{
  "id": 73693,
  "title": "perfect generalization",
  "url": "/competitions/quickdraw-doodle-recognition/discussion/73693",
  "author_name": "hengck23",
  "post_date": "2018-12-05T00:18:05.535000",
  "votes": 41,
  "comment_count": 17,
  "views": 0,
  "content": "<p>what a number!</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/433278/10826/same.png\" alt=\"enter image description here\"></p>",
  "messages": [
    {
      "id": 433278,
      "postDate": "2018-12-05T00:18:05.537Z",
      "content": "<p>what a number!</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/433278/10826/same.png\" alt=\"enter image description here\"></p>",
      "rawMarkdown": "what a number!\n\n  ![enter image description here][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/433278/10826/same.png",
      "votes": 41
    },
    {
      "id": 433392,
      "postDate": "2018-12-05T02:55:12.790Z",
      "content": "<p>Thank you Heng for your contribution. Hopefully you can share your source code solution to us! :-)</p>",
      "rawMarkdown": "Thank you Heng for your contribution. Hopefully you can share your source code solution to us! :-)\n",
      "votes": 8
    },
    {
      "id": 435554,
      "postDate": "2018-12-08T08:29:03.213Z",
      "content": "<p>Hi Heng, is it possible for you to share your source code solution using pytorch? :-)</p>",
      "rawMarkdown": "Hi Heng, is it possible for you to share your source code solution using pytorch? :-)",
      "votes": 5
    },
    {
      "id": 433492,
      "postDate": "2018-12-05T06:01:42.450Z",
      "content": "<p>@Heng thanks for your active participation in the discussion forums which helped lot of us in improving our models. </p>",
      "rawMarkdown": "@Heng thanks for your active participation in the discussion forums which helped lot of us in improving our models. ",
      "votes": 1
    },
    {
      "id": 434204,
      "postDate": "2018-12-06T03:38:24.360Z",
      "content": "<p>thanks for your ideas and code during this competition</p>",
      "rawMarkdown": "thanks for your ideas and code during this competition"
    },
    {
      "id": 433494,
      "postDate": "2018-12-05T06:03:00.923Z",
      "content": "<p>interesting!!</p>",
      "rawMarkdown": "interesting!!"
    },
    {
      "id": 433458,
      "postDate": "2018-12-05T05:15:33.867Z",
      "content": "<p><a href=\"/hengck23\">@hengck23</a>, I would like to thank you for all your sharing and insights during the competition, together with the template of  beluga ( <a href=\"/gaborfodor\">@gaborfodor</a> ), they are invaluable sources and help us improve tremendously. Our medal is tributed to both (you &amp; beluga) of your kindness!</p>",
      "rawMarkdown": "@hengck23, I would like to thank you for all your sharing and insights during the competition, together with the template of  beluga ( @gaborfodor ), they are invaluable sources and help us improve tremendously. Our medal is tributed to both (you &amp; beluga) of your kindness!"
    },
    {
      "id": 433333,
      "postDate": "2018-12-05T01:29:36.703Z",
      "content": "<p>Well, it seems that the final result is similar to training_dataset_accuracy\nThanks for your sharing in this competition!</p>",
      "rawMarkdown": "Well, it seems that the final result is similar to training_dataset_accuracy\nThanks for your sharing in this competition!"
    },
    {
      "id": 433323,
      "postDate": "2018-12-05T01:24:25.347Z",
      "content": "<p>wow! what a coincidence!</p>",
      "rawMarkdown": "wow! what a coincidence!"
    },
    {
      "id": 433303,
      "postDate": "2018-12-05T00:57:51.390Z",
      "content": "<p>Thanks for your sharing in this competition!</p>",
      "rawMarkdown": "Thanks for your sharing in this competition!"
    },
    {
      "id": 433294,
      "postDate": "2018-12-05T00:40:02.807Z",
      "content": "<p>My score went up from <strong>public LB: 0.93711</strong> to <strong>private LB: 0.93863</strong>, interesting...</p>",
      "rawMarkdown": "My score went up from **public LB: 0.93711** to **private LB: 0.93863**, interesting..."
    },
    {
      "id": 433290,
      "postDate": "2018-12-05T00:31:43.123Z",
      "content": "<p>Congrats @Hengck23 on another very strong finish.</p>\n\n<p>I am surprised at how close my ensemble scored as well, i.e. public <strong>LB=0.92573</strong>  compared to the private <strong>LB=0.92353</strong> since it is only a simple average method. It is in step with the scores of the individual models.</p>",
      "rawMarkdown": "Congrats @Hengck23 on another very strong finish.\n\nI am surprised at how close my ensemble scored as well, i.e. public **LB=0.92573**  compared to the private **LB=0.92353** since it is only a simple average method. It is in step with the scores of the individual models."
    },
    {
      "id": 434195,
      "postDate": "2018-12-06T03:24:33.020Z",
      "rawMarkdown": "",
      "votes": 8,
      "isDeleted": true
    },
    {
      "id": 433664,
      "postDate": "2018-12-05T10:06:57.937Z",
      "rawMarkdown": "",
      "isDeleted": true
    },
    {
      "id": 434155,
      "postDate": "2018-12-06T01:20:07.830Z",
      "content": "<p>Thanks for your ideas! </p>",
      "rawMarkdown": "Thanks for your ideas! "
    },
    {
      "id": 433347,
      "postDate": "2018-12-05T01:46:13.697Z",
      "content": "<p>Thank you for your sharing!</p>",
      "rawMarkdown": "Thank you for your sharing!"
    },
    {
      "id": 433309,
      "postDate": "2018-12-05T01:04:36.783Z",
      "content": "<p>Thank you for the shared ideas.</p>",
      "rawMarkdown": "Thank you for the shared ideas."
    },
    {
      "id": 433288,
      "postDate": "2018-12-05T00:30:58.163Z",
      "content": "<p>@Heng Thanks for your ideas.</p>",
      "rawMarkdown": "@Heng Thanks for your ideas."
    }
  ],
  "comments": [
    {
      "id": 433392,
      "author_name": "",
      "author_url": "",
      "post_date": "2018-12-05T02:55:12.790000",
      "content": "<p>Thank you Heng for your contribution. Hopefully you can share your source code solution to us! :-)</p>",
      "votes": 8,
      "replies": []
    },
    {
      "id": 435554,
      "author_name": "",
      "author_url": "",
      "post_date": "2018-12-08T08:29:03.213000",
      "content": "<p>Hi Heng, is it possible for you to share your source code solution using pytorch? :-)</p>",
      "votes": 5,
      "replies": []
    },
    {
      "id": 433492,
      "author_name": "Rajesh Shreedhar",
      "author_url": "",
      "post_date": "2018-12-05T06:01:42.450000",
      "content": "<p>@Heng thanks for your active participation in the discussion forums which helped lot of us in improving our models. </p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 434204,
      "author_name": "good good study",
      "author_url": "",
      "post_date": "2018-12-06T03:38:24.360000",
      "content": "<p>thanks for your ideas and code during this competition</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433494,
      "author_name": "souraj",
      "author_url": "",
      "post_date": "2018-12-05T06:03:00.923000",
      "content": "<p>interesting!!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433458,
      "author_name": "Neuron Engineer",
      "author_url": "",
      "post_date": "2018-12-05T05:15:33.867000",
      "content": "<p><a href=\"/hengck23\">@hengck23</a>, I would like to thank you for all your sharing and insights during the competition, together with the template of  beluga ( <a href=\"/gaborfodor\">@gaborfodor</a> ), they are invaluable sources and help us improve tremendously. Our medal is tributed to both (you &amp; beluga) of your kindness!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433333,
      "author_name": "Dilapsky Lee",
      "author_url": "",
      "post_date": "2018-12-05T01:29:36.703000",
      "content": "<p>Well, it seems that the final result is similar to training_dataset_accuracy\nThanks for your sharing in this competition!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433323,
      "author_name": "Soonhwan Kwon",
      "author_url": "",
      "post_date": "2018-12-05T01:24:25.347000",
      "content": "<p>wow! what a coincidence!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433303,
      "author_name": "Yiheng Wang",
      "author_url": "",
      "post_date": "2018-12-05T00:57:51.390000",
      "content": "<p>Thanks for your sharing in this competition!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433294,
      "author_name": "Yuming Huang",
      "author_url": "",
      "post_date": "2018-12-05T00:40:02.807000",
      "content": "<p>My score went up from <strong>public LB: 0.93711</strong> to <strong>private LB: 0.93863</strong>, interesting...</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433290,
      "author_name": "YaGana Sheriff-Hussaini",
      "author_url": "",
      "post_date": "2018-12-05T00:31:43.123000",
      "content": "<p>Congrats @Hengck23 on another very strong finish.</p>\n\n<p>I am surprised at how close my ensemble scored as well, i.e. public <strong>LB=0.92573</strong>  compared to the private <strong>LB=0.92353</strong> since it is only a simple average method. It is in step with the scores of the individual models.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 434195,
      "author_name": "",
      "author_url": "",
      "post_date": "2018-12-06T03:24:33.020000",
      "content": "",
      "votes": 8,
      "replies": []
    },
    {
      "id": 433664,
      "author_name": "",
      "author_url": "",
      "post_date": "2018-12-05T10:06:57.937000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 434155,
      "author_name": "Tommy Jiang",
      "author_url": "",
      "post_date": "2018-12-06T01:20:07.830000",
      "content": "<p>Thanks for your ideas! </p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433347,
      "author_name": "cab",
      "author_url": "",
      "post_date": "2018-12-05T01:46:13.697000",
      "content": "<p>Thank you for your sharing!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433309,
      "author_name": "Yibing Wu",
      "author_url": "",
      "post_date": "2018-12-05T01:04:36.783000",
      "content": "<p>Thank you for the shared ideas.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433288,
      "author_name": "wqk",
      "author_url": "",
      "post_date": "2018-12-05T00:30:58.163000",
      "content": "<p>@Heng Thanks for your ideas.</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "433278": "what a number!\n\n  ![enter image description here][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/433278/10826/same.png",
    "433392": "Thank you Heng for your contribution. Hopefully you can share your source code solution to us! :-)\n",
    "435554": "Hi Heng, is it possible for you to share your source code solution using pytorch? :-)",
    "433492": "@Heng thanks for your active participation in the discussion forums which helped lot of us in improving our models. ",
    "434204": "thanks for your ideas and code during this competition",
    "433494": "interesting!!",
    "433458": "@hengck23, I would like to thank you for all your sharing and insights during the competition, together with the template of  beluga ( @gaborfodor ), they are invaluable sources and help us improve tremendously. Our medal is tributed to both (you &amp; beluga) of your kindness!",
    "433333": "Well, it seems that the final result is similar to training_dataset_accuracy\nThanks for your sharing in this competition!",
    "433323": "wow! what a coincidence!",
    "433303": "Thanks for your sharing in this competition!",
    "433294": "My score went up from **public LB: 0.93711** to **private LB: 0.93863**, interesting...",
    "433290": "Congrats @Hengck23 on another very strong finish.\n\nI am surprised at how close my ensemble scored as well, i.e. public **LB=0.92573**  compared to the private **LB=0.92353** since it is only a simple average method. It is in step with the scores of the individual models.",
    "434195": "",
    "433664": "",
    "434155": "Thanks for your ideas! ",
    "433347": "Thank you for your sharing!",
    "433309": "Thank you for the shared ideas.",
    "433288": "@Heng Thanks for your ideas."
  }
}