{
  "id": 118759,
  "title": "Are you predicting multiple short answers?",
  "url": "/competitions/tensorflow2-question-answering/discussion/118759",
  "author_name": "",
  "post_date": "2019-11-24T09:11:17.769551900Z",
  "votes": 7,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Can we submit multiple candidates for short answers?\nI mean can we output multiple <code>start_token:end_token</code> separated by space. For example, <code>start_token_c1:end_token_c1 start_token_c2:end_token_c2</code> for a short answer in submission file?</p>\n\n<p>From the evaluation script link I can see they check all the short answer candidates <a href=\"https://github.com/google-research-datasets/natural-questions/blob/master/nq_eval.py\">https://github.com/google-research-datasets/natural-questions/blob/master/nq_eval.py</a>.\nCan someone clarify this?</p>",
  "messages": [
    {
      "id": "680226",
      "postDate": "11/24/2019 09:11:17",
      "content": "<p>Can we submit multiple candidates for short answers?\nI mean can we output multiple <code>start_token:end_token</code> separated by space. For example, <code>start_token_c1:end_token_c1 start_token_c2:end_token_c2</code> for a short answer in submission file?</p>\n\n<p>From the evaluation script link I can see they check all the short answer candidates <a href=\"https://github.com/google-research-datasets/natural-questions/blob/master/nq_eval.py\">https://github.com/google-research-datasets/natural-questions/blob/master/nq_eval.py</a>.\nCan someone clarify this?</p>",
      "rawMarkdown": "Can we submit multiple candidates for short answers?\nI mean can we output multiple `start_token:end_token` separated by space. For example, `start_token_c1:end_token_c1 start_token_c2:end_token_c2` for a short answer in submission file?\n\nFrom the evaluation script link I can see they check all the short answer candidates https://github.com/google-research-datasets/natural-questions/blob/master/nq_eval.py.\nCan someone clarify this?",
      "votes": null
    },
    {
      "id": "681290",
      "postDate": "11/26/2019 00:05:48",
      "content": "<p>Thanks for your question! You can predict only one short answer. If you identify <em>any</em> of the correct short answers, your answer is scored as correct.</p>",
      "rawMarkdown": "Thanks for your question! You can predict only one short answer. If you identify _any_ of the correct short answers, your answer is scored as correct.",
      "votes": null
    },
    {
      "id": "681474",
      "postDate": "11/26/2019 06:35:01",
      "content": "<p>Thanks for the answer</p>",
      "rawMarkdown": "Thanks for the answer",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 681290,
      "author_name": "philculliton",
      "author_url": "",
      "post_date": "11/26/2019 00:05:48",
      "content": "<p>Thanks for your question! You can predict only one short answer. If you identify <em>any</em> of the correct short answers, your answer is scored as correct.</p>",
      "votes": null,
      "replies": [
        {
          "id": 681474,
          "author_name": "axel81",
          "author_url": "",
          "post_date": "11/26/2019 06:35:01",
          "content": "<p>Thanks for the answer</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "680226": "Can we submit multiple candidates for short answers?\nI mean can we output multiple `start_token:end_token` separated by space. For example, `start_token_c1:end_token_c1 start_token_c2:end_token_c2` for a short answer in submission file?\n\nFrom the evaluation script link I can see they check all the short answer candidates https://github.com/google-research-datasets/natural-questions/blob/master/nq_eval.py.\nCan someone clarify this?",
    "681290": "Thanks for your question! You can predict only one short answer. If you identify _any_ of the correct short answers, your answer is scored as correct.",
    "681474": "Thanks for the answer"
  },
  "source": "meta"
}