.8740234375,
106
],
"flags": {},
"order": 2,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
64
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "EmptyLatentImage"
},
"widgets_values": [
768,
512,
1
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 17,
"type": "CLIPTextEncode",
"pos": [
-74.14247131347656,
-654.4688720703125
],
"size": [
341.8740234375,
152.28765869140625
],
"flags": {
"collapsed": false
},
"order": 12,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 21
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
40
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"A beautiful analog-style photograph,table top"
],
"color": "#223",
"bgcolor": "#335"
},
{
"id": 42,
"type": "EmptyLatentImage",
"pos": [
-74.14247131347656,
-414.10546875
],
"size": [
341.8740234375,
106
],
"flags": {},
"order": 3,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
63
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "EmptyLatentImage"
},
"widgets_values": [
768,
768,
1
],
"color": "#223",
"bgcolor": "#335"
},
{
"id": 13,
"type": "CLIPTextEncode",
"pos": [
247.21710205078125,
380.4800109863281
],
"size": [
341.8740234375,
152.28765869140625
],
"flags": {},
"order": 11,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 15
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
44
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"RAW photo of an apple"
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 44,
"type": "EmptyLatentImage",
"pos": [
247.21710205078125,
612.2808837890625
],
"size": [
341.8740234375,
106
],
"flags": {},
"order": 4,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
65
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "EmptyLatentImage"
},
"widgets_values": [
768,
512,
1
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 38,
"type": "PrimitiveNode",
"pos": [
-74.14247131347656,
188.72134399414062
],
"size": [
341.8740234375,
82
],
"flags": {},
"order": 5,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "INT",
"type": "INT",
"widget": {
"name": "end_at_step"
},
"links": [
46,
47,
48,
50
]
}
],
"title": "end_at_step",
"properties": {
"Run widget replace on values": false
},
"widgets_values": [
10,
"fixed"
]
},
{
"id": 33,
"type": "LatentCompositeMasked",
"pos": [
833.8609008789062,
-348.69537353515625
],
"size": [
270,
146
],
"flags": {},
"order": 17,
"mode": 0,
"inputs": [
{
"name": "destination",
"type": "LATENT",
"link": 52
},
{
"name": "source",
"type": "LATENT",
"link": 53
},
{
"name": "mask",
"shape": 7,
"type": "MASK",
"link": 54
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
55
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "LatentCompositeMasked"
},
"widgets_values": [
0,
0,
true
]
},
{
"id": 40,
"type": "LatentCompositeMasked",
"pos": [
1161.2132568359375,
-135.64053344726562
],
"size": [
270,
146
],
"flags": {},
"order": 18,
"mode": 0,
"inputs": [
{
"name": "destination",
"type": "LATENT",
"link": 55
},
{
"name": "source",
"type": "LATENT",
"link": 57
},
{
"name": "mask",
"shape": 7,
"type": "MASK",
"link": 56
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
58,
69
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "LatentCompositeMasked"
},
"widgets_values": [
0,
0,
true
]
},
{
"id": 8,
"type": "VAEDecode",
"pos": [
2287.659423828125,
38.063751220703125
],
"size": [
177.93228149414062,
46
],
"flags": {},
"order": 21,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 51
},
{
"name": "vae",
"type": "VAE",
"link": 13
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"slot_index": 0,
"links": [
10
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "VAEDecode"
},
"widgets_values": []
},
{
"id": 10,
"type": "PreviewImage",
"pos": [
2515.8095703125,
38.063751220703125
],
"size": [
582.5722045898438,
626.802978515625
],
"flags": {},
"order": 23,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 10
}
],
"outputs": [],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "PreviewImage"
},
"widgets_values": []
},
{
"id": 35,
"type": "LoadImageMask",
"pos": [
837.768310546875,
-145.2086639404297
],
"size": [
214.080078125,
330
],
"flags": {},
"order": 6,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MASK",
"type": "MASK",
"links": [
54
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "LoadImageMask"
},
"widgets_values": [
"banana (2).png",
"red",
"image"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 34,
"type": "LoadImageMask",
"pos": [
1166.2705078125,
73.07683563232422
],
"size": [
214.080078125,
330
],
"flags": {},
"order": 7,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MASK",
"type": "MASK",
"links": [
56
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "LoadImageMask"
},
"widgets_values": [
"apple (2).png",
"red",
"image"
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 45,
"type": "VAEDecode",
"pos": [
1454.0460205078125,
-376.2972106933594
],
"size": [
177.93228149414062,
46
],
"flags": {},
"order": 20,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 69
},
{
"name": "vae",
"type": "VAE",
"link": 66
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"slot_index": 0,
"links": [
67
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "VAEDecode"
},
"widgets_values": []
},
{
"id": 46,
"type": "PreviewImage",
"pos": [
1674.0460205078125,
-376.2972106933594
],
"size": [
210,
258
],
"flags": {},
"order": 22,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 67
}
],
"outputs": [],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "PreviewImage"
},
"widgets_values": []
},
{
"id": 7,
"type": "CLIPTextEncode",
"pos": [
-78.64060974121094,
875.3549194335938
],
"size": [
341.8740234375,
104.42913055419922
],
"flags": {},
"order": 10,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 5
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
38,
41,
45,
61
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"text, watermark, low quality"
]
},
{
"id": 41,
"type": "CLIPTextEncode",
"pos": [
1516.0802001953125,
126.6052017211914
],
"size": [
341.8740234375,
152.28765869140625
],
"flags": {},
"order": 13,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 59
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
60
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"RAW photo of a beautiful analog-style photograph,wood table,an apple,a banan"
]
},
{
"id": 4,
"type": "CheckpointLoaderSimple",
"pos": [
-613.933349609375,
38.063751220703125
],
"size": [
315,
98
],
"flags": {},
"order": 8,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"slot_index": 0,
"links": [
39,
42,
43,
62
]
},
{
"name": "CLIP",
"type": "CLIP",
"slot_index": 1,
"links": [
3,
5,
15,
21,
59
]
},
{
"name": "VAE",
"type": "VAE",
"slot_index": 2,
"links": []
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CheckpointLoaderSimple",
"models": [
{
"name": "v1-5-pruned-emaonly-fp16.safetensors",
"url": "https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/resolve/main/v1-5-pruned-emaonly-fp16.safetensors?download=true",
"directory": "checkpoints"
}
]
},
"widgets_values": [
"📷-v1.x\\real-dream-15.safetensors"
]
},
{
"id": 36,
"type": "KSamplerAdvanced",
"pos": [
438.809326171875,
-654.4688720703125
],
"size": [
304.748046875,
334
],
"flags": {},
"order": 16,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 42
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 40
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 41
},
{
"name": "latent_image",
"type": "LATENT",
"link": 63
},
{
"name": "end_at_step",
"type": "INT",
"widget": {
"name": "end_at_step"
},
"link": 48
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
52
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "KSamplerAdvanced"
},
"widgets_values": [
"enable",
777,
"fixed",
20,
8,
"dpmpp_2m",
"karras",
0,
10,
"enable"
],
"color": "#223",
"bgcolor": "#335"
},
{
"id": 32,
"type": "KSamplerAdvanced",
"pos": [
438.809326171875,
-228.59242248535156
],
"size": [
304.748046875,
334
],
"flags": {},
"order": 14,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 39
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 37
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 38
},
{
"name": "latent_image",
"type": "LATENT",
"link": 64
},
{
"name": "end_at_step",
"type": "INT",
"widget": {
"name": "end_at_step"
},
"link": 46
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
53
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "KSamplerAdvanced"
},
"widgets_values": [
"enable",
777,
"fixed",
20,
8,
"dpmpp_2m",
"karras",
0,
10,
"enable"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 37,
"type": "KSamplerAdvanced",
"pos": [
760.1702270507812,
294.8553466796875
],
"size": [
304.748046875,
334
],
"flags": {},
"order": 15,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 43
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 44
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 45
},
{
"name": "latent_image",
"type": "LATENT",
"link": 65
},
{
"name": "end_at_step",
"type": "INT",
"widget": {
"name": "end_at_step"
},
"link": 47
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
57
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "KSamplerAdvanced"
},
"widgets_values": [
"enable",
777,
"fixed",
20,
8,
"dpmpp_2m",
"karras",
0,
10,
"enable"
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 39,
"type": "KSamplerAdvanced",
"pos": [
1953.4892578125,
38.063751220703125
],
"size": [
304.748046875,
334
],
"flags": {},
"order": 19,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 62
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 60
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 61
},
{
"name": "latent_image",
"type": "LATENT",
"link": 58
},
{
"name": "start_at_step",
"type": "INT",
"widget": {
"name": "start_at_step"
},
"link": 50
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
51
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "KSamplerAdvanced"
},
"widgets_values": [
"disable",
777,
"fixed",
20,
8,
"dpmpp_2m",
"karras",
10,
20,
"disable"
]
}
],
"links": [
[
3,
4,
1,
6,
0,
"CLIP"
],
[
5,
4,
1,
7,
0,
"CLIP"
],
[
10,
8,
0,
10,
0,
"IMAGE"
],
[
13,
12,
0,
8,
1,
"VAE"
],
[
15,
4,
1,
13,
0,
"CLIP"
],
[
21,
4,
1,
17,
0,
"CLIP"
],
[
37,
6,
0,
32,
1,
"CONDITIONING"
],
[
38,
7,
0,
32,
2,
"CONDITIONING"
],
[
39,
4,
0,
32,
0,
"MODEL"
],
[
40,
17,
0,
36,
1,
"CONDITIONING"
],
[
41,
7,
0,
36,
2,
"CONDITIONING"
],
[
42,
4,
0,
36,
0,
"MODEL"
],
[
43,
4,
0,
37,
0,
"MODEL"
],
[
44,
13,
0,
37,
1,
"CONDITIONING"
],
[
45,
7,
0,
37,
2,
"CONDITIONING"
],
[
46,
38,
0,
32,
4,
"INT"
],
[
47,
38,
0,
37,
4,
"INT"
],
[
48,
38,
0,
36,
4,
"INT"
],
[
50,
38,
0,
39,
4,
"INT"
],
[
51,
39,
0,
8,
0,
"LATENT"
],
[
52,
36,
0,
33,
0,
"LATENT"
],
[
53,
32,
0,
33,
1,
"LATENT"
],
[
54,
35,
0,
33,
2,
"MASK"
],
[
55,
33,
0,
40,
0,
"LATENT"
],
[
56,
34,
0,
40,
2,
"MASK"
],
[
57,
37,
0,
40,
1,
"LATENT"
],
[
58,
40,
0,
39,
3,
"LATENT"
],
[
59,
4,
1,
41,
0,
"CLIP"
],
[
60,
41,
0,
39,
1,
"CONDITIONING"
],
[
61,
7,
0,
39,
2,
"CONDITIONING"
],
[
62,
4,
0,
39,
0,
"MODEL"
],
[
63,
42,
0,
36,
3,
"LATENT"
],
[
64,
43,
0,
32,
3,
"LATENT"
],
[
65,
44,
0,
37,
3,
"LATENT"
],
[
66,
47,
0,
45,
1,
"VAE"
],
[
67,
45,
0,
46,
0,
"IMAGE"
],
[
69,
40,
0,
45,
0,
"LATENT"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 0.5290024565035728,
"offset": [
713.933349609375,
754.4688720703125
]
},
"frontendVersion": "1.24.1",
"VHS_latentpreview": false,
"VHS_latentpreviewrate": 0,
"VHS_MetadataImage": true,
"VHS_KeepIntermediate": true
},
"version": 0.4
}
-
- まず複数の画像（バナナ画像、リンゴ画像、背景画像など）を生成する
-
- それぞれのlatentを、マスクを使って1枚分のlatentに貼り合わせる
-
- その合成されたlatentに対して、残りのサンプリングステップを回す
各オブジェクトを別々の条件で生成でき、最後に「一枚の画像」としてなじませられます。
ただし、これを使うくらいなら複数回Inpaintingして、最後に全体をimage2imageしたほうが確実な場面も多いです。
Latent Couple / Attention Couple
Latent Couple
潜在空間を領域ごとに完全に分割し、それぞれ別々の設定（プロンプト・LoRAなど）で生成してから結合する方法です。
- 各領域にまったく別の設定を使えるという点では理想的
- 領域の数だけ画像を生成しているのと同じなので、計算量が大きい
現時点でComfyUIに直接の実装はありません
Attention Couple
Lantent CoupleではUNet丸々計算していましたが、こちらはCross-Attention層のみ計算します
その分計算量は大分少なくなりますが、LoRAの領域別指定はできません
{
"id": "bd31dc25-50da-43d7-b1cb-e53f5d146f11",
"revision": 0,
"last_node_id": 19,
"last_link_id": 25,
"nodes": [
{
"id": 8,
"type": "VAEDecode",
"pos": [
1573.7244873046875,
560.4592895507812
],
"size": [
150.13829040527344,
46
],
"flags": {},
"order": 14,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 7
},
{
"name": "vae",
"type": "VAE",
"link": 25
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"slot_index": 0,
"links": [
10
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "VAEDecode"
},
"widgets_values": []
},
{
"id": 6,
"type": "CLIPTextEncode",
"pos": [
466.21844482421875,
52.61830520629883
],
"size": [
306.9939880371094,
121.5467300415039
],
"flags": {},
"order": 6,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 3
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
4,
13
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"A beautiful analog-style photograph,table top"
],
"color": "#223",
"bgcolor": "#335"
},
{
"id": 12,
"type": "SolidMask",
"pos": [
466.21844482421875,
234.42247009277344
],
"size": [
306.9939880371094,
106
],
"flags": {},
"order": 11,
"mode": 0,
"inputs": [
{
"name": "width",
"type": "INT",
"widget": {
"name": "width"
},
"link": 16
},
{
"name": "height",
"type": "INT",
"widget": {
"name": "height"
},
"link": 18
}
],
"outputs": [
{
"name": "MASK",
"type": "MASK",
"links": [
14
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "SolidMask"
},
"widgets_values": [
1,
768,
768
],
"color": "#223",
"bgcolor": "#335"
},
{
"id": 17,
"type": "LoadImageMask",
"pos": [
466.21844482421875,
582.4840698242188
],
"size": [
306.9939880371094,
330
],
"flags": {},
"order": 0,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MASK",
"type": "MASK",
"links": [
21
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "LoadImageMask"
},
"widgets_values": [
"banana (2).png",
"red",
"image"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 15,
"type": "CLIPTextEncode",
"pos": [
466.21844482421875,
400.6799011230469
],
"size": [
306.9939880371094,
121.5467300415039
],
"flags": {},
"order": 8,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 19
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
20
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"RAW photo of a banana"
],
"color": "#322",
"bgcolor": "#533"
},
{
"id": 18,
"type": "CLIPTextEncode",
"pos": [
466.21844482421875,
972.7415161132812
],
"size": [
306.9939880371094,
121.5467300415039
],
"flags": {},
"order": 9,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 22
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
23
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"RAW photo of an apple"
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 14,
"type": "PrimitiveNode",
"pos": [
-6.389673233032227,
859.0269165039062
],
"size": [
210,
82
],
"flags": {},
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "INT",
"type": "INT",
"widget": {
"name": "height"
},
"links": [
17,
18
]
}
],
"title": "height",
"properties": {
"Run widget replace on values": false
},
"widgets_values": [
768,
"fixed"
]
},
{
"id": 16,
"type": "LoadImageMask",
"pos": [
466.21844482421875,
1154.545654296875
],
"size": [
306.9939880371094,
330
],
"flags": {},
"order": 2,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MASK",
"type": "MASK",
"links": [
24
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "LoadImageMask"
},
"widgets_values": [
"apple (2).png",
"red",
"image"
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 7,
"type": "CLIPTextEncode",
"pos": [
466.21844482421875,
1544.8031005859375
],
"size": [
306.9939880371094,
89.91178131103516
],
"flags": {},
"order": 7,
"mode": 0,
"inputs": [
{
"name": "clip",
"type": "CLIP",
"link": 5
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"slot_index": 0,
"links": [
6
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CLIPTextEncode"
},
"widgets_values": [
"ow qualitytext, watermark, "
]
},
{
"id": 10,
"type": "AttentionCouplePPM",
"pos": [
986.5342407226562,
560.4592895507812
],
"size": [
186.96835327148438,
186
],
"flags": {},
"order": 12,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 12
},
{
"name": "base_cond",
"type": "CONDITIONING",
"link": 13
},
{
"name": "base_mask",
"type": "MASK",
"link": 14
},
{
"name": "cond_1",
"type": "CONDITIONING",
"link": 20
},
{
"name": "mask_1",
"type": "MASK",
"link": 21
},
{
"name": "cond_2",
"type": "CONDITIONING",
"link": 23
},
{
"name": "mask_2",
"type": "MASK",
"link": 24
},
{
"name": "cond_3",
"type": "CONDITIONING",
"link": null
},
{
"name": "mask_3",
"type": "MASK",
"link": null
}
],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"links": [
11
]
}
],
"properties": {
"cnr_id": "comfyui-ppm",
"ver": "7a28475ba26ebb76687c6cc343437b6deff36232",
"Node name for S&R": "AttentionCouplePPM"
},
"widgets_values": [],
"color": "#432",
"bgcolor": "#653"
},
{
"id": 11,
"type": "PreviewImage",
"pos": [
1766.4737548828125,
560.4592895507812
],
"size": [
529.1884155273438,
593.6753540039062
],
"flags": {},
"order": 15,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 10
}
],
"outputs": [],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "PreviewImage"
},
"widgets_values": []
},
{
"id": 19,
"type": "VAELoader",
"pos": [
1577.5760498046875,
560.4592895507812
],
"size": [
268.0599060058594,
58.97004699707031
],
"flags": {
"collapsed": true
},
"order": 3,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "VAE",
"type": "VAE",
"links": [
25
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "VAELoader"
},
"widgets_values": [
"sdxl_vae.safetensors"
]
},
{
"id": 4,
"type": "CheckpointLoaderSimple",
"pos": [
-109.82740020751953,
560.4592895507812
],
"size": [
315,
98
],
"flags": {},
"order": 4,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"slot_index": 0,
"links": [
12
]
},
{
"name": "CLIP",
"type": "CLIP",
"slot_index": 1,
"links": [
3,
5,
19,
22
]
},
{
"name": "VAE",
"type": "VAE",
"slot_index": 2,
"links": []
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "CheckpointLoaderSimple",
"models": [
{
"name": "v1-5-pruned-emaonly-fp16.safetensors",
"url": "https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/resolve/main/v1-5-pruned-emaonly-fp16.safetensors?download=true",
"directory": "checkpoints"
}
]
},
"widgets_values": [
"📷-XL\\RealVisXL_V4.0.safetensors"
]
},
{
"id": 3,
"type": "KSampler",
"pos": [
1216.113525390625,
560.4592895507812
],
"size": [
315,
262
],
"flags": {},
"order": 13,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 11
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 4
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 6
},
{
"name": "latent_image",
"type": "LATENT",
"link": 2
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"slot_index": 0,
"links": [
7
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "KSampler"
},
"widgets_values": [
13022,
"fixed",
20,
8,
"dpmpp_2m",
"karras",
1
]
},
{
"id": 5,
"type": "EmptyLatentImage",
"pos": [
946.1537475585938,
846.269287109375
],
"size": [
227.35647583007812,
106
],
"flags": {},
"order": 10,
"mode": 0,
"inputs": [
{
"name": "width",
"type": "INT",
"widget": {
"name": "width"
},
"link": 15
},
{
"name": "height",
"type": "INT",
"widget": {
"name": "height"
},
"link": 17
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"slot_index": 0,
"links": [
2
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.44",
"Node name for S&R": "EmptyLatentImage"
},
"widgets_values": [
768,
768,
1
]
},
{
"id": 13,
"type": "PrimitiveNode",
"pos": [
-6.389673233032227,
729.040283203125
],
"size": [
210,
82
],
"flags": {},
"order": 5,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "INT",
"type": "INT",
"widget": {
"name": "width"
},
"links": [
15,
16
]
}
],
"title": "width",
"properties": {
"Run widget replace on values": false
},
"widgets_values": [
768,
"fixed"
]
}
],
"links": [
[
2,
5,
0,
3,
3,
"LATENT"
],
[
3,
4,
1,
6,
0,
"CLIP"
],
[
4,
6,
0,
3,
1,
"CONDITIONING"
],
[
5,
4,
1,
7,
0,
"CLIP"
],
[
6,
7,
0,
3,
2,
"CONDITIONING"
],
[
7,
3,
0,
8,
0,
"LATENT"
],
[
10,
8,
0,
11,
0,
"IMAGE"
],
[
11,
10,
0,
3,
0,
"MODEL"
],
[
12,
4,
0,
10,
0,
"MODEL"
],
[
13,
6,
0,
10,
1,
"CONDITIONING"
],
[
14,
12,
0,
10,
2,
"MASK"
],
[
15,
13,
0,
5,
0,
"INT"
],
[
16,
13,
0,
12,
0,
"INT"
],
[
17,
14,
0,
5,
1,
"INT"
],
[
18,
14,
0,
12,
1,
"INT"
],
[
19,
4,
1,
15,
0,
"CLIP"
],
[
20,
15,
0,
10,
3,
"CONDITIONING"
],
[
21,
17,
0,
10,
4,
"MASK"
],
[
22,
4,
1,
18,
0,
"CLIP"
],
[
23,
18,
0,
10,
5,
"CONDITIONING"
],
[
24,
16,
0,
10,
6,
"MASK"
],
[
25,
19,
0,
8,
1,
"VAE"
]
],
"groups": [],
"config": {},
"extra": {
"ds": {
"scale": 1.0308761329305212,
"offset": [
-632.0231183731564,
-69.89356683867605
]
},
"frontendVersion": "1.24.1",
"VHS_latentpreview": false,
"VHS_latentpreviewrate": 0,
"VHS_MetadataImage": true,
"VHS_KeepIntermediate": true
},
"version": 0.4
}
雑コラのリファイン (おすすめ)
雑なコラージュ画像を作って、それをもとに自然な絵になるように作り直させる方法です。
非常に直感的に位置を指定できますし、生成する物自体も適当なオブジェクトを貼り付けておけばよいので、実際のところかなりおすすめの方法です。
詳しくは → 雑コラのリファイン