File size: 4,092 Bytes
747f527 083e849 747f527 083e849 747f527 083e849 | 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 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 | {
"name": "SpatialEval",
"release_date": "2024-09-09",
"subsets": {
"vqa": {
"language": [
"en"
],
"modalities": [
"single_image_start"
],
"task_type": "multiple_choice_qa",
"score_pipeline": [
"exact-match",
"rule-match"
],
"score_protocol": {
"reference": "official@github.com/jiayuww/SpatialEval evals/evaluation.py — task-specific regex extraction (extract_answer_from_text_spatialmap L26 /_mazenav L140 /_spatialgrid L205) then lowercase substring match of the oracle answer ('ref_ans.lower() in model_answer.lower()', L286-288); no LLM anywhere.",
"note": "Official grading matches the oracle ANSWER CONTENT by substring after regex extraction; the mm-eval copy stores the option letter (oracle_option, e.g. 'C') as ground truth with options embedded in the question text ('Available options:\\nA. 4...'), keeping oracle_answer/oracle_full_answer in extra. Scoring the published copy is therefore letter-selection matching, not content-substring matching. Official also defines a text-only tqa mode not shipped by the org (vqa/vtqa only)."
},
"prompt_template": "<image>{{ question }}",
"prompt_template_source": {
"origin": "source_column",
"reference": "https://huggingface.co/datasets/MilaWang/SpatialEval",
"notes": "Tier 2: SpatialEval text column pre-renders the question with embedded option list; oracle_option holds the gold letter. Template prepends <image> and uses text verbatim."
},
"mapping_from_source": {
"media": {
"from": "image",
"type": "list",
"min_items": 1,
"max_items": 1
},
"id": {
"from": "id"
},
"question": {
"from": "text"
},
"answer": {
"from": "oracle_option",
"optional": true
},
"extra": {
"oracle_full_answer": {
"from": "oracle_full_answer"
},
"oracle_answer": {
"from": "oracle_answer"
}
},
"source": {
"format": "huggingface",
"url": {
"test": "https://huggingface.co/datasets/MilaWang/SpatialEval"
}
}
}
},
"vtqa": {
"language": [
"en"
],
"modalities": [
"single_image_start"
],
"task_type": "multiple_choice_qa",
"score_pipeline": [
"exact-match",
"rule-match"
],
"score_protocol": {
"reference": "official@github.com/jiayuww/SpatialEval evals/evaluation.py — same rule-based regex extraction + lowercase substring match of the oracle answer (run with --mode 'vtqa'); no LLM.",
"note": "Same as SpatialEval/vqa: published GT is the option letter while the official script substring-matches oracle answer content; tqa mode not shipped."
},
"prompt_template": "<image>{{ question }}",
"prompt_template_source": {
"origin": "source_column",
"reference": "https://huggingface.co/datasets/MilaWang/SpatialEval",
"notes": "Tier 2: SpatialEval text column pre-renders the question with embedded option list; oracle_option holds the gold letter. Template prepends <image> and uses text verbatim."
},
"mapping_from_source": {
"media": {
"from": "image",
"type": "list",
"min_items": 1,
"max_items": 1
},
"id": {
"from": "id"
},
"question": {
"from": "text"
},
"answer": {
"from": "oracle_option",
"optional": true
},
"extra": {
"oracle_full_answer": {
"from": "oracle_full_answer"
},
"oracle_answer": {
"from": "oracle_answer"
}
},
"source": {
"format": "huggingface",
"url": {
"test": "https://huggingface.co/datasets/MilaWang/SpatialEval"
}
}
}
}
}
} |