{
  "id": 71132,
  "title": "Is is allowed to use predictions from current Google model?",
  "url": "/competitions/quickdraw-doodle-recognition/discussion/71132",
  "author_name": "Ogurtsov",
  "post_date": "2018-11-10T14:57:09.741000",
  "votes": 2,
  "comment_count": 1,
  "views": 0,
  "content": "<p>From rules:</p>\n\n<blockquote>\n  <p>Entrants may re-annotate images in the training set, but may not hand-label predictions, including having human observers rate and evaluate the test data set.</p>\n</blockquote>\n\n<p>So, is it allowed to use predictions from current Google model for test data as additional feature?</p>",
  "messages": [
    {
      "id": 418751,
      "postDate": "2018-11-10T14:57:09.740Z",
      "content": "<p>From rules:</p>\n\n<blockquote>\n  <p>Entrants may re-annotate images in the training set, but may not hand-label predictions, including having human observers rate and evaluate the test data set.</p>\n</blockquote>\n\n<p>So, is it allowed to use predictions from current Google model for test data as additional feature?</p>",
      "rawMarkdown": "From rules:\n\n&gt; Entrants may re-annotate images in the training set, but may not hand-label predictions, including having human observers rate and evaluate the test data set.\n\nSo, is it allowed to use predictions from current Google model for test data as additional feature?",
      "votes": 2
    },
    {
      "id": 418894,
      "postDate": "2018-11-10T20:19:19.577Z",
      "content": "<p>I'm not entirely sure about which predictions are you talking about.</p>\n\n<ul>\n<li>If it's \"recognized\" column in the CSV-s, sure it's allowed. Note that is not available for the test set.</li>\n<li>If you plan to use any other pretrained models or external data you should link them in the external data thread.</li>\n</ul>",
      "rawMarkdown": "I'm not entirely sure about which predictions are you talking about.\n\n- If it's \"recognized\" column in the CSV-s, sure it's allowed. Note that is not available for the test set.\n- If you plan to use any other pretrained models or external data you should link them in the external data thread.\n",
      "votes": 1
    }
  ],
  "comments": [
    {
      "id": 418894,
      "author_name": "Miha Skalic",
      "author_url": "",
      "post_date": "2018-11-10T20:19:19.577000",
      "content": "<p>I'm not entirely sure about which predictions are you talking about.</p>\n\n<ul>\n<li>If it's \"recognized\" column in the CSV-s, sure it's allowed. Note that is not available for the test set.</li>\n<li>If you plan to use any other pretrained models or external data you should link them in the external data thread.</li>\n</ul>",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "418751": "From rules:\n\n&gt; Entrants may re-annotate images in the training set, but may not hand-label predictions, including having human observers rate and evaluate the test data set.\n\nSo, is it allowed to use predictions from current Google model for test data as additional feature?",
    "418894": "I'm not entirely sure about which predictions are you talking about.\n\n- If it's \"recognized\" column in the CSV-s, sure it's allowed. Note that is not available for the test set.\n- If you plan to use any other pretrained models or external data you should link them in the external data thread.\n"
  }
}