yomi-chen commited on
Commit
205bf26
·
verified ·
1 Parent(s): 2da2925

Use full-image-only asset judge protocol v2

Browse files
EditableDesignBench.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c06a19b16a4069c0e43b1146c6bb1295da98b96346ef1306ad2d657f2a48ee15
3
- size 523877268
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52172a86b9deb24cd1a911a69264faaa6db8114441a5905b5b799e8cd8df3c53
3
+ size 523877409
README.md CHANGED
@@ -83,10 +83,14 @@ The MLLM judge makes four independent 0–5 scoring requests per task:
83
  | Text | Instruction following (IF), text accuracy (TA), visual consistency (VC), layout preservation (LP) |
84
  | Assets | Instruction following (IF), asset-state accuracy (OA), visual coherence (VC), non-target layout/content preservation (LP) |
85
 
86
- Asset judges receive the full source and prediction plus matching source-defined detail views, the original instruction and the operation. They do not receive the reference target, hidden replacement material, or the editing model's name. Natural movement/scale variations are accepted where the instruction leaves them unspecified. Unrelated text and graphics remain protected, including inside rectangular editing regions.
87
 
88
  Asset results include per-dimension and per-operation scores and a joint qualification rate: all four dimensions must score at least 4. Its denominator is all selected tasks; missing model outputs count as unqualified. API failures remain pending and are retried rather than converted into model scores. Scores and coverage are reported together. Asset criteria and the joint rate are project-specific extensions, not an official TextEditBench asset protocol. Report the two tracks separately.
89
 
 
 
 
 
90
  ## Release 2.0.0
91
 
92
  This release updates the existing 400 text tasks to the locally validated text revision 2, correcting 20 target layouts that could overlap or clip neighboring content. Task IDs are retained and source PNGs are unchanged; three movement instructions were updated together with their references. It adds the 200-task asset track, with 18 enlargements and 22 reductions among the scaling tasks. Re-run changed instructions rather than treating predictions generated for the old task as current outputs.
 
83
  | Text | Instruction following (IF), text accuracy (TA), visual consistency (VC), layout preservation (LP) |
84
  | Assets | Instruction following (IF), asset-state accuracy (OA), visual coherence (VC), non-target layout/content preservation (LP) |
85
 
86
+ Asset judges receive exactly two native-resolution full images (source and prediction), the original instruction and the operation. The current asset MLLM protocol is editabledesignbench-asset-mllm-v2; no detail crops or masks are read or supplied to the judge. They do not receive the reference target, hidden replacement material, or the editing model's name. Natural movement/scale variations are accepted where the instruction leaves them unspecified. Unrelated text and graphics remain protected, including inside rectangular editing regions.
87
 
88
  Asset results include per-dimension and per-operation scores and a joint qualification rate: all four dimensions must score at least 4. Its denominator is all selected tasks; missing model outputs count as unqualified. API failures remain pending and are retried rather than converted into model scores. Scores and coverage are reported together. Asset criteria and the joint rate are project-specific extensions, not an official TextEditBench asset protocol. Report the two tracks separately.
89
 
90
+ ## Scoring protocol update (2026-11-04)
91
+
92
+ Asset judging now uses two full images instead of full images plus detail crops. Re-evaluate asset MLLM scores with the updated code; old and new protocols use separate cache identities. Task data, pixel metrics and text judging are unchanged.
93
+
94
  ## Release 2.0.0
95
 
96
  This release updates the existing 400 text tasks to the locally validated text revision 2, correcting 20 target layouts that could overlap or clip neighboring content. Task IDs are retained and source PNGs are unchanged; three movement instructions were updated together with their references. It adds the 200-task asset track, with 18 enlargements and 22 reductions among the scaling tasks. Re-run changed instructions rather than treating predictions generated for the old task as current outputs.
assets/evaluation_protocol.json CHANGED
@@ -20,13 +20,13 @@
20
  "changed_pixels_outside_region": 0
21
  },
22
  "semantic_judge": {
23
- "implementation_status": "Implemented in the associated EditableDesignBench GitHub evaluation scripts.",
24
  "evaluation_repository": "https://github.com/Ch3nSir/EditableDesignBench",
25
  "scripts": [
26
  "scripts/mllm_evaluate.py",
27
  "scripts/asset_judge_profile.py"
28
  ],
29
- "judge_version": "editabledesignbench-asset-mllm-v1",
30
  "criteria": {
31
  "IF": "Requested asset operation is completed on the intended unit.",
32
  "OA": "Requested asset state, category, count, attributes and visible identity are correct.",
@@ -43,18 +43,8 @@
43
  "inputs": [
44
  "original natural-language instruction and operation",
45
  "native full source image",
46
- "native full candidate image",
47
- "source-derived detail crop",
48
- "candidate detail at identical crop coordinates and enlargement"
49
  ],
50
- "crop_policy": {
51
- "source": "source/instruction-derived allowed_edit_mask only",
52
- "context_margin_pixels": 32,
53
- "minimum_detail_short_side": 512,
54
- "maximum_upscaled_detail_long_side": 1536,
55
- "resampling": "Lanczos",
56
- "full_images": "native RGB without resizing"
57
- },
58
  "reference_target_supplied": false,
59
  "replacement_material_supplied": false,
60
  "model_identity_supplied": false,
@@ -71,6 +61,20 @@
71
  "text_TA_applicable": false,
72
  "official_TextEditBench_asset_metric": false,
73
  "model_results_included": false,
74
- "release_note": "The asset scorer is implemented in the associated evaluation repository; this dataset release contains no model predictions or model evaluation results."
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75
  }
76
  }
 
20
  "changed_pixels_outside_region": 0
21
  },
22
  "semantic_judge": {
23
+ "implementation_status": "Implemented; asset judging uses two native-resolution full images. Previous detail-view scores require re-evaluation.",
24
  "evaluation_repository": "https://github.com/Ch3nSir/EditableDesignBench",
25
  "scripts": [
26
  "scripts/mllm_evaluate.py",
27
  "scripts/asset_judge_profile.py"
28
  ],
29
+ "judge_version": "editabledesignbench-asset-mllm-v2",
30
  "criteria": {
31
  "IF": "Requested asset operation is completed on the intended unit.",
32
  "OA": "Requested asset state, category, count, attributes and visible identity are correct.",
 
43
  "inputs": [
44
  "original natural-language instruction and operation",
45
  "native full source image",
46
+ "native full candidate image"
 
 
47
  ],
 
 
 
 
 
 
 
 
48
  "reference_target_supplied": false,
49
  "replacement_material_supplied": false,
50
  "model_identity_supplied": false,
 
61
  "text_TA_applicable": false,
62
  "official_TextEditBench_asset_metric": false,
63
  "model_results_included": false,
64
+ "release_note": "The asset scorer is implemented in the associated evaluation repository; this dataset release contains no model predictions or model evaluation results.",
65
+ "version": "editabledesignbench-asset-mllm-v2",
66
+ "message_template_revision": "asset-full-image-number-score-v3",
67
+ "views": [
68
+ "full source",
69
+ "full prediction"
70
+ ],
71
+ "detail_crops_supplied": false,
72
+ "masks_supplied": false,
73
+ "image_policy": {
74
+ "image_count": 2,
75
+ "full_images": "native RGB without resizing",
76
+ "detail_crops": false,
77
+ "masks_read": false
78
+ }
79
  }
80
  }