{
  "id": 73381,
  "title": "What's your expectation on the magnitude of incoming shake up? :D",
  "url": "/competitions/quickdraw-doodle-recognition/discussion/73381",
  "author_name": "",
  "post_date": "2018-12-02T14:45:47.663000",
  "votes": 15,
  "comment_count": 4,
  "views": 0,
  "content": "",
  "messages": [
    {
      "id": 431579,
      "postDate": "2018-12-02T14:45:47.663Z",
      "rawMarkdown": "",
      "votes": 15
    },
    {
      "id": 431607,
      "postDate": "2018-12-02T15:38:07.443Z",
      "content": "<p>Something we observed in our local experiments is that one form of encoding has a superior public LB than another while sharing the same local scores. So it would be interesting to see how it turns out.</p>",
      "rawMarkdown": "Something we observed in our local experiments is that one form of encoding has a superior public LB than another while sharing the same local scores. So it would be interesting to see how it turns out.",
      "votes": 1,
      "replies": [
        {
          "id": 431624,
          "postDate": "2018-12-02T16:16:51.913Z",
          "content": "<p>I totally agree on your point!</p>",
          "rawMarkdown": "I totally agree on your point!",
          "votes": 10
        }
      ]
    },
    {
      "id": 433331,
      "postDate": "2018-12-05T01:28:39.220Z",
      "content": "<p>Shaked a just little :D</p>",
      "rawMarkdown": "Shaked a just little :D"
    },
    {
      "id": 432805,
      "postDate": "2018-12-04T10:32:29.243Z",
      "content": "<p>Test set is quite large, so if the public lb 9% data has same ditribution as othet 91% shakeup will be limited. At least i hope so )</p>",
      "rawMarkdown": "Test set is quite large, so if the public lb 9% data has same ditribution as othet 91% shakeup will be limited. At least i hope so )"
    }
  ],
  "comments": [
    {
      "id": 431607,
      "author_name": "Peiyuan Liao",
      "author_url": "",
      "post_date": "2018-12-02T15:38:07.443000",
      "content": "<p>Something we observed in our local experiments is that one form of encoding has a superior public LB than another while sharing the same local scores. So it would be interesting to see how it turns out.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 431624,
          "author_name": "",
          "author_url": "",
          "post_date": "2018-12-02T16:16:51.913000",
          "content": "<p>I totally agree on your point!</p>",
          "votes": 10,
          "replies": []
        }
      ]
    },
    {
      "id": 433331,
      "author_name": "Soonhwan Kwon",
      "author_url": "",
      "post_date": "2018-12-05T01:28:39.220000",
      "content": "<p>Shaked a just little :D</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 432805,
      "author_name": "Nickolay Safronov",
      "author_url": "",
      "post_date": "2018-12-04T10:32:29.243000",
      "content": "<p>Test set is quite large, so if the public lb 9% data has same ditribution as othet 91% shakeup will be limited. At least i hope so )</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "431579": "",
    "431607": "Something we observed in our local experiments is that one form of encoding has a superior public LB than another while sharing the same local scores. So it would be interesting to see how it turns out.",
    "433331": "Shaked a just little :D",
    "432805": "Test set is quite large, so if the public lb 9% data has same ditribution as othet 91% shakeup will be limited. At least i hope so )"
  }
}