wikeeyang commited on
Commit
a62265e
·
verified ·
1 Parent(s): 651b642

Upload 3 files

Browse files
Files changed (4) hide show
  1. .gitattributes +1 -0
  2. README.md +45 -0
  3. v1-example-workflow.json +900 -0
  4. v1-sample.jpg +3 -0
.gitattributes CHANGED
@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
  Ernie-Image-Flex-V1-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text
 
 
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
  Ernie-Image-Flex-V1-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text
37
+ v1-sample.jpg filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ language:
4
+ - en
5
+ - zh
6
+ library_name: diffusers
7
+ pipeline_tag: text-to-image
8
+ tasks:
9
+ - text-to-image-synthesis
10
+ base_model:
11
+ - baidu/ERNIE-Image
12
+ frameworks: PyTorch
13
+ base_model_relation: finetune
14
+ ---
15
+ # Ernie-Image-Flex V1.0
16
+
17
+ 本模型为百度 [ERNIE-Image](https://huggingface.co/baidu/ERNIE-Image) 的微调模型,相比 [ERNIE-Image-Turbo](https://huggingface.co/baidu/ERNIE-Image-Turbo) 在以下几方面有所改善和提升 (The model is a fine-tuned version of Baidu's [ERNIE-Image](https://huggingface.co/baidu/ERNIE-Image), with improvements and enhancements over [ERNIE-Image-Turbo](https://huggingface.co/baidu/ERNIE-Image-Turbo) in the following aspects):
18
+
19
+ 1. 基本去除 [ERNIE-Image-Turbo](https://huggingface.co/baidu/ERNIE-Image-Turbo) 8 步模型油腻较重、噪点过强、颜色偏黄的问题。(Essentially eliminates the issues of heavy oiliness, excessive noise, and yellowish color found in the 8-step [ERNIE-Image-Turbo](https://huggingface.co/baidu/ERNIE-Image-Turbo) model.)
20
+ 2. 相比 [ERNIE-Image](https://huggingface.co/baidu/ERNIE-Image) 需要 cfg=4.0 / steps=50 出图很慢,本模型只需 cfg=1.0 / steps=16 - 25,且可使用 Turbo LoRA 加速至 step=8,有更好的弹性尺度。(Compared to [ERNIE-Image](https://huggingface.co/baidu/ERNIE-Image), which requires cfg=4.0 / steps=50 and is very slow to generate images, this model only needs cfg=1.0 / steps=16-25, and can be accelerated to step=8 using Turbo LoRA, offering better scalability.)
21
+ 3. 更好的人像构图与人脸优化,画面更干净,复杂构图时肢体相比 [ERNIE-Image-Turbo](https://huggingface.co/baidu/ERNIE-Image-Turbo) 不容易崩。(Improved portrait composition and face optimization, resulting in cleaner images, and limbs are less likely to deform in complex compositions compared to [ERNIE-Image-Turbo](https://huggingface.co/baidu/ERNIE-Image-Turbo).)
22
+ 4. 中英文文字渲染能力得到较好延续,比官方原版略有损耗。(Continues to render Chinese and English text well, with slightly less accuracy than the official original version.)
23
+ 5. 量化模型采用全新的分层/分块/混合精度量化机制,使 8bit 模型与 bf16 模型效果更接近。(The quantized model uses a new hierarchical / block / mixed-precision quantization mechanism, making the 8bit model closer in effect to the bf16 model.)
24
+
25
+ 本模型依旧为纯净底模,可用于微调或 LoRA 训练,也可叠加各种 Turbo / Distill / NSFW / SFW 的 LoRA 使用。(Pure base model, suitable for fine-tuning or LoRA training, and can also be combined with various Turbo / Distill / SFW / NSFW LoRAs.)
26
+
27
+ 感谢:**@亮亮rayne**,[@GuangyuanSD](https://huggingface.co/GuangyuanSD) 在本模型调制优化过程中给与技术、构图、美学等方面的指导和支持。
28
+
29
+ Much thanks to: **@亮亮rayne**, [@GuangyuanSD](https://huggingface.co/GuangyuanSD) for their guidance and support in technical, compositional, and aesthetic aspects during the finetuning of the model.
30
+
31
+ 感谢:[@T8star-Aix](https://space.bilibili.com/385085361),[@AI-KSK](https://space.bilibili.com/110353151),[@黑鹤001](https://space.bilibili.com/515231056),[@大麦茶看AI](https://space.bilibili.com/310599757),[@joeMultimedia](https://space.bilibili.com/612740284) , [山竹老师AIGC](https://space.bilibili.com/158424637) , [CommintAI](https://space.bilibili.com/158429672) , [雪煮松风](https://space.bilibili.com/692784900) , [冰缘排骨](https://space.bilibili.com/3546582953757144) , [活力网TV](https://space.bilibili.com/92055552) , [HyLan_L](https://space.bilibili.com/591877893) , 等一众大佬对本人微调各种模型的技术评测、使用介绍、社区推荐及进一步微调提升等方面的建议和支持。
32
+
33
+ Much thanks to: [@T8star-Aix](https://space.bilibili.com/385085361), [@AI-KSK](https://space.bilibili.com/110353151), [@黑鹤001](https://space.bilibili.com/515231056), [@大麦茶看AI](https://space.bilibili.com/310599757), [@joeMultimedia](https://space.bilibili.com/612740284), [山竹老师AIGC](https://space.bilibili.com/158424637), [CommintAI](https://space.bilibili.com/158429672), [雪煮松风](https://space.bilibili.com/692784900), [冰缘排骨](https://space.bilibili.com/3546582953757144), [活力网TV](https://space.bilibili.com/92055552), [HyLan_L](https://space.bilibili.com/591877893), and many other experts for their advice and support in technical evaluation, usage introduction, community recommendations, and further fine-tuning improvements of various models I worked on.
34
+
35
+ <p align="center">
36
+ <img src="v1-sample.jpg" width="1200"/>
37
+ <p>
38
+
39
+
40
+
41
+ 基础工作流请参见(Base Workflow):[v1-example-workflow.json](https://huggingface.co/wikeeyang/Ernie-Image-Flex-V1/resolve/master/v1-example-workflow.json)
42
+
43
+ Also on: [Modelscope](https://modelscope.cn/models/wikeeyang/Ernie-Image-Flex-V1)
44
+
45
+ The model under Apache 2.0.
v1-example-workflow.json ADDED
@@ -0,0 +1,900 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "id": "9ae6082b-c7f4-433c-9971-7a8f65a3ea65",
3
+ "revision": 0,
4
+ "last_node_id": 207,
5
+ "last_link_id": 506,
6
+ "nodes": [
7
+ {
8
+ "id": 172,
9
+ "type": "CLIPTextEncode",
10
+ "pos": [
11
+ -205.62151998421714,
12
+ 999.3425944916831
13
+ ],
14
+ "size": [
15
+ 333.7879242887998,
16
+ 88
17
+ ],
18
+ "flags": {
19
+ "collapsed": true
20
+ },
21
+ "order": 9,
22
+ "mode": 0,
23
+ "inputs": [
24
+ {
25
+ "localized_name": "clip",
26
+ "name": "clip",
27
+ "type": "CLIP",
28
+ "link": 505
29
+ },
30
+ {
31
+ "localized_name": "text",
32
+ "name": "text",
33
+ "type": "STRING",
34
+ "widget": {
35
+ "name": "text"
36
+ },
37
+ "link": null
38
+ }
39
+ ],
40
+ "outputs": [
41
+ {
42
+ "localized_name": "CONDITIONING",
43
+ "name": "CONDITIONING",
44
+ "type": "CONDITIONING",
45
+ "links": [
46
+ 288
47
+ ]
48
+ }
49
+ ],
50
+ "properties": {
51
+ "cnr_id": "comfy-core",
52
+ "ver": "0.19.3",
53
+ "Node name for S&R": "CLIPTextEncode",
54
+ "ue_properties": {
55
+ "widget_ue_connectable": {},
56
+ "input_ue_unconnectable": {},
57
+ "version": "7.8"
58
+ }
59
+ },
60
+ "widgets_values": [
61
+ "低分辨率,低画质,肢体畸形,手指畸形,画面过饱和,蜡像感,人脸无细节,过度光滑,画面具有AI感。构图混乱。文字模糊,扭曲"
62
+ ]
63
+ },
64
+ {
65
+ "id": 106,
66
+ "type": "VAEDecode",
67
+ "pos": [
68
+ 79.08411671096368,
69
+ 438.4285628963098
70
+ ],
71
+ "size": [
72
+ 173.0576818117222,
73
+ 46
74
+ ],
75
+ "flags": {},
76
+ "order": 11,
77
+ "mode": 0,
78
+ "inputs": [
79
+ {
80
+ "localized_name": "samples",
81
+ "name": "samples",
82
+ "type": "LATENT",
83
+ "link": 134
84
+ },
85
+ {
86
+ "localized_name": "vae",
87
+ "name": "vae",
88
+ "type": "VAE",
89
+ "link": 506
90
+ }
91
+ ],
92
+ "outputs": [
93
+ {
94
+ "localized_name": "IMAGE",
95
+ "name": "IMAGE",
96
+ "type": "IMAGE",
97
+ "slot_index": 0,
98
+ "links": [
99
+ 499
100
+ ]
101
+ }
102
+ ],
103
+ "properties": {
104
+ "cnr_id": "comfy-core",
105
+ "ver": "0.3.64",
106
+ "Node name for S&R": "VAEDecode",
107
+ "ue_properties": {
108
+ "widget_ue_connectable": {},
109
+ "version": "7.7",
110
+ "input_ue_unconnectable": {}
111
+ },
112
+ "enableTabs": false,
113
+ "tabWidth": 65,
114
+ "tabXOffset": 10,
115
+ "hasSecondTab": false,
116
+ "secondTabText": "Send Back",
117
+ "secondTabOffset": 80,
118
+ "secondTabWidth": 65
119
+ },
120
+ "widgets_values": []
121
+ },
122
+ {
123
+ "id": 104,
124
+ "type": "EmptyFlux2LatentImage",
125
+ "pos": [
126
+ -446.0879876688805,
127
+ 1045.7868837003746
128
+ ],
129
+ "size": [
130
+ 210,
131
+ 106
132
+ ],
133
+ "flags": {},
134
+ "order": 0,
135
+ "mode": 0,
136
+ "inputs": [
137
+ {
138
+ "localized_name": "width",
139
+ "name": "width",
140
+ "type": "INT",
141
+ "widget": {
142
+ "name": "width"
143
+ },
144
+ "link": null
145
+ },
146
+ {
147
+ "localized_name": "height",
148
+ "name": "height",
149
+ "type": "INT",
150
+ "widget": {
151
+ "name": "height"
152
+ },
153
+ "link": null
154
+ },
155
+ {
156
+ "localized_name": "batch_size",
157
+ "name": "batch_size",
158
+ "type": "INT",
159
+ "widget": {
160
+ "name": "batch_size"
161
+ },
162
+ "link": null
163
+ }
164
+ ],
165
+ "outputs": [
166
+ {
167
+ "localized_name": "LATENT",
168
+ "name": "LATENT",
169
+ "type": "LATENT",
170
+ "links": [
171
+ 138
172
+ ]
173
+ }
174
+ ],
175
+ "properties": {
176
+ "cnr_id": "comfy-core",
177
+ "ver": "0.18.1",
178
+ "Node name for S&R": "EmptyFlux2LatentImage",
179
+ "ue_properties": {
180
+ "widget_ue_connectable": {},
181
+ "version": "7.7",
182
+ "input_ue_unconnectable": {}
183
+ },
184
+ "enableTabs": false,
185
+ "tabWidth": 65,
186
+ "tabXOffset": 10,
187
+ "hasSecondTab": false,
188
+ "secondTabText": "Send Back",
189
+ "secondTabOffset": 80,
190
+ "secondTabWidth": 65
191
+ },
192
+ "widgets_values": [
193
+ 768,
194
+ 1376,
195
+ 1
196
+ ]
197
+ },
198
+ {
199
+ "id": 136,
200
+ "type": "Seed (rgthree)",
201
+ "pos": [
202
+ -227.8991345790711,
203
+ 1045.509133855802
204
+ ],
205
+ "size": [
206
+ 228.4669349069087,
207
+ 130
208
+ ],
209
+ "flags": {},
210
+ "order": 1,
211
+ "mode": 0,
212
+ "inputs": [],
213
+ "outputs": [
214
+ {
215
+ "dir": 4,
216
+ "name": "SEED",
217
+ "shape": 3,
218
+ "type": "INT",
219
+ "links": [
220
+ 176
221
+ ]
222
+ }
223
+ ],
224
+ "properties": {
225
+ "cnr_id": "rgthree-comfy",
226
+ "ver": "683836c46e898668936c433502504cc0627482c5",
227
+ "randomMax": 1125899906842624,
228
+ "randomMin": 0,
229
+ "ue_properties": {
230
+ "widget_ue_connectable": {},
231
+ "input_ue_unconnectable": {},
232
+ "version": "7.8"
233
+ }
234
+ },
235
+ "widgets_values": [
236
+ 472693926650286,
237
+ "",
238
+ "",
239
+ "okay"
240
+ ]
241
+ },
242
+ {
243
+ "id": 107,
244
+ "type": "KSampler",
245
+ "pos": [
246
+ 25.410264645482513,
247
+ 528.1531174838914
248
+ ],
249
+ "size": [
250
+ 230.55144887224515,
251
+ 646.3688441575066
252
+ ],
253
+ "flags": {},
254
+ "order": 10,
255
+ "mode": 0,
256
+ "inputs": [
257
+ {
258
+ "localized_name": "model",
259
+ "name": "model",
260
+ "type": "MODEL",
261
+ "link": 496
262
+ },
263
+ {
264
+ "localized_name": "positive",
265
+ "name": "positive",
266
+ "type": "CONDITIONING",
267
+ "link": 137
268
+ },
269
+ {
270
+ "localized_name": "negative",
271
+ "name": "negative",
272
+ "type": "CONDITIONING",
273
+ "link": 288
274
+ },
275
+ {
276
+ "localized_name": "latent_image",
277
+ "name": "latent_image",
278
+ "type": "LATENT",
279
+ "link": 138
280
+ },
281
+ {
282
+ "localized_name": "seed",
283
+ "name": "seed",
284
+ "type": "INT",
285
+ "widget": {
286
+ "name": "seed"
287
+ },
288
+ "link": 176
289
+ },
290
+ {
291
+ "localized_name": "steps",
292
+ "name": "steps",
293
+ "type": "INT",
294
+ "widget": {
295
+ "name": "steps"
296
+ },
297
+ "link": null
298
+ },
299
+ {
300
+ "localized_name": "cfg",
301
+ "name": "cfg",
302
+ "type": "FLOAT",
303
+ "widget": {
304
+ "name": "cfg"
305
+ },
306
+ "link": null
307
+ },
308
+ {
309
+ "localized_name": "sampler_name",
310
+ "name": "sampler_name",
311
+ "type": "COMBO",
312
+ "widget": {
313
+ "name": "sampler_name"
314
+ },
315
+ "link": null
316
+ },
317
+ {
318
+ "localized_name": "scheduler",
319
+ "name": "scheduler",
320
+ "type": "COMBO",
321
+ "widget": {
322
+ "name": "scheduler"
323
+ },
324
+ "link": null
325
+ },
326
+ {
327
+ "localized_name": "denoise",
328
+ "name": "denoise",
329
+ "type": "FLOAT",
330
+ "widget": {
331
+ "name": "denoise"
332
+ },
333
+ "link": null
334
+ }
335
+ ],
336
+ "outputs": [
337
+ {
338
+ "localized_name": "LATENT",
339
+ "name": "LATENT",
340
+ "type": "LATENT",
341
+ "slot_index": 0,
342
+ "links": [
343
+ 134
344
+ ]
345
+ }
346
+ ],
347
+ "properties": {
348
+ "cnr_id": "comfy-core",
349
+ "ver": "0.3.64",
350
+ "Node name for S&R": "KSampler",
351
+ "ue_properties": {
352
+ "widget_ue_connectable": {},
353
+ "version": "7.7",
354
+ "input_ue_unconnectable": {}
355
+ },
356
+ "enableTabs": false,
357
+ "tabWidth": 65,
358
+ "tabXOffset": 10,
359
+ "hasSecondTab": false,
360
+ "secondTabText": "Send Back",
361
+ "secondTabOffset": 80,
362
+ "secondTabWidth": 65
363
+ },
364
+ "widgets_values": [
365
+ 252902017622294,
366
+ "fixed",
367
+ 8,
368
+ 1,
369
+ "ddim",
370
+ "simple",
371
+ 1
372
+ ]
373
+ },
374
+ {
375
+ "id": 194,
376
+ "type": "LoraLoaderModelOnly",
377
+ "pos": [
378
+ -449.7085279406747,
379
+ 398.6827410507062
380
+ ],
381
+ "size": [
382
+ 447.70438853371843,
383
+ 82
384
+ ],
385
+ "flags": {},
386
+ "order": 7,
387
+ "mode": 0,
388
+ "inputs": [
389
+ {
390
+ "localized_name": "model",
391
+ "name": "model",
392
+ "type": "MODEL",
393
+ "link": 503
394
+ },
395
+ {
396
+ "localized_name": "lora_name",
397
+ "name": "lora_name",
398
+ "type": "COMBO",
399
+ "widget": {
400
+ "name": "lora_name"
401
+ },
402
+ "link": null
403
+ },
404
+ {
405
+ "localized_name": "strength_model",
406
+ "name": "strength_model",
407
+ "type": "FLOAT",
408
+ "widget": {
409
+ "name": "strength_model"
410
+ },
411
+ "link": null
412
+ }
413
+ ],
414
+ "outputs": [
415
+ {
416
+ "localized_name": "MODEL",
417
+ "name": "MODEL",
418
+ "type": "MODEL",
419
+ "links": [
420
+ 496
421
+ ]
422
+ }
423
+ ],
424
+ "properties": {
425
+ "cnr_id": "comfy-core",
426
+ "ver": "0.20.1",
427
+ "Node name for S&R": "LoraLoaderModelOnly",
428
+ "ue_properties": {
429
+ "widget_ue_connectable": {},
430
+ "input_ue_unconnectable": {},
431
+ "version": "7.8"
432
+ }
433
+ },
434
+ "widgets_values": [
435
+ "ErnieImage-Turbo-Extract_lora_rank64.safetensors",
436
+ 0.4
437
+ ]
438
+ },
439
+ {
440
+ "id": 187,
441
+ "type": "Note",
442
+ "pos": [
443
+ -448.4846463075421,
444
+ 269.06227530443016
445
+ ],
446
+ "size": [
447
+ 230.95844933738942,
448
+ 88
449
+ ],
450
+ "flags": {},
451
+ "order": 2,
452
+ "mode": 0,
453
+ "inputs": [],
454
+ "outputs": [],
455
+ "properties": {
456
+ "ue_properties": {
457
+ "widget_ue_connectable": {},
458
+ "version": "7.8",
459
+ "input_ue_unconnectable": {}
460
+ }
461
+ },
462
+ "widgets_values": [
463
+ "Turbo LoRA Weight: 0.1 - 0.5\nSteps: 20 - 8\nNo Turbo LoRA Steps: 16 - 25"
464
+ ],
465
+ "color": "#432",
466
+ "bgcolor": "#653"
467
+ },
468
+ {
469
+ "id": 189,
470
+ "type": "PreviewImage",
471
+ "pos": [
472
+ 283.9696826482286,
473
+ 438.36456706295013
474
+ ],
475
+ "size": [
476
+ 386.852479982109,
477
+ 734.4512715495057
478
+ ],
479
+ "flags": {},
480
+ "order": 12,
481
+ "mode": 0,
482
+ "inputs": [
483
+ {
484
+ "localized_name": "images",
485
+ "name": "images",
486
+ "type": "IMAGE",
487
+ "link": 499
488
+ }
489
+ ],
490
+ "outputs": [],
491
+ "properties": {
492
+ "cnr_id": "comfy-core",
493
+ "ver": "0.19.0",
494
+ "Node name for S&R": "PreviewImage",
495
+ "ue_properties": {
496
+ "widget_ue_connectable": {},
497
+ "input_ue_unconnectable": {},
498
+ "version": "7.8"
499
+ }
500
+ },
501
+ "widgets_values": []
502
+ },
503
+ {
504
+ "id": 108,
505
+ "type": "CLIPTextEncode",
506
+ "pos": [
507
+ -449.2537356792463,
508
+ 525.3475439421027
509
+ ],
510
+ "size": [
511
+ 449.1001613396379,
512
+ 478.566299927341
513
+ ],
514
+ "flags": {},
515
+ "order": 8,
516
+ "mode": 0,
517
+ "inputs": [
518
+ {
519
+ "localized_name": "clip",
520
+ "name": "clip",
521
+ "type": "CLIP",
522
+ "link": 504
523
+ },
524
+ {
525
+ "localized_name": "text",
526
+ "name": "text",
527
+ "type": "STRING",
528
+ "widget": {
529
+ "name": "text"
530
+ },
531
+ "link": null
532
+ }
533
+ ],
534
+ "outputs": [
535
+ {
536
+ "localized_name": "CONDITIONING",
537
+ "name": "CONDITIONING",
538
+ "type": "CONDITIONING",
539
+ "links": [
540
+ 137
541
+ ]
542
+ }
543
+ ],
544
+ "properties": {
545
+ "cnr_id": "comfy-core",
546
+ "ver": "0.3.73",
547
+ "Node name for S&R": "CLIPTextEncode",
548
+ "ue_properties": {
549
+ "widget_ue_connectable": {},
550
+ "version": "7.7",
551
+ "input_ue_unconnectable": {}
552
+ },
553
+ "enableTabs": false,
554
+ "tabWidth": 65,
555
+ "tabXOffset": 10,
556
+ "hasSecondTab": false,
557
+ "secondTabText": "Send Back",
558
+ "secondTabOffset": 80,
559
+ "secondTabWidth": 65
560
+ },
561
+ "widgets_values": [
562
+ "这是一个采用**极端特写**视角的超写实商业静物摄影,画面主体聚焦于一只深埋于厚重雪层中的银灰色铝制易拉罐,镜头几乎贴近罐身,营造出极致的冷冽质感。罐体呈**正面**视角展示,垂直贯穿画面中央,仿佛刚刚被强行按入冰原之中。罐身表面覆盖着一层细腻的颗粒状冰霜,金属光泽在冷调光线下呈现出哑光与高光交织的质感。\n\n画面主体采用了一种**冷冽的蓝白色调滤镜**,赋予了整个场景一种近乎零度的视觉温度。在灯光效果上,光源似乎来自正上方的漫射冷光,在罐体凹陷处形成了深邃的阴影,强化了冰雪堆砌的立体感;同时,罐身侧面反射着微弱的天光,与周围洁白蓬松的碎冰块形成了明暗对比,凸显了罐身被冰雪侵蚀的纹理。拍摄角度为**平视镜头**,视线与罐体高度平行,让观者仿佛正低头凝视这片冰雪世界。构图风格严格遵循**对称构图**,罐体作为绝对的中轴线,将画面一分为二,四周不规则分布的碎冰块则起到了平衡视觉重心的作用。\n\n背景虽由细碎的雪块和冰渣构成,但通过极浅的景深处理,背景呈现出**柔焦般的朦胧感**,细节虽在虚化中依稀可见,却未抢夺主体的锋芒,成功将视觉重心锁定在中央的易拉罐上。\n\n画面中包含了清晰的白色文字元素,易拉罐顶部印有“RELAX AND ENJOY”的字样,底部则是醒目的黑色大写品牌名“WHITE LION”以及下方的“KEEP COOL”小字,罐身中部还绘有一个白色的狮子图案剪影。易拉罐底部完全被雪掩埋,仅露出上半部分,底部边缘的积雪微微隆起,暗示着巨大的下压力。若将易拉罐拟人化,它呈现出一种**半侧卧**的慵懒姿态,头部(拉环端)微微上扬,身体顺着雪槽微微倾斜,头部与身体形成一条流畅的**15度夹角**,仿佛刚刚在雪地里翻滚了一遭。\n\n罐身设计简约而现代,银灰色的底色上点缀着深褐色的品牌字体和清新的白色的狮子插画,字体边缘带有轻微的冰晶纹理,增强了真实感。易拉罐的质感呈现出金属与冰霜的混合质地,既保留了铝罐的坚硬冷硬,又增添了冰雪的蓬松柔软。在罐体与周围雪堆的互动中,雪块被挤压变形��呈现出一种被强力插入的物理张力。\n\n在画面的最底端,叠加着白色的无衬线字体说明,上方是巨大的“KEEPS COOL”,下方则是较小的“Relaxation white lion drink”,字体干净利落,与上方充满质感的易拉罐形成虚实对比,强化了这是一张旨在传达清凉与放松情绪的广告海报。\n\n图片上部有一行深灰色、冷色调的粗体艺术化的中文字“白狮冰啤”,在每个文字的实线和边缘上覆盖雪花、雪粒,强调一种极致严寒的冰镇效果,从而带来强烈的视觉冲击。字体下方的右侧靠边缘处,有个暗红色的商标“S.T.A.R®”."
563
+ ],
564
+ "color": "#232",
565
+ "bgcolor": "#353"
566
+ },
567
+ {
568
+ "id": 204,
569
+ "type": "UnetLoaderGGUF",
570
+ "pos": [
571
+ -900.7283041130668,
572
+ 527.1773361424006
573
+ ],
574
+ "size": [
575
+ 421.361456984701,
576
+ 58
577
+ ],
578
+ "flags": {},
579
+ "order": 4,
580
+ "mode": 0,
581
+ "inputs": [
582
+ {
583
+ "localized_name": "unet_name",
584
+ "name": "unet_name",
585
+ "type": "COMBO",
586
+ "widget": {
587
+ "name": "unet_name"
588
+ },
589
+ "link": null
590
+ }
591
+ ],
592
+ "outputs": [
593
+ {
594
+ "localized_name": "MODEL",
595
+ "name": "MODEL",
596
+ "type": "MODEL",
597
+ "links": null
598
+ }
599
+ ],
600
+ "properties": {
601
+ "ue_properties": {
602
+ "widget_ue_connectable": {},
603
+ "input_ue_unconnectable": {}
604
+ },
605
+ "cnr_id": "comfyui-gguf",
606
+ "ver": "6ea2651e7df66d7585f6ffee804b20e92fb38b8a",
607
+ "Node name for S&R": "UnetLoaderGGUF"
608
+ },
609
+ "widgets_values": [
610
+ "Ernie-Image-Flex-V1-Q8_0.gguf"
611
+ ]
612
+ },
613
+ {
614
+ "id": 207,
615
+ "type": "VAELoader",
616
+ "pos": [
617
+ -895.0165865388439,
618
+ 891.7775347591978
619
+ ],
620
+ "size": [
621
+ 414.69775161967084,
622
+ 58
623
+ ],
624
+ "flags": {},
625
+ "order": 6,
626
+ "mode": 0,
627
+ "inputs": [
628
+ {
629
+ "localized_name": "vae_name",
630
+ "name": "vae_name",
631
+ "type": "COMBO",
632
+ "widget": {
633
+ "name": "vae_name"
634
+ },
635
+ "link": null
636
+ }
637
+ ],
638
+ "outputs": [
639
+ {
640
+ "localized_name": "VAE",
641
+ "name": "VAE",
642
+ "type": "VAE",
643
+ "links": [
644
+ 506
645
+ ]
646
+ }
647
+ ],
648
+ "properties": {
649
+ "ue_properties": {
650
+ "widget_ue_connectable": {},
651
+ "input_ue_unconnectable": {}
652
+ },
653
+ "cnr_id": "comfy-core",
654
+ "ver": "0.20.1",
655
+ "Node name for S&R": "VAELoader"
656
+ },
657
+ "widgets_values": [
658
+ "Flux2-vae.safetensors"
659
+ ]
660
+ },
661
+ {
662
+ "id": 205,
663
+ "type": "UNETLoader",
664
+ "pos": [
665
+ -898.8244331167627,
666
+ 623.3251876984084
667
+ ],
668
+ "size": [
669
+ 421.36140469204565,
670
+ 82
671
+ ],
672
+ "flags": {},
673
+ "order": 3,
674
+ "mode": 0,
675
+ "inputs": [
676
+ {
677
+ "localized_name": "unet_name",
678
+ "name": "unet_name",
679
+ "type": "COMBO",
680
+ "widget": {
681
+ "name": "unet_name"
682
+ },
683
+ "link": null
684
+ },
685
+ {
686
+ "localized_name": "weight_dtype",
687
+ "name": "weight_dtype",
688
+ "type": "COMBO",
689
+ "widget": {
690
+ "name": "weight_dtype"
691
+ },
692
+ "link": null
693
+ }
694
+ ],
695
+ "outputs": [
696
+ {
697
+ "localized_name": "MODEL",
698
+ "name": "MODEL",
699
+ "type": "MODEL",
700
+ "links": [
701
+ 503
702
+ ]
703
+ }
704
+ ],
705
+ "properties": {
706
+ "ue_properties": {
707
+ "widget_ue_connectable": {},
708
+ "input_ue_unconnectable": {}
709
+ },
710
+ "cnr_id": "comfy-core",
711
+ "ver": "0.20.1",
712
+ "Node name for S&R": "UNETLoader"
713
+ },
714
+ "widgets_values": [
715
+ "Ernie-Image-Flex-V1-bf16.safetensors",
716
+ "default"
717
+ ]
718
+ },
719
+ {
720
+ "id": 206,
721
+ "type": "CLIPLoader",
722
+ "pos": [
723
+ -897.8724453259554,
724
+ 743.2719888542628
725
+ ],
726
+ "size": [
727
+ 419.45753369574163,
728
+ 106
729
+ ],
730
+ "flags": {},
731
+ "order": 5,
732
+ "mode": 0,
733
+ "inputs": [
734
+ {
735
+ "localized_name": "clip_name",
736
+ "name": "clip_name",
737
+ "type": "COMBO",
738
+ "widget": {
739
+ "name": "clip_name"
740
+ },
741
+ "link": null
742
+ },
743
+ {
744
+ "localized_name": "type",
745
+ "name": "type",
746
+ "type": "COMBO",
747
+ "widget": {
748
+ "name": "type"
749
+ },
750
+ "link": null
751
+ },
752
+ {
753
+ "localized_name": "device",
754
+ "name": "device",
755
+ "shape": 7,
756
+ "type": "COMBO",
757
+ "widget": {
758
+ "name": "device"
759
+ },
760
+ "link": null
761
+ }
762
+ ],
763
+ "outputs": [
764
+ {
765
+ "localized_name": "CLIP",
766
+ "name": "CLIP",
767
+ "type": "CLIP",
768
+ "links": [
769
+ 504,
770
+ 505
771
+ ]
772
+ }
773
+ ],
774
+ "properties": {
775
+ "ue_properties": {
776
+ "widget_ue_connectable": {},
777
+ "input_ue_unconnectable": {}
778
+ },
779
+ "cnr_id": "comfy-core",
780
+ "ver": "0.20.1",
781
+ "Node name for S&R": "CLIPLoader"
782
+ },
783
+ "widgets_values": [
784
+ "Ministral-3-3B-TE-Only.safetensors",
785
+ "flux2",
786
+ "default"
787
+ ]
788
+ }
789
+ ],
790
+ "links": [
791
+ [
792
+ 134,
793
+ 107,
794
+ 0,
795
+ 106,
796
+ 0,
797
+ "LATENT"
798
+ ],
799
+ [
800
+ 137,
801
+ 108,
802
+ 0,
803
+ 107,
804
+ 1,
805
+ "CONDITIONING"
806
+ ],
807
+ [
808
+ 138,
809
+ 104,
810
+ 0,
811
+ 107,
812
+ 3,
813
+ "LATENT"
814
+ ],
815
+ [
816
+ 176,
817
+ 136,
818
+ 0,
819
+ 107,
820
+ 4,
821
+ "INT"
822
+ ],
823
+ [
824
+ 288,
825
+ 172,
826
+ 0,
827
+ 107,
828
+ 2,
829
+ "CONDITIONING"
830
+ ],
831
+ [
832
+ 496,
833
+ 194,
834
+ 0,
835
+ 107,
836
+ 0,
837
+ "MODEL"
838
+ ],
839
+ [
840
+ 499,
841
+ 106,
842
+ 0,
843
+ 189,
844
+ 0,
845
+ "IMAGE"
846
+ ],
847
+ [
848
+ 503,
849
+ 205,
850
+ 0,
851
+ 194,
852
+ 0,
853
+ "MODEL"
854
+ ],
855
+ [
856
+ 504,
857
+ 206,
858
+ 0,
859
+ 108,
860
+ 0,
861
+ "CLIP"
862
+ ],
863
+ [
864
+ 505,
865
+ 206,
866
+ 0,
867
+ 172,
868
+ 0,
869
+ "CLIP"
870
+ ],
871
+ [
872
+ 506,
873
+ 207,
874
+ 0,
875
+ 106,
876
+ 1,
877
+ "VAE"
878
+ ]
879
+ ],
880
+ "groups": [],
881
+ "config": {},
882
+ "extra": {
883
+ "frontendVersion": "1.43.17",
884
+ "workflowRendererVersion": "LG",
885
+ "VHS_latentpreview": false,
886
+ "VHS_latentpreviewrate": 0,
887
+ "VHS_MetadataImage": true,
888
+ "VHS_KeepIntermediate": true,
889
+ "ue_links": [],
890
+ "ds": {
891
+ "scale": 0.8769226950000015,
892
+ "offset": [
893
+ 1011.5538907405809,
894
+ -154.09019654263307
895
+ ]
896
+ },
897
+ "links_added_by_ue": []
898
+ },
899
+ "version": 0.4
900
+ }
v1-sample.jpg ADDED

Git LFS Details

  • SHA256: 16eee4ffcc9e46c226a5a4fb3f0a8d25a8fd61d3402f65fdf0092c46df3bd557
  • Pointer size: 132 Bytes
  • Size of remote file: 3.67 MB