{
  "id": 3789,
  "title": "Release format for the 2nd augmented training set and full public leaderboard days",
  "url": "/competitions/flight/discussion/3789",
  "author_name": "",
  "post_date": "2013-02-08T06:21:35.553Z",
  "votes": null,
  "comment_count": 15,
  "views": 8105,
  "content": "<p>On Saturday, February 9th UTC (precise time unknown), we'll be releasing the full public leaderboard day files in addition to the 2nd Augmented training set.</p>\r\n<p>I have two choices for releasing these sets:</p>\r\n<ul>\r\n<li>I could upload to additional archives (PublicLeaderboardTrainSet.zip, AugmentedTrainSet2.Zip), which contains the two corresponding sets\r\n</li><li>I could upload Train.zip, which would contain all the days between November 12, 2012 and February 6, 2013\r\n</li></ul>\r\n<div>I'm leaning towards the second one since it makes the I/O for your final models a bit simpler (even though you'll have to download a file that is ~2x larger), but wanted to give you a chance to provide input. Please reply to this thread if you have any\r\n thoughts or preferences on this.</div>\r\n<div>We will not be releasing the data in both possible formats.</div>",
  "messages": [
    {
      "id": "20169",
      "postDate": "02/08/2013 06:21:35",
      "content": "<p>On Saturday, February 9th UTC (precise time unknown), we'll be releasing the full public leaderboard day files in addition to the 2nd Augmented training set.</p>\r\n<p>I have two choices for releasing these sets:</p>\r\n<ul>\r\n<li>I could upload to additional archives (PublicLeaderboardTrainSet.zip, AugmentedTrainSet2.Zip), which contains the two corresponding sets\r\n</li><li>I could upload Train.zip, which would contain all the days between November 12, 2012 and February 6, 2013\r\n</li></ul>\r\n<div>I'm leaning towards the second one since it makes the I/O for your final models a bit simpler (even though you'll have to download a file that is ~2x larger), but wanted to give you a chance to provide input. Please reply to this thread if you have any\r\n thoughts or preferences on this.</div>\r\n<div>We will not be releasing the data in both possible formats.</div>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20173",
      "postDate": "02/08/2013 06:35:54",
      "content": "<p><span style=\"font-size:14px; line-height:1.4em\">I'd prefer choice nr. 1 - the training set for the public leaderboard was partitioned too so it shouldn't be much of an issue.</span></p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20174",
      "postDate": "02/08/2013 06:38:11",
      "content": "<p>Hi Ben!<br>\r\nI prefer the second option. One file - train.7z.</p>\r\n<p>All the best,</p>\r\n<p>Alex.</p>\r\n<p><br>\r\n&nbsp;</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20175",
      "postDate": "02/08/2013 06:52:55",
      "content": "<p>Downloading the augmented training set failed twice for me (corrupted zips) and took a lot of time even when it finally succeeded (by going for the 7z instead). So I guess a&nbsp;single&nbsp;jumbo file containing all the training data would be an efficient way to\r\n weed out competitors with less than stellar ISPs. ;)</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20179",
      "postDate": "02/08/2013 08:14:44",
      "content": "<p><span style=\"font-size:14px; line-height:1.4em\">I'd prefer choice nr. 1. Thanks, Xavier</span></p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20180",
      "postDate": "02/08/2013 08:48:11",
      "content": "<p>Nr. 1. seems better for me (without fix broadband connection at home :)</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20181",
      "postDate": "02/08/2013 09:10:03",
      "content": "",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20182",
      "postDate": "02/08/2013 09:10:17",
      "content": "",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20185",
      "postDate": "02/08/2013 14:29:57",
      "content": "<p><span>I'd prefer choice nr. 1. Thanks!</span></p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20186",
      "postDate": "02/08/2013 15:08:18",
      "content": "<p><span><span>Thx for asking, I'd also prefer choice #1</span></span></p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20187",
      "postDate": "02/08/2013 15:16:38",
      "content": "",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20188",
      "postDate": "02/08/2013 15:16:52",
      "content": "<p>I would prefer #1.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20192",
      "postDate": "02/08/2013 16:31:57",
      "content": "<p>It seems no chances to downlolad a single archive, so I need 1.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20200",
      "postDate": "02/08/2013 20:26:20",
      "content": "<p>I prefer #1. I've had problems downloading larger files, so it would be safer if there were two smaller ones.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20203",
      "postDate": "02/08/2013 21:35:31",
      "content": "<p>Well that's virtually unanimous. Two separate archives containing only the new data it is.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "20215",
      "postDate": "02/09/2013 04:16:18",
      "content": "<p>yaaay</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 20173,
      "author_name": "pprett",
      "author_url": "",
      "post_date": "02/08/2013 06:35:54",
      "content": "<p><span style=\"font-size:14px; line-height:1.4em\">I'd prefer choice nr. 1 - the training set for the public leaderboard was partitioned too so it shouldn't be much of an issue.</span></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20174,
      "author_name": "alexxanderlarko",
      "author_url": "",
      "post_date": "02/08/2013 06:38:11",
      "content": "<p>Hi Ben!<br>\r\nI prefer the second option. One file - train.7z.</p>\r\n<p>All the best,</p>\r\n<p>Alex.</p>\r\n<p><br>\r\n&nbsp;</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20175,
      "author_name": "hyperborean",
      "author_url": "",
      "post_date": "02/08/2013 06:52:55",
      "content": "<p>Downloading the augmented training set failed twice for me (corrupted zips) and took a lot of time even when it finally succeeded (by going for the 7z instead). So I guess a&nbsp;single&nbsp;jumbo file containing all the training data would be an efficient way to\r\n weed out competitors with less than stellar ISPs. ;)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20179,
      "author_name": "xavierconort",
      "author_url": "",
      "post_date": "02/08/2013 08:14:44",
      "content": "<p><span style=\"font-size:14px; line-height:1.4em\">I'd prefer choice nr. 1. Thanks, Xavier</span></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20180,
      "author_name": "gaborfodor",
      "author_url": "",
      "post_date": "02/08/2013 08:48:11",
      "content": "<p>Nr. 1. seems better for me (without fix broadband connection at home :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20181,
      "author_name": "alexxanderlarko",
      "author_url": "",
      "post_date": "02/08/2013 09:10:03",
      "content": "",
      "votes": null,
      "replies": []
    },
    {
      "id": 20182,
      "author_name": "alexxanderlarko",
      "author_url": "",
      "post_date": "02/08/2013 09:10:17",
      "content": "",
      "votes": null,
      "replies": []
    },
    {
      "id": 20185,
      "author_name": "zhanpengfang",
      "author_url": "",
      "post_date": "02/08/2013 14:29:57",
      "content": "<p><span>I'd prefer choice nr. 1. Thanks!</span></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20186,
      "author_name": "jay670790",
      "author_url": "",
      "post_date": "02/08/2013 15:08:18",
      "content": "<p><span><span>Thx for asking, I'd also prefer choice #1</span></span></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20187,
      "author_name": "manishm",
      "author_url": "",
      "post_date": "02/08/2013 15:16:38",
      "content": "",
      "votes": null,
      "replies": []
    },
    {
      "id": 20188,
      "author_name": "manishm",
      "author_url": "",
      "post_date": "02/08/2013 15:16:52",
      "content": "<p>I would prefer #1.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20192,
      "author_name": "nedelko",
      "author_url": "",
      "post_date": "02/08/2013 16:31:57",
      "content": "<p>It seems no chances to downlolad a single archive, so I need 1.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20200,
      "author_name": "budiansky",
      "author_url": "",
      "post_date": "02/08/2013 20:26:20",
      "content": "<p>I prefer #1. I've had problems downloading larger files, so it would be safer if there were two smaller ones.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20203,
      "author_name": "benhamner",
      "author_url": "",
      "post_date": "02/08/2013 21:35:31",
      "content": "<p>Well that's virtually unanimous. Two separate archives containing only the new data it is.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 20215,
      "author_name": "iamivo",
      "author_url": "",
      "post_date": "02/09/2013 04:16:18",
      "content": "<p>yaaay</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "20169": "",
    "20173": "",
    "20174": "",
    "20175": "",
    "20179": "",
    "20180": "",
    "20181": "",
    "20182": "",
    "20185": "",
    "20186": "",
    "20187": "",
    "20188": "",
    "20192": "",
    "20200": "",
    "20203": "",
    "20215": ""
  },
  "source": "meta"
}