ComfyPilotComfyPilot
工作流 / LLM / Qwen3.5：文本生成
Qwen3.5：文本生成

Qwen3.5：文本生成

Qwen3.5: Text Generation

使用Qwen3.5模型分析输入图像并生成描述性文本提示词。此工作流执行图像描述和反向提示词工程。

LLM 5 个节点 1 个输入 198 次使用 ComfyUI ≥ 0.19.0 ComfyUI 6 KB

下载工作流 JSON 查看官方来源

下载后把文件拖进 ComfyUI 窗口即可打开。

需要的权重

这个工作流引用了 1 个权重文件，其中 1 个能在本站模型库里找到。

qwen3.5_4b_bf16.safetensors 在模型库查看

节点构成

CLIPLoader × 1
LoadImage × 1
TextGenerate × 1
PreviewAny × 1
MarkdownNote × 1

官方标注的模型

Qwen3.5

工作流 JSON

和上面「下载工作流 JSON」是同一份文件。

1{
2 "id": "a3b7ba33-76af-48e2-906b-211db01b1bb9",
3 "revision": 0,
4 "last_node_id": 6,
5 "last_link_id": 3,
6 "nodes": [
7 {
8 "id": 1,
9 "type": "CLIPLoader",
10 "pos": [
11 -820.0000811524385,
12 3579.99989518941
13 ],
14 "size": [
15 430,
16 150
17 ],
18 "flags": [],
19 "order": 0,
20 "mode": 0,
21 "showAdvanced": false,
22 "inputs": [],
23 "outputs": [
24 {
25 "name": "CLIP",
26 "type": "CLIP",
27 "links": [
28 1
29 ]
30 }
31 ],
32 "properties": {
33 "enableTabs": false,
34 "tabWidth": 65,
35 "tabXOffset": 10,
36 "hasSecondTab": false,
37 "secondTabText": "Send Back",
38 "secondTabOffset": 80,
39 "secondTabWidth": 65,
40 "cnr_id": "comfy-core",
41 "ver": "0.18.1",
42 "Node name for S&R": "CLIPLoader",
43 "models": [
44 {
45 "name": "qwen3.5_4b_bf16.safetensors",
46 "url": "https://huggingface.co/Comfy-Org/Qwen3.5/resolve/main/text_encoders/qwen3.5_4b_bf16.safetensors",
47 "directory": "text_encoders"
48 }
49 ]
50 },
51 "widgets_values": [
52 "qwen3.5_4b_bf16.safetensors",
53 "stable_diffusion",
54 "default"
55 ]
56 },
57 {
58 "id": 2,
59 "type": "LoadImage",
60 "pos": [
61 -810.0000738820444,
62 3819.999892124373
63 ],
64 "size": [
65 420,
66 370
67 ],
68 "flags": [],
69 "order": 1,
70 "mode": 0,
71 "inputs": [],
72 "outputs": [
73 {
74 "name": "IMAGE",
75 "type": "IMAGE",
76 "links": [
77 2
78 ]
79 },
80 {
81 "name": "MASK",
82 "type": "MASK",
83 "links": null
84 }
85 ],
86 "properties": {
87 "enableTabs": false,
88 "tabWidth": 65,
89 "tabXOffset": 10,
90 "hasSecondTab": false,
91 "secondTabText": "Send Back",
92 "secondTabOffset": 80,
93 "secondTabWidth": 65,
94 "cnr_id": "comfy-core",
95 "ver": "0.18.1",
96 "Node name for S&R": "LoadImage"
97 },
98 "widgets_values": [
99 "man_with_red_hat.png",
100 "image"
101 ]
102 },
103 {
104 "id": 3,
105 "type": "TextGenerate",
106 "pos": [
107 -310,
108 3580
109 ],
110 "size": [
111 430,
112 482.65625
113 ],
114 "flags": [],
115 "order": 3,
116 "mode": 0,
117 "inputs": [
118 {
119 "name": "clip",
120 "type": "CLIP",
121 "link": 1
122 },
123 {
124 "name": "image",
125 "shape": 7,
126 "type": "IMAGE",
127 "link": 2
128 }
129 ],
130 "outputs": [
131 {
132 "name": "generated_text",
133 "type": "STRING",
134 "links": [
135 3
136 ]
137 }
138 ],
139 "properties": {
140 "enableTabs": false,
141 "tabWidth": 65,
142 "tabXOffset": 10,
143 "hasSecondTab": false,
144 "secondTabText": "Send Back",
145 "secondTabOffset": 80,
146 "secondTabWidth": 65,
147 "cnr_id": "comfy-core",
148 "ver": "0.18.1",
149 "Node name for S&R": "TextGenerate"
150 },
151 "widgets_values": [
152 "Describe this image in extreme detail, including subject appearance, clothing, pose, expression, background elements, lighting, camera angle, composition, color palette, and overall aesthetic style. Use precise, descriptive language suitable for AI image generation prompts. Do not use vague terms. Only return the prompts itself. Focus on 90s retro tech aesthetic and photorealistic quality.",
153 256,
154 "on",
155 0.7,
156 64,
157 0.95,
158 0.05,
159 1.05,
160 0,
161 0,
162 false
163 ]
164 },
165 {
166 "id": 4,
167 "type": "PreviewAny",
168 "pos": [
169 179.99985675588255,
170 3579.99989518941
171 ],
172 "size": [
173 500,
174 600
175 ],
176 "flags": [],
177 "order": 4,
178 "mode": 0,
179 "inputs": [
180 {
181 "name": "source",
182 "type": "*",
183 "link": 3
184 }
185 ],
186 "outputs": [],
187 "properties": {
188 "enableTabs": false,
189 "tabWidth": 65,
190 "tabXOffset": 10,
191 "hasSecondTab": false,
192 "secondTabText": "Send Back",
193 "secondTabOffset": 80,
194 "secondTabWidth": 65,
195 "cnr_id": "comfy-core",
196 "ver": "0.18.1",
197 "Node name for S&R": "PreviewAny"
198 },
199 "widgets_values": [
200 null,
201 null,
202 null
203 ]
204 },
205 {
206 "id": 6,
207 "type": "MarkdownNote",
208 "pos": [
209 -1420,
210 3580
211 ],
212 "size": [
213 560,
214 680
215 ],
216 "flags": [],
217 "order": 2,
218 "mode": 0,
219 "inputs": [],
220 "outputs": [],
221 "properties": [],
222 "widgets_values": [
223 "Related PR: https://github.com/Comfy-Org/ComfyUI/pull/12771\n## Model Links\n\n- [Hugging Face:Comfy-Org/Qwen3.5](https://huggingface.co/Comfy-Org/Qwen3.5)\n- [ModelScope:Comfy-Org/Qwen3.5](https://modelscope.cn/models/Comfy-Org/Qwen3.5)\n\n**text_encoders**\n\n- [qwen3.5_4b_bf16.safetensors](https://huggingface.co/Comfy-Org/Qwen3.5/resolve/main/text_encoders/qwen3.5_4b_bf16.safetensors) (8.68 GB)\n\n\n## Model Storage Location\n\n```\n📂 ComfyUI/\n├── 📂 models/\n│ └── 📂 text_encoders/\n│ └── qwen3.5_4b_bf16.safetensors\n```\n\n## Report Issue\n\nNote: Please update ComfyUI first ([guide](https://docs.comfy.org/installation/update_comfyui)) and prepare required models. Desktop/Cloud updates follow stable releases, so some nightly-supported models may not be available yet.\n\n- Cannot run / runtime errors: [ComfyUI/issues](https://github.com/comfyanonymous/ComfyUI/issues)\n- UI / frontend issues: [ComfyUI_frontend/issues](https://github.com/Comfy-Org/ComfyUI_frontend/issues)\n- Workflow issues: [workflow_templates/issues](https://github.com/Comfy-Org/workflow_templates/issues)\n"
224 ],
225 "color": "#222",
226 "bgcolor": "#000"
227 }
228 ],
229 "links": [
230 [
231 1,
232 1,
233 0,
234 3,
235 0,
236 "CLIP"
237 ],
238 [
239 2,
240 2,
241 0,
242 3,
243 1,
244 "IMAGE"
245 ],
246 [
247 3,
248 3,
249 0,
250 4,
251 0,
252 "STRING"
253 ]
254 ],
255 "groups": [],
256 "config": [],
257 "extra": {
258 "ds": {
259 "scale": 0.637255983510033,
260 "offset": [
261 1670.5994309815867,
262 -3061.81244596459
263 ]
264 },
265 "frontendVersion": "1.42.8",
266 "VHS_latentpreview": false,
267 "VHS_latentpreviewrate": 0,
268 "VHS_MetadataImage": true,
269 "VHS_KeepIntermediate": true
270 },
271 "version": 0.4
272}