Datasets:
File size: 1,608 Bytes
2da2925 6b2b925 2da2925 6b2b925 2da2925 6b2b925 2da2925 6b2b925 2da2925 6b2b925 2da2925 6b2b925 2da2925 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 | {
"dataset_name": "EditableDesignBench",
"release_version": "4.0.0",
"release_date": "2026-11-08",
"license": "cc-by-nc-4.0",
"split": "test",
"tasks": 100,
"operations": {
"delete": 20,
"insert": 20,
"move": 20,
"replace": 20,
"scale": 20
},
"domains": {
"poster": 39,
"scientific_figure": 3,
"slides": 5,
"webpage": 53
},
"source_designs": 34,
"selection_manifest": "asset_subset.json",
"selection_only": true,
"case_payloads_changed": false,
"instructions_changed": false,
"parent_release": "3.0.0",
"parent_hf_revision": "v3.0.0",
"split_policy": "Group source_id across both tracks when constructing any split or estimating uncertainty.",
"version": "asset100-v1",
"complete": true,
"validation_passed": true,
"validation_scope": "All selected PNGs checked; original unchanged geometry/rendering validation retained.",
"expected_cases": 100,
"cases": 100,
"asset_families": 40,
"generated_materials_referenced": 39,
"scale_directions": {
"down": 12,
"up": 8
},
"changed_pixels_outside_eval_masks": 0,
"changed_pixels_outside_allowed_masks": 0,
"input_protocol": "asset-edit-single-source-v1",
"errors": [],
"expected_operations": {
"delete": 20,
"insert": 20,
"move": 20,
"replace": 20,
"scale": 20
},
"mask_policy": "Binary unions of recorded four-corner rectangles including image-layer shadow extent.",
"scoring_note": "Asset-semantic judging uses one joint four-criterion request; foreground exact pixel identity is not the instruction-only success criterion."
}
|