metadata: migrate score_type -> score_pipeline (atomic stage contract; see mm-eval scorer docs/en/SCORING.md)

metadata.json

@@ -1,85 +1,88 @@
 {
- "name": "GMAI-MMBench",
- "release_date": "2024-08-12",
- "subsets": {
-  "default": {
-   "language": [
-    "en"
-   ],
-   "modalities": [
-    "single_image_start"
-   ],
-   "task_type": "multiple_choice_qa",
-   "score_type": "llm_extract",
-   "score_protocol": {
-    "reference": "vlmevalkit@vlmeval/dataset/image_mcq.py:745 class GMAIMMBenchDataset(ImageMCQDataset) — evaluate() -> mcq_vanilla_eval (rule can_infer then GPT judge extraction); authors (OpenGVLab) ship GMAI-MMBench through VLMEvalKit as its official eval. Confirmed subclass at line 745 in clone.",
-    "note": "Medical MCQ A-E (E often empty). Official aggregation reports accuracy grouped by clinical VQA task / department / modality / perceptual granularity, not just a single headline — group scores are not reproduced by the per-sample scorer."
-   },
-   "prompt_template": "<image>{{ question }}\nA. {{ options.A }}\nB. {{ options.B }}\nC. {{ options.C }}{% if options.D %}\nD. {{ options.D }}{% endif %}{% if options.E %}\nE. {{ options.E }}{% endif %}\nAnswer with the option's letter from the given choices directly.",
-   "prompt_template_source": {
-    "origin": "official",
-    "reference": "https://huggingface.co/datasets/OpenGVLab/GMAI-MMBench",
-    "notes": "Tier 1: GMAI-MMBench is published as MCQ A-E (E often empty), letter answer; medical multimodal eval."
-   },
-   "mapping_from_source": {
-    "media": {
-     "from": "image",
-     "type": "list",
-     "min_items": 1,
-     "max_items": 1
-    },
-    "id": {
-     "from": "index"
-    },
-    "question": {
-     "from": "question"
-    },
-    "answer": {
-     "from": "answer",
-     "optional": true
-    },
-    "extra": {
-     "A": {
-      "from": "A"
-     },
-     "B": {
-      "from": "B"
-     },
-     "C": {
-      "from": "C"
-     },
-     "D": {
-      "from": "D"
-     },
-     "E": {
-      "from": "E"
-     },
-     "category": {
-      "from": "category"
-     },
-     "clinical_VQA_task": {
-      "from": "clinical VQA task"
-     },
-     "department": {
-      "from": "department"
-     },
-     "perceptual_granularity": {
-      "from": "perceptual granularity"
-     },
-     "modality": {
-      "from": "modality"
-     },
-     "original_task": {
-      "from": "original task"
-     }
-    },
-    "source": {
-     "format": "tsv",
-     "url": {
-      "val": "https://huggingface.co/datasets/OpenGVLab/GMAI-MMBench"
-     }
+  "name": "GMAI-MMBench",
+  "release_date": "2024-08-12",
+  "subsets": {
+    "default": {
+      "language": [
+        "en"
+      ],
+      "modalities": [
+        "single_image_start"
+      ],
+      "task_type": "multiple_choice_qa",
+      "score_pipeline": [
+        "rule-match",
+        "llm-match"
+      ],
+      "score_protocol": {
+        "reference": "vlmevalkit@vlmeval/dataset/image_mcq.py:745 class GMAIMMBenchDataset(ImageMCQDataset) — evaluate() -> mcq_vanilla_eval (rule can_infer then GPT judge extraction); authors (OpenGVLab) ship GMAI-MMBench through VLMEvalKit as its official eval. Confirmed subclass at line 745 in clone.",
+        "note": "Medical MCQ A-E (E often empty). Official aggregation reports accuracy grouped by clinical VQA task / department / modality / perceptual granularity, not just a single headline — group scores are not reproduced by the per-sample scorer."
+      },
+      "prompt_template": "<image>{{ question }}\nA. {{ options.A }}\nB. {{ options.B }}\nC. {{ options.C }}{% if options.D %}\nD. {{ options.D }}{% endif %}{% if options.E %}\nE. {{ options.E }}{% endif %}\nAnswer with the option's letter from the given choices directly.",
+      "prompt_template_source": {
+        "origin": "official",
+        "reference": "https://huggingface.co/datasets/OpenGVLab/GMAI-MMBench",
+        "notes": "Tier 1: GMAI-MMBench is published as MCQ A-E (E often empty), letter answer; medical multimodal eval."
+      },
+      "mapping_from_source": {
+        "media": {
+          "from": "image",
+          "type": "list",
+          "min_items": 1,
+          "max_items": 1
+        },
+        "id": {
+          "from": "index"
+        },
+        "question": {
+          "from": "question"
+        },
+        "answer": {
+          "from": "answer",
+          "optional": true
+        },
+        "extra": {
+          "A": {
+            "from": "A"
+          },
+          "B": {
+            "from": "B"
+          },
+          "C": {
+            "from": "C"
+          },
+          "D": {
+            "from": "D"
+          },
+          "E": {
+            "from": "E"
+          },
+          "category": {
+            "from": "category"
+          },
+          "clinical_VQA_task": {
+            "from": "clinical VQA task"
+          },
+          "department": {
+            "from": "department"
+          },
+          "perceptual_granularity": {
+            "from": "perceptual granularity"
+          },
+          "modality": {
+            "from": "modality"
+          },
+          "original_task": {
+            "from": "original task"
+          }
+        },
+        "source": {
+          "format": "tsv",
+          "url": {
+            "val": "https://huggingface.co/datasets/OpenGVLab/GMAI-MMBench"
+          }
+        }
+      }
     }
-   }
   }
- }
-}
+}