{
  "id": 30621,
  "title": "Important: Labeling error in the \"additional\" dataset",
  "url": "/competitions/intel-mobileodt-cervical-cancer-screening/discussion/30621",
  "author_name": "",
  "post_date": "2017-03-24T22:59:55.245292800Z",
  "votes": 14,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hi all,</p>\n\n<p>We've pulled the \"additional\" dataset out of the data download since we have discovered there were bugs in the data labeling process. We have located where the inconsistencies are from, and will update the data in a few days. </p>\n\n<p>Very sorry for the inconvenience. In the mean time, the train and test images have the correct labels that you can trust. </p>\n\n<p>Kaggle admins</p>",
  "messages": [
    {
      "id": "170302",
      "postDate": "03/24/2017 22:59:55",
      "content": "<p>Hi all,</p>\n\n<p>We've pulled the \"additional\" dataset out of the data download since we have discovered there were bugs in the data labeling process. We have located where the inconsistencies are from, and will update the data in a few days. </p>\n\n<p>Very sorry for the inconvenience. In the mean time, the train and test images have the correct labels that you can trust. </p>\n\n<p>Kaggle admins</p>",
      "rawMarkdown": "Hi all,\n\nWe've pulled the \"additional\" dataset out of the data download since we have discovered there were bugs in the data labeling process. We have located where the inconsistencies are from, and will update the data in a few days. \n\nVery sorry for the inconvenience. In the mean time, the train and test images have the correct labels that you can trust. \n\nKaggle admins",
      "votes": null
    },
    {
      "id": "170317",
      "postDate": "03/25/2017 02:53:18",
      "content": "<p>Hi Wendy,</p>\n\n<p>would it also be possible to split the additional data into multiple file e.g 5GB each - 20GB+ files are pretty hard to download here in the UK if you are not sitting on a fat pipe like an university - unfortunately can't use torrent.</p>",
      "rawMarkdown": "Hi Wendy,\n\nwould it also be possible to split the additional data into multiple file e.g 5GB each - 20GB+ files are pretty hard to download here in the UK if you are not sitting on a fat pipe like an university - unfortunately can't use torrent.",
      "votes": null
    },
    {
      "id": "170333",
      "postDate": "03/25/2017 06:44:28",
      "content": "<p>Thanks a ton!!!! This would be one heck of a relief! :D</p>",
      "rawMarkdown": "Thanks a ton!!!! This would be one heck of a relief! :D",
      "votes": null
    },
    {
      "id": "170442",
      "postDate": "03/25/2017 20:57:37",
      "content": "<p>Will do! Thanks for the feedback. </p>",
      "rawMarkdown": "Will do! Thanks for the feedback.",
      "votes": null
    },
    {
      "id": "170561",
      "postDate": "03/26/2017 12:45:32",
      "content": "<p>would be great if you could provide a shell script for moving the files, so we don't need to download again\n(or just a csv with the correct labels for each file, if possible including the test files ;)</p>",
      "rawMarkdown": "would be great if you could provide a shell script for moving the files, so we don't need to download again\n(or just a csv with the correct labels for each file, if possible including the test files ;)",
      "votes": null
    },
    {
      "id": "170828",
      "postDate": "03/27/2017 18:56:14",
      "content": "<p>Will do. </p>",
      "rawMarkdown": "Will do.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 170317,
      "author_name": "fppkaggle",
      "author_url": "",
      "post_date": "03/25/2017 02:53:18",
      "content": "<p>Hi Wendy,</p>\n\n<p>would it also be possible to split the additional data into multiple file e.g 5GB each - 20GB+ files are pretty hard to download here in the UK if you are not sitting on a fat pipe like an university - unfortunately can't use torrent.</p>",
      "votes": null,
      "replies": [
        {
          "id": 170442,
          "author_name": "wendykan",
          "author_url": "",
          "post_date": "03/25/2017 20:57:37",
          "content": "<p>Will do! Thanks for the feedback. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 170333,
      "author_name": "yadavsarthak",
      "author_url": "",
      "post_date": "03/25/2017 06:44:28",
      "content": "<p>Thanks a ton!!!! This would be one heck of a relief! :D</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 170561,
      "author_name": "steelrose",
      "author_url": "",
      "post_date": "03/26/2017 12:45:32",
      "content": "<p>would be great if you could provide a shell script for moving the files, so we don't need to download again\n(or just a csv with the correct labels for each file, if possible including the test files ;)</p>",
      "votes": null,
      "replies": [
        {
          "id": 170828,
          "author_name": "wendykan",
          "author_url": "",
          "post_date": "03/27/2017 18:56:14",
          "content": "<p>Will do. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "170302": "Hi all,\n\nWe've pulled the \"additional\" dataset out of the data download since we have discovered there were bugs in the data labeling process. We have located where the inconsistencies are from, and will update the data in a few days. \n\nVery sorry for the inconvenience. In the mean time, the train and test images have the correct labels that you can trust. \n\nKaggle admins",
    "170317": "Hi Wendy,\n\nwould it also be possible to split the additional data into multiple file e.g 5GB each - 20GB+ files are pretty hard to download here in the UK if you are not sitting on a fat pipe like an university - unfortunately can't use torrent.",
    "170333": "Thanks a ton!!!! This would be one heck of a relief! :D",
    "170442": "Will do! Thanks for the feedback.",
    "170561": "would be great if you could provide a shell script for moving the files, so we don't need to download again\n(or just a csv with the correct labels for each file, if possible including the test files ;)",
    "170828": "Will do."
  },
  "source": "meta"
}