{
  "fetched_at": "2026-10-20T06:52:39Z",
  "source_url": "https://www.kaggle.com/competitions/ViTextVQA-evaluation/data",
  "pages": [
    {
      "name": "data-description",
      "content": "ViTextVQA contains over 16,000 images and over 50,000 questions with answers. The dataset is designed to evaluate the ability of AI models to comprehend text within images and answer questions based on that understanding.\nThe purpose of ViTextVQA is to provide a benchmark for evaluating the reading comprehension ability of Visual Question Answering (VQA) models in the Vietnamese language. As a developing country, Vietnam is still in need of resources and benchmarks to advance research in AI and machine learning.\n\nplease contact 21521333@gm.uit.edu.vn or haryquan.minh@gmail.com for providing dataset",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "**Measuring System Performance in Question-Answering Tasks Using EM**\n\nExact Match (EM) score is a metric commonly used in NLP tasks, particularly in Question Answering systems. It measures the percentage of predictions made by a model that exactly match the target or ground truth answers.\n\nIn the context of question answering, an answer is considered to be correct only if it matches the reference answer exactly. The EM score is a straightforward evaluation metric that provides a clear indication of how often a model produces answers that are exactly right.\n\nThe formula for EM score is:\n\n$$\n\\textbf{EM} = \\frac{\\text{Count of answers predicted exactly}}{\\text{Count of answers}}\n$$\n\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "\n## Submission File\nFor each ID in the test set, you must predict a answer for the question. The file should contain a header and have the following format:\n\n```\nID,Answer\n2,tôi yêu bạn\n5,anh yêu em\n6,i love you\netc.\n```\n\nParticipants are required to submit predictions for both the development (dev) set and the test set. The results for the dev set will be used for public leaderboard rankings, while the results for the test set will be used for private leaderboard rankings. The dev set results are evaluated and ranked before the test set results.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "This is a page where you can include rules that participants must accept before joining. You may wish to include rules like:\n\n - Maintain integrity and honesty in your work.\n - Give your best effort and stay dedicated.\n - Enjoy the challenge and have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "Welcome to ViTextVQA (Vietnamese Text-based Visual Question Answering) dataset! This dataset is the first high-quality large-scale dataset in Vietnamese specializing in understanding text appearing in images.\n\nThis is the dedicated portal for submitting and evaluating the Exact Match (EM) scores of your models. Here, you can test your model’s performance on our dataset and see how it compares to other models in the community.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}