{
  "id": 305255,
  "title": "Are annotations allowed?",
  "url": "/competitions/happy-whale-and-dolphin/discussion/305255",
  "author_name": "",
  "post_date": "2022-02-04T13:02:57.659298100Z",
  "votes": 22,
  "comment_count": 6,
  "views": 0,
  "content": "<p>Is it allowed to annotate the whale part of the <strong>train data images?</strong> I would like to try object detection.</p>\n<p>In the rules, it says</p>\n<blockquote>\n  <p>Submissions may not use or incorporate information from hand labeling or human prediction of the validation dataset or test data records.</p>\n</blockquote>",
  "messages": [
    {
      "id": "1675754",
      "postDate": "02/04/2022 13:02:57",
      "content": "<p>Is it allowed to annotate the whale part of the <strong>train data images?</strong> I would like to try object detection.</p>\n<p>In the rules, it says</p>\n<blockquote>\n  <p>Submissions may not use or incorporate information from hand labeling or human prediction of the validation dataset or test data records.</p>\n</blockquote>",
      "rawMarkdown": "Is it allowed to annotate the whale part of the **train data images?** I would like to try object detection.\n\nIn the rules, it says\n\n > Submissions may not use or incorporate information from hand labeling or human prediction of the validation dataset or test data records.",
      "votes": null
    },
    {
      "id": "1675969",
      "postDate": "02/04/2022 15:15:57",
      "content": "<p>You can annotate the training images. Any test images must be evaluated by your algorithm without any human intervention, so no hand labeling is allowed for the test images.</p>",
      "rawMarkdown": "You can annotate the training images. Any test images must be evaluated by your algorithm without any human intervention, so no hand labeling is allowed for the test images.",
      "votes": null
    },
    {
      "id": "1676155",
      "postDate": "02/04/2022 18:17:30",
      "content": "<p>So, pseudo-labeling is allowed right?</p>",
      "rawMarkdown": "So, pseudo-labeling is allowed right?",
      "votes": null
    },
    {
      "id": "1676925",
      "postDate": "02/05/2022 11:57:46",
      "content": "<p>Thank you!</p>",
      "rawMarkdown": "Thank you!",
      "votes": null
    },
    {
      "id": "1713828",
      "postDate": "03/06/2022 12:26:49",
      "content": "<p><a href=\"https://www.kaggle.com/inversion\" target=\"_blank\">@inversion</a> <br>\nMust we make the annotated or some hand-labeled training dataset publicly available ? <br>\n(Just to be sure, I don't use any private annotated data at this time.)</p>",
      "rawMarkdown": "inversion \nMust we make the annotated or some hand-labeled training dataset publicly available ? \n(Just to be sure, I don't use any private annotated data at this time.)",
      "votes": null
    },
    {
      "id": "1715198",
      "postDate": "03/07/2022 17:48:30",
      "content": "<p>No - if you annotate the training images yourself you are not required to make that publicly available.</p>",
      "rawMarkdown": "No - if you annotate the training images yourself you are not required to make that publicly available.",
      "votes": null
    },
    {
      "id": "1715512",
      "postDate": "03/08/2022 04:56:06",
      "content": "<p><a href=\"https://www.kaggle.com/addisonhoward\" target=\"_blank\">@addisonhoward</a> OK, thank you for your clear response !</p>",
      "rawMarkdown": "addisonhoward OK, thank you for your clear response !",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1675969,
      "author_name": "inversion",
      "author_url": "",
      "post_date": "02/04/2022 15:15:57",
      "content": "<p>You can annotate the training images. Any test images must be evaluated by your algorithm without any human intervention, so no hand labeling is allowed for the test images.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1676155,
          "author_name": "chankhavu",
          "author_url": "",
          "post_date": "02/04/2022 18:17:30",
          "content": "<p>So, pseudo-labeling is allowed right?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1676925,
          "author_name": "shinmurashinmura",
          "author_url": "",
          "post_date": "02/05/2022 11:57:46",
          "content": "<p>Thank you!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1713828,
          "author_name": "yoichi7yamakawa",
          "author_url": "",
          "post_date": "03/06/2022 12:26:49",
          "content": "<p><a href=\"https://www.kaggle.com/inversion\" target=\"_blank\">@inversion</a> <br>\nMust we make the annotated or some hand-labeled training dataset publicly available ? <br>\n(Just to be sure, I don't use any private annotated data at this time.)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1715198,
          "author_name": "addisonhoward",
          "author_url": "",
          "post_date": "03/07/2022 17:48:30",
          "content": "<p>No - if you annotate the training images yourself you are not required to make that publicly available.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1715512,
          "author_name": "yoichi7yamakawa",
          "author_url": "",
          "post_date": "03/08/2022 04:56:06",
          "content": "<p><a href=\"https://www.kaggle.com/addisonhoward\" target=\"_blank\">@addisonhoward</a> OK, thank you for your clear response !</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1675754": "Is it allowed to annotate the whale part of the **train data images?** I would like to try object detection.\n\nIn the rules, it says\n\n > Submissions may not use or incorporate information from hand labeling or human prediction of the validation dataset or test data records.",
    "1675969": "You can annotate the training images. Any test images must be evaluated by your algorithm without any human intervention, so no hand labeling is allowed for the test images.",
    "1676155": "So, pseudo-labeling is allowed right?",
    "1676925": "Thank you!",
    "1713828": "inversion \nMust we make the annotated or some hand-labeled training dataset publicly available ? \n(Just to be sure, I don't use any private annotated data at this time.)",
    "1715198": "No - if you annotate the training images yourself you are not required to make that publicly available.",
    "1715512": "addisonhoward OK, thank you for your clear response !"
  },
  "source": "meta"
}