nodes/fblissjr-ComfyUI-QwenImageWanBridge
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
Files
f43ef4cf859cd3689a7b7551b615b556c4fc23fc
fblissjr-ComfyUI-QwenImageW…/example_workflows/qwen_vl_image_edit_example_updated.json
T
Fred Bliss 4a091e9298 Fix coordinate handling and make object_ref optional 
- Fixed generateTokens to properly handle different region types (bounding_box, polygon, object_reference)
- Added checkbox UI to control includeObjectRef for boxes/polygons
- Object references now correctly handle 2 coordinates instead of expecting 4
- Updated Python code to respect includeObjectRef flag from JavaScript
- Fixed coordinate format issues (x1,y1,x2,y2 for boxes, x,y pairs for polygons)
- Restored updateBasePromptWithTokens function for editable prompts
- Default behavior keeps object_ref (checkbox checked) for backward compatibility
- Without object_ref, generates pure spatial tokens like <|box_start|>0.1,0.2,0.3,0.4<|box_end|>
2025-09-06 09:33:32 -05:00

990 lines
17 KiB
JSON
Raw Blame History

{
"id": "qwen-vl-image-edit-example",
"revision": 0,
"last_node_id": 115,
"last_link_id": 219,
"nodes": [
{
"id": 1,
"type": "Note",
"pos": [
-400,
-200
],
"size": [
400,
150
],
"flags": {},
"order": 0,
"mode": 0,
"inputs": [],
"outputs": [],
"properties": {
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
"IMPROVED QWEN2.5-VL WORKFLOW\n\nThis workflow uses the new proper Qwen2.5-VL nodes with real vision processing:\n- QwenVLLoader: Loads actual Qwen2.5-VL model\n- QwenVLTextEncoder: Real multimodal encoding\n- Exact system prompts from DiffSynth-Studio\n\nRequires: pip install transformers"
],
"color": "#223",
"bgcolor": "#335"
},
{
"id": 37,
"type": "UNETLoader",
"pos": [
20,
20
],
"size": [
330,
90
],
"flags": {},
"order": 4,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"slot_index": 0,
"links": [
196
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "UNETLoader",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
"qwen_image_edit_fp8_e4m3fn.safetensors",
"fp8_e4m3fn"
]
},
{
"id": 39,
"type": "VAELoader",
"pos": [
20,
340
],
"size": [
330,
60
],
"flags": {},
"order": 1,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "VAE",
"type": "VAE",
"slot_index": 0,
"links": [
134,
136,
167
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "VAELoader",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
"qwen_image_vae.safetensors"
]
},
{
"id": 66,
"type": "ModelSamplingAuraFlow",
"pos": [
860,
-145
],
"size": [
300,
58
],
"flags": {},
"order": 10,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 217
}
],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"links": [
154
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "ModelSamplingAuraFlow",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
3.5
]
},
{
"id": 73,
"type": "LoraLoaderModelOnly",
"pos": [
395,
50
],
"size": [
270,
82
],
"flags": {},
"order": 8,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 196
}
],
"outputs": [
{
"name": "MODEL",
"type": "MODEL",
"links": [
170,
217
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "LoraLoaderModelOnly",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
"Qwen-Image-Lightning-4steps-V1.0.safetensors",
1
]
},
{
"id": 76,
"type": "LoadImage",
"pos": [
20,
460
],
"size": [
280,
320
],
"flags": {},
"order": 2,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
185,
194,
200
]
},
{
"name": "MASK",
"type": "MASK",
"links": [
201
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "LoadImage",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
"view.png",
"image"
]
},
{
"id": 77,
"type": "VAEEncode",
"pos": [
620,
680
],
"size": [
140,
46
],
"flags": {},
"order": 6,
"mode": 0,
"inputs": [
{
"name": "pixels",
"type": "IMAGE",
"link": 202
},
{
"name": "vae",
"type": "VAE",
"link": 136
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
209
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "VAEEncode",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": []
},
{
"id": 80,
"type": "ConditioningZeroOut",
"pos": [
600,
580
],
"size": [
200,
26
],
"flags": {},
"order": 11,
"mode": 0,
"inputs": [
{
"name": "conditioning",
"type": "CONDITIONING",
"link": 212
}
],
"outputs": [
{
"name": "CONDITIONING",
"type": "CONDITIONING",
"links": [
208
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "ConditioningZeroOut",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": []
},
{
"id": 81,
"type": "CFGNorm",
"pos": [
870,
15
],
"size": [
270,
58
],
"flags": {},
"order": 12,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 154
}
],
"outputs": [
{
"name": "patched_model",
"type": "MODEL",
"links": [
206
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "CFGNorm",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
1
]
},
{
"id": 84,
"type": "VAEDecodeTiled",
"pos": [
1225,
760
],
"size": [
210,
150
],
"flags": {},
"order": 14,
"mode": 0,
"inputs": [
{
"name": "samples",
"type": "LATENT",
"link": 210
},
{
"name": "vae",
"type": "VAE",
"link": 167
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
205
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "VAEDecodeTiled",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
512,
64,
64,
8
]
},
{
"id": 106,
"type": "ImageResizeKJv2",
"pos": [
130,
840
],
"size": [
270,
340
],
"flags": {},
"order": 7,
"mode": 0,
"inputs": [
{
"name": "image",
"type": "IMAGE",
"link": 200
},
{
"name": "mask",
"shape": 7,
"type": "MASK",
"link": 201
}
],
"outputs": [
{
"name": "IMAGE",
"type": "IMAGE",
"links": [
202
]
},
{
"name": "width",
"type": "INT",
"links": null
},
{
"name": "height",
"type": "INT",
"links": null
},
{
"name": "mask",
"type": "MASK",
"links": null
}
],
"properties": {
"cnr_id": "comfyui-kjnodes",
"ver": "1.1.4",
"Node name for S&R": "ImageResizeKJv2",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
1328,
1328,
"lanczos",
"pad",
"0, 0, 0",
"center",
2,
"cpu"
]
},
{
"id": 108,
"type": "SaveImage",
"pos": [
1435,
155
],
"size": [
270,
270
],
"flags": {},
"order": 15,
"mode": 0,
"inputs": [
{
"name": "images",
"type": "IMAGE",
"link": 205
}
],
"outputs": [],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "SaveImage",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
"qwen_vl_edit"
]
},
{
"id": 109,
"type": "KSampler",
"pos": [
874.0758666992188,
114.94629669189453
],
"size": [
300,
556
],
"flags": {},
"order": 13,
"mode": 0,
"inputs": [
{
"name": "model",
"type": "MODEL",
"link": 206
},
{
"name": "positive",
"type": "CONDITIONING",
"link": 213
},
{
"name": "negative",
"type": "CONDITIONING",
"link": 208
},
{
"name": "latent_image",
"type": "LATENT",
"link": 209
}
],
"outputs": [
{
"name": "LATENT",
"type": "LATENT",
"links": [
210
]
}
],
"properties": {
"cnr_id": "comfy-core",
"ver": "0.3.51",
"Node name for S&R": "KSampler",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
668523011998122,
"randomize",
4,
1,
"euler",
"simple",
1
]
},
{
"id": 111,
"type": "QwenVLTextEncoder",
"pos": [
410,
250
],
"size": [
400,
280
],
"flags": {},
"order": 9,
"mode": 0,
"inputs": [
{
"name": "qwen_model",
"type": "QWEN_VL_MODEL",
"link": 219
},
{
"name": "edit_image",
"shape": 7,
"type": "IMAGE",
"link": 194
}
],
"outputs": [
{
"name": "conditioning",
"type": "CONDITIONING",
"links": [
212,
213
]
},
{
"name": "vision_features",
"type": "LATENT",
"links": null
}
],
"title": "Qwen2.5-VL Text Encoder (NEW)",
"properties": {
"aux_id": "fblissjr/ComfyUI-QwenImageWanBridge",
"ver": "2439687525b8950a5d3ed63042f6e5b96adc80b0",
"Node name for S&R": "QwenVLTextEncoder",
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
"Transform the white coffee mug to float 20px above the wooden table's surface while maintaining its size and orientation, preserving all other elements in the kitchen scene.",
"image_edit",
true
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 112,
"type": "Note",
"pos": [
850,
760
],
"size": [
350,
200
],
"flags": {},
"order": 3,
"mode": 0,
"inputs": [],
"outputs": [],
"properties": {
"ue_properties": {
"widget_ue_connectable": {},
"version": "7.0.1"
}
},
"widgets_values": [
"KEY IMPROVEMENTS:\n\n1. QwenVLLoader loads actual Qwen2.5-VL model\n2. QwenVLTextEncoder does real vision processing\n3. Vision tokens actually work now!\n4. Exact system prompts from DiffSynth-Studio\n5. Proper multimodal fusion\n\nThe old CLIPLoader approach is replaced with proper transformers integration."
],
"color": "#232",
"bgcolor": "#353"
},
{
"id": 115,
"type": "QwenVLLoader",
"pos": [
29.912580490112305,
176.76263427734375
],
"size": [
270,
106
],
"flags": {},
"order": 5,
"mode": 0,
"inputs": [],
"outputs": [
{
"name": "model",
"type": "QWEN_VL_MODEL",
"links": [
219
]
}
],
"properties": {
"aux_id": "fblissjr/ComfyUI-QwenImageWanBridge",
"ver": "2439687525b8950a5d3ed63042f6e5b96adc80b0",
"Node name for S&R": "QwenVLLoader",
"ue_properties": {
"widget_ue_connectable": {
"model_name": true,
"precision": true,
"device": true
},
"version": "7.0.1"
}
},
"widgets_values": [
"qwen_2.5_vl_7b.safetensors",
"fp16",
"cuda"
]
}
],
"links": [
[
134,
39,
0,
77,
1,
"VAE"
],
[
136,
39,
0,
77,
1,
"VAE"
],
[
154,
66,
0,
81,
0,
"MODEL"
],
[
162,
88,
0,
86,
0,
"MODEL"
],
[
167,
39,
0,
84,
1,
"VAE"
],
[
169,
86,
0,
87,
0,
"MODEL"
],
[
170,
73,
0,
88,
0,
"MODEL"
],
[
185,
76,
0,
79,
0,
"IMAGE"
],
[
194,
76,
0,
111,
1,
"IMAGE"
],
[
196,
37,
0,
73,
0,
"MODEL"
],
[
200,
76,
0,
106,
0,
"IMAGE"
],
[
201,
76,
1,
106,
1,
"MASK"
],
[
202,
106,
0,
77,
0,
"IMAGE"
],
[
204,
87,
0,
66,
0,
"MODEL"
],
[
205,
84,
0,
108,
0,
"IMAGE"
],
[
206,
81,
0,
109,
0,
"MODEL"
],
[
208,
80,
0,
109,
2,
"CONDITIONING"
],
[
209,
77,
0,
109,
3,
"LATENT"
],
[
210,
109,
0,
84,
0,
"LATENT"
],
[
212,
111,
0,
80,
0,
"CONDITIONING"
],
[
213,
111,
0,
109,
1,
"CONDITIONING"
],
[
217,
73,
0,
66,
0,
"MODEL"
],
[
219,
115,
0,
111,
0,
"QWEN_VL_MODEL"
]
],
"groups": [
{
"id": 1,
"title": "Step 1 - Load Models (IMPROVED)",
"bounding": [
10,
-50,
370,
480
],
"color": "#3f789e",
"font_size": 24,
"flags": {}
},
{
"id": 2,
"title": "Step 2 - Image Preprocessing",
"bounding": [
10,
800,
400,
400
],
"color": "#3f789e",
"font_size": 24,
"flags": {}
},
{
"id": 3,
"title": "Step 3 - Multimodal Encoding (NEW)",
"bounding": [
380,
160,
450,
470
],
"color": "#a1309b",
"font_size": 24,
"flags": {}
},
{
"id": 4,
"title": "Step 4 - Generation Pipeline",
"bounding": [
850,
-400,
600,
1100
],
"color": "#3f789e",
"font_size": 24,
"flags": {}
}
],
"config": {},
"extra": {
"ds": {
"scale": 1.0764329258859686,
"offset": [
698.0757143334406,
279.76412109968516
]
},
"ue_links": [],
"links_added_by_ue": [],
"frontendVersion": "1.25.9",
"VHS_latentpreview": true,
"VHS_latentpreviewrate": 0,
"VHS_MetadataImage": true,
"VHS_KeepIntermediate": true
},
"version": 0.4
}
Reference in New Issue View Git Blame Copy Permalink