{
  "id": 19306,
  "title": "model file size incorrect",
  "url": "/competitions/second-annual-data-science-bowl/discussion/19306",
  "author_name": "",
  "post_date": "2016-03-04T01:14:14.763Z",
  "votes": null,
  "comment_count": 3,
  "views": 470,
  "content": "<p>the model package size reported in the uploaded model page is different from what I see in local, (for example, locally the .tar.gz file is 146M but once I uploaded it says it is 139M), is it something I should worry about?</p>",
  "messages": [
    {
      "id": "110244",
      "postDate": "03/04/2016 01:14:14",
      "content": "<p>the model package size reported in the uploaded model page is different from what I see in local, (for example, locally the .tar.gz file is 146M but once I uploaded it says it is 139M), is it something I should worry about?</p>",
      "rawMarkdown": "the model package size reported in the uploaded model page is different from what I see in local, (for example, locally the .tar.gz file is 146M but once I uploaded it says it is 139M), is it something I should worry about?",
      "votes": null
    },
    {
      "id": "110251",
      "postDate": "03/04/2016 02:02:47",
      "content": "<p>I believe it's normal due the diffences between the Kaggle's storage system and your computer.</p>\n\n<p>You also can see differences transfering files between Windows and Linux and maybe Mac.</p>\n\n<p>The  the same file can have different sizes in these three systems</p>",
      "rawMarkdown": "I believe it's normal due the diffences between the Kaggle's storage system and your computer.\r\n\r\nYou also can see differences transfering files between Windows and Linux and maybe Mac.\r\n\r\nThe  the same file can have different sizes in these three systems",
      "votes": null
    },
    {
      "id": "110270",
      "postDate": "03/04/2016 07:46:30",
      "content": "<pre><code>139*1024^2=146M\n</code></pre>",
      "rawMarkdown": "139*1024^2=146M",
      "votes": null
    },
    {
      "id": "110305",
      "postDate": "03/04/2016 14:38:38",
      "content": "<p>As others have pointed out, the difference is mebibytes vs megabytes:</p>\n\n<p><a href=\"http://physics.nist.gov/cuu/Units/binary.html\">http://physics.nist.gov/cuu/Units/binary.html</a></p>\n\n<p>We may move to using megabytes in the future, as that seems to be the preference (I believe OSX/Linux use 1000, Windows uses 1024). In either case, your best bet to verify your model upload is to download the file from the site and check that it extracts. Don't just rely on the size.</p>",
      "rawMarkdown": "As others have pointed out, the difference is mebibytes vs megabytes:\r\n\r\nhttp://physics.nist.gov/cuu/Units/binary.html\r\n\r\nWe may move to using megabytes in the future, as that seems to be the preference (I believe OSX/Linux use 1000, Windows uses 1024). In either case, your best bet to verify your model upload is to download the file from the site and check that it extracts. Don't just rely on the size.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 110251,
      "author_name": "alvaroosvaldo",
      "author_url": "",
      "post_date": "03/04/2016 02:02:47",
      "content": "<p>I believe it's normal due the diffences between the Kaggle's storage system and your computer.</p>\n\n<p>You also can see differences transfering files between Windows and Linux and maybe Mac.</p>\n\n<p>The  the same file can have different sizes in these three systems</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 110270,
      "author_name": "nagadomi",
      "author_url": "",
      "post_date": "03/04/2016 07:46:30",
      "content": "<pre><code>139*1024^2=146M\n</code></pre>",
      "votes": null,
      "replies": []
    },
    {
      "id": 110305,
      "author_name": "wcukierski",
      "author_url": "",
      "post_date": "03/04/2016 14:38:38",
      "content": "<p>As others have pointed out, the difference is mebibytes vs megabytes:</p>\n\n<p><a href=\"http://physics.nist.gov/cuu/Units/binary.html\">http://physics.nist.gov/cuu/Units/binary.html</a></p>\n\n<p>We may move to using megabytes in the future, as that seems to be the preference (I believe OSX/Linux use 1000, Windows uses 1024). In either case, your best bet to verify your model upload is to download the file from the site and check that it extracts. Don't just rely on the size.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "110244": "the model package size reported in the uploaded model page is different from what I see in local, (for example, locally the .tar.gz file is 146M but once I uploaded it says it is 139M), is it something I should worry about?",
    "110251": "I believe it's normal due the diffences between the Kaggle's storage system and your computer.\r\n\r\nYou also can see differences transfering files between Windows and Linux and maybe Mac.\r\n\r\nThe  the same file can have different sizes in these three systems",
    "110270": "139*1024^2=146M",
    "110305": "As others have pointed out, the difference is mebibytes vs megabytes:\r\n\r\nhttp://physics.nist.gov/cuu/Units/binary.html\r\n\r\nWe may move to using megabytes in the future, as that seems to be the preference (I believe OSX/Linux use 1000, Windows uses 1024). In either case, your best bet to verify your model upload is to download the file from the site and check that it extracts. Don't just rely on the size."
  },
  "source": "meta"
}