bensblueprints commited on
Commit
d5fdb6d
·
verified ·
1 Parent(s): de0c848

Mirror saved portrait workflows and recovery source c0a1479f

Browse files
workflows/portraits/costumes300-033-pasta-maker/REPORT.md CHANGED
@@ -1,8 +1,8 @@
1
  # Pasta Maker — development-roster report
2
 
3
- Status: **in_progress**. Updated 2026-10-16T20:52:56Z.
4
 
5
- 3 / 5 qualifying input records pass the current shared recipe. 15 source photos excluded by the input-quality gate. 1 candidate recipes; 3 renders.
6
 
7
  The independent output critic stays fixed. Acceptance: original input >=7; complete round arithmetic mean >=8 across all four scores for every qualifying dog. Individual dimension minima and the 85% pass fraction do not gate this policy.
8
  Each successful dog check is checkpointed privately. A recipe change invalidates earlier passes until those dogs are checked again. One exact recipe must be checked on the full qualifying roster and average >=8 to lock this costume. Incomplete rounds cannot pass.
@@ -11,7 +11,7 @@ All submitted photos are development data under the owner-requested policy. Hist
11
 
12
  | Candidate | Dogs checked | Dogs passing | Lowest body | Strategy | Pose policy |
13
  | --- | ---: | ---: | ---: | --- | --- |
14
- | 1 | 3 | 3 | 9 | fresh-latent | natural-sitting |
15
 
16
  Exact images, per-dog checkpoints, native job IDs, critiques and unsuccessful graphs remain private. Accepted exports are parameterized and omit customer identities and media paths.
17
  Difficult references remain provisional for another catalog pass; technical failures require inspection. The submission deadline does not become an acceptance shortcut.
 
1
  # Pasta Maker — development-roster report
2
 
3
+ Status: **accepted**. Updated 2026-10-16T20:55:13Z.
4
 
5
+ 5 / 5 qualifying input records pass the current shared recipe. 15 source photos excluded by the input-quality gate. 1 candidate recipes; 5 renders.
6
 
7
  The independent output critic stays fixed. Acceptance: original input >=7; complete round arithmetic mean >=8 across all four scores for every qualifying dog. Individual dimension minima and the 85% pass fraction do not gate this policy.
8
  Each successful dog check is checkpointed privately. A recipe change invalidates earlier passes until those dogs are checked again. One exact recipe must be checked on the full qualifying roster and average >=8 to lock this costume. Incomplete rounds cannot pass.
 
11
 
12
  | Candidate | Dogs checked | Dogs passing | Lowest body | Strategy | Pose policy |
13
  | --- | ---: | ---: | ---: | --- | --- |
14
+ | 1 | 5 | 5 | 9 | fresh-latent | natural-sitting |
15
 
16
  Exact images, per-dog checkpoints, native job IDs, critiques and unsuccessful graphs remain private. Accepted exports are parameterized and omit customer identities and media paths.
17
  Difficult references remain provisional for another catalog pass; technical failures require inspection. The submission deadline does not become an acceptance shortcut.
workflows/portraits/costumes300-033-pasta-maker/status.json CHANGED
@@ -1,23 +1,23 @@
1
  {
2
  "reference_id": "costumes300-033-pasta-maker",
3
- "status": "in_progress",
4
  "target_hardware": "RTX PRO 6000 Blackwell 96 GB",
5
  "baseline_sha256": "fb4d6d0e6fa230cad49dd67a9dcd02825668549273a00de8810b368240cbd525",
6
  "baseline_is_reference_specific": false,
7
  "customer_identity_transfer_validated": false,
8
  "requires_design_review": false,
9
- "accepted_workflow": null,
10
  "latest_provisional_workflow": null,
11
  "last_run_version": "roster-screening-v6-costumes300-033-pasta-maker",
12
- "last_updated_utc": "2026-10-16T20:52:56Z",
13
- "known_roster_identity_transfer_validated": false,
14
  "evaluation_policy": "qualified-inputs-screening-five-v6",
15
  "unseen_dog_validation": false,
16
  "ai_estimate_only": true,
17
  "rounds_started": 1,
18
- "generation_candidates": 3,
19
  "roster_entries": 5,
20
- "passing_current_recipe": 3,
21
  "required_average": 8,
22
  "current_recipe_average": 9.0,
23
  "candidate_revisions": 1,
 
1
  {
2
  "reference_id": "costumes300-033-pasta-maker",
3
+ "status": "accepted",
4
  "target_hardware": "RTX PRO 6000 Blackwell 96 GB",
5
  "baseline_sha256": "fb4d6d0e6fa230cad49dd67a9dcd02825668549273a00de8810b368240cbd525",
6
  "baseline_is_reference_specific": false,
7
  "customer_identity_transfer_validated": false,
8
  "requires_design_review": false,
9
+ "accepted_workflow": "workflow.json",
10
  "latest_provisional_workflow": null,
11
  "last_run_version": "roster-screening-v6-costumes300-033-pasta-maker",
12
+ "last_updated_utc": "2026-10-16T20:55:13Z",
13
+ "known_roster_identity_transfer_validated": true,
14
  "evaluation_policy": "qualified-inputs-screening-five-v6",
15
  "unseen_dog_validation": false,
16
  "ai_estimate_only": true,
17
  "rounds_started": 1,
18
+ "generation_candidates": 5,
19
  "roster_entries": 5,
20
+ "passing_current_recipe": 5,
21
  "required_average": 8,
22
  "current_recipe_average": 9.0,
23
  "candidate_revisions": 1,
workflows/portraits/costumes300-033-pasta-maker/workflow.json ADDED
@@ -0,0 +1,416 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "format": "consistent-pet-average-recipe-v2",
3
+ "template_id": "costumes300-033-pasta-maker",
4
+ "status": "locked_on_average",
5
+ "thresholds": {
6
+ "round_average": 8
7
+ },
8
+ "recipe": {
9
+ "steps": 48,
10
+ "cfg": 4.5,
11
+ "grow": 48,
12
+ "feather": 12,
13
+ "original_megapixels": 1.2,
14
+ "face_megapixels": 1.0,
15
+ "design_megapixels": 0.8,
16
+ "strategy": "fresh-latent",
17
+ "pose_policy": "natural-sitting",
18
+ "focus": [
19
+ "identity",
20
+ "fur",
21
+ "body"
22
+ ],
23
+ "instruction": "Match the original individual face exactly: muzzle length, skull, eye shape, ears, age and asymmetric markings. Keep the face naturally lit and attached to its own neck. Every visible animal surface must use the ORIGINAL dog coat distribution and texture, including forelegs, paws, hindquarters and tail. The erased template animal contributes no coat information. Rebuild a connected body at the original dog's breed, age and build. Fit the costume to that body. Use compact rounded canine paws with short grouped toes, no long fingers or human hands. Respect perspective and genuinely hidden limbs. Use a relaxed natural canine sitting posture with the front paws under the shoulders, bent hindquarters, and proportions matching the source dog. Preserve the costume and scene."
24
+ },
25
+ "scene": "Dog wearing a beige apron with red stripes, sitting at a rustic wooden kitchen table with fresh pasta, tomatoes, basil, and a pasta machine. Background includes copper pots, jars, and a sunlit window with flowers. Medium: photorealistic, warm lighting, shallow depth of field.",
26
+ "clothing": "beige apron with red stripes and tied waist",
27
+ "graph": {
28
+ "9": {
29
+ "class_type": "SaveImage",
30
+ "inputs": {
31
+ "filename_prefix": "__OUTPUT_PREFIX__",
32
+ "images": [
33
+ "repair-composite",
34
+ 0
35
+ ]
36
+ }
37
+ },
38
+ "76": {
39
+ "class_type": "LoadImage",
40
+ "inputs": {
41
+ "image": "__DOG_ORIGINAL__"
42
+ }
43
+ },
44
+ "81": {
45
+ "class_type": "LoadImage",
46
+ "inputs": {
47
+ "image": "__COSTUME_REFERENCE__"
48
+ }
49
+ },
50
+ "92:102": {
51
+ "class_type": "KSamplerSelect",
52
+ "inputs": {
53
+ "sampler_name": "euler"
54
+ }
55
+ },
56
+ "92:103": {
57
+ "class_type": "SamplerCustomAdvanced",
58
+ "inputs": {
59
+ "noise": [
60
+ "92:105",
61
+ 0
62
+ ],
63
+ "guider": [
64
+ "92:114",
65
+ 0
66
+ ],
67
+ "sampler": [
68
+ "92:102",
69
+ 0
70
+ ],
71
+ "sigmas": [
72
+ "92:115",
73
+ 0
74
+ ],
75
+ "latent_image": [
76
+ "92:109",
77
+ 0
78
+ ]
79
+ }
80
+ },
81
+ "92:104": {
82
+ "class_type": "VAEDecode",
83
+ "inputs": {
84
+ "samples": [
85
+ "92:103",
86
+ 0
87
+ ],
88
+ "vae": [
89
+ "92:107",
90
+ 0
91
+ ]
92
+ }
93
+ },
94
+ "92:105": {
95
+ "class_type": "RandomNoise",
96
+ "inputs": {
97
+ "noise_seed": "__DOG_SEED__"
98
+ }
99
+ },
100
+ "92:106": {
101
+ "class_type": "UNETLoader",
102
+ "inputs": {
103
+ "unet_name": "flux-2-klein-base-4b-fp8.safetensors",
104
+ "weight_dtype": "default"
105
+ }
106
+ },
107
+ "92:107": {
108
+ "class_type": "VAELoader",
109
+ "inputs": {
110
+ "vae_name": "full_encoder_small_decoder.safetensors"
111
+ }
112
+ },
113
+ "92:108": {
114
+ "class_type": "GetImageSize",
115
+ "inputs": {
116
+ "image": [
117
+ "92:110",
118
+ 0
119
+ ]
120
+ }
121
+ },
122
+ "92:109": {
123
+ "class_type": "EmptyFlux2LatentImage",
124
+ "inputs": {
125
+ "batch_size": 1,
126
+ "width": 832,
127
+ "height": 1040
128
+ }
129
+ },
130
+ "92:110": {
131
+ "class_type": "ImageScaleToTotalPixels",
132
+ "inputs": {
133
+ "upscale_method": "lanczos",
134
+ "megapixels": 1.2,
135
+ "resolution_steps": 1,
136
+ "image": [
137
+ "76",
138
+ 0
139
+ ]
140
+ }
141
+ },
142
+ "92:85": {
143
+ "class_type": "ImageScaleToTotalPixels",
144
+ "inputs": {
145
+ "upscale_method": "lanczos",
146
+ "megapixels": 0.8,
147
+ "resolution_steps": 1,
148
+ "image": [
149
+ "81",
150
+ 0
151
+ ]
152
+ }
153
+ },
154
+ "92:111": {
155
+ "class_type": "CLIPLoader",
156
+ "inputs": {
157
+ "clip_name": "qwen_3_4b.safetensors",
158
+ "type": "flux2",
159
+ "device": "default"
160
+ }
161
+ },
162
+ "92:113": {
163
+ "class_type": "CLIPTextEncode",
164
+ "inputs": {
165
+ "text": "__DOG_NEUTRAL_REPAIR_PROMPT__",
166
+ "clip": [
167
+ "92:111",
168
+ 0
169
+ ]
170
+ }
171
+ },
172
+ "92:87": {
173
+ "class_type": "CLIPTextEncode",
174
+ "inputs": {
175
+ "text": "",
176
+ "clip": [
177
+ "92:111",
178
+ 0
179
+ ]
180
+ }
181
+ },
182
+ "92:114": {
183
+ "class_type": "CFGGuider",
184
+ "inputs": {
185
+ "cfg": 4.5,
186
+ "model": [
187
+ "92:106",
188
+ 0
189
+ ],
190
+ "positive": [
191
+ "92:130",
192
+ 0
193
+ ],
194
+ "negative": [
195
+ "92:128",
196
+ 0
197
+ ]
198
+ }
199
+ },
200
+ "92:115": {
201
+ "class_type": "Flux2Scheduler",
202
+ "inputs": {
203
+ "steps": 48,
204
+ "width": 832,
205
+ "height": 1040
206
+ }
207
+ },
208
+ "92:125": {
209
+ "class_type": "ReferenceLatent",
210
+ "inputs": {
211
+ "conditioning": [
212
+ "92:87",
213
+ 0
214
+ ],
215
+ "latent": [
216
+ "92:126",
217
+ 0
218
+ ]
219
+ }
220
+ },
221
+ "92:126": {
222
+ "class_type": "VAEEncode",
223
+ "inputs": {
224
+ "pixels": [
225
+ "92:110",
226
+ 0
227
+ ],
228
+ "vae": [
229
+ "92:107",
230
+ 0
231
+ ]
232
+ }
233
+ },
234
+ "92:127": {
235
+ "class_type": "ReferenceLatent",
236
+ "inputs": {
237
+ "conditioning": [
238
+ "92:113",
239
+ 0
240
+ ],
241
+ "latent": [
242
+ "92:126",
243
+ 0
244
+ ]
245
+ }
246
+ },
247
+ "92:128": {
248
+ "class_type": "ReferenceLatent",
249
+ "inputs": {
250
+ "conditioning": [
251
+ "detailnegative",
252
+ 0
253
+ ],
254
+ "latent": [
255
+ "92:129",
256
+ 0
257
+ ]
258
+ }
259
+ },
260
+ "92:129": {
261
+ "class_type": "VAEEncode",
262
+ "inputs": {
263
+ "pixels": [
264
+ "92:85",
265
+ 0
266
+ ],
267
+ "vae": [
268
+ "92:107",
269
+ 0
270
+ ]
271
+ }
272
+ },
273
+ "92:130": {
274
+ "class_type": "ReferenceLatent",
275
+ "inputs": {
276
+ "conditioning": [
277
+ "detailpositive",
278
+ 0
279
+ ],
280
+ "latent": [
281
+ "92:129",
282
+ 0
283
+ ]
284
+ }
285
+ },
286
+ "detailimage": {
287
+ "class_type": "LoadImage",
288
+ "inputs": {
289
+ "image": "__DOG_FACE__"
290
+ }
291
+ },
292
+ "detailscale": {
293
+ "class_type": "ImageScaleToTotalPixels",
294
+ "inputs": {
295
+ "image": [
296
+ "detailimage",
297
+ 0
298
+ ],
299
+ "upscale_method": "lanczos",
300
+ "megapixels": 1.0,
301
+ "resolution_steps": 1
302
+ }
303
+ },
304
+ "detaillatent": {
305
+ "class_type": "VAEEncode",
306
+ "inputs": {
307
+ "pixels": [
308
+ "detailscale",
309
+ 0
310
+ ],
311
+ "vae": [
312
+ "92:107",
313
+ 0
314
+ ]
315
+ }
316
+ },
317
+ "detailpositive": {
318
+ "class_type": "ReferenceLatent",
319
+ "inputs": {
320
+ "conditioning": [
321
+ "92:127",
322
+ 0
323
+ ],
324
+ "latent": [
325
+ "detaillatent",
326
+ 0
327
+ ]
328
+ }
329
+ },
330
+ "detailnegative": {
331
+ "class_type": "ReferenceLatent",
332
+ "inputs": {
333
+ "conditioning": [
334
+ "92:125",
335
+ 0
336
+ ],
337
+ "latent": [
338
+ "detaillatent",
339
+ 0
340
+ ]
341
+ }
342
+ },
343
+ "repair-base": {
344
+ "class_type": "LoadImage",
345
+ "inputs": {
346
+ "image": "__TEMPLATE_LAYOUT__"
347
+ }
348
+ },
349
+ "repair-mask": {
350
+ "class_type": "LoadImageMask",
351
+ "inputs": {
352
+ "image": "__SUBJECT_MASK__",
353
+ "channel": "red"
354
+ }
355
+ },
356
+ "repair-latent": {
357
+ "class_type": "VAEEncodeForInpaint",
358
+ "inputs": {
359
+ "pixels": [
360
+ "repair-base",
361
+ 0
362
+ ],
363
+ "vae": [
364
+ "92:107",
365
+ 0
366
+ ],
367
+ "mask": [
368
+ "repair-mask",
369
+ 0
370
+ ],
371
+ "grow_mask_by": 0
372
+ }
373
+ },
374
+ "repair-composite": {
375
+ "class_type": "ImageCompositeMasked",
376
+ "inputs": {
377
+ "destination": [
378
+ "repair-base",
379
+ 0
380
+ ],
381
+ "source": [
382
+ "92:104",
383
+ 0
384
+ ],
385
+ "x": 0,
386
+ "y": 0,
387
+ "resize_source": false,
388
+ "mask": [
389
+ "repair-mask",
390
+ 0
391
+ ]
392
+ }
393
+ }
394
+ },
395
+ "evaluation": {
396
+ "policy": "qualified-inputs-screening-five-v6",
397
+ "screening_sample": true,
398
+ "qualified_pool_entries": 17,
399
+ "untested_qualifying_entries": 12,
400
+ "roster_entries": 5,
401
+ "passing_entries": 5,
402
+ "checked_entries": 5,
403
+ "average_score": 9.0,
404
+ "score_sum": 180,
405
+ "score_count": 20,
406
+ "required_average": 8,
407
+ "same_recipe_for_all": true,
408
+ "unseen_dog_validation": false,
409
+ "ai_estimate_only": true,
410
+ "input_minimum": 7,
411
+ "excluded_input_entries": 15
412
+ },
413
+ "production_enabled": false,
414
+ "binding": "Bind original dog, real face crop, reference assets and seed through repair_graph and apply_recipe. Do not add dog-specific review corrections.",
415
+ "sha256": "9123f4ce13bc164ccdb3745958a5807e546f6ad612dbd26fb8fcad225be69560"
416
+ }
workflows/portraits/costumes300-034-afternoon-tea-host/REPORT.md ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Afternoon Tea Host — development-roster report
2
+
3
+ Status: **in_progress**. Updated 2026-10-16T20:58:59Z.
4
+
5
+ 3 / 5 qualifying input records pass the current shared recipe. 15 source photos excluded by the input-quality gate. 1 candidate recipes; 3 renders.
6
+
7
+ The independent output critic stays fixed. Acceptance: original input >=7; complete round arithmetic mean >=8 across all four scores for every qualifying dog. Individual dimension minima and the 85% pass fraction do not gate this policy.
8
+ Each successful dog check is checkpointed privately. A recipe change invalidates earlier passes until those dogs are checked again. One exact recipe must be checked on the full qualifying roster and average >=8 to lock this costume. Incomplete rounds cannot pass.
9
+
10
+ All submitted photos are development data under the owner-requested policy. Historical holdout runs remain immutable; this new result is not independent validation on unseen dogs. AI estimates do not establish universal or perfect performance.
11
+
12
+ | Candidate | Dogs checked | Dogs passing | Lowest body | Strategy | Pose policy |
13
+ | --- | ---: | ---: | ---: | --- | --- |
14
+ | 1 | 3 | 3 | 7 | fresh-latent | source-compatible |
15
+
16
+ Exact images, per-dog checkpoints, native job IDs, critiques and unsuccessful graphs remain private. Accepted exports are parameterized and omit customer identities and media paths.
17
+ Difficult references remain provisional for another catalog pass; technical failures require inspection. The submission deadline does not become an acceptance shortcut.
workflows/portraits/costumes300-034-afternoon-tea-host/status.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "reference_id": "costumes300-034-afternoon-tea-host",
3
- "status": "pending_gpu_refinement",
4
  "target_hardware": "RTX PRO 6000 Blackwell 96 GB",
5
  "baseline_sha256": "fb4d6d0e6fa230cad49dd67a9dcd02825668549273a00de8810b368240cbd525",
6
  "baseline_is_reference_specific": false,
@@ -8,6 +8,27 @@
8
  "requires_design_review": false,
9
  "accepted_workflow": null,
10
  "latest_provisional_workflow": null,
11
- "last_run_version": null,
12
- "last_updated_utc": null
13
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  {
2
  "reference_id": "costumes300-034-afternoon-tea-host",
3
+ "status": "in_progress",
4
  "target_hardware": "RTX PRO 6000 Blackwell 96 GB",
5
  "baseline_sha256": "fb4d6d0e6fa230cad49dd67a9dcd02825668549273a00de8810b368240cbd525",
6
  "baseline_is_reference_specific": false,
 
8
  "requires_design_review": false,
9
  "accepted_workflow": null,
10
  "latest_provisional_workflow": null,
11
+ "last_run_version": "roster-screening-v6-costumes300-034-afternoon-tea-host",
12
+ "last_updated_utc": "2026-10-16T20:58:59Z",
13
+ "known_roster_identity_transfer_validated": false,
14
+ "evaluation_policy": "qualified-inputs-screening-five-v6",
15
+ "unseen_dog_validation": false,
16
+ "ai_estimate_only": true,
17
+ "rounds_started": 1,
18
+ "generation_candidates": 3,
19
+ "roster_entries": 5,
20
+ "passing_current_recipe": 3,
21
+ "required_average": 8,
22
+ "current_recipe_average": 8.833333333333334,
23
+ "candidate_revisions": 1,
24
+ "screening_sample": true,
25
+ "qualified_pool_entries": 17,
26
+ "input_minimum": 7,
27
+ "output_thresholds": {
28
+ "identity": 8,
29
+ "fur": 8,
30
+ "body": 7,
31
+ "template": 8
32
+ },
33
+ "excluded_input_entries": 15
34
+ }