{
    "@context": "https://huggingface.co/datasets/croissant/context.jsonld",
    "@type": "Dataset",
    "name": "MMCR",
    "description": "A benchmark dataset for evaluating multimodal compositional reasoning.",
    "license": "https://creativecommons.org/licenses/by/4.0/",
    "creator": {
      "@type": "Person",
      "name": "Luning Zhang"
    },
    "distribution": {
      "@type": "DataDownload",
      "contentUrl": "https://www.kaggle.com/datasets/zhangluning1217/mmcr-benchmark/download",
      "encodingFormat": "application/zip"
    },
    "keywords": ["multimodal", "reasoning", "benchmark", "image-text", "compositionality"],
    "variableMeasured": [
      {
        "@type": "PropertyValue",
        "name": "image",
        "description": "The image input involved in the reasoning process.",
        "valueType": "string"
      },
      {
        "@type": "PropertyValue",
        "name": "question",
        "description": "A natural language question about the image.",
        "valueType": "string"
      },
      {
        "@type": "PropertyValue",
        "name": "answer",
        "description": "The correct answer based on the image and question.",
        "valueType": "string"
      }
    ]
  }