{
  "id": 193285,
  "title": "Leaderboard scoring speed up idea ",
  "url": "/competitions/rsna-str-pulmonary-embolism-detection/discussion/193285",
  "author_name": "Dewei Chen",
  "post_date": "2020-10-26T10:30:56.701000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I tried to submit just now. This submission just a copy of a full inference of my model. Data size is (152703, 2) id + label. Fill other labels with 0.5 The Public LB come out with 3 mins</p>\n<p>The main idea is because when we submit our kernel, Kaggle will be using <strong>real data</strong> (private + public dataset) replace <strong>fake data</strong> (what we used) </p>\n<p>Hence, I think we can submit a full <strong>152703</strong> prediction file, and using this data instead of the linking Id's label.  This could be used to get a result faster.</p>\n<p>NOTICE: We still need to <strong>CHECK</strong> the rule, to make sure this method is valid.</p>",
  "messages": [
    {
      "id": 1060569,
      "postDate": "2020-10-26T10:30:56.700Z",
      "content": "<p>I tried to submit just now. This submission just a copy of a full inference of my model. Data size is (152703, 2) id + label. Fill other labels with 0.5 The Public LB come out with 3 mins</p>\n<p>The main idea is because when we submit our kernel, Kaggle will be using <strong>real data</strong> (private + public dataset) replace <strong>fake data</strong> (what we used) </p>\n<p>Hence, I think we can submit a full <strong>152703</strong> prediction file, and using this data instead of the linking Id's label.  This could be used to get a result faster.</p>\n<p>NOTICE: We still need to <strong>CHECK</strong> the rule, to make sure this method is valid.</p>",
      "rawMarkdown": "I tried to submit just now. This submission just a copy of a full inference of my model. Data size is (152703, 2) id + label. Fill other labels with 0.5 The Public LB come out with 3 mins\n\nThe main idea is because when we submit our kernel, Kaggle will be using **real data** (private + public dataset) replace **fake data** (what we used) \n\nHence, I think we can submit a full **152703** prediction file, and using this data instead of the linking Id's label.  This could be used to get a result faster.\n\n\nNOTICE: We still need to **CHECK** the rule, to make sure this method is valid.",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1060569": "I tried to submit just now. This submission just a copy of a full inference of my model. Data size is (152703, 2) id + label. Fill other labels with 0.5 The Public LB come out with 3 mins\n\nThe main idea is because when we submit our kernel, Kaggle will be using **real data** (private + public dataset) replace **fake data** (what we used) \n\nHence, I think we can submit a full **152703** prediction file, and using this data instead of the linking Id's label.  This could be used to get a result faster.\n\n\nNOTICE: We still need to **CHECK** the rule, to make sure this method is valid."
  }
}