{
  "id": 69259,
  "title": "Follow CV than LB",
  "url": "/competitions/human-protein-atlas-image-classification/discussion/69259",
  "author_name": "Hanke Chen",
  "post_date": "2018-10-22T04:27:58.162000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<blockquote>\n  <p>This leaderboard is calculated with approximately 29% of the test\n  data. The final results will be based on the other 71%, so the final\n  standings may be different.</p>\n</blockquote>\n\n<p>I think it is pretty self-explanatory. And according to this notebook <a href=\"https://www.kaggle.com/alexanderliao/exploratory-data-analysis/notebook\">Here</a>, </p>\n\n<pre><code># File sizes\ntrain                         14.0321076GB (124288 files)\ntest                          4.6823579GB (46808 files)\ntrain.csv                     0.0012744GB\nsample_submission.csv         0.0004564GB\n</code></pre>\n\n<p>The shakeup is giant in LB. Any more thoughts?</p>",
  "messages": [
    {
      "id": 407963,
      "postDate": "2018-10-22T04:27:58.163Z",
      "content": "<blockquote>\n  <p>This leaderboard is calculated with approximately 29% of the test\n  data. The final results will be based on the other 71%, so the final\n  standings may be different.</p>\n</blockquote>\n\n<p>I think it is pretty self-explanatory. And according to this notebook <a href=\"https://www.kaggle.com/alexanderliao/exploratory-data-analysis/notebook\">Here</a>, </p>\n\n<pre><code># File sizes\ntrain                         14.0321076GB (124288 files)\ntest                          4.6823579GB (46808 files)\ntrain.csv                     0.0012744GB\nsample_submission.csv         0.0004564GB\n</code></pre>\n\n<p>The shakeup is giant in LB. Any more thoughts?</p>",
      "rawMarkdown": "&gt; This leaderboard is calculated with approximately 29% of the test\n&gt; data. The final results will be based on the other 71%, so the final\n&gt; standings may be different.\n\nI think it is pretty self-explanatory. And according to this notebook [Here][1], \n\n    # File sizes\n    train                         14.0321076GB (124288 files)\n    test                          4.6823579GB (46808 files)\n    train.csv                     0.0012744GB\n    sample_submission.csv         0.0004564GB\n\nThe shakeup is giant in LB. Any more thoughts?\n\n  [1]: https://www.kaggle.com/alexanderliao/exploratory-data-analysis/notebook"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "407963": "&gt; This leaderboard is calculated with approximately 29% of the test\n&gt; data. The final results will be based on the other 71%, so the final\n&gt; standings may be different.\n\nI think it is pretty self-explanatory. And according to this notebook [Here][1], \n\n    # File sizes\n    train                         14.0321076GB (124288 files)\n    test                          4.6823579GB (46808 files)\n    train.csv                     0.0012744GB\n    sample_submission.csv         0.0004564GB\n\nThe shakeup is giant in LB. Any more thoughts?\n\n  [1]: https://www.kaggle.com/alexanderliao/exploratory-data-analysis/notebook"
  }
}